Global HIV Strategic Information Working Group
Biobehavioural Survey Guidelines For Populations At Risk For HIV
© World Health Organization 2017 Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). The mark “CDC” is owned by the US Dept. of Health and Human Services and is used with permission. Use of this logo is not an endorsement by HHS or CDC of any particular product, service, or enterprise. Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization. Suggested citation WHO, CDC, UNAIDS, FHI 360. Biobehavioral survey guidelines for Populations at Risk for HIV. Geneva: World Health Organization; 2017. Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data CIP data are available at http://apps.who.int/iris. Sales, rights and licensing To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see http://www.who.int/about/licensing. Third-party materials If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. Printed in Switzerland ISBN: 978-92-4-151301-2
Authorship and acknowledgements Authors Abu Abdul-Quader, Mark Berry, Trista Bingham, Janet Burnett, Maxia Dong, Amy Drake, Avi Hakim, Wolfgang Hladik, Angele Marandet, Anne McIntyre, Chris Murrill, Joyce Neal and Nita Patel of the US Centers for Disease Control and Prevention (CDC); Rajatashuvra Adhikary (formerly of FHI 360); Tobi Saidel of Partnership for Epidemic Analysis (PEMA) Partners; and Angela Kelly-Hanku of the University of New South Wales and of the Papua New Guinea Institute of Medical Research. Editors Abu Abdul-Quader, Mark Berry, Trista Bingham, Janet Burnett, Dana Dolan, Maxia Dong, Amy Drake, Avi Hakim, Wolfgang Hladik, Angele Marandet, Anne McIntyre, Chris Murrill, Joyce Neal and Nita Patel of CDC; Rajatashuvra Adhikary (formerly of FHI 360), Johannes van Dam and Steve Mills of FHI 360; staff of the Joint United Nations Programme on HIV/AIDS (UNAIDS); Jesus Garcia Calleja of WHO; Thomas Rehle of the Human Sciences Research Council (HSRC); Tobi Saidel of PEMA Partners; and Ted Alcorn of the Bill & Melinda Gates Foundation. Reviewers Maxia Dong, Shahul Ebrahim, Avi Hakim, Wolfgang Hladik, Amy Herman-Roloff, Andrea Kim, Rachel Kwezi, Sheryl Lyss, John Macom, Chris Murrill, Patrick Nadol, Sanny Chen Northbrook, Bharat Parekh, Nita Patel, Dimitri Prybylski, Ray Shiraishi and Peter Young of CDC; Rajatashuvra Adhikary (formerly of FHI 360), Timothy Mastro, Mike Merrigan, Steve Mills and Johannes van Dam of FHI 360; staff of UNAIDS; Jesus Garcia Calleja of WHO; Thomas Rehle: HSRC and University of Cape Town; Tobi Saidel of PEMA Partners; and Angela Kelly-Hanku of the Kirby Institute, University of New South Wales, and Sexual and Reproductive Health Unit, Papua New Guinea Institute of Medical Research. We would like to thank the individuals who contributed to this document: Ashley Burson and Laura Porter of CDC; Helen Coelho, Amanda Geller, Seseni Nu, Betty Treschitta and Almeta West of ICF International, Vanessa Brown of the Office of the US Global AIDS Coordinator; Maria Au, and Tisha Wheeler of the United States Agency for International Development (USAID), and Emily Crawford (formerly with USAID) The findings, conclusions and recommendations in this report do not necessarily represent the official position of the CDC. Funding to support this work is from the United States President’s Emergency Plan for AIDS Relief (PEPFAR) and the Bill & Melinda Gates Foundation.
Foreword To address a public health problem, you first have to measure it accurately. Biobehavioural surveys have proven to be invaluable tools for measuring and addressing HIV, which remains the world’s biggest public health challenge. This current iteration of the Biobehavioural survey guidelines is a welcome addition to the list of useful documents targeting those who plan to conduct biobehavioural surveys of HIV and HIV-risk behaviours in their countries. The guidelines can be applied across different countries, and to this end the document provides questionnaire modules that can be adapted to various contexts. The guidelines are presented in a logical and coherent manner, covering all survey aspects, from conceptualization of the survey to dissemination of the report and data use. The major focus is on key populations, which are often hidden and difficult to measure as part of general population-based surveys. This is particularly important because key populations are at high risk for HIV, and for exclusion from HIV and other health services. Estimating the size of these populations and their burden of HIV disease is extremely challenging, and these guidelines are a valuable resource for survey specialists as they undertake the surveys. The guidelines fill a gap in providing tools for surveying HIV prevalence in key populations, and the included questionnaires may also inform general population surveys. The 2000 Behavioural Surveillance Survey guidelines, while still useful, needed to be updated with newer survey methodology techniques and to incorporate biomarker testing. The guidelines will also serve as a textbook for students interested in working for research institutions that embark on epidemiological surveys. Currently, many researchers undertake surveys using country-specific indicators. These guidelines standardize the conduct of biobehavioural surveys to permit comparisons between as well as within countries over time. The use of common indicators allows for uniformity in the measurement of items and production of data that can be used by various global, regional, national and local actors in planning prevention and treatment services, tracking progress in the provision of HIV prevention and treatment services, and identifying gaps in access to services. The appendix on indicators will help scientists and data specialists to harmonize data management with a view to collaborating across countries using common yardsticks. The authors of these guidelines considered key aspects of surveys, from survey planning, design, data collection, analysis, presentation of results and dissemination of reports to data use. These guidelines are a must-have for anyone planning to conduct surveillance, whether experienced or not. It is our hope that the guidelines will help to refine measurement of HIV and help countries to address the unmet needs of their communities, to further reduce the toll of the epidemic. Olive Shisana Hon Professor, University of Cape Town President and CEO, Evidence Based Solutions and Chris Beyrer Desmond M. Tutu Professor of Public Health and Human Rights, Johns Hopkins University, Bloomberg School of Public Health
Foreword │ 1
Preface Biobehavioural surveys (BBS) provide specific populationlevel estimates for the burden of HIV disease and HIV-related risk factors, and estimates for the coverage of prevention and treatment services for populations at increased risk for HIV. These key populations include men who have sex with men, sex workers, people who inject drugs, transgender individuals, prisoners and other vulnerable populations at increased risk for HIV infection. For many of these stigmatized and socially marginalized populations, there are no conventional sampling frames, meaning that complex sampling designs are needed for these populations. The most frequently used survey guidelines and tools to date are the Behavioral surveillance surveys, issued in 2000. However, new HIV prevention, care and treatment policies – coupled with the emergence of new data needs, methods and technologies – warranted a thorough update of the 2000 publication. Thus, the US Centers for Disease Control and Prevention, FHI 360, the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the World Health Organization (WHO) are publishing these new Biobehavioural survey guidelines for populations at risk for HIV. This revised publication outlines the latest approaches and methodologies, and includes updated questionnaires for planning and conducting BBS. The new BBS guidelines are a comprehensive resource that covers all survey aspects, from conceptualization, planning, undertaking and costing of the survey to dissemination of a report and use of data. The ultimate goal of these guidelines is to facilitate the collection of high-quality survey data for informed public health action. The new guidelines: • update the overall approach and methodology of BBS in light of advances made during the past two decades; • improve the quality of BBS by providing comprehensive guidance, particularly for sampling methods and the collection of interview and biomarker data; • increase the relevance of survey data for public health programming by ensuring the collection of representative and actionable data; and • promote the use of survey findings to improve service delivery, monitoring and evaluation, and policy development. This document includes several new topics and features: • formative assessment covers the initial collection of information about a population, to inform how best to prepare and conduct a BBS. 2 │ Preface
• respondent driven sampling covers this peer-driven chain-referral sampling method, which is particularly useful for hard-to-sample populations, and is currently viewed as the most suitable probability-based sampling design. • biomarker considerations covers the entire range of biological measurements, from HIV serology to viral load, HIV recency, and biomarkers of other sexually transmitted infections. The document emphasizes the potential of population-level, aggregate viral-load metrics, such as prevalence of unsuppressed viral load. • population size estimation provides guidance on using integrated methods to estimate the number of members of a population. • questionnaire modules are included for collecting data on a wide range of topics, including exposure to and uptake of HIV-related services. Accompanying this publication will be electronic ready-to-use questionnaires, with the aim of keeping the questionnaires up-to-date as standards and indicators change. • indicators appendix lists standard and newly proposed indicators for both data and biomarker-related metrics. These guidelines, built on the lessons learned by experts and implementers from around the world, are intended to serve as a one-stop resource for survey planning and implementation, and dissemination of findings. They are expected to improve the quality of survey data through better survey design and implementation, and to promote the standardization of data measures to improve the comparability of survey data. Most of all, the guidelines should make it easier to plan and conduct BBS. By providing sample documents and guidance on every step of the process, we hope that these guidelines strengthen the capacity of public health workers to collect the information they need in a timely manner, allowing for an effective and strategic response to stop the HIV epidemic among key populations. Dr Shannon Hader Director, Division of Global HIV and TB, Center for Global Health/CDC Dr Gottfried Hirnschall Director, Department of HIV/AIDS and Global Hepatitis Programme, WHO Dr Luiz Loures Deputy Executive Director, Programme Branch, UNAIDS. Dr Timothy Mastro Chief Science Officer, FHI 360
Abbreviations and Acronyms ACASI ART ARV BBS CAB CAPI CASI CCS CDC CI CRC CT DBS DEFF DFA DNA EIA EMoS EPS FSW FP GoC HBV HCV HIV HPV HSRC HSV-2 ID IEC IRB MoS MSM NAAT NG NGO PCR PEP PMTCT PPS PrEP PSU PWID QA audio computer-assisted self-interview antiretroviral therapy antiretroviral biobehavioural survey community advisory board computer-assisted personal interview computer-assisted self-interview conventional cluster sampling Centers for Disease Control and Prevention confidence interval capture-recapture method Chlamydia trachomatis dried blood spot design effect direct fluorescent antibody deoxyribonucleic acid enzyme immunoassay estimated measure of size equal probability sampling female sex worker family planning game of contacts hepatitis B virus hepatitis C virus human immunodeficiency virus human papillomavirus Human Sciences Research Council herpes simplex virus-2 identification number information, education and communication institutional review board measure of size men who have sex with men nucleic acid amplification test Neisseria gonorrhoeae nongovernmental organization polymerase chain reaction post-exposure prophylaxis prevention of mother-to-child transmission probability proportional to size pre-exposure prophylaxis primary sampling unit people who inject drugs quality assurance QDS REC RDS RNA RPR RS RT SOP SI SRS STD/STI SW TB TG TLS UN UNAIDS UPC VDRL VL WB YCS Questionnaire Development System research ethics committee respondent-driven sampling ribonucleic acid rapid plasma reagin random start rapid test standard operating procedure sampling interval simple random sampling sexually transmitted disease/sexually transmitted infection sex worker tuberculosis transgender person time-location sampling United Nations Joint United Nations Programme on HIV/AIDS unique participant code venereal disease research laboratory viral load western blot Y chromosomal sequences
Table of Contents Foreword Preface 1 2 11 17 18 18 18 20 20
1. Introduction A. Survey preparation 1 Stakeholders A-1.1 Rationale for including stakeholders A-1.2 Stakeholder contributions A-1.3 Stakeholder engagement process A-1.4 References 2 Planning considerations before starting a biobehavioural survey A-2.1 Reflect on the HIV epidemic A-2.2 Consider epidemiological factors when choosing target population for BBS A-2.3 Conduct an initial review of information about proposed target population A-2.4 Ensure adherence to ethical standards A-2.5 Ensure funding is sufficient to meet objectives A-2.6 Consider timing and intervals between BBS rounds A-2.7 References 3 Ethical consideration in planning and conducting surveys A-3.1 International guidelines for human subjects research A-3.2 Fundamental ethical principles A-3.3 Research ethics committees A-3.4 Sampling of participants A-3.5 Obtaining voluntary informed consent A-3.6 Ensuring privacy and confidentiality A-3.7 Personal identifying information A-3.8 Data security A-3.9 Reimbursement to participants A-3.10 Return of test results A-3.11 Treatment and referral A-3.12 Adverse events, protocol deviations and protocol amendments A-3.13 Research ethics training A-3.14 Special considerations for incarcerated populations A-3.15 Web-based surveys A-3.16 References 4 Formative assessment A-4.1 The purpose of formative assessment A-4.2 Formative assessment objectives A-4.3 Conducting a formative assessment A-4.4 Optimal use of information gathered through the formative assessment
21 21 21 22 23 23 23 24
25 25 26 26 26 26 28 29 29 29 29 29 30 30 30 31 31
32 32 33 35 35
A-4.5 Ongoing formative assessment A-4.6 References 5 Questionnaire development A-5.1 Key steps in questionnaire development A-5.2 Summary A-5.3 References 6 Biomarker considerations A-6.1 Biomarker selection A-6.2 Resource availability A-6.3 Ethical considerations A-6.4 Providing test results to participants A-6.5 On-site or off-site testing A-6.6 Treatment and referral A-6.7 Testing for HIV biomarkers A-6.8 Testing for non-HIV biomarkers A-6.9 Selection of tests and testing algorithms A-6.10 Development of standard operating procedures A-6.11 Specimen type, collection, labelling, processing and transport A-6.12 Laboratory data management A-6.13 Laboratory procedures A-6.14 Specimen repository A-6.15 Quality assurance (assessment) and quality control A-6.16 Supply chain considerations A-6.17 Biosafety and biosecurity A-6.18 References 7 Preparing biological data-collection instruments A-7.1 Specimen tracking forms A-7.2 Specimen transport logs A-7.3 Refrigerator or freezer temperature logs A-7.4 Test results form 8 Eligibility for participation A-8.1 Defining the population A-8.2 Verifying eligibility A-8.3 Elements of eligibility A-8.4 Exclusion criteria 9 Sampling strategy A-9.1 Concepts used in sampling, and types of sampling A-9.2 Nonprobability sampling methods A-9.3 Cluster sampling overview A-9.4 Procedures for two-stage and multistage sampling A-9.5 Conventional cluster sampling versus time-location cluster sampling A-9.6 Time-location sampling A-9.7 Respondent-driven sampling A-9.8 References
35 35
36 36 43 43
45 45 47 47 47 48 48 48 50 54 55 55 59 59 59 59 60 60 60
62 63 63 63 63
65 65 66 66 67
68 68 72 74 76 79 83 92 101
10 Sample size requirements A-10.1 Considerations for determining sample size A-10.2 Additional considerations A-10.3 References 11 Population size estimation methods used with surveys A-11.1 Population size estimation based on surveys A-11.2 Emerging methods A-11.3 Selecting a method A-11.4 References 12 Supplemental studies A-12.1 Qualitative studies A-12.2 Cohort studies A-12.3 Partner surveys A-12.4 References 13 Data management A-13.1 Data documentation A-13.2 Data dictionary A-13.3 Unique participant identifier A-13.4 Data-quality checks A-13.5 Skip patterns A-13.6 Data entry A-13.7 Data confidentiality A-13.8 Data access and use A-13.9 Data backup A-13.10 Data archiving and version control A-13.11 Data security and storage 14 Staff selection A-14.1 Considerations for field staff selection A-14.2 Field staff roles and responsibilities: non-method specific A-14.3 Field staff roles and responsibilities: method specific A-14.4 Considerations for staff payment 15 Standard operating procedures A-15.1 How to write a standard operating procedure A-15.2 BBS activities that require an SOP A-15.3 Sampling-specific SOPs A-15.4 SOPs for checklists, logbooks labels and forms 16 Training and training methods A-16.1 Training A-16.2 Training methods
103 103 107 108
109 112 116 117 119
121 121 130 131 133
135 135 135 136 136 137 137 137 137 138 138 138
140 140 141 142 142
143 143 143 144 144
146 146 147
B. Survey implementation and quality assurance 1 Data collection B-1.1 Surveys using nonprobability sampling methods B-1.2 Conventional cluster sampling B-1.3 Implementation of time-location sampling B-1.4 Respondent-driven sampling 2 Survey monitoring B-2.1 Observing data collection B-2.2 Laboratory monitoring 3 Preparing data for analysis B-3.1 Identifying and correcting data errors
148 149 149 152 156 167
173 174 174
176 176
C. Data analysis and use 1 Data analysis C-1.1 Developing a data analysis plan C-1.2 Preparing a dataset for analysis C-1.3 Analytical methods for BBS data C-1.4 Combining multiple surveys and subsetting data for analysis C-1.5 Comparing surveys C-1.6 Suggested resources C-1.7 References 2 Use and dissemination of survey findings C-2.1 Potential uses for BBS data C-2.2 Identifying a target audience C-2.3 Tailoring a message C-2.4 Determining a dissemination strategy C-2.5 Additional considerations for BBS data interpretation C-2.6 Relevant websites C-2.7 References 3 Evaluating survey implementation C-3.1 Data collection for evaluation C-3.2 Using evaluation findings 4 Data sharing for public use C-4.1 Data Sharing C-4.2 References
185 186 186 187 193 196 196 199 199
200 200 201 202 202 203 204 204
205 205 206
207 207 208
Supplemental Materials Section I: Appendices I-1: Protocol Checklist I-2: Sample Budget Calculator and Gantt Chart (Survey Timeline) I-3: Sample Consent Form I-4: Survey Design Checklist I-5: Brief Demographic Survey for Key Informants and Focus Groups I-6: Formative Interview Guide for Gatekeepers I-7: Formative Assessment Questionnaire Guide for Target Population I-8: Formative Mapping I-9: Safety and Security Guide I-10: Specimens Used for HIV Testing I-11: Material Transfer Agreement I-12: Interpretation of Hepatitis Serologic Test Results I-13: Specimen Tracking Form I-14: Specimen Transportation Driver Log I-15: Comparison of HIV Testing Technologies I-16: Specimen Temperature Monitoring Logs I-17: Sample HIV Rapid Testing Results Log I-18: Equal Probability Sampling (EPS) and Probability Proportional to Size (PPS) Cluster Sampling Guide I-19: Survey Implementation & Quality Assurance – Data Collection - TLS I-20: Venue Observation Forms I-21: Unique Participant Code (UPC) I-22: Sample Size Requirements to Detect Differences I-23: Calculator for the Number of Unique Objects for Estimating Population Size I-24: Qualitative Interview Guide I-25: Qualitative Interview for Target Population (FSW) Consent Form I-26: Formative Assessment Consent Form for Key Informants I-27: Data Management SOP for RDS Survey Using QDS I-28: Sample Data Dictionary I-29: Staff Data Agreement I-30: Specimen Handling (Collection, Processing, and Storage) SOP I-31: Recruitment Training Script I-32: RDS Participant Checklist I-33: Data Error Audit Log I-34: Sample Table of Contents for Survey Report I-35: Sample Press Release I-36: Data Presentation Guide I-37: Data Use (Confidentiality) Agreement 1 2 10 30 33 39 41 46 59 66 63 69 72 74 76 78 80 84 86 91 94 99 100 101 102 109 110 112 114 116 118 120 122 125 126 128 130 132 133 136 140 151 154 156 159 162 171
Section II: Questionnaire Modules 1. Questionnaire Parameters 2. Eligibility 3. RDS Recruiter-Recruit Relationship 4. RDS Personal Network Size 5. Size Estimation 6. ACASI Tutorial 7. Demographics 8. Transgender Characteristics
9. Sex Work Characteristics 10. Clients Of Sex Workers Characteristics 11. Incarceration Characteristics 12. Transport Worker Characteristics 13. Reproductive Health 14. Sexual History 15. Recall Sexual Behavior 16. Partner Concurrency 17. Last Sex Act 18. Condom Use/Accessibility 19. Lubricant Use 20. Microbicides 21. Alcohol Use 22. Non-Injection Drug Use 23. Injection Drug Use 24. Drug Overdose 25. Sexually Transmitted Infections 26. Hepatitis B and C 27. Services Uptake 28. PEP and PrEP 29. Shame, Stigma, Harassment & Discrimination 30. Physical Violence 31. Sexual Violence: Coercion and Rape 32. Internalized Homophobia 33. Depression 34. Suicide Ideation and Attempts 35. Social Support 36. Social Cohesion, Social Participation and Collective Agency 37. Game of Contacts 38. HIV Knowledge and Perceptions 39. Questionnaire Feedback/Interview Status 40. RDS, Peer Recruitment
176 185 188 196 200 208 218 239 243 249 256 261 264 270 277 295 301 306 309 341 346 369 383 403 406 410 415 431 437 439 442 444 451 454 454 455 456 457 460 460 461 463 464 465 466 467 470
Section III: Indicators Introduction What are indicators? How are data collected for indicators? How are indicators used? Indicator selection for these guidelines Indicators Sexual debut Condom use at last sex Consistent condom use Buying sex Selling sex Unprotected receptive anal intercourse Alcohol use Injecting drug use
Received clean needles or syringes Sharing injecting equipment Sharing drug-preparation equipment Use of sterile injecting equipment Received medication for drug dependency Drug overdose Drug overdose prevention training Targeted information, education and communications Received condoms Never tested for HIV HIV test in last 12 months Tuberculosis screening Received tuberculosis treatment Currently in care CD4 count Currently on antiretroviral therapy Prevention of mother-to-child transmission Hepatitis B virus testing Hepatitis C virus testing Discussed pre-exposure prophylaxis Taken pre-exposure prophylaxis Current use of modern family planning methods Antenatal clinic access and HIV testing Seeking care for sexually transmitted infection symptoms Sexually transmitted infection screening Received sexually transmitted infection treatment Received lubricants HIV knowledge and perceptions Incarceration Health-care stigma Discrimination and social exclusion Violence Condom use at last sex (sex workers only) Biomarker Indicators Eligible but not on HIV treatment HIV treatment coverage Viral load suppression Unsuppressed viral load Population viral load HIV/sexually transmitted infection coinfection HIV/hepatitis coinfection HIV prevalence Sexually transmitted infection prevalence
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 504 505 506 507 508 509 510 511 512 513 514 515
Section IV: Glossary
Introduction
Background Rationale for the guideline Objectives Target audience Relationship to other resources Structure References
1 2 3 4 5 6 7
Introduction
A Survey preparation
B Survey implementation & quality assurance
C Data analysis and use
Foreword │ 11
Background An understanding of HIV burden, risk factors, and coverage of prevention and treatment services is critical for combatting the HIV epidemic (1). Thus, biobehavioural surveys (BBS) assessing these parameters are integral components of a national HIV strategy and surveillance system. These guidelines outline the latest approaches and methodologies for planning and conducting such surveys. Countries that implement repeated BBS can monitor changes in their populations’ risks for HIV, determinants of those risks, and access to prevention and treatment over time. Many individuals at high risk for HIV are socially marginalized and may not identify themselves as such when accessing services. This makes it difficult to track them in HIV programme registers, and impedes efforts to assess the effectiveness of services. Such key populations include sex workers (SW), men who have sex with men (MSM), transgender women (TG) and people who inject drugs (PWID), all of whom are at increased risk for HIV infection compared with the population at large (2-7). These guidelines describe sampling designs and methods for conducting surveys among key populations for which ready-made sampling frames are not typically available. As affirmed by the Joint United Nations Programme on HIV/AIDS (UNAIDS)/WHO in the publication Guidelines for second generation HIV surveillance: an update: know your epidemic (8) HIV surveillance among these key populations is a priority in all epidemic settings. In countries where the majority of HIV transmission takes place among members of the general population, data from HIV surveys among the general population should be used together with data from BBS among key populations to inform the epidemic response. This is due to the realization that countries previously categorized as having ‘generalized’ epidemics in fact experience ‘mixed’ epidemics, where HIV epidemics exist simultaneously among the general population and key populations. Separate surveys for key populations are warranted in such settings because key population members are usually not well identified or covered in general population surveys. Although these guidelines focus on key populations, they are also applicable to other populations, such as clients of SW, prisoners, transportation workers and migrants. The term “target population” is used here to denote any survey group deemed to be at high risk for HIV infection, including key populations. HIV control efforts directed at key and other high-risk populations can have a substantial impact on the epidemic. HIV policies and programmes will be more effective if they are informed by accurate measures of
HIV prevalence and incidence among those populations, trends in their HIV-related risk behaviours, and the extent to which they access prevention and treatment services. This makes surveys critical tools for monitoring the HIV epidemic and evaluating HIV control efforts. UNAIDS and other groups have projected that 28–65% of new HIV infections in the People’s Republic of China, the Dominican Republic, Kenya, Mozambique, Nigeria and Peru are among key populations and their partners (2-7). Up to 30% of all HIV infections worldwide are due directly or indirectly to sex work, which can involve male, female and transgender individuals, demonstrating the importance of considering all genders when planning BBS (9). Additionally, global prevalence of HIV among MSM is 19 times higher than in the general population, and prevalence among TG is 48 times higher (10-12). Meanwhile, PWID represent an estimated 10% of all HIV infections globally (13) and are 22 times more likely to be HIV-infected compared to the general population (14). For SW, HIV transmission is more likely to occur due to inconsistent condom use with multiple partners. MSM and TG are at high risk because HIV transmission is five times more likely to occur during unprotected, receptive anal sex than during vaginal sex (15). Among PWID, sharing nonsterile needles or syringes is a highly efficient way to transmit HIV (16). Elevated risk for HIV transmission and acquisition among key populations is primarily related to four factors: • efficient transmission of infection via unprotected anal sex and needle-sharing behaviours; • frequent exposure to infection via multiple sex partners if not consistently using condoms; • high HIV prevalence among networks of sexual or drug-injecting partners; and • inferior access to quality health care compared to individuals in the general population.
Key Populations and Risk Being an SW, MSM, TG or PWID alone does not increase the individual risk for getting or transmitting HIV. Rather, risk is elevated when individuals practise “unsafe” behaviours (i.e. have vaginal or receptive anal sex without a condom or share nonsterile injecting equipment) with partners who have a different, or discordant, HIV status from them. The underlying reason that key populations are at higher risk for HIV is that HIV prevalence among their sexual or needle-sharing network is already high. Although members of the general population also engage in frequent vaginal and anal sex without condoms, their chance of having an HIV-infected partner is much lower than that of members of key populations.
I │ Introduction
Policies that criminalize behaviours of key populations or limit their access to services may further elevate their risk of acquiring and transmitting HIV. For example, more than two thirds of sub-Saharan African countries criminalize same-sex practices (17). Criminalized populations receive less than 25% of intended humanrights-focused interventions, and may consequently lack access to services necessary for protecting themselves (18). In countries where the behaviours of SW, MSM, TG or PWID are criminalized, and those laws are enforced, the negative impact on access to services is even greater (19). Conversely, laws that prohibit discrimination against certain populations are correlated with better access to prevention services by the populations. Sociostructural factors such as stigma, harassment, marginalization and lack of social support can contribute to higher transmission of HIV and other infections. This is particularly apparent among sex workers, sexual minorities and persons who inject drugs (20, 21). Social marginalization or rejection can force individuals into isolation, unemployment, poverty and unstable living conditions or homelessness. It can also expose them to physical and sexual violence, and limit their access to HIV outreach commodities and educational messages (22). Fear of police harassment, arrest or discrimination by health-care providers may also discourage members of key populations from accessing prevention and treatment services. In an assessment of the HIV prevention services available to MSM in 147 low- and middle-income countries, 55% were unable to document provision of even the most basic HIV-related services (23). It is estimated that, worldwide, only 4% of PWID with HIV who are eligible for treatment are on antiretroviral therapy (ART) (24).
hepatitis), qualitative interviews and population size estimation activities. The ethics of conducting surveys with HIV testing – particularly the return of test results to participants – have also evolved. National governments, donors and United Nations (UN) agencies increasingly require data on access to HIV services, particularly because of the important role of antiretroviral therapy (ART) in HIV treatment and prevention. Engaging target populations and other stakeholders has become more important for survey planning and implementation, as has providing third-party researchers broader access to survey data. These guidelines aim to facilitate BBS design and implementation in this context of evolving survey methods and measures. The ultimate purpose of the document is to facilitate the collection of high-quality survey data that will lead to informed public health action.
Objectives These guidelines aim to: • update the overall approach and methodology of BBS in light of advances made during the past decade; • improve the quality of BBS by providing comprehensive guidance, particularly for sampling methods and the collection of interview and biomarker data; • increase the relevance of BBS for public health programming by ensuring the collection of representative and actionable data; and • promote the use of survey findings to improve service delivery, monitoring and evaluation, and policy development.
Rationale for the guideline In 2000, Family Health International (now known as FHI 360) issued a resource for HIV survey protocol development, implementation and data interpretation entitled Guidelines for repeated behavioral surveys in populations at risk for HIV (25). Informally known as the “Red Book” because of the colour of its cover, that document has been widely used as a guide for conducting BBS among populations at risk for HIV. The HIV landscape has changed significantly since the arrival of combination antiretroviral treatment in the 2000s in low- and middle-income countries. In addition, data needs and survey methods have changed substantially in the years since then. New sampling methods have emerged, such as respondent-driven sampling (RDS). Surveys increasingly include tests for biomarkers (e.g. HIV serostatus, CD4+ T-cell count, viral load, other sexually transmitted infections or viral
Target audience These guidelines are written for policy-makers, organizations and technical staff who are planning or conducting BBS among key populations and other at-risk populations. The document is applicable for all settings, but particularly for low- and middle-income countries. We assume that the reader possesses a basic understanding of epidemiological and survey principles, sampling, inference and measures of risk.
Relationship to other resources There are many excellent resources related to HIV surveys (26), population size estimation (27), and second-generation HIV surveillance (1, 8). This document supplements and complements these resources.
Introduction │ II
However, BBS are just one data source informing key population HIV epidemics; others include population size estimation, mapping, HIV case surveillance, cohort studies, programme-based data and, further upstream, data on policy, stigma and discrimination. These guidelines are not intended for general population or household-based surveys; other resources are available for those purposes.
Structure These guidelines are intended to provide an all-in-one resource for the conduct of BBS. Written in a concise manner, they focus on the “how to” and practical aspects of survey implementation – from formative assessment, survey design and protocol development to survey conduct, data analysis and dissemination of findings. The core of the guidelines is divided into three principal sections, which are subdivided into chapters. Section A, Survey preparation, is the largest part of the guidelines, and begins by discussing stakeholder involvement, formative assessment, human subjects considerations, survey instrument development, biomarker considerations, survey eligibility criteria, and population size estimation methods. Other chapters cover principal sampling strategies, sample size calculation, data management and data-collection procedures. Section A also covers practical aspects of survey preparation – including the development of standard operating procedures (SOP), ensuring adequate human resources and conducting appropriate staff training. Section B, Survey implementation and quality assurance, focuses on monitoring and field supervision during survey implementation and data management, and specific cluster survey and RDS-related considerations for survey conduct. Section C, Data analysis and use, addresses data analysis, data use, preparation of anonymized individual data files for data sharing, and evaluation of the completed survey.
The document also includes extensive supplementary materials: Appendices (Section I), Questionnaire modules (Section II), Indicators (Section III) and Glossary (Section IV). The collection of data through standardized data instruments is the core of any survey activity. The data instruments have undergone considerable modifications compared with those in the original guidelines; for example, they now include questions on the continuum of care. Rather than complete, standalone data instruments for distinct target populations, the instruments are organized as modules focused on particular topics. This provides users with more flexibility to construct data instruments according to local priorities and the needs of each target population. The data instruments’ core variables have also been made available in ready-to-use electronic form. Indicators and other standardized data measurements facilitate comparison of the HIV epidemic between different time points, locations or populations. The recommended reference sheets for data measurements are harmonized with indicators used by the UN and other global organizations engaged in HIV funding and programming. This section also includes additional data measurements recommended by the guidelines’ authors. The supplementary materials include an array of resources, including references for further reading and examples of protocol checklists, survey consent language and data analysis commands for commonly used software packages. We hope these updated guidelines will facilitate the collection of more detailed, accurate and relevant data among populations at risk for HIV and thus facilitate more effective HIV services, advocacy and policy-making.
References 1 UNAIDS/WHO Working Group on Global HIV/AIDS and STI surveillance. Guidelines for second generation HIV surveillance: the next decade. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/WHO Working Group on Global HIV/AIDS and STI surveillance; 2000 (http://www.who.int/hiv/pub/surveillance/pub3/en/index. html, accessed 4 August 2016). PAHO. Modos de transmisión del VIH en América Latina. Pan American Health Organization (PAHO); 2009 (http:// www.unaids.org/en/media/unaids/contentassets/documents/countryreport/2009/20090810_MOT_Peru_es.pdf, accessed 3 August 2016). UNAIDS. Country snapshot: China: HIV and men who have sex with men. Joint United Nations Programme on HIV/ AIDS (UNAIDS); 2012 (http://www.unaids.org.cn/pics/20121227100804.pdf, accessed 4 August 2016).
2
3
III │ Introduction
4
UNAIDS/COPRESIDA/DIGECITSS. HIV modes of transmission model: analysis of the distribution of new HIV infections in the Dominican Republic and recommendations for prevention. Santo Domingo, República Dominicana: Joint United Nations Programme on HIV/AIDS (UNAIDS), Consejo Presidencial del SIDA (COPRESIDA), Dirección General de Infecciones de Transmisión Sexual y SIDA (DIGECITSS); 2010 (http://www.unaids.org/en/media/unaids/ contentassets/documents/countryreport/2010/201011_MOT_DominicanRepublic_en.pdf, accessed 3 August 2016).
5 UNAIDS/World Bank. New HIV infections by modes of transmission in West Africa: a multi-country analysis. Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Bank; 2010 (http://www.unaids.org/sites/default/files/ en/media/unaids/contentassets/documents/countryreport/2010/201003_MOT_West_Africa_en.pdf, accessed 4 August 2016). 6 World Bank/UNAIDS. Kenya: HIV prevention response and modes of transmission analysis. World Bank/ Joint United Nations Programme on HIV/AIDS (UNAIDS); 2009 (http://siteresources.worldbank.org/INTHIVAIDS/ Resources/375798-1103037153392/KenyaMOT22March09Final.pdf, accessed 4 August 2016). World Bank/UNAIDS. Analysis of modes of HIV transmission and national response to HIV and AIDS. Maputo, Mozambique: World Bank/Joint United Nations Programme on HIV/AIDS (UNAIDS); 2010 (http://regist2.virologyeducation.com/4thINTEREST/docs/14_Fazenda.pdf, accessed 4 August 2016). UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines for second generation HIV surveillance: an update: know your epidemic. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/ World Health Organization (WHO); 2013 (http://apps.who.int/iris/bitstream/10665/85511/1/9789241505826_eng. pdf, accessed 3 August 2016). Baral S, Beyrer C, Muessig K, Poteat T, Wirtz AL, Decker MR et al. Burden of HIV among female sex workers in lowincome and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12(7):538–549.
7
8
9
10 Baral S, Sifakis F, Cleghorn F, Beyrer C. Elevated risk for HIV infection among men who have sex with men in lowand middle-income countries 2000–2006: a systematic review. PLoS Med. 2007;4(12):e339 (http://www.ncbi.nlm. nih.gov/pubmed/18052602, accessed 4 August 2016). 11 Baral SD, Poteat T, Stromdahl S, Wirtz AL, Guadamuz TE, Beyrer C. Worldwide burden of HIV in transgender women: a systematic review and meta-analysis. Lancet Infect Dis. 2013;13(3):214–222. 12 Beyrer C, Baral SD, van Griensven F, Goodreau SM, Chariyalertsak S, Wirtz AL et al. Global epidemiology of HIV infection in men who have sex with men. Lancet. 2012;380(9839):367–377. 13 Mathers BM, Degenhardt L, Phillips B, Wiessing L, Hickman M, Strathdee SA et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733–1745. 14 UNAIDS. Global report: UNAIDS report on the global AIDS epidemic 2012. Joint United Nations Programme on HIV/AIDS (UNAIDS); 2012 (http://www.unaids.org/sites/default/files/media_asset/20121120_UNAIDS_Global_ Report_2012_with_annexes_en_1.pdf, accessed 4 August 2016). 15 Varghese B, Maher JE, Peterman TA, Branson BM, Steketee RW. Reducing the risk of sexual HIV transmission: quantifying the per-act risk for HIV on the basis of choice of partner, sex act, and condom use. Sex Transm Dis. 2002;29(1):38–43. 16 Kaplan EH, Heimer R. A model-based estimate of HIV infectivity via needle sharing. J Acquir Immune Defic Syndr. 1992;5(11):1116–1118. 17 Itaborahy LP, Zhu J. State-sponsored homophobia: a world survey of laws: criminalisation, protection and recognition of same-sex love. International Lesbian Gay Bisexual Trans and Intersex Association (ILGA); 2013 (http://old.ilga.org/Statehomophobia/ILGA_State_Sponsored_Homophobia_2013.pdf, accessed 4 August 2016).
Introduction │ IV
18 UNDP/The Global Fund/UNAIDS. Analysis of key human rights programmes in Global Fund-supported HIV programmes. New York: United Nations Development Programme (UNDP)/The Global Fund/Joint United Nations Programme on HIV/AIDS (UNAIDS); 2011 (http://www.undp.org/content/dam/aplaws/publication/en/publications/ hiv-aids/analysis-of-key-human-rights-programmes-in-global-fund-supported-hiv-programmes/Analysis%20of%20 Key%20HRTS%20Programmes%20in%20GF-Supported%20HIV%20Programmes.pdf, accessed 4 August 2016). 19 PAHO. Improving access of key populations to comprehensive HIV health services: towards a Caribbean consensus. Washington D.C.: Pan American Health Organization (PAHO); 2011 (http://www.paho.org/hq/index. php?option=com_docman&task=doc_download&gid=16974&Itemid=, accessed 4 August 2016). 20 Altice FL, Kamarulzaman A, Soriano VV, Schechter M, Friedland GH. Treatment of medical, psychiatric, and substance-use comorbidities in people infected with HIV who use drugs. Lancet. 2010;376(9738):367–387. 21 Rhodes T, Singer M, Bourgois P, Friedman SR, Strathdee SA. The social structural production of HIV risk among injecting drug users. Soc Sci Med. 2005;61(5):1026–1044. 22 Centers for Disease Control and Prevention. Integrated prevention services for HIV infection, viral hepatitis, sexually transmitted diseases, and tuberculosis for persons who use drugs illicitly: summary guidance from CDC and the U.S. Department of Health and Human Services. MMWR. 2012;61(RR05):1–40 (http://www.cdc.gov/mmwr/pdf/rr/ rr6105.pdf, accessed 8 August 2016). 23 UNAIDS. The Prevention Gap Report. Geneva: UNAIDS; 2016. (http://www.unaids.org/sites/default/files/media_ asset/2016-prevention-gap-report_en.pdf, accessed 31 August 2017) 24 Harm Reduction International (2016) ‘Global State of Harm Reduction 25 Amon J, Brown T, Hogle J, MacNeil J, Magnani R, Mills S et al. Behavioral surveillance surveys BSS: guidelines for repeated behavioral surveys in populations at risk of HIV. Arlington: Family Health International (FHI); 2000 (http:// www.who.int/hiv/strategic/en/bss_fhi2000.pdf, accessed 3 August 2016). 26 UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines on surveillance among populations most at risk for HIV. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2011 (http://www.unaids.org/en/media/unaids/contentassets/restore/20110518_Surveillance_among_ most_at_risk.pdf, accessed 26 June 2016). 27 UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines on estimating the size of populations most at risk to HIV. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2010 (http://www.unaids.org/sites/default/files/media_asset/2011_Estimating_Populations_en_0.pdf, accessed 4 August 2016).
V │ Introduction
A. Survey preparation
1. Stakeholders 2. Planning considerations before starting a biobehavioural survey 3. Ethical consideration in planning and conducting surveys 4. Formative assessment 5. Questionnaire development 6. Biomarker considerations 7. Preparing biological data-collection instruments 8. Eligibility for participation 9. Sampling strategy 10. Sample size requirements 11. Population size estimation methods used with surveys 12. Supplemental studies 13. Data management 14. Staff selection 15. Standard operating procedures 16. Training and training methods
16 19 23 30 34 43 60 63 66 101 107 121 135 140 143 146
A Survey preparation
Data analysis and use
C
1. Data analysis 2. Use and dissemination of survey findings 3. Evaluating survey implementation 4. Data sharing for public use
186 200 206 208
1. Data collection 149 2. Survey monitoring 173 3. Preparing data for analysis 176
B Survey implementation & quality assurance
Section
A Stakeholders may include: • agencies of the local or national government; for example, different ministries, provincial health authorities, national reference laboratories or HIV control programmes (as appropriate); • development agencies; • academic institutions; • service providers; • nongovernmental organizations (NGOs) and community-based organizations; • donors; • community leaders; and • members of the target population.
1. Stakeholders
This chapter describes the role that stakeholders play in survey planning and preparation. Stakeholders are individuals or organizations with an interest in or concern for the implementation and results of a biobehavioural survey (BBS).
Key terms Community-based An organization largely composed of organization: members from a given population (e.g. men who have sex with men), often working in advocacy or providing services. Nongovernmental An organization founded by citizens organization and not part of the government and (NGO): not a for-profit company. Stakeholder: A party or group of people with an interest in a given activity, who can contribute to it or is affected by it.
A-1.1 Rationale for including stakeholders Various stakeholders can contribute to the successful implementation of BBS at every stage of the process. They can help to ensure that BBS design is consistent with on-the-ground realities, encourage target population participation, and provide resources or services to facilitate survey implementation. The primary consumers of the survey findings, including the national HIV programme and donors, should be involved in the planning and implementation of the BBS. Early stakeholder engagement often leads to greater use of BBS results to inform services, policy and advocacy. Target population support for the survey is crucial, and investigators planning a BBS should engage influential members early in the process, both to seek their input on survey content and to communicate how the data may be used. Investigators can also learn about the local context from service providers including clinicians, NGOs and other civil society organizations that offer outreach services. Because of their close relationship with the target population, these stakeholders may also be able to encourage participation. Local academic institutions may be able to offer guidance on sampling strategies, data analysis and report development. 18 1. Stakeholders
Investigators may also benefit from engaging legislators and politicians, who can allocate resources for surveys, and use the findings to advocate for policies that benefit the target population. Such stakeholders can also improve access to services and prevent harassment of the target population during the BBS. Although stakeholders make important contributions, national officials should play a lead role in choosing the target populations, variables of interest and types of analyses, and in disseminating the data. Often, leadership by a national body increases the likelihood that the BBS will be a high-quality survey, will serve the needs of the country as a whole and will result in data that are widely used.
A-1.2 Stakeholder contributions Individuals and groups may be involved in all aspects of survey planning and implementation, from financial support to dissemination of findings (see Table A-1.1). Stakeholder dialogue is essential when there may be a difference of opinion about the scope (e.g. survey location, data and biomarker measures) or objectives of a BBS. For example, stakeholders may have different views on the prioritization of one target group or the inclusion of biomarkers related or unrelated to HIV. B. Survey implementation & quality assurance C. Data analysis and use
A. Survey preparation
Dialogue among stakeholders can clarify these issues, and can be useful in considering options and making decisions. Common stakeholder expectations are essential for a successful survey. Stakeholder contributions might include: • defining the scope of the survey; • defining the populations to be surveyed;
• providing input on the data and biomarkers to be collected; • raising funds for the survey; • promoting participation in and acceptance of the survey; and • employing survey findings to develop and advocate for policies.
Table A-1.1 Potential stakeholders to consider for BBS planning and implementation activities Stakeholder • Members of the target population • Community leaders
What they offer • Ensure the survey is perceived as legitimate by the target population • Provide investigators with social context • Provide information on the needs of target populations • Provide insight into: - survey content - appropriateness of sampling strategy - survey design - wording of the questionnaire • Provide social context • Encourage participation by the target population • Provide information on the needs of target populations • Facilitate collaboration with the target population
• Nongovernmental organizations • Community-based organizations
• Ensure investigators have the necessary approvals from the national, provincial and local • Ministry of Health government • National HIV/AIDS programme or committee • Provide legal and social context • Provide input on survey design • Coordinate with existing government services for target populations • Ensure that different government entities (including law enforcement, clinics and laboratories) are informed of the survey, as appropriate, and do not obstruct survey implementation • Integrate surveillance activities into the national monitoring and evaluation plan • Minimize risk that efforts are duplicated by different organizations • Disseminate results • HIV service providers and policymakers • Provide information on the needs of target populations • Ensure that survey results provide actionable information on a timely basis • Guide strategic planning and advocacy • Facilitate biological specimen collection, cold chain and storage • Provide funding; formulate information needs • Provide guidance on sampling strategies • Support survey implementation and data analysis • Draw on experience implementing similar surveys in the past • Provide information on the needs of target populations • Facilitate collaboration with the target population • Improve efficiency by integrating STI and HIV surveillance activities and survey findings • Ensure appropriate biomarkers are collected • Depending on country or local context, ensure the safety and security of surveyors and participants, • For incarcerated populations, ensure proper ethical protections and procedures are in place
• Donors • Academic institutions • Research organizations • Staff engaged in surveillance for sexually transmitted infections (STIs) • Police • Ministry of Security • Correctional officers (jails/prisons) Source: UNAIDS/WHO 2011 (1)
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Stakeholders
19
A-1.3 Stakeholder engagement process It may be helpful to hold regular consultations with stakeholders throughout all stages of the BBS, to ensure that they contribute ideas and exchange experiences. This can be achieved by forming a working group of stakeholders that meets regularly. The group can convene to review key areas such as the scope of the
survey, methods for implementation, and the plan for disseminating and employing the resulting data (2). These meetings may result in changes to questionnaires, testing approaches or methods that are best explored during the formative assessment or before the BBS is implemented. In some situations, it may be necessary to coordinate between stakeholder groups with different areas of interest, such as budget and technical stakeholder groups.
A-1.4 References 1 UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines on surveillance among populations most at risk for HIV. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2011 (http://www.unaids.org/en/media/unaids/contentassets/restore/20110518_Surveillance_among_ most_at_risk.pdf, accessed 26 June 2016). Amon J, Brown T, Hogle J, MacNeil J, Magnani R, Mills S et al. Behavioral surveillance surveys BSS: guidelines for repeated behavioral surveys in populations at risk of HIV. Arlington: Family Health International (FHI); 2000 (http:// www.who.int/hiv/strategic/en/bss_fhi2000.pdf, accessed 3 August 2016).
2
20
1. Stakeholders
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms
2. Planning considerations before starting a biobehavioural survey
Before committing to conducting a biobehavioural survey (BBS), investigators should establish what is already known about the HIV epidemic in the country and within the target population, and assess whether a BBS is warranted. This initial data review should serve to quickly gather and synthesize preliminary information justifying the BBS; it need not be extensive. This chapter describes the process for conducting this initial review.
Risk: In ethical terms, the exposure to injury, loss or harm, expressed in terms of the probability and magnitude of that harm. Risks to subjects must be minimized and must be reasonable in relation to anticipated benefits to subjects and the importance of the expected knowledge. Time-location Recruits individuals (target population sampling members) from specific locations (TLS): (venues) during specific time periods; thus, TLS surveys represent only those who frequent specific venues. Vulnerability: The (relative) powerlessness to resist or survive the effects of an adverse environment.
A-2.2 Consider epidemiological factors when choosing target population for BBS When considering a BBS with a target population, investigators should first consider the role that the group plays in the local HIV epidemic. This may be done by examining risk factors, vulnerability, population size and HIV prevalence, as discussed below.
A-2.2.1 Risk factors Most BBS are conducted among populations who are defined by their risk factors and behaviours (e.g. commercial sex, anal sex or injecting drug use), or because investigators suspect that the population exhibits risk behaviours more frequently than the general population. For target populations other than key populations, it is important to assess what is known about the behaviours of those populations before committing to a BBS of them. For example, where data on clients of SW are not available, truck drivers have been studied as a proxy, although not all truck drivers may be clients of SW. Although not representative of all clients, surveys of truck drivers may shed light on HIV transmission dynamics along truck routes. Before conducting a survey on truck drivers, it is important to gather available information on the drivers and how their HIV-related risk behaviours resemble or differ from those of other populations.
A-2.1 Reflect on the HIV epidemic Before conducting a BBS of key populations – sex workers (SW), men who have sex with men (MSM), transgender women (TG) and people who inject drugs (PWID) – it is best to conduct a review, to ensure that the BBS will not duplicate efforts and to confirm that it is feasible. A lack of data on other populations does not imply that a BBS should be conducted; other factors should be taken into account, including risk behaviour, ability to reach the sample size and HIV prevalence. The methods presented below are not intended as a replacement for discussions with stakeholders. Rather, they are a means for producing an evidence base to guide decisions about conducting a BBS. This evidence should be reviewed with all relevant stakeholders, including members of the proposed target population, before making a final decision about the BBS.
A-2.2.2 Vulnerability Vulnerability refers to factors typically outside an individual’s control that indirectly raise an individual’s risk for HIV infection – including unequal opportunities, social exclusion, unemployment or precarious employment (1). Other populations are deemed vulnerable due to
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
2. Planning considerations before starting a biobehavioural survey
21
poor access to prevention, care and treatment services. A target population may be of interest because of its vulnerabilities, although these may only be apparent in the context of other groups. For example, migrant workers can experience social exclusion, arrest and unequal employment opportunities. In contrast, there is little reason to think that teachers face vulnerabilities that make them more susceptible to HIV; in some cases, teaching may even represent a position of power.
testing data can provide valuable information, especially if the testing facility serves specific populations. Information should also be sought from stakeholders including officials from the ministry of health, donors, health-care workers, laboratory officials, NGOs and community-based organizations that work with the population, other groups that interact regularly with them and, most importantly, members of the proposed target population. For example, brothel managers may know whether truck drivers access SW. Health-care workers may be able to describe their clientele and indicate if the proposed target population has a high burden of HIV. Once sufficient information has been gathered, investigators can synthesize it and share the findings with other stakeholders. The group can weigh the factors described above and determine the value of conducting a BBS with the proposed target population.
A-2.2.3 Population size The target population needs to be sufficiently large to ensure an adequate sample size can be reached. As a general rule, a BBS intended to be representative of a target population must have a sample of at least 500 respondents. Thus, it is not possible to conduct a BBS in a small community. A town of 5000 people is unlikely to have 500 MSM; even if it did and the survey team could identify all of them, it is unlikely they would all be eligible for the survey or willing to participate.
A-2.2.4 HIV prevalence HIV prevalence indicates the proportion of individuals in a population who have HIV at a given time (1). Generally speaking, BBS should be prioritized for populations with the highest HIV prevalence and impact on the overall epidemic. HIV prevalence is generally higher among SW, MSM, TG and PWID in mixed and concentrated epidemics alike (2-7). However, BBS on other populations may also be warranted if the data review suggests that they also have an elevated prevalence of HIV.
A-2.3.1 Consider security concerns for the target population and the survey team Where it is determined that a BBS with a given population will be valuable, investigators must then consider the security dimensions of survey implementation; for example, the survey may draw undue attention to the entire target population. Investigators should weigh the potential for risk to participants and the larger target population as a whole against the benefits of the survey,2 and the safety of the survey team and participants should be a top priority in planning and implementation. If the threat of danger to the survey team and participants is agreed to be too high after consulting with local experts and community-based organizations, then it may be necessary to cancel the survey. If the survey is conducted, then safety procedures must be in place to reduce the chances of placing staff and participants in danger, to reduce the harm that could be done to staff and participants, and to respond quickly and appropriately in case a safety issue does occur. These procedures should include plans for dealing with hostile visitors or participants, reporting incidents through chain of command, conducting a pre-survey checklist for safety, emergency security actions and data security.
A-2.3 Conduct an initial review of information about proposed target population Investigators can begin by reviewing published and unpublished information about the target population. Report summaries and journal abstracts can be accessed free of charge online.1 For countries and populations where reports do not exist, literature from neighbouring countries or similar populations may provide useful insights. Investigators can also review service data; such data are especially important if there are few or no reports or journal articles about the local target population. HIV
A-2.3.2 Develop a survey protocol Next, the survey protocol and consent forms must be developed. A protocol specifies the survey objectives, defines the population that will be surveyed and the
1 2
For example, see www.unaids.org and www.pubmed.gov.
Information on human rights is available at: http://www.who.int/gender-equity-rights/en/, and information on health and human rights is available at http:// www.who.int/hhr/HHR%20linkages.pdf?ua=1).
22
2. Planning considerations before starting a biobehavioural survey
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
geographical area in which it will be conducted, and describes all the procedures involved (including sampling, obtaining consent, conducting interviews and providing counselling, collecting and testing biological specimen, and managing and analysing data) (8). At this time, to ensure that the survey objectives are satisfied when the data are collected and later analysed, investigators should develop comprehensive plans for data management (Chapter A-13) and data analysis (Chapter C-1) along with the survey protocol. Guidance and a checklist for developing a protocol are included (Appendix I-1). When developing the protocol it is important to consider the following: • what questions about the target populations need to be answered? • what information is needed to address these questions? • should the size of the target population be estimated? • should a supplemental study be conducted (e.g. a qualitative, cohort or partner study)? • what biomarkers should be collected? • how will the data be used? Suggested key objectives for biobehavioral surveys include: • to estimate the proportion of suppressed HIV viral load (among people living with HIV) • to estimate the prevalence of HIV infection • to estimate access to and uptake of HIV prevention, care and treatment services • to examine correlates of HIV infection • to estimate the population size
A-2.5 Ensure funding is sufficient to meet objectives The scope of a BBS is dependent on the financial resources available. Core budget components usually include: • administrative costs, including overhead and management; • staff costs, including investigators, data collectors, administrative support team and drivers; • laboratory consumables and equipment costs, including gloves, specimen collection instruments and assays; • travel costs, including transportation to the survey site if that is not in the same city as the investigators, or international technical advisors or investigators; and • documentation and dissemination-related expenses, including report writing, editing, layout, and national and sub-national disseminations. When estimating a BBS budget, important considerations include the target sample size, duration of survey, number of survey staff and biomarkers on which data will be collected. Investigators should estimate the number of HIV-infected participants and the follow-up tests that will be provided for each (e.g. CD4 and viral load). Costs will also vary by survey design; that is, conventional cluster sampling (CCS), respondent-driven sampling (RDS) and time-location sampling (TLS). A budget template to facilitate resource planning and alignment with the protocol can be found in Appendix I-2.
A-2.4 Ensure adherence to ethical standards Participants who are afraid to share information about risk behaviours or HIV/sexually transmitted infection (STI) status may drop out of the survey or falsify their responses. To successfully survey a marginalized population, these biases must be minimized by obtaining fully informed consent from participants and ensuring their absolute confidentiality (9). Therefore, all survey protocols should be assessed by ethical review boards or human subjects oversight committees to ensure that they adhere to national and international laws and ethical standards (10). More information on ethical considerations of survey conduct and the human subjects review process can be found in Chapter A-3.
A-2.6 Consider timing and intervals between BBS rounds Ideally, investigators should conduct BBS at intervals in order to capture changes in risk behaviours and the HIV/STI epidemic over time. The Joint United Nations Programme on HIV/AIDS (UNAIDS)/WHO Guidelines for second generation HIV surveillance recommends conducting a BBS every 1–3 years in key populations and every 3–5 years in other target populations, as warranted (9). One year is not typically long enough for changes in HIV/STI prevalence to be detected. Depending on survey objectives, local resources and context, investigators are advised to consider waiting 18–24 months in order to yield more practical results.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
2. Planning considerations before starting a biobehavioural survey
23
Regardless of the time interval between BBS rounds, the benefits of regular data collection include: • information for continuously reevaluating service coverage and needs; • comparison estimates (or trend data, if three or more rounds of BBS have been implemented) for examining the effects of interventions over time; and • the development of survey implementation and management capacity among national and local institutions. Trend data from repeated BBS yield useful information for monitoring the epidemic. Although methodological consistency is important to produce these data, investigators should make advisable changes to variables, assays or
methods where necessary rather than seeking to maintain the same instruments over time. It is possible to make such changes to improve the relevance or quality of survey results and still preserve trend analyses. Circumstances may exist where BBS or subsequent rounds of surveying are not advised. For example, national priorities may shift away from key populations, limited financial or human resources may be devoted to other public health priorities, or the HIV/STI epidemic or patterns of risk behaviours may change. The decision to stop conducting BBS of a population must be considered in the context of the local epidemiological, political and financial environment. In circumstances where the HIV prevalence of the target population is lower than that of the general population, or where funding is no longer available, it may not be appropriate or feasible to continue BBS.
A-2.7 References 1 UNAIDS. UNAIDS terminology guidelines. Joint United Nations Programme on HIV/AIDS (UNAIDS); 2011 (http:// files.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2011/JC2118_terminologyguidelines_en.pdf, accessed 4 August 2016). Baral S, Beyrer C, Muessig K, Poteat T, Wirtz AL, Decker MR et al. Burden of HIV among female sex workers in lowincome and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12(7):538–549. Baral S, Sifakis F, Cleghorn F, Beyrer C. Elevated risk for HIV infection among men who have sex with men in lowand middle-income countries 2000–2006: a systematic review. PLoS Med. 2007;4(12):e339 (http://www.ncbi.nlm. nih.gov/pubmed/18052602, accessed 4 August 2016). Beyrer C, Baral SD, van Griensven F, Goodreau SM, Chariyalertsak S, Wirtz AL et al. Global epidemiology of HIV infection in men who have sex with men. Lancet. 2012;380(9839):367–377. Mathers BM, Degenhardt L, Phillips B, Wiessing L, Hickman M, Strathdee SA et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733–1745. UNAIDS. Global report: UNAIDS report on the global AIDS epidemic 2012. Joint United Nations Programme on HIV/AIDS (UNAIDS); 2012 (http://www.unaids.org/sites/default/files/media_asset/20121120_UNAIDS_Global_ Report_2012_with_annexes_en_1.pdf, accessed 4 August 2016). WHO. Prevention and treatment of HIV and other sexually transmitted infections for sex workers in low-and middle-income countries: recommendations for a public health approach. Geneva, World Health Organization (WHO). 2012 (http://apps.who.int/iris/bitstream/10665/77745/1/9789241504744_eng.pdf, accessed 25 August 2016). CDC. Developing a protocol. Atlanta, GA: Centers for Disease Control and Prevention (CDC); 2013 (http://www.cdc. gov/globalhealth/healthprotection/fetp/training_modules/13/developing-protocols_pw_final_09262013.pdf, accessed 1 November 2016).
2 3
4 5 6
7
8
9 UNAIDS/WHO Working Group on Global HIV/AIDS and STI surveillance. Guidelines for second generation HIV surveillance: the next decade. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/WHO Working Group on Global HIV/AIDS and STI surveillance; 2000 (http://www.who.int/hiv/pub/surveillance/pub3/en/index. html, accessed 4 August 2016). 10 WHO/UNAIDS. Guiding principles on ethical issues in HIV surveillance. Geneva: World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS); 2013 (http://www.who.int/hiv/pub/ surveillance/2013package/module2/en/, accessed 3 August 2016). 2. Planning considerations before starting a biobehavioural survey B. Survey implementation & quality assurance
24
A. Survey preparation
C. Data analysis and use
Section
A Key terms
3. Ethical consideration in planning and conducting surveys
This chapter describes the ethical considerations of survey conduct and the human subjects review process. As with any research on people (also known as human subjects), biobehavioural surveys (BBS) have the potential to harm the individuals involved. BBS often involve members of marginalized, criminalized or otherwise vulnerable populations, and may collect sensitive data about stigmatized issues and behaviours. Such surveys must be designed and conducted ethically to ensure that participants can freely give informed consent, that their privacy and confidentiality are protected, and that the survey’s benefits outweigh the potential harm to participants. The international community has developed ethical principles and guidance for the protection of human subjects that are applicable to both experimental research and surveys. Investigators and staff, who take responsibility for the well-being and safety of survey participants, must follow these ethical principles throughout the conducting of a survey.
Anonymous: Describes something having no known name or identity. Achieved by removing all personal identifying information from an interview record or biological specimen. Assent: An agreement by persons who cannot legally agree (consent) to survey participation (e.g. minors). Usually, consent by parents or caregivers is necessary in addition to assent by a minor potential survey participant. Autonomous: Freedom from external control or influence; independence. Beneficence: The effort to secure the well-being of survey participants, both by avoiding or minimizing harm and by maximizing possible benefits. Confidential: The expectation or promise that information will be protected from disclosure to others. Consent: Permission given by an individual to proceed with a specific test or procedure, with an understanding of the risks, benefits, limitations and potential implications of the procedure itself and its results. Fingerprint Fingerprint scanners, together with the codes: appropriate software, convert a finger print image into an alphanumeric code without storing the actual image. If such codes cannot be used to reconstruct the fingerprint images they may be regarded as anonymous ID numbers. These codes are anonymous, but can nevertheless be used to link different visits of the same participants. Justice: The fair selection of research participants and the fair distribution of risks and benefits for the study participants.
A-3.1 International guidelines for human subjects research International guidelines for human subjects research have been published over the past century in a handful of key documents. As the direct result of reported abuses of human subjects in biomedical experiments during the Second World War, the international community drafted the Nuremburg code (1947) (1) – a set of 10 requirements for the legitimate conduct of experimental research involving human subjects. The Declaration of Helsinki (published in 1964, last updated in 2013) (2) was primarily developed to guide physicians, but is considered by many to be the first global standard for ethical principles for human subjects research. The Belmont Report, Ethical principles and guidelines for the protection of human subjects of research (1979) (3), published by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, lays out three fundamental principles for the ethical conduct of human subjects research: respect for persons, beneficence and justice. The International Ethical guidelines for biomedical research involving human subjects (published in 1982, last updated in 2002) (4) describes how ethical principles can be applied in practice, particularly in developing countries. UNAIDS/ WHO also published the Guiding principles on ethical issues in HIV surveillance in 2013 (5).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Ethical consideration in planning and conducting surveys
25
A-3.2 Fundamental ethical principles Respect for persons “Respect for persons” is the principle that all individuals should be treated as autonomous (i.e. capable of making informed decisions) and their choices should be respected. For potential participants to make informed decisions to participate in a survey, survey staff must adequately inform them about the survey’s objectives, its possible risks and benefits, and their right to refuse to be involved in or to withdraw from the survey at any time without penalty. Survey participants should decide whether to participate voluntarily, without undue influence or coercion or consequence, and with full knowledge of the procedures involved in participation. Vulnerable persons may be unable to make informed decisions or be at a higher risk of being coerced. Studies involving such individuals must offer extra protections or precautions to ensure that their decision to participate is made voluntarily. Examples of vulnerable persons include individuals whose behaviours are stigmatized or criminalized – for example, people living with HIV, men who have sex with men (MSM), sex workers (SW), people who inject drugs (PWID) and transgender persons (TG) – and minors, those with mental or behavioural disabilities, subordinate members of hierarchal groups, elderly persons, pregnant women, incarcerated adults, ethnic minorities and displaced persons. Beneficence “Beneficence” is the principle that the goal of the research should be the welfare of the subject. The researcher should strive to “do no harm” and to maximize possible benefits while minimizing possible harms. Ethically, surveys can only be conducted if the public health benefit is likely to outweigh the personal risks of the participants involved. Investigators are responsible for participants’ physical, mental and social well-being throughout their participation in the survey. They should be open with potential participants about the risks, discomforts and burdens that may be involved in the survey, and monitor and seek to minimize these factors throughout the process. Justice “Justice” is the principle that the risks and benefits of the research should be shared by the individual participant and the larger community. Investigators must ensure that the selection of survey participants is equitable (fair) and nonexploitive, meaning that participants are not selected because of their easy availability, compromised position (vulnerability) or manipulability (i.e. can be tricked into survey participation).
A-3.3 Research ethics committees Research ethics committees (RECs), sometimes known as institutional review boards (IRBs), are responsible for reviewing research and surveys, and ensuring that they are designed and conducted in accordance with the three fundamental ethical principles. Investigators should always submit their survey protocol to an REC for review and approval before survey implementation begins. During protocol review, RECs should ensure that: • the survey design and conduct support the safety and well-being of participants; • the recruitment strategy is appropriate and equitable; • informed consent processes meet ethical standards and are sufficiently described; • the possible benefits to the target population outweigh the risks; • risks to participants are monitored and minimized; and • privacy and confidentiality are protected.
A-3.4 Sampling of participants Investigators should sample participants in a way that both protects participants and respects ethical principles. To help avoid unintentionally coercing individuals into joining the survey, investigators should assure potential participants that there is no penalty for declining participation. Investigators should also recruit participants in a way that separates or minimizes the link between survey participation and any services the potential participant is eligible to receive. Reimbursement for time should be reasonable (i.e. not too high). Where possible, survey-related benefits (e.g. free serological testing with results provided) should be offered to potential participants even if they decline to participate.
A-3.5 Obtaining voluntary informed consent “Voluntary informed consent” occurs when an autonomous person with a good understanding of the survey actively decides to participate in the survey without being influenced or pressured. Consent is a process that starts before survey participation and continues until participation is complete. It is the investigator’s responsibility to ensure that the potential participant receives enough information to give voluntary informed consent. Information should be provided in a manner that is appropriate for the culture, language ability, education level and general contextual environment (see Appendix I-3).
26
3. Ethical consideration in planning and conducting surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
The informed consent form serves as an agreement between the investigator and the survey participant. It represents a pledge by the investigator as to what the survey participant can expect in terms of survey process, benefits, rights, risks, costs and potential harm. It must be written in simple, clear language that is native to or well understood by the participant, and be at a reading level that is appropriate for the population being surveyed. It should use active voice, short words and short sentences, and should not use coercive language. During the informed consent process, potential participants must be given the following information: • a description of the survey and the organization conducting it, the survey objectives, the populations being invited to participate and the number of expected participants; • a clear indication that participation is voluntary and that participants have the right to withdraw from the survey at any time without penalty or loss of benefits; • a description of the participant’s role in the survey, including duration and frequency of participation, the procedures that the participant will undergo, the general nature of the interview questions, and the procedures for specimen testing; • a description of the possible risks, burdens and discomforts of participation; • a description of the direct benefits to the participant, or the indirect benefits to the community or society that are expected to come from the survey; • measures made to protect the privacy and confidentiality of the participant and any information that the person provides; • documentation that the protocol was reviewed and approved by an REC; and • contact information and procedures for contacting the overseeing REC and principal investigator in case of further questions. Informed consent can be given orally or can be written (e.g. initials, signature, mark or thumbprint). Particularly in surveys of stigmatized populations, investigators and participants may prefer to use oral informed consent because it does not produce a written record that could compromise participant anonymity. Some countries, however, require written informed consent. Anonymity and confidentiality can still be preserved in these cases. For example, participants can be asked to mark the informed consent form with an “X” to avoid providing any personal identifying information. If needed, the participant’s unique survey identification number (ID) can be used to link the informed consent form with datacollection forms and biological specimens.
Voluntary agreement component of consent form I understand what it means to join the survey. I understand my rights and risks. I had time to ask questions about the survey. I understand that I can join the survey at my free will. I understand that I can leave the survey at any time. Have all your questions been answered?
□ □ □ □ □
YES
□ □ □ □ □
NO
Do you agree to do an interview? YES NO
Do you agree to the blood draw and testing? YES NO
Do you agree to the urine sample and testing? YES NO
Do you agree to the swab sample and testing? YES NO
Do you agree that we may keep the leftover blood/ urine/swab for future testing?
□
YES
□
NO
If a national government requires that participants sign their name on the informed consent form, procedures must be put in place to ensure confidentiality of the information. This may include not linking informed consent forms to participant data and specimens, or using unique survey codes to link the forms, data and specimens; storing the informed consent forms securely and separately from data and biological specimens collected during the survey, as well as after the survey is complete; and minimizing access to the informed consent forms. Participants must give informed consent for each element of the survey, including the interview, each specimen collected, and any storage or future testing that might be conducted on the stored specimens. Participants may agree to some procedures but refuse others. A single form can be used to obtain informed consent for all elements as long as the procedures to which a participant is consenting are clear (see the list of possible procedures in the box). Separate consent should be obtained for specimen storage and future unspecified testing. Since BBS are generally conducted anonymously, it is not typically possible to return results of future testing to participants. The consent form should indicate whether this is the case, and whether the participants should expect to benefit in some way from this testing.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Ethical consideration in planning and conducting surveys
27
Only adults are able to provide informed consent. The age of majority that defines adulthood should be considered during survey planning because it differs by country. Some individuals who have reached the age of majority are unable to provide consent; for example, adults with cognitive (mental) disabilities or people who are under the influence of drugs or alcohol. For these individuals, informed consent must be given by their legally authorized representative (e.g. parent or guardian), in accordance with national laws. And individuals unable to provide informed consent must still give assent, indicating their agreement to participate. In cases where survey participants are too young to give their own informed consent but obtaining informed consent from their parent is inappropriate or could cause harm to the participant, investigators can seek a waiver for parental informed consent. An applicable situation would be a survey involving drug users aged under 18 years who are living and working away from their parents. The request for waiver must be submitted to, and reviewed and approved by, the governing REC or RECs for each protocol. Although it is often advisable to provide participants with a copy of their consent form, investigators should use caution when doing so in surveys of stigmatized populations. Possession of the form might itself pose a threat to the participant’s safety by revealing the person’s risk group identity (e.g. MSM or PWID). Instead, investigators can provide participants with minimal information, such as the contact information of the investigators and appropriate RECs.
Site selection and recruitment It is the investigator’s responsibility to ensure the security of survey participants, particularly when the survey is conducted among vulnerable or marginalized populations. The investigator should choose a site and opening hours that are convenient for participants, and a site where they will feel safe from harassment by the general public or law enforcement. As far as possible, participants should not be visible to outsiders when participating in the survey. Responses to interview questions should not be audible to others at the survey site. The site itself and its location should be chosen to be discreet and to not draw unwanted attention, although whether this should be a busy location or a quiet place will depend on the context (e.g. the location could be near a frequented space such as a church, mosque or market, or on a quiet residential street with few pedestrians). When selecting survey sites, investigators should consult with the target population and with individuals who provide services to them. For surveys involving populations that engage in illegal activities, investigators should seek guidance from the target population about whether to sensitize local law enforcement forces, to help minimize threats of harassment and raids during survey conduct. Nature of data collected To further respect the privacy of individuals, the survey should only seek to obtain and record information that is needed and will be used. For example, information about a participant’s personal hobbies or home address are unrelated to the individual’s HIV/sexually transmitted infection (STI) risk behaviours and should not be included in a survey. Confidentiality “Confidentiality” refers to the protection of the private information provided by the participant. During surveys, participants provide information with the expectation that it will only be disclosed according to processes outlined during informed consent. Survey staff should sign data confidentiality agreements that explain the procedures used to ensure data confidentiality. Data dissemination and use When disseminating results, care must be taken to ensure that the information does not reveal the identity of any individual participant or group of participants. For example, if the survey involved population mapping, the exact locations of hotspots should not be published because this might jeopardize participants’ anonymity. Instead, these data could be aggregated at a neighbourhood or higher level, depending on the population or hotspot numbers.
A-3.6 Ensuring privacy and confidentiality Protection of a person’s privacy and data confidentiality are essential for surveys. It is the investigator’s responsibility to ensure that both are protected during and after the survey. Privacy refers to the ability of an individual or group to control access to themselves or to information about themselves. The nature of privacy can vary, depending on the individual and the context in which the information is provided. For example, protecting a child’s privacy may require the presence of the child’s parent or legal guardian; whereas protecting a young person’s privacy may be better achieved by withholding notice of survey participation from the person’s parent or legal guardian. Aspects of privacy related to survey conduct include site selection and recruitment processes, the nature of the data collected, and data dissemination and use.
28
3. Ethical consideration in planning and conducting surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-3.7 Personal identifying information “Personal identifying information” is information that can be used on its own or in combination with other information to reveal a person’s identity. Among its many forms are names, phone numbers, physical addresses, email addresses, social security or other identifiable numbers, fingerprint or other images, and voice recordings. Surveys should generally be conducted anonymously, meaning that personal identifiers should not be recorded on any forms, including informed consent forms, datacollection forms, biological specimens or referral forms. There should be no way to identify a participant from the survey data. Unique survey codes should be used in all records, and constructed without the use of personal identifying information such as complete date of birth.
reimbursement is generally given both for survey participation and for the efforts of participants to recruit others for the survey.
A-3.10 Return of test results In general, test results that may be of use for the participant’s health care should be returned. Procedures for returning clinically meaningful test results to participants must be clearly described during informed consent. Confidentiality is best upheld using anonymous, linked specimen testing. For participants who wish to know the results of their HIV or STI tests, processes must be in place to allow easy return of test results and appropriate post-test counselling. Where rapid testing is feasible, results can be returned during the initial survey contact. It is important to follow local testing algorithms for confirmation of test results, where needed, to comply with national policies. Where rapid testing is not available, it may be possible to have participants return to the survey site after a specified time frame in order to obtain their test results. In this case, it is important to implement measures to verify a participant’s identity at the second visit. In some settings, it may be possible to return test results to a clinic or health facility of the participant’s choosing.
A-3.8 Data security Data collected during the survey should be managed in a way that protects the privacy of participants and the confidentiality of the information they provided. Investigators should develop clear data security and datamanagement in standard operating procedures (SOP), and should ensure that the SOP are understood by all involved in the survey, and are applied throughout data collection, transfer and storage. Access to collected data should be limited to staff on a need-to-know basis. Paperbased data forms should be stored in a locked room or file cabinet, and electronic data should be stored on password-protected computers or in password-protected files. Any data shared outside the survey team – such as with a third party conducting secondary analysis – should be anonymized.
A-3.11 Treatment and referral Participants with test results indicating infection (e.g. HIVseropositive) should be provided with information about appropriate treatment and care services. For conditions that can be treated safely by survey staff, same-day treatment should be provided (e.g. oral treatment with antibiotics for select STIs), in line with national policies and IRB approval. For infections that require complex or long-term treatment (e.g. HIV or viral hepatitis), efforts need to be made for a facilitated process of effective linkage to nearby health facilities, ideally to those with experience working with the survey population. Investigators should develop agreements with these service providers to ensure that referred survey participants are accepted. Investigators must ensure that the act of referral does not disclose information about the participant’s risk group (e.g. MSM). For example, whereas it may be common to give participants a referral sheet with a list of clinics that have experience working with their population, such documents can indicate that a person participated in the survey but should not indicate which populations the survey targeted.
A-3.9 Reimbursement to participants Participants may be given reimbursement for travel costs, time spent, lost wages or other expenses incurred for participating in the survey. Additionally, they may be given free medical care. The reimbursement and medical care offered should not be so substantial that persons feel they need to participate. Investigators should be particularly sensitive to this situation in resource-limited settings, where even a small amount of reimbursement can have a major impact on a person’s decision to participate in a survey. The type and quantity of reimbursement should be decided in consultation with the survey population and other stakeholders, and must be justified and approved by the governing RECs. In respondent-driven sampling (RDS) surveys,
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Ethical consideration in planning and conducting surveys
29
A-3.12 Adverse events, protocol deviations and protocol amendments Adverse events An adverse event is a negative event that happens to a participant during the conduct of the survey, regardless of whether it is related to survey participation. Adverse events may include unwarranted disclosure of a survey participant’s identity, arrest or violence related to survey participation, errors in post-test counselling (communication of a wrong result), or allergic reaction to an antibiotic administered for a diagnosed STI. “Serious adverse events” are events that are lifethreatening, require hospitalization, or result in death or persistent or significant disability. “Unexpected adverse events” are those that were not cited by the investigator in the protocol as a possible consequence of participation. Reporting of adverse events Investigators should promptly report unexpected adverse events to the governing REC according to the REC’s guidelines. Often, the REC will require an initial informal notification via email within a few days. Subsequently, the REC may expect a formal incident report within a given time, often 1 week. The investigator should develop or use the REC’s adverse event forms to document and report such events, including a description of the adverse event, the date it occurred, responses taken to the event, and the date it was reported to the governing REC. Protocol deviations Protocol deviations are unintended survey procedures or events that were not described in, or that differ from, the approved protocol (e.g. redefinition of the sampling domain, or survey participation by an individual that did not meet eligibility criteria). Protocol deviations may or may not put participants at risk. They may even be necessary in emergency situations. However, all protocol deviations should be reported to the overseeing REC according to its guidelines, and may ultimately require that the protocol be amended. Protocol amendments All approved survey materials must be used as approved by the governing REC. This applies not only to the protocol but also to informed consent documents, data-collection tools and recruitment materials. If investigators wish to alter some of the procedures described in the approved protocol, a protocol amendment must be prepared and submitted to the REC for review and approval. The proposed changes can take effect only after approval is given.
A-3.13 Research ethics training Investigators and survey staff interacting with participants should receive training in human subjects research ethics. The curricula for such training generally cover ethical principles of research, REC roles and responsibilities, participant privacy and confidentiality, voluntary informed consent procedures, and reporting of adverse events and deviations from the protocol. Examples of relevant courses include the Collaborative Institutional Training Initiative course (United States of America, USA), the National Institutes of Health’s Protecting Human Research Participants course (USA), Training and Resources in Research Ethics Evaluation (Switzerland, available in English, French, German and Portuguese), the Epigeum course (United Kingdom), the Fundacion Bioetica course (Spain) and FHI 360’s Research Ethics Training Curriculum (USA).
A-3.14 Special considerations for incarcerated populations Incarcerated populations are especially vulnerable; hence, special considerations must be taken into account when surveying this population. Incarcerated populations tend to be of lower socioeconomic status, have lower education attainment and reading level, and have higher rates of infectious disease, chronic illness and mental illness than the general population. They may also have higher rates of substance and alcohol use, and be more prone to injury, violence and abuse. Within a prison setting, inmates generally have poorer access to health and legal services, and little or no personal privacy. Two major considerations for surveys among incarcerated populations are the ability of prisoners to provide informed consent and their right to privacy, confidentiality and autonomy. Right to privacy, confidentiality and autonomy Maintaining privacy is an enormous challenge in prison settings, because inmate movements are closely monitored by both prison officials and other inmates. Providing health information and services to a prisoner in a confidential setting may be equally challenging. In addition, prisoners may fear repercussions from prison officials or other inmates, including threats, physical or sexual abuse, punishment or lengthened prison terms, as a result of their decision to participate or not. Special care must be taken in designing the survey’s methods for recruitment, testing and provision of test results, and follow-up care (treatment and referrals) to ensure that potential participants’ right to privacy, confidentiality and autonomy is upheld throughout the process. It is the investigator’s responsibility to consult with local experts to ensure that the methods used are tailored and acceptable to the local context.
30
3. Ethical consideration in planning and conducting surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Prisoners may also feel coerced into participating by prison authorities or compelled to participate in order to access the health services offered as part of the survey. This is particularly relevant for surveys conducting HIV and STI testing in prisons, where routine access to such testing is often limited or nonexistent. Where possible, investigators should seek to minimize the discrepancy in health services offered to participants and nonparticipants. For example, confidential HIV and STI testing services could be made available to all inmates during the time of the survey, but only data from those selected to participate would be recorded and reported by the survey team. Ability to provide informed consent Incarcerated populations may read at lower levels. Therefore, investigators must adjust the language used in informed consent forms to ensure that participants are able to fully understand the survey’s purpose, risks and benefits, and procedures, and the steps taken to ensure privacy and confidentiality.
A-3.15 Web-based surveys Investigators increasingly use the internet to conduct surveys, and current guidelines may not address all of the issues raised by the unique characteristics of an internet survey. In addition to the ethical considerations applicable to conventional surveys, investigators must consider the following aspects: • IRBs may regard internet protocol (IP) addresses as personal identifiers. Surveys that strive to maintain anonymity need to take measures to prevent the unintended collection of IP addresses, or else take measures to protect the confidentiality of the individuals associated with these IP addresses. • internet-based surveys may face particular challenges to assess eligibility criteria; for example, it may be difficult to assess age, and this could lead to the enrolment of minors in a survey designed to be limited to adults.
A-3.16 References 1 2 The Nuremberg Code. 1947 (https://history.nih.gov/research/downloads/nuremberg. pdf, accessed 25 August 2016). WMA. Declaration of Helsinki – ethical principles for medical research involving human subjects (2013 update). World Medical Association (WMA); 1964 (https://www.wma.net/policies-post/wma-declaration-of-helsinkiethical-principles-for-medical-research-involving-human-subjects/ accessed 22 May 2017. The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Belmont Report: ethical principles and guidelines for the protection of human subjects of research 1979 (http:// www.hhs.gov/ohrp/regulations-and-policy/belmont-report/index.html, accessed 25 August 2016). CIOMS/WHO. International ethical guidelines for biomedical research involving human subjects. Prepared by the Council for International Organizations of Medical Sciences (CIOMS) in collaboration with World Health Organization. Geneva, Switzerland: CIOMS, Council for International Organizations of Medical Sciences (CIOMS)/ World Health Organization (WHO); 2002 (http://cioms.ch/publications/layout_guide2002.pdf, accessed 25 August 2016). WHO/UNAIDS. Guiding principles on ethical issues in HIV surveillance. Geneva: World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS); 2013 (http://www.who.int/hiv/pub/ surveillance/2013package/module2/en/, accessed 3 August 2016).
3
4
5
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Ethical consideration in planning and conducting surveys
31
Section
A
4. Formative assessment
This chapter describes formative assessment, one of the most important parts of the survey process. A good formative assessment will inform the design and implementation of a biobehavioural survey (BBS), including how participants are recruited and operational processes are carried out. The assessment also provides an opportunity for investigators to develop relationships with the target population, build trust and solicit their input on how the survey should be carried out. A BBS that is informed by the findings of a formative assessment is often more acceptable to the target population, more likely to succeed and less likely to result in unintentional harm to the target population than a survey that has not been subject to such an assessment.
Key terms Formative The collection of initial information about assessment: a population to inform the preparation and conduct of a survey of that population. Free listing: Asking respondents to list as many items as will fit in a particular domain (e.g. all the different drugs that are taken by local drug users). Pile sorting: Asking the respondent to group items from the same domain, often using cards; for example, sorting the names of the locally used drugs that are either stimulants or depressants. Venue: In terms of BBS, a place where the target population congregates (e.g. bars or brothels).
A-4.1 The purpose of formative assessment A formative assessment is used to collect initial information about a target population, gain that population’s support for a BBS, and inform the development and conduct of the survey itself. Formative assessment is most important in settings where little is known about the target population. However, even when there is abundant knowledge about the target population, a formative assessment can help to re-engage them, identify changes since the last survey and learn from the population’s experiences with previous surveys. In all settings, formative assessment can provide immediate feedback on service availability. Formative assessment should be conducted after the target population has been selected, and in every location and for every population in which a BBS is planned. Qualitative methods are most commonly used for formative assessment data collection, but quantitative methods may also be used. The results of the formative assessment may shape the design of the BBS, particularly in settings where the investigators have little or no pre-existing relationship with the target population. It is thus good practice to develop separate protocols for the formative assessment and the actual BBS, so that findings from the formative assessment
can be used to develop or modify the BBS protocol. For example, if the formative assessment shows that the target population is socially well connected but does not congregate at specific locations, investigators may choose to use respondent-driven sampling (RDS). Conversely, if the formative assessment shows that the population is not socially well connected but does congregate at specific locations, the investigators may choose to use timelocation sampling (TLS). Investigators should consult with members of the target population and meaningfully engage them in all stages of the BBS process, beginning with the formative assessment. It is also good to include community-based organizations – such as those that serve or are composed of people who inject drugs (PWID) or sex workers (SW) – in survey design and implementation. Often, trust that investigators build with target population members during the formative assessment will result in greater willingness on the part of the population to connect investigators with other key informants and venues where the target population
32
4. Formative assessment
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
congregates. The formative assessment is also a good opportunity to identify potential survey staff, venues for TLS, or seeds for a BBS using RDS.1
A-4.2 Formative assessment objectives This subsection describes the objectives of formative assessment, which are to: • understand the target population and context; • identify existing services and gaps; • inform survey methods (e.g. sampling strategy, questionnaire and biological specimen collection); and • engage stakeholders. These objectives are discussed below.
Characterize the target population Before conducting a BBS, investigators should be able to roughly describe the target population, including the sex, age range, nationality, marital status, ethnicity, neighbourhood of residence, employment and subgroups. For RDS surveys, this information will assist investigators in selecting diverse seeds. And for all sampling methods, this information can later be compared to the BBS sample to identify subpopulations that are under-sampled or excluded. Understand the legal and social context of the target population, and how has it changed since the last survey Investigators should be able to place the target population in their social context based, for example, on their demographic characteristics, risk behaviours and laws that affect them. For instance, investigators might observe that PWID have started using new injection methods, or that their social visibility is increasing or decreasing. Legal prohibitions on commercial sex, homosexuality, or injecting drug use can discourage participation in the survey, affect the openness with which the team conducts the survey and have ethical implications. Accordingly, the legal context must be well understood so that investigators can ensure the safety and security of their staff and participants. Identify where investigators locate members of the target population and engage them in the survey and in understanding how members of the target population interact with one another To choose the most appropriate sampling methodology, investigators must understand where and how members of the target population interact with one another, including interaction between subgroups. For example, knowing that SW are largely based in brothels is not sufficient for determining whether it will be more appropriate to recruit them through TLS. Brothel-based SW can be very different from street-based SW, and cluster sampling methods may yield a sample that does not represent both populations. Similar challenges may exist in circumstances where the target population is ethnically diverse, or includes both native and foreign members but the sampling frame captures only one of these groups. Determine whether target sample size is achievable During the formative assessment, investigators should solicit rough estimates of the size of the target population or the share of the total population they represent. If the target population is too small to obtain a representative sample, investigators may opt to conduct a nonrepresentative BBS or a qualitative survey of risk behaviours and service needs. For RDS
A-4.2.1 Understand the target population and context A BBS is more likely to succeed if investigators have a collaborative and respectful relationship with members of the target population. The relationship between investigators and the target population may take time to develop, and may be threatened if the expectations of the investigators and the target population are far apart. It is important to spend sufficient time and effort to reach a common understanding with the target population about key issues of the survey, and to maintain this collaboration throughout the survey until final results are disseminated. During the formative assessment, investigators should learn about the social and legal environment of the target population, the language and questions that are most appropriate for use in the survey questionnaire, and the health service needs of the target population. Formative assessment should answer the following questions: • who is the target population? • what is the legal and social context of the target population, and how has it changed since the last survey? • where can investigators locate members of the target population and engage them in the survey? How do members of the target population interact with one another? • was the target sample size achievable? • What are the sexual and drug-taking practices of the target population, and how do they talk about them?
1
In RDS surveys, “seeds” are the individuals who are chosen by the investigators and represent the start of the recruitment chains.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
4. Formative assessment
33
surveys, investigators should inquire about participants’ social network sizes, to determine whether they are large enough for peer-referral sampling. Methods for estimating population size for programme planning, monitoring and epidemic modelling are described in Section A-11. Understand the sexual and drug-taking practices of the target population, and how the population talks about those practices During the formative assessment, investigators should learn about the sexual and drug-taking behaviours of the target population, and how they talk about those relationships and practices. The information gained will allow investigators to adapt standard data-collection tools, such as those presented here, to the local context. BBS results are only valid if the participants accurately understand and respond to the survey questions. Participants and investigators may use terms differently. For example, participants may use the term “bisexual” to describe having had more than one sexual partner, whereas investigators may use it to describe someone who identifies as being sexually attracted to men and women. Similarly, individuals with “sugar daddies” (i.e. sexual partners who regularly give them money or gifts) may not perceive this as a “transactional” sexual relationship and may use another term for this activity, or not label the activity with any term at all. Other categorizations – for example, “regular”, “casual” and “commercial” – may mean little to them. Similarly, target populations may use specific local terminology to describe sexual or drug-related behaviours, or may redefine terms (e.g. using “shooting up” to indicate injecting drugs).
and survey methods – and mobilize and engage the target population. The formative assessment can help investigators to understand where the target population is located, how members interact, whether they are willing to recruit one another into a survey and whether there is another method for identifying population members (e.g. a list kept by service providers). This will help investigators to select the optimal sampling strategy. For more on sampling strategies, see Section A-9.1. The formative assessment can also help gauge the target population’s willingness to provide biological specimens for HIV and sexually transmitted infection (STI) testing, the preferred location and timing of the survey, characteristics of survey staff, coupon design for RDS surveys, the desired method of reimbursement for participation, and whether authorities should be involved in or informed of the survey. The assessment may also yield seeds for an RDS survey, venues for TLS, and an understanding of local terminology; for example, terminology about various sex partner types or how population members differentiate between paying and nonpaying partners. The formative assessment can also reveal barriers to survey participation and can help to improve the accuracy with which the results are interpreted. For example, whereas BBS data may suggest that older men who have sex with men (MSM) make up a minority of the target population, formative assessment results may show that older MSM are really just less willing to participate in the survey than younger MSM. Similarly, the formative assessment may help to explain why SW who operate on the street participate at a different rate than those in brothels or guesthouses. Finally, the formative assessment is an opportunity for investigators to learn the terminology that the target population uses to describe the behaviours they engage in. This will allow investigators to design a survey questionnaire that is relevant to and easily understood by the target population.
A-4.2.2 Identify existing services and gaps Formative assessment findings about service accessibility can prompt action long before final results from the BBS are available. The formative assessment can yield information about the target population’s health needs, the services provided to them, and their access to those services. Service providers can respond to gaps in service; for example, by improving access to HIV testing, condoms and lubricants. Later, final results from the BBS can be used to fine-tune or expand services. An understanding of the available services is also needed for efficient referral of participants during the survey.
A-4.2.4 Engage stakeholders Engaging a broad group of stakeholders will improve the acceptability, quality and credibility of survey results. It may also help to ensure that the survey does not duplicate research already undertaken. Also, such broad engagement will build national and local interest from all sectors of society, particularly the target population, so that all partners will be willing to act on the findings of the BBS. Investigators should seek to engage stakeholders throughout the entire survey process, from the formative assessment to the BBS, and finally to the data-use stage.
A-4.2.3 Inform survey methods (e.g. sampling strategy, questionnaire and biological specimen collection) As noted above, the formative assessment should inform the design of the BBS – specifically the sampling strategy
34
4. Formative assessment
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-4.3 Conducting a formative assessment The structure and extensiveness of the formative assessment will depend in part on how much is already known about the target population, and on the strength of the existing relationship between the investigators and the target population. The first step in the formative process is usually a review of existing publications about the target population, including those describing their sociocultural history and their epidemiological characteristics. Investigators should review peer-reviewed manuscripts and the “grey” literature (e.g. surveillance reports, programme reports and conference abstracts). Where little or no country-specific data are available, investigators may review data on target populations from the same region. Over the past decade, there has been an increasing appreciation for qualitative methods in the formative assessment phase of a BBS (1, 2). Two methods are now typically used: key-informant interviews and focus groups. Other participatory methods include free listing and pile sorting – which may help to establish appropriate language for use in survey questions, informed consent forms and transition text in the surveys – and mapping and ranking (3). More information on qualitative methods, including interview guide development and data analysis, can be found in Chapter A-12.
assessment results, to ensure that the results are interpreted accurately. Formative data may also help to improve service provision and policy-making. For example, service providers may believe that they are providing services to MSM; however, if MSM cannot name a single organization that targets them, that disconnect is important to recognize, and points to an opportunity for MSM to be further engaged in identifying the necessary services and their delivery. Investigators should start working on the BBS protocol as soon as the formative assessment has helped to identify the most appropriate recruitment method. Meanwhile, formative assessment findings can be disseminated through a report, and investigators should present them to key stakeholders.
A-4.5 Ongoing formative assessment Ideally, formative assessment is an ongoing activity that should continue during the BBS. Investigators should regularly consult the target population in order to identify issues that might affect participation or survey conduct. For example, police may have begun targeting the population with arrests, or reimbursement for participation may be too low to engage a sufficient sample size; conversely, reimbursement may be too high, encouraging ineligible individuals to fake membership in the target population in order to participate. This consultation can include conducting exit interviews of survey participants, having an anonymous suggestion box for participants to leave comments, or discussions with stakeholders, including community-based organization staff members who work with the target population. Exit interviews and the questions that will be asked should be included in the formative assessment protocol. Findings from ongoing formative assessment can be used to revise and strengthen survey procedures.
A-4.4 Optimal use of information gathered through the formative assessment Data from the formative assessment can inform all aspects of BBS protocol development. The findings of the BBS should also be reviewed in light of the formative
A-4.6 References 1 Johnston LG, Whitehead S, Simic-Lawson M, Kendall C. Formative research to optimize respondent-driven sampling surveys among hard-to-reach populations in HIV behavioral and biological surveillance: lessons learned from four case studies. AIDS Care. 2010;22(6):784–792. Simic M, Johnston LG, Platt L, Baros S, Andjelkovic V, Novotny T et al. Exploring barriers to 'respondent-driven sampling' in sex worker and drug-injecting sex worker populations in Eastern Europe. J Urban Health. 2006;83(6):i6–i15. International HIV AIDS Alliance. Tools together now! 100 participatory tools to mobilize communities for HIV/AIDS. Brighton: International HIV AIDS Alliance; 2006 (http://www.aidsalliance.org/assets/000/000/370/229-Toolstogether-now_original.pdf?1405520036, accessed 4 August 2016).
2 3
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
4. Formative assessment
35
Section
A
5. Questionnaire development
This chapter provides an overview of various methods for collecting survey data, and for developing and testing questionnaires before implementation. The questionnaire module section (Section II) of these guidelines contains standardized instruments for collecting survey data among key populations – men who have sex with men (MSM), people who inject drugs (PWID), sex workers (SW) and transgender persons (TG) – and other vulnerable populations. These instruments have been field tested and, where indicated, the modules use validated questions and scores. Investigators are strongly encouraged to use these “established” questionnaires whenever possible, for two main reasons. First, questionnaire development is a difficult process; it requires extensive preparatory work, including cognitive testing and evaluation. The instruments provided use questions, time references and skip patterns that have been tested and are known to produce high-quality data. Second, these instruments include questions that have been used throughout the world; thus, their continued use will enable researchers to compare survey results across countries. The standardized instruments provided with these guidelines were designed to be self-administered using electronic data-collection methods, which may complement or replace paper-based data collection. These instruments may be used in face-to-face or computer-assisted self-interviews, will work on multiple electronic platforms (e.g. tablets, laptops or smartphones), may be adapted to a variety of local contexts and may be modified for application using paper-based data-collection methods. When planning a survey, investigators should be familiar with the advantages and disadvantages of various datacollection methods in order to choose the method most appropriate for the circumstance.
Key terms Cognitive: Relating to thought processes and gaining knowledge (through activities such as thinking, understanding, learning and remembering) by the use of reasoning, intuition or perception. Method to investigate how well a question performs and whether respondents understand the question correctly.
Cognitive testing:
Distal With regard to HIV, a structural (not biological determinant: or behavioural) variable that affects HIV risk (e.g. stigma and discrimination, community mobilization, violence and poverty). Proximal With regard to HIV, a biological or behavioural determinant: variable that directly facilitates HIV acquisition or transmission. Recall bias: Systematic error due to differences in accuracy or completeness of remembered past events. Saliency: Degree to which a topic or event resonates with a potential survey participant. Salient: Notable or important. Syndemic: Two or more endemics or epidemics occurring simultaneously and acting synergistically.
1. decide on methods for questionnaire (interview) administration. 2. determine investigation topics (questionnaire domains). 3. develop and adapt the questionnaire. 4. translate and back-translate the questionnaire. 5. conduct cognitive testing. 6. pretest the questionnaire. 7. train interviewers. 8. pilot test the survey tool.
A-5.1 Key steps in questionnaire development The key steps in questionnaire development, described in detail below, are as follows:
A-5.1.1 Decide on methods for questionnaire (interview) administration Questionnaire delivery mode The questionnaire may be delivered by an interviewer or may be self-administered by the participants themselves.
36
5. Questionnaire development
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Some standardized questionnaires are designed for a specific delivery mode, whereas others may need to be modified. When deciding on the method of delivery, investigators should consider factors such as the length and complexity of the questionnaire, the literacy of a typical participant and the sensitivity of the questions. There are trade-offs to each of these factors. Formative assessment and, if applicable, pilot testing of different interview administration methods may help investigators to make an informed decision well in advance of survey implementation, thus allowing time for modification. Interview questions and instructions should be clear. For self-administered interviews, questions and instructions need to leave no room for misunderstanding because participants may not feel comfortable asking for clarification. Misunderstanding of the questions may result in invalid answers. Data collection mode There are two data collection modes – paper based and electronic – as discussed below.
Paper-based data collection Paper-based questionnaires may be delivered by an interviewer or may be self-administered. When using paper-based questionnaires to collect data, additional time should be allotted for training interviewers. Also, the interviewers should have previous experience collecting data from complex surveys on paper. To complete self-administered paper-based questionnaires, participants must be moderately literate and able to read, understand and follow instructions for completing the form. The number and complexity of “skip” instructions should be limited, especially for self-administered questionnaires. Data may be entered electronically either shortly after the interview or later in batches. Double data entry is strongly recommended to limit data entry errors. Electronic data collection Electronic data collection instruments can also be delivered by an interviewer or self-administered. They may be developed using a variety of existing software; for example, the Questionnaire Development System (QDS™ Nova Research, Bethesda, MD) and Open Data Kit (ODK). Table A-5.1 summarizes various systems of electronic data collection, and their advantages and disadvantages.
Table A-5.1 Systems of electronic data collection, and advantages and disadvantages System Computerassisted self-interview (CASI)
Mode of delivery Advantages • Participants • Allows participants to answer self-administer sensitive personal questions privately the questionnaire on a computer • Eliminates the need for subsequent data entry and the risk of introducing errors • Ensures questionnaire is delivered consistently – participants all experience the questions in the exact same way • Flexible and convenient: - can be completed with a mouse or a touch screen - can be segmented so that some sections are self-administered and some are administered by an interviewer - handles skip patterns automatically - can be offered in multiple languages - works for single site or multisite studies - data are readily exported for analysis • Same as CASI but with the addition of an audio component, so participants can listen to questions (e.g. through a headset) as well as read them • More appropriate than CASI for collecting data from participants with limited reading ability; otherwise, same advantages as CASI
Disadvantages • Requires participants to have a certain degree of literacy • May be less suitable for participants who are unfamiliar with or uncomfortable using a computer • May not offer sufficient opportunities for participants to seek clarification about individual questions, resulting in misinterpretation of their meaning and invalid responses
Audio computerassisted self-interview (ACASI)
• Same disadvantages as CASI
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
5. Questionnaire development
37
System
Mode of delivery Advantages
Disadvantages
Computer• A face-to-face assisted personal interview in interview which the (CAPI) interviewer immediately enters the response data into the computer
• May result in response bias if • Eliminates the need for subsequent participants are not comfortable data entry and the risk of introducing being honest with the interviewer errors • Ensures questionnaire is delivered consistently – participants all experience the questions in the exact same way • Interviewer is immediately available to clarify questions and responses • Flexible and convenient: - works with either a mouse or a touch screen. - can be segmented so that some sections are interviewer-administered and some are administered in private (via CASI or ACASI) - handles skip patterns automatically - can be offered in multiple languages - works for single site or multisite studies - data are readily exported for analysis • Same disadvantages as CAPI
• Same advantages as CAPI Handheld• Similar to CAPI, assisted personal a face-to-face interview (HAPI) interview in which the interviewer immediately enters the response data into a handheld computer (e.g. personal digital assistant (PDA) or mobile telephone) Web-based • A questionnaire is displayed on a website and participants log on and enter their answersa b • Allows participants to answer sensitive personal questions privately on a computer • Participants can complete the questionnaire on their own schedule and over several sessions, if necessary • Handles skip patterns automatically
• Requires participants to have a certain degree of literacy • May be less suitable for participants who are unfamiliar with or uncomfortable using a computer or the internet, and inappropriate for populations entirely without access to them • The clarity of the questions and instructions are especially important as participants may have even less opportunity to seek clarification from survey investigators • Potential participants must be given the web address and invited to complete the survey through some type of contact, email list, or advertisement • Requires a website host that is always accessible and fast enough to handle multiple participants simultaneously
a
For online survey, see http://www.orau.gov/cdcynergy/soc2web/Content/activeinformation/tools/toolscontent/surveyselfadministered_internet.htm (accessed 17 June 2014) Dillman 2000 (1)
b
Several commercial vendors offer web-based survey development services, including Survey Monkey1 and GfK Knowledge Networks.2 1 2
http://www.SurveyMonkey.com http://www.knowledgenetworks.com/ganp/
38
5. Questionnaire development
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Proximal and distal determinants of HIV The spread of HIV is due to both “proximal” and “distal” or “structural” determinants (2-4). Proximal determinants of HIV are biological or behavioural variables that directly facilitate the acquisition or transmission of HIV. Biological variables include HIV viral load and subtype, and the presence of other infections such as sexually transmitted infections (STIs). Behavioural variables include the type (e.g. vaginal or anal) and frequency of sex, condom use, and the types and number of sex partners. Distal or structural determinants can be just as powerful as proximal variables, but they affect HIV risk through the proximal variables. These determinants include factors such as stigma and discrimination, community mobilization, violence and poverty. The prevalence and relationship of distal determinants to risk behaviours, and ultimately to HIV prevalence and viral load suppression, are important for informing HIV interventions (5-7). An emerging recognition of “syndemics” has helped to show that HIV epidemics do not occur in isolation. Rather, they often occur in tandem with other epidemics and need to be addressed comprehensively (8).i The best known syndemics co-occurring with HIV are STI, tuberculosis (TB) and injecting drug use. Equally important syndemics include psychosocial issues, alcohol use and violence (9, 10).
A-5.1.2 Determine investigation topics (questionnaire domains) Investigation topics All biobehavioural surveys (BBS) should collect standard information on demographics, risk behaviours and access to HIV testing, prevention and health-care services. BBS data-collection instruments may include questions to assess both proximal and distal determinants of HIV. Questionnaire domains may include but are not limited to: • demographics • sexual and reproductive health • sexual behaviour • alcohol and drug use • HIV services uptake - access to counselling and testing - knowledge of serostatus - access to care and treatment - retention in care • access to other health-care services - general - STI - tuberculosis (TB) (HIV coinfected) • mental health, including depression • psychosocial support • shame, stigma and discrimination in the community and health facility and violence • knowledge of HIV prevention and transmission Additional domains may be added depending on the target population, survey context, and information needed to understand and respond to the HIV epidemic.
A-5.1.3 Develop and adapt the questionnaire General considerations When developing or adapting the questionnaire and selecting the method of administration, investigators must consider the points listed below: • choice of variables – The data obtained should be those that are necessary to better understand the target population and respond to the epidemic. • recall period – Participants are typically asked to recall incidents or behaviours occurring within a specified time period. Shortening this time period may reduce recall bias (i.e. systematic error due to differences in accuracy or completeness of remembering events) but may also limit the ability to measure the specified behaviour, because some participants will not have engaged in a particular behaviour during the shorter time period. • order of questions – Questionnaires should open with easy, salient and nonthreatening questions (11). The interview should move from more general questions to more specific ones, and questions that reference dates should be asked in consistent chronological order. When changing topics, the interviewer should use transitional phrases to allow participants to adjust their thinking. • length of the interview – The questionnaire should be kept as short as possible by removing questions that are redundant, yield the same answer from all participants, or are unlikely to be analysed or used (12). The saliency of the topic to the survey participant (see text box) will be a major determinant of the length of the questionnaire. Individuals on HIV treatment have more questions to answer than those who are HIV negative. Longer interviews may result in higher
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
5. Questionnaire development
39
survey refusal rates or participant fatigue, and unhappy participants may discourage others from participating. Survey investigators should pretest the questionnaire to determine how much time participants are willing to devote to the interview and adjust the length accordingly. • validation rules – To minimize errors in recording data, investigators should build validation rules into the questionnaire for continuous variables (i.e. responses that take the form of numbers, including age or number of sex partners). For electronic interviews these rules (e.g. lower or upper limits) can be programmed into the software so that it automatically detects invalid responses. Validation rules can also be used with paper-based data collection. This requires substantial interviewer training to detect invalid responses. • mandatory fields – It is almost always better to record a response such as “Don’t know”, “Refuse to answer” or “Not applicable” than to leave a response field blank. Blank response fields may be misinterpreted as missing data. • comment and open-text fields – Occasionally, it is necessary to ask an open-ended question and record the response as text; for example, in the field “Specify if other”. But open text can be difficult and time consuming to clean and categorize, so these fields should be used sparingly if at all, especially if they have little analytical value. Adapting the data-collection instrument for local use The standardized data-collection instrument in these guidelines comprises several questionnaire modules. Each module covers a specific topic and includes
questions deemed important for better understanding the epidemic or for satisfying reporting requirements. Within each module, questions are designated as “core” if the information being collected is considered essential (e.g. standard demographic information or information required to characterize risk) or actionable (e.g. able to guide public health action), and “secondary” if the information is useful but not essential. Questions that are core for some target populations may be secondary or not applicable for others. Survey investigators may build their questionnaire by choosing modules that are relevant to their country and the target population. This modular approach allows investigators to include additional topic areas by adding locally relevant modules (e.g. hepatitis and reproductive history) as appropriate, or omit modules that are irrelevant. All questions and instructions should be designed so that they may be self-administered using computer-assisted self-interview (CASI) or audio computer-assisted self-interview (ACASI). Respondents can read questions, and with ACASI also listen to them through a headset, and choose response options. Alternatively, survey investigators may opt to use the questionnaires for face-to-face interviews. The questionnaire modules included with these guidelines are listed on the next page. The description and rationale for each questionnaire module is provided in the introduction to the module.
Saliency Saliency is the degree to which a topic or event is relevant to a survey participant (13). Topics of great importance to the participant are said to have a high degree of saliency. Saliency operates at two levels, that of the question and that of the survey as a whole. In regards to individual questions, saliency refers to the importance of an event or action in a person’s life. Important events are better remembered than those with less saliency. For the survey as a whole, saliency refers to the degree to which the subject matter of the survey is relevant to the target population. If the questions being asked are of great interest to the typical member of the target population, the survey is said to be highly salient. Saliency has implications for survey investigators. Gaining cooperation or attaining a high response rate is more difficult when the saliency is low, because the target population has little motivation to participate. It is much easier to achieve a high response rate with a questionnaire that is highly salient and has a low participant burden (i.e. takes little time to complete, and is easy and straightforward to understand) than with one that has low saliency and a high participant burden. In the latter case, it may be necessary to give participants more reimbursement for their time in order to improve the response rate.
40
5. Questionnaire development
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Questionnaire Modules 1. Questionnaire parameters 2. Eligibility 3. RDS recruiter–recruit relationship 4. Respondent-driven sampling (RDS) personal network size 5. Size estimation 6. ACASI tutorial 7. Demographics 8. Transgender characteristics 9. Sex work characteristics 10. Clients of sex workers characteristics 11. Incarceration characteristics 12. Transport work characteristics 13. Reproductive health 14. Sexual history 15. Recall sexual behaviour 16. Partner concurrency 17. Last sex act 18. Condom use/accessibility 19. Lubricant use 20. Microbicide use 21. Alcohol use 22. Noninjecting drug use 23. Injecting drug use 24. Drug overdose 25. Sexually transmitted infections 26. Hepatitis B virus/hepatitis C virus (HBV/HCV) 27. Services uptake 28. Post-exposure prophylaxis/pre-exposure prophylaxis (PEP/PrEP) 29. Shame, stigma, harassment and discrimination 30. Physical violence 31. Sexual violence 32. Internalized homophobia 33. Depression 34. Suicide ideation and attempts 35. Social support 36. Social cohesion, social participation and collective agency 37. Game of contacts 38. HIV knowledge and perceptions 39. Questionnaire feedback/interview status 40. RDS peer recruitment Data measures Survey investigators should review existing data reporting requirements and indicators such as those from the Joint United Nations Programme on HIV/AIDS (UNAIDS); the United States President’s Emergency Plan for AIDS Relief (PEPFAR); and the Global Fund to Fight AIDS, Tuberculosis and Malaria, to: • determine whether the questionnaire collects the information needed; • identify gaps and determine whether new questions are needed; and
• consider how the data will be tabulated or analysed. It is important to know which denominators will be used for each variable. For example, the denominator used to calculate the share of the population that used a condom at last sex with a casual partner should include only those who reported having sex with a casual partner. Accordingly, the survey must include a question that asks whether the participant has had sex with a casual partner. Substituting in the number of participants who reported having sex with any type of partner could result in an underestimate of the indicator. For more information see Section 3, which provides an overview of data measures and suggests how they should be used. That section includes reference sheets for a core set of data measures considered critical for understanding the context in which key populations live and make decisions, and for monitoring and evaluating programmatic responses.
A-5.1.4 Translate and back-translate the questionnaire Many surveys will use more than one interview language. Translators should have knowledge of health terminology, colloquial phrases and jargon used by the survey population. When translating questionnaires into other languages, it is recommended to back translate the questionnaire into the original language. Staff not involved with the translation should do the back translation without using the original-language questionnaire. The two questionnaires (i.e. original and back-translated) should then be compared, and any question and answer phrasings that differ in their meaning should be noted. The translation should be changed as necessary until the two questionnaires fully agree in their question and answer phrasing.
A-5.1.5 Conduct cognitive testing Often, survey investigators want to add a few questions to the standardized questionnaire. Also, in some cases, a country may want to collect information for which no standardized questions exist. New, previously unused questions for a country should be evaluated for validity and to determine what information they are actually capturing. The goals of evaluation are to assess how respondents interpret survey questions and identify potential response problems that could introduce errors in the survey data. Questions should be designed based on the results of this evaluation, not on expert opinion (14). After a formative assessment has been conducted to identify topics of inclusion in the questionnaire, questions should be cognitively tested to ensure that they are properly understood (15). The primary purpose of cognitive testing is to investigate how well questions 5. Questionnaire development
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
41
perform when asked of survey participants; that is, to test whether participants understand the question correctly and can provide accurate answers (16). Cognitive testing ensures that a survey question captures the investigators’ intent and makes sense to participants. Cognitive testing includes the following steps: • administer draft questionnaires and collect additional information about responses: - facilitate participants to share their thoughts while hearing and responding to questions; - identify specific problems and answer specific questions; • use information to: - evaluate the quality of the question; - understand whether the question gathers the intended information; • look at question-answering from the participant’s perspective in order to understand: - the thought processes used to answer questions; - how questions are interpreted; and - how participants understand concepts. Cognitive testing can help to ensure that survey questions capture the intended information. If they do not, investigators can modify questions well before they start the survey. Investigators should identify and document what each question should measure, and compare it to what it does measure. For example, for investigators, the term “sex” may refer to only vaginal and anal sex, but for participants it may also include oral sex. This process may thus identify subtle but important differences in interpretation. It should be well documented and capable of replication. Cognitive testing should answer the following questions: • how do respondents interpret the survey question? • do respondents interpret questions differently than intended? • how do respondents determine their response? • does the question have the same meaning in all the languages in which it is asked? • does the question collect the same type of information across cultures, languages, subgroups or settings (e.g. urban and rural)?
investigators to identify and address problems before they occur in actual survey implementation. Pretesting can also identify problems introduced by translation or the adaptation of an existing questionnaire to the local context. Survey staff should pretest all data-collection forms with diverse members of the target population, ideally under typical field conditions. Pretesting activities and findings should be documented to inform the revision of data-collection tools.
A-5.1.7 Train interviewers If the survey is to use face-to-face interviews, investigators should develop an interview guide for the interviewers and their supervisors. The interview guide should address each interview question, explaining its rationale and its intended meaning, skip patterns, legal value definitions, appropriate background computation (if applicable) and informational messages. This will be helpful for training and to clarify any ambiguities or misunderstandings that may arise during implementation. Interviewers should be thoroughly trained to ensure that the interview process is as standardized as possible. All interviewers should administer the questionnaire or other data-collection instruments in the same way to avoid introducing bias in how participants respond. This standardization should extend to how the questions are read and explained, and to the attitude the interviewers present to the participants. This can greatly influence the given responses – especially if the interviewer asks about illegal or stigmatized behaviour. To increase the likelihood of obtaining honest responses, interviewers should assure participants of their confidentiality, and should be thoroughly trained in open and nonjudgemental interview techniques. They should also be trained to accurately record responses. The amount of training required will depend on who conducts the interviews. Experienced or professional interviewers may be familiar with features of complex questionnaires such as coding responses and skip patterns, and need less training to effectively administer the questionnaire. But peer interviewers drawn from the target population may be perceived as less judgemental. Although peer interviewers from the target population may need more training, they may also be perceived as less judgemental. Training should include substantial role-playing so that interviewers become familiar and comfortable with the questions and skip patterns. Supervisors should also participate in this role play to become familiar with the interview process and to learn how to maintain quality control during survey implementation. More information on training can be found in Chapter A-16.
A-5.1.6 Pretest the questionnaire Whether survey questionnaires are self-administered or implemented through face-to-face interviews, datacollection instruments must first be pretested. Pretesting is different from cognitive testing; its purpose is for trained interviewers to practise using the instruments as a means for improving data collection. Pretesting helps survey
42
5. Questionnaire development
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-5.1.8 Pilot test the survey After interviewers have been trained, the entire survey process should be pilot tested before actual data collection begins. This allows unforeseen challenges to be identified and corrected before the survey starts. The interview component should assess the feasibility of conducting the interview, the duration of the interview, and whether completed questionnaires and other survey instruments or forms are stored and transported as planned.
A-5.2 Summary Data-collection instrument pretesting, questionnaire adaptation, interviewer training and survey procedure pretesting are all essential components of survey preparation and should be included in BBS protocols. Adequate advance time for these components needs to be incorporated into the survey preparation timeline. To help ensure the quality of data collection and improve the survey instrument, investigators should: • adapt data-collection instruments to the local context; • develop and test software as warranted;
• ascertain the reading level of the target population to determine what interview mode should be used for self-administered questionnaires (CASI, ACASI or other); • determine whether responses vary depending on interview mode – oral responses given during a face-to-face interview may differ from responses entered privately in a self-administered interview, and questions that are read (as in CASI) may be understood and answered differently than those that are heard (as in ACASI); • take the time to translate the questions and then to back translate them into the original language, thus ensuring that complex concepts are interpretable in a commonly understood manner; • check questionnaires – electronic or paper based – for logic and validity of response choices and skip patterns; • introduce artificial errors and assess quality control by measuring the proportion of those errors identified during pretesting; • pretest (electronic) questionnaires, and make improvements and corrections as needed; • verify that the questions are clear to survey participants, understood as intended, and answerable; and • examine the test data produced with these instruments.
A-5.3 References 1 2 3 4 Dillman DA. Mail and internet surveys: The tailored design method. Wiley New York. 2000. Baral S, Logie CH, Grosso A, Wirtz AL, Beyrer C. Modified social ecological model: a tool to guide the assessment of the risks and risk contexts of HIV epidemics. BMC Public Health. 2013;13:482. Hayes R, Kapiga S, Padian N, McCormack S, Wasserheit J. HIV prevention research: taking stock and the way forward. AIDS. 2010;24(Suppl 4):S81–S92. Shannon K, Goldenberg SM, Deering KN, Strathdee SA. HIV infection among female sex workers in concentrated and high prevalence epidemics: why a structural determinants framework is needed. Curr Opin HIV AIDS. 2014;9(2):174–182. Chersich MF, Luchters S, Ntaganira I, Gerbase A, Lo YR, Scorgie F et al. Priority interventions to reduce HIV transmission in sex work settings in sub-Saharan Africa and delivery of these services. J Int AIDS Soc. 2013;16(1):17980.
5
6 Mayer KH, Wheeler DP, Bekker L-G, Grinsztejn B, Remien RH, Sandfort TGM et al. Overcoming biological, behavioral, and structural vulnerabilities: new directions in research to decrease HIV transmission in men who have sex with men. J Acquir Immune Defic Syndr. 2013;63(Suppl 2):S161–S167. 7 Wheeler T, Kiran U, Dallabetta G, Jayaram M, Chandrasekaran P, Tangri A et al. Learning about scale, measurement and community mobilisation: reflections on the implementation of the Avahan HIV/AIDS initiative in India. J Epidemiol Community Health. 2012;66 Suppl 2:ii16–25 (http://www.ncbi.nlm.nih.gov/pubmed/22945907, accessed 4 August 2016).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
5. Questionnaire development
43
8
Drobnik A, Pinchoff J, Bushnell G, Ly S, Yuan J, Varma JK et al. Matching HIV, tuberculosis, viral hepatitis, and sexually transmitted diseases surveillance data, 2000–2010: identification of infectious disease syndemics in New York City. J Public Health Manag Pract. 2014;20(5):506–512 (http://www.ncbi.nlm.nih.gov/pubmed/24335712, accessed 4 August 2016). Russell BS, Eaton LA, Petersen-Williams P. Intersecting epidemics among pregnant women: alcohol use, interpersonal violence, and HIV infection in South Africa. Curr HIV/AIDS Rep. 2013;10(1):103–110 (http://www. ncbi.nlm.nih.gov/pubmed/23233038, accessed 4 August 2016).
9
10 Santos GM, Do T, Beck J, Makofane K, Arreola S, Pyun T et al. Syndemic conditions associated with increased HIV risk in a global sample of men who have sex with men. Sex Transm Infect. 2014;90(3):250–253 (http://www.ncbi. nlm.nih.gov/pubmed/24431183, accessed 4 August 2016). 11 Sudman S, Bradburn NM. Asking questions: A practical guide to questionnaire design. San Francisco, California, USA, Jossey-Bass. 1989. 12 Sudman S. Estimating response to follow-ups in mail surveys. Public Opin Q. 1983;46(4):582–584 (https://www. ncbi.nlm.nih.gov/pubmed/10278173, accessed 1 November 2016). 13 Mooney G. In: Lavrakas PJ (ed), Encyclopedia of survey research methods, Thousand Oaks, California, SAGE Publications. 2008. 14 Willis GB. Cognitive interviewing: a tool for improving questionnaire design. Thousand Oaks, California, SAGE. 2005. 15 Campanelli P. Testing survey questions: New directions in cognitive interviewing. Bulletin of Sociological Methodology. 1997;55(1):5–17 (http://bms.sagepub.com/content/55/1/5, accessed 30 August 2016). 16 Collins D. Pretesting survey instruments: an overview of cognitive methods. Qual Life Res. 2003;12(3):229–238.
44
Questionnaire development
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms Algorithm: Step-by-step procedure.
6. Biomarker considerations
This chapter focuses on selecting biomarkers for measurement in biobehavioural surveys (BBS) and the laboratory considerations most relevant to them. It discusses everything from selecting biomarkers that are appropriate to the target population and local setting, to the practicalities of dealing with specimens.
A-6.1 Biomarker selection Survey investigators should choose biomarkers based on the biological information needed to characterize the local epidemic and risk factors for HIV (see Table A-6.3). Linkage of biomarkers with participant characteristics and their reported behaviours allows investigators to explore associations between them (e.g. HIV status by age group), and provides useful information for policy and service planning. Biomarker data collected in repeated surveys may be used to monitor trends in biomarker prevalence, which may then be explained using linkage with behavioural data. Because biological testing technology is constantly evolving, survey investigators should consult with laboratory advisors, national authorities, the Centers for Disease Control and Prevention (CDC), WHO and test kit manufacturers for the most up-to-date information regarding biomarker assays; specimen, instrumentation and commodity requirements; and SOPs. Collaboration with laboratory experts and the national reference or public health laboratory is strongly recommended, and should be established during the survey design phase. Biomarker data may be qualitative (e.g. a categorical outcome, such as positive or negative) or quantitative (e.g. a measure on a continuous scale, such as 10 000 viral copies/mL). Going beyond measuring clinical and molecular biomarkers for estimating HIV prevalence, some markers (e.g. CD4+ T-cell count and viral load) are particularly useful for staging the HIV infection, measuring the potential for HIV transmission, and assessing viral suppression or treatment failure to measure antiretroviral therapy (ART) programme impact.
Biological the data derived from testing biological data or specimens. “biodata”: Biological Biological material, such as blood, urine specimen: or saliva, collected from a person. Biomarker: A biological characteristic measured through laboratory or other tests, generally using a biological specimen such as blood for testing; examples include HIV antibody or viral load. Biosafety: Containment principles, technologies and practices implemented to prevent unintentional exposure to pathogens and toxins, or their accidental release. Biosecurity: Protection, control and accountability practices for biological materials within laboratories that prevent their unauthorized access, loss, theft, misuse, diversion or intentional release. Dried blood Small volume of blood, collected from spot (DBS): a participant via a finger prick or other means, spotted onto a filter paper card, dried and transferred to a laboratory for analysis. Plasma: Liquid portion of blood in which red and white blood cells, platelets, nutrients, waste products, antibodies, clotting factors, hormones and proteins are suspended. When anticoagulated whole blood is centrifuged (or allowed to settle), red cells go to the bottom, and white cells and platelets to the middle, leaving the yellowish plasma at the top. Quality A set of activities conducted as part of assessment: quality assurance to evaluate a laboratory or testing site.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
45
Biological specimens and biomarkers The terms “biological specimens” and “biomarkers” are sometimes used interchangeably, but in fact have different meanings. In this document, “biological specimen” refers to the actual biological material collected from a survey participant, such as blood, urine or saliva. A “biomarker”, often derived from a biological specimen through testing, is a measurable factor associated with health or a particular medical condition. Examples include HIV antibodies or viral load as markers of HIV infection; C-reactive protein, a marker of acute inflammation; CD4+ T-lymphocyte count, an indicator of immune function; and hepatitis B virus (HBV) surface antigen or antibodies to HBV core and surface antigen (see Appendix I-12). Biomarkers may also include genetic test results, such as HIV drug resistance or HIV subtype (1). Some biomarkers do not require biological specimens but are derived from other measures (e.g. blood pressure to assess hypertension or the ratio of height to weight to identify malnutrition in children under five years of age).
Key terms Quality A range of activities that enable assurance: laboratories to achieve and maintain high levels of accuracy and proficiency. Quality assurance: • establishes standard operating procedures (SOPs) for each step of the laboratory testing process, from specimen handling to instrument performance validation; • defines administrative requirements, such as mandatory record-keeping, data evaluation and internal audits to monitor adherence to SOPs; • specifies corrective actions when problems are identified, documentation, and the persons responsible for carrying out corrective actions; and • sustains high-quality employee performance. Serum: Fluid that rises to the top when coagulated (clotted) blood has been centrifuged or allowed to separate. Clearer than plasma because it contains fewer proteins such as clotting factors, which are held in the clot at the bottom of the tube. Venipuncture: Puncture of a vein with a needle to collect blood (also referred to as phlebotomy) or to administer intravenous therapy. Viral load: Concentration of HIV in the blood expressed as the number of viral copies per millilitre. In clinical care, viral load measurement is used to monitor the effectiveness of antiretroviral treatment in suppressing viral load. Whole blood: Blood that includes both the cellular elements (red and white blood cells), platelets and plasma.
When selecting biomarkers for inclusion in a BBS, investigators should assess the need for biological information about a population, taking into account previously collected data, and the information needs of different disease control programmes. For example, sexually transmitted infection (STI) programmes may be interested in measuring the prevalence of syphilis in men who have sex with men (MSM), and national hepatitis programmes may be interested in the rate of hepatitis C virus (HCV) among people who inject drugs (PWID).
A-6.1.1 Target population and local context The appropriate biomarkers to collect during a BBS will depend on the target population and the local context. For example, STI biomarkers could be a higher priority for surveys among MSM, sex workers (SW) and transgender persons (TG) than among PWID. Testing for HCV could be a higher priority for a survey among PWID than among SW. However, HCV testing may be warranted where injecting drug use is common among SW. Also important to consider is the cost of tests and the local capacity for conducting them, and the feasibility and legality of exporting samples if testing can only be done outside of the country. When selecting biomarkers and laboratory methods, investigators should consider, but not be limited by, comparability with previous surveys. For example, as testing technology improves, investigators should consider using improved methods that may offer greater sensitivity and specificity rather than using only the methods employed in previous surveys. When interpreting results, however, investigators need to consider the impact that switching methods may have
on analysis of differences or trends (2). Furthermore, investigators may decide to discontinue collection of biomarker data that are no longer needed or used.
A-6.1.2 Data needed for action Investigators need to consider how the biological data will contribute to local knowledge, and whether the country or programme has the ability to act on the data obtained. Examples of how biological and behavioural data may be combined to assess need for public health action are shown in Table A-6.1.
46
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table A-6.1 Surveys that collect both biological and behavioural data may indicate levels of risk, and highlight the need for public health action Biomarker and behavioural data HIV or STI biomarkers and data on risk behaviour HIV viral load and data on risk behaviour CD4 count and HIV testing history HIV status and data on receipt of prior HIV test results HBV infection status and data on injection behaviour
May indicate Risk for HIV acquisition or transmission HIV transmission potential Delay in testing; eligibility for treatment Awareness of HIV serostatus Risk for HBV acquisition or transmission, and potential need for immunization
A country may not have the resources to immediately act on survey findings that indicate high levels of risk behaviour or disease burden (e.g. by providing hepatitis treatment or vaccination). Nevertheless, the data may provide the foundation to advocate for implementing or strengthening prevention, care and treatment efforts.
A-6.2 Resource availability Investigators must consider available resources, such as budget, laboratory infrastructure and human capacity when selecting which biomarkers to include in a BBS. If specimens can be stored in a repository, investigators may be able to perform additional testing or facilitate future shipment to an external laboratory when funding becomes available. Collecting, handling, transporting and storing specimens may have significant costs in addition to the costs of biomarker test kits, reagents and equipment. Investigators also must consider the human resources required for specimen collection, processing and testing, as well as for providing counselling, and referrals to care and treatment. During the survey planning phase, survey investigators should collaborate with laboratory experts to determine local capacity for testing and storing specimens, where tests will be performed, and what testing platform and assays will be used.
specified or indefinite period, and conduct unspecified future tests on the specimens. The survey protocol should specify the type and quantity of specimens to be collected (e.g. 10 mL of whole blood or 15 dried blood spots), amount to be stored (e.g. volume or number of spots), and the length of time specimens will be stored. It is generally considered unethical to obtain consent for future unspecified testing if investigators know at the time of consent what future tests they will conduct. Protocols that seek to identify bacterial STIs must include a plan to provide treatment or referral to a health facility willing to provide treatment to survey participants. Protocols should also describe how investigators will measure successful referral of participants to HIV or STI care and treatment (e.g. participants may be given vouchers to present at referral sites). Vouchers can be collected and counted to measure the proportion of referred participants who presented to referral sites. In practice, it can be difficult to ensure successful referrals, and even harder to document them when they do occur. Documenting successful referrals may be particularly difficult because there may be multiple clinics from which participants may choose to receive care, including clinics not included on the referral list; also, some participants may leave the area or simply be lost to follow-up.
A-6.4 Providing test results to participants Survey investigators should give participants all test results that could affect their clinical care. Results that are of no health consequence to the participant (e.g. illicit drug metabolites detected in urine, or prostatespecific antigen in vaginal swabs), or that have utility at the population level but limited accuracy at the individual level (e.g. recent HIV infection), may not need to be returned to the participant. The ethical obligation to provide test results and connect survey participants to care is especially important for key populations who
A-6.3 Ethical considerations As part of the informed consent process, participants must be told what specimens are being collected and how; which tests will be performed; the risks and benefits of testing; if, when and how test results will be returned; and which treatment or referral services will be provided. Also, when applicable, survey investigators must obtain informed consent to store specimens for a
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
47
may have less access to health services than members of the general population. Where health services for key populations do not exist, investigators should identify and train willing providers to work with these populations. Surveys can return results while maintaining the anonymity of participants by returning results either on the same day the specimen is collected or through a system whereby participants can anonymously verify their identities (e.g. with a unique code or through a coded fingerprint scan) and collect the result at a later date. Surveys that return results days or weeks after sample collection often find that some participants cannot be reached or do not return for their results, meaning that those participants miss an opportunity to learn their HIV or STI status. Rapid testing is thus ideal for ensuring that participants obtain results. Additionally, it may be logistically challenging to return results of some assays that are performed off-site or at a much later date (e.g. testing for HIV drug resistance). One option for making such results available to participants includes giving them a phone number that they can call after a specified period to obtain the results. This may be feasible even after the survey has closed.
the collection of personal identifiers that may complicate data security procedures. A material transfer agreement (Appendix I-11) should be used when specimens are sent off-site for testing.
A-6.6 Treatment and referral If survey investigators test for HIV, they assume responsibility for returning results to participants and referring them to HIV services. Successful linkage of PLHIV to treatment is extremely important ethically and to combat the epidemic. Referral IDs linked to survey IDs may be used to track successful linkage to health or other services. The use of referral IDs can prevent service providers from connecting patient names to survey IDs. To maintain participant anonymity, delete referral IDs from the dataset prior to releasing a public dataset. Ensure that the names of people who were successfully referred are not communicated to survey staff. If STI biomarkers are assessed, investigators should provide STI treatment or referral to treatment, as recommended by local health authorities, based on test results. Providing on-site treatment for bacterial STIs is preferable to referral because some participants may not follow through with the referral. For some STIs, such as herpes simplex virus-2 (HSV-2) infection, treatment may not be feasible or easily available. Special consideration should be paid to treatment of participants who are pregnant or have allergies, mode of drug administration (oral, intramuscular or intravenous), managing allergic reactions and length of treatment (single or multiple doses). If national treatment guidelines do not exist, investigators should consult the most recent WHO or other international treatment guidelines.
A-6.5 On-site or off-site testing Surveys that use rapid tests and other point-of-care technology to test for biomarkers on-site (or near the site of specimen collection) should provide participants with same-day results (e.g. for HIV, CD4+ T-cell count and syphilis). Investigators are strongly encouraged to use the national standard for diagnostic testing rather than a testing algorithm recommended only for surveillance purposes. If diagnostic testing is performed, the counsellor may be able to provide the participant with a final result based on rapid testing. Some national standards, however, may require that rapid test results be confirmed with another test that cannot be performed on-site. In this case, counsellors will need to ask the participant to return to obtain the final result or to visit another facility for confirmatory testing. For some biomarkers, including HIV, it is important to confirm positive or reactive test results by including a specific test (e.g. western blot or similar) to ensure that accurate results are returned to participants. For specimens that are tested off-site (at a central or reference laboratory) a unique code or identifier can be used to return results to participants together with posttest counselling at a later date, either at the survey site or at nearby health facilities. A disadvantage of this approach is that participants may forget or decide not to return for their results, or may be unavailable during a repeat visit in the case of time-location sampling. Investigators could potentially use technologies such as mobile phones to return test results or send reminders, but their use implies 48 6. Biomarker considerations
A-6.7 Testing for HIV biomarkers HIV biomarkers include viral components (e.g. proteins, commonly referred to as HIV antigens) and nucleic acids (e.g. RNA and DNA) that indicate presence of HIV, and antibodies produced as part of the immune response to HIV infection. Antibody- and antigen-based assays are the most commonly used HIV diagnostic tests. HIV genomic sequencing is used for HIV drug resistance or molecular epidemiology. CD4+ T-cell assay results indicate level of immunosuppression. Viral load is an indicator of virus transmission potential and the effectiveness of antiretroviral treatment. Providing additional test results, such as CD4+ T-cell count and viral load, to HIV-infected survey participants together with a referral to the nearest health facility offering ART may facilitate linkage of participants to treatment. Table A-6.2 lists commonly measured HIV-related biomarkers and their utility.
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table A-6.2 HIV-related biomarkers and tests, and their usefulness in BBS Biomarker HIV antibody
Test
Usefulness in BBS Presence indicates HIV exposure and infection; HIV seroprevalence is the proportion of the population with HIV antibody. Can be used to validate self-reported HIV status and measure the proportion of sample population undiagnosed or unaware of their status. An exception is infants who may have maternal antibodies to HIV but who are not themselves infected with HIV.
RT EIA WB
Rapid testing facilitates provision of same-day test results. Some RTs also detect HIV antigens, while others can distinguish between recent and long-term infection. Laboratory-based method that is more efficient than RTs for testing a large number of batched specimens, but same-day results may not be feasible. Confirms EIA results (if required by national guidelines). Other highly specific rapid confirmatory tests are now available that may be used to confirm positive results, HIV1 and HIV-2 dually reactive specimens, or serve as a tiebreaker for discordant results, and can discriminate between HIV-1 and HIV-2 infections.
Tests for recent HIV Distinguish recent from long-term HIV-1 infection to estimate population-level infection incidence; tests are performed on HIV-1-seropositive specimens and the testing algorithm should include viral load testing to improve the predictive value of a test for recent infection. These tests are used to estimate HIV incidence at the population level, and usually are not recommended for individual diagnosis. Most BBS cannot achieve sample sizes large enough to estimate assay-based HIV incidence with adequate precision. HIV antigen (e.g. p24 antigen) Presence indicates HIV infection earlier than antibody detection. WB RT CD4+ Tlymphocytes HIV RNA CD4+ T-cell count Rarely used in surveys. Combination assays that independently distinguish results for HIV-1 p24 antigen and HIV antibodies in a single test are now commercially available. Indicates level of immunosuppression and ART eligibility; approximates duration of infection in treatment-naive individuals. CD4 testing requires a whole blood specimen, which should be processed within 24 hours of collection. Measures amount of virus in the blood. Indicator of treatment effectiveness and transmission potential; can be used to determine population (summary) VL and proportion of HIV-infected persons with suppressed (e.g. <1000 copies/mL) VL. Definition of viral suppression may vary based on factors such as clinical guidelines and assay used. Common definitions include: • population VL: among those HIV+ (diagnosed or not) • diagnosed VL: among those diagnosed with HIV • treatment VL: among those on ART • indicator for final step of HIV cascade: percentage on ART who are virally suppressed. HIV drug resistance (HIVDR) mutations HIV subtype Genomic sequencing Indicates level of resistance to different ARVs. HIVDR may be reported and used at the individual level to help guide patient treatment, or at the population level to estimate levels of resistance to different classes of ARVs. Identifies clusters of HIV infection in a population; useful in describing subepidemics (e.g. when HIV subtypes among key populations differ from those in the general population). Detects presence of ARV metabolites. Marker of ART or use of PrEP. May validate clinical or self-reported information about ARV use (e.g. ART adherence, or whether or not participant is currently on ART). Percentage on ART is an HIV cascade indicator.
VL test
Genotyping
ARV metabolites
HPLC combined with mass spectrophotometry
ART, antiretroviral therapy; ARV, antiretroviral; EIA, enzyme immunoassay; HIVDR, HIV drug resistance; HPLC, high-performance liquid chromatography; PrEP, pre-exposure prophylaxis; RT, rapid test; VL, viral load; WB, western blot
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
49
A-6.8 Testing for non-HIV biomarkers
Hepatitis B (caused by HBV) and hepatitis C (caused by HCV) are bloodborne infections that are easily transmitted through contact with blood, including by BBS often include other biomarkers, particularly for STIs. sharing injecting equipment, and occasionally through Many STIs are transmitted much more easily than HIV, sex. These infections, as well as other viral STIs such as and such infections are often easily treated and less likely HSV-2, have a tendency to become chronic infections, to become chronic. Thus, STI prevalence is often a more and treatment may be difficult or expensive to obtain, or sensitive indicator of recent sexual risk behaviour than HIV prevalence. Some STI measurements in surveys can serve as: the infection may not be curable. Measurement of these biomarkers is less useful for correlation with sexual risk behaviours, although prevalence of HBV and HCV may be • an early warning system for potential HIV acquisition useful indicators for monitoring the impact of prevention or transmission and emergence of HIV in new groups or programmes among drug users. Measures of HSV-2 geographical areas; may be useful for estimating the lifetime exposure to • a risk assessment tool for HIV prevention programmes; unprotected sex, especially among young people. and • a means to detect independent STI epidemics (e.g. syphilis or gonorrhoea) that need public health attention. Table A-6.3 lists commonly measured non-HIV biomarkers and their utility.
Table A-6.3 Non-HIV biomarkers and tests, and their usefulness in BBS Non-HIV biomarker Chlamydia trachomatis (CT) Nucleic acid amplification test (NAAT)
Test
Usefulness Causative agent of chlamydia. All listed tests detect active infection with CT. Useful for validating self-reported sexual behaviour; monitoring trends in sexual risk behaviour; and measuring prevalence of CT in the survey population. NAAT is highly sensitive and specific and requires closed systems and trained laboratory staff. Involves noninvasive specimen collection (urine or swabs – vaginal or rectal). Current standard of testing in high-income countries but increasingly also used in resource-constrained settings. Many NAATs include both CT and NG probes allowing testing for both organisms with a single specimen. For surveillance purposes (e.g. BBS) NAAT should be considered over other test methods. DFA was previously used for clinical diagnosis, but is currently not recommended for routine testing of genital tract specimens. Procedure requires experienced microscopist and is labour intensive and time consuming. Not useful for BBS. EIA is used in clinical diagnosis. Detects chlamydial antigen. Requires experienced technician and collection of adequate specimen. Also requires confirmatory assay (with a blocking antibody or direct fluorescent antibody test) to eliminate false positive result. None of the EIAs are as sensitive or specific as the NAATs, and costs are typically high. Impractical for BBS. Although culture was formerly the reference standard against which all other tests have been compared, few laboratories have capacity for culture of this delicate organism. Culture methods for C. trachomatis are insensitive, difficult to standardize, technically demanding, and expensive. Maintaining viability of organisms during transport and storage in diverse settings is a serious challenge. Impractical for BBS. Current serological tests have limited if any value for screening due to crossreactive antibodies. Causative agent of gonorrhoea. All tests listed detect active infection with NG, indicating recent unprotected sexual intercourse. Useful for validating selfreported sexual behaviour; monitoring trends in sexual risk behaviour; and measuring prevalence of NG in the survey population.
Direct fluorescent antibody (DFA) Enzyme immunoassay (EIA)
Cell culture
Serology
Neisseria gonorrhoeae (NG)
50
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Non-HIV biomarker
Test NAAT
Usefulness NAAT is highly sensitive and specific and requires closed systems and trained laboratory staff. Involves noninvasive specimen collection (urine or swabs – vaginal, rectal or oropharyngeal). Current standard of testing in high-income countries, increasingly being used in resource-constrained settings. Many NAATs include both CT and NG probes allowing testing for both organisms with a single specimen. For surveillance purposes (e.g. BBS) NAAT should be considered over other test methods. GS or GV is useful for diagnosis of symptomatic disease in men with urethritis, but not commonly used in BBS. Useful for distinguishing NG from CT in symptomatic men through detection of intracellular diplococci. Less useful in women, and not useful in asymptomatic men or women. Requires microscope and experienced technician. Culture had been the reference standard against which all other tests have been compared; however, NG is an obligate anaerobe and maintaining the viability of organisms during transport and storage in the diverse settings is a serious challenge. Cultures require enriched media. Because of the difficulty in transporting NG, culture is not very sensitive (i.e. most of the organisms die during transport). Culture is currently needed for studying antibiotic resistance. Detects NG antigen in genital secretions. Same caveats as for CT. DFA testing is difficult, expensive and not very sensitive. See above caveats for CT. EIA is less sensitive than NAAT and not recommended. There is no serological assay available for NG. Causative agent of syphilis
Gram stain or gentian violet stain (GS or GV)
Culture
DFA EIA Serology Treponema Pallidum (T. pallidum) Serological tests – nontreponemal and treponemal antibodies
Serological tests may be used to detect current or past syphilis. Nontreponemal antibodies may indicate active (current) syphilis. Treponemal antibodies may indicate either active or past (resolved) syphilis. Specimens reactive for nontreponemal antibodies are usually confirmed by testing for treponemal antibodies. Useful for estimating prevalence and for diagnosis and treatment of syphilis among key populations. Infection with nonsexually transmitted treponematoses (e.g. yaws and pinta) will result in a positive serological test for T. pallidum and should be considered where such infections are endemic. Currently, there are no routine laboratory methods to distinguish endemic treponematoses from each other or from syphilis.
Nontreponemal tests – indirect method that detect biomarkers released during cellular damage caused by T. pallidum. These include: • VDRL (venereal disease research laboratory) slide test • Unheated serum reagin (USR) test • RPR (rapid plasma reagin) card test • TRUST (toluidine red unheated serum test) • rapid immunoassays
Can be useful for screening; rapid, simple and inexpensive, but require subjective interpretation by an experienced laboratory technician. Nontreponemal tests can be used as a marker of sexual activity. Reactive results may be confirmed by a treponemal test (see below). Because one of the objectives of BBS may be to detect active infections, reactive nontreponemal tests should be confirmed with a treponemal test. Likewise, reactive treponemal test results should be confirmed with a nontreponemal test (e.g. RPR or VDRL) to confirm current syphilis infection. VDRL and USR must be read under a microscope. RPR and TRUST do not require a microscope, but do require trained personnel using standardized equipment and specialized reagents.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
51
Non-HIV biomarker
Test Treponemal tests – detect T. pallidum antibodies that occur as a direct result of infection • TPPA (T. pallidum particle agglutination) • FTA-Abs (fluorescent treponemal antibody absorption) • TPHA (T. pallidum particle haemagglutination assay) • Rapid immunoassays • Combination rapid assays (treponemal and nontreponemal) Dark-field microscopy
Usefulness A reactive treponemal test indicates current or past infection and may not differentiate between the two. Treponemal test results may remain reactive for life, even with treatment. Because a reactive test result only indicates exposure to TP at some time in a person’s life, it needs to be confirmed with a nontreponemal test. Treponemal tests do not differentiate venereal syphilis from endemic syphilis (yaws and pinta). Traditionally treponemal tests were used mainly as confirmatory tests to verify reactive screening test results. However, a rapid treponemal test or EIA can be used for screening in certain situations (e.g. population prevalence <1% or screening in antenatal women). Individuals with reactive test results would then be treated presumptively or have a follow-up RPR or VDRL to identify active infection. Rapid immunoassays are available that simultaneously test for treponemal and nontreponemal antibodies, simplifying the detection of both current and past infection. Useful for diagnosis when lesions are present, but impractical for BBS. Requires examination by trained technicians immediately (within 20 minutes) after specimen is collected onto a slide, and a microscope with dark-field illumination. Dark-field examination of lesion exudates for motile spirochetes may be helpful, but if a patient has self-medicated or if there are only a few organisms present, spirochetes may not be seen. Causative agent of trichomoniasis, the most prevalent nonviral STI; trichomoniasis prevalence.
Trichomonas vaginalis (T. vaginalis) Direct microscopic examination (DME) – wet-mount preparation of vaginal discharge to assess presence of motile protozoa Nucleic acid detection (Polymerase chain reaction-PCR) Broth culture
DME is inexpensive and easy; however, sensitivity is highly variable (38% to 82%) and is dependent upon inoculum size and skill of examiner. Requires cervical smears, and thus often impractical for BBS unless an examination is conducted.
Several PCR-based diagnostic assays using vaginal and urine specimens exist; these vary in their sensitivity and specificity. Can be useful in BBS using urine (men) or vaginal swab (women) samples. Sensitivity varies by culture technique. Less expensive and more convenient than cell cultures, but insensitive compared with PCR. Could be used for BBS, but may be costly. Causative agent of chancroid; facilitates HIV transmission and acquisition; extremely delicate organism and thus difficult to culture. Clinically, chancroid can appear similar to syphilis in causing genital ulcer syndrome; however, chancroid is typically painful and associated with more lymphadenopathy than syphilis.
Haemophilus ducreyi
Nucleic acid detection with or without amplification Serology DFA (direct fluorescent antibody) Culture
See description of PCR above.
Serological tests have limited sensitivity in individuals, but may be useful in population studies. See description of and caveats about DFA above. Culture had been the reference standard against which all other tests have been compared; however, with development of PCR it has been realized that culture is only about 75% sensitive at best. H. ducreyi is difficult to transport and culture. Cultures require highly enriched media. Not practical for BBS.
52
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Non-HIV biomarker Herpes simplex virus-2 (HSV-2)
Test
Usefulness HSV-2 is a causative agent of genital herpes and a cofactor of HIV infection; indicator of population risk for ever having unprotected sex. Estimation of HSV-2 prevalence may be most useful for surveys in young age groups.
EIA (for HSV antibody) and other serological tests
EIAs can aid in the diagnosis of HSV: a positive result suggests recent infection, but cannot distinguish between primary or reactivated infection. A negative result does not rule out infection. Costly and difficult; not useful in screening. Several type-specific serological assays exist, although some cross-reactivity may occur between HSV-1 and HSV-2. Serology is preferred for screening populations and tests perform reasonably well on a population level compared with individual diagnosis. DFA testing requires scraping cells from the ulcer base in symptomatic patients. Can be useful in distinguishing HSV-1 from HSV-2, but not practical for surveys. PCR is sensitive and specific in symptomatic patients, and can be used to detect asymptomatic viral shedding in infected persons. Not useful for screening. Rarely used. Costly and difficult. Specimen collection must be done during acute phase of infection when skin lesions are present. Not practical for surveys. HBV prevalence – indication of acute or past infection, chronic carriage, or immunization depending on tests used (see Appendix I-12); prevalence of HIV-HBV coinfection. Assess impact of prevention, care, control and treatment programmes.
DFA (direct fluorescent antibody) PCR Western blot Viral culture Hepatitis B virus (HBV)
EIA Nucleic acid testing Hepatitis C virus (HCV) HCV RNA
HBsAg, Anti-HBc, IgM anti-HBc, Anti-HBs HBV DNA HCV prevalence; common among PWID; prevalence of HIV-HCV coinfection. Assess impact of prevention, care, control and treatment programmes. May be used as indicator of injecting drug use, but need to understand current and historical modes of HCV transmission in the local setting. With the availability of HCV treatments, monitoring the HCV testing, care and treatment cascade is becoming more important.
EIA or rapid diagnostic tests (RDTs) HCV RNA Sputum-smear microscopy
Detect anti-HCV antibody (cannot distinguish new or reinfection, chronic infection or past resolved infection). Biomarker for viremia; indicates current or chronic HCV infection; distinguishes those with current infection from those whose infection has resolved
Mycobacterium tuberculosis
Causative agent of tuberculosis (TB); prevalence of TB or TB disease among those with HIV; TB disease among those with HIV indicates severe immunosuppression and is an AIDS-defining condition. Most BBS cannot achieve sample sizes large enough to estimate TB incidence or prevalence in a population, but testing may be appropriate in some instances, such as in BBS conducted among prisoners.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
53
Non-HIV biomarker
Test • Sputum smear microscopy: conventional light microscopy used to examine direct smears stained with Ziehl–Neelsen, with or without specific sputum-processing methods; or • Fluorescence microscopy; or • Rapid molecular test with high accuracy, such as the Xpert MTB/RIF test (Cepheid, Sunnyvale, CA); or • Any rapid test recommended by WHO
Usefulness • Sputum smear microscopy (including fluorescence microscopy) is highly specific in the diagnosis of pulmonary TB among persons in areas with a very high prevalence of TB and can identify those who are most infectious. Sensitivity of detection is low, however, for sputum specimens with fewer than 10,000 bacteria/mL specimen and among those with extrapulmonary TB and in persons with HIV-TB coinfection. • Molecular tests provide rapid results for diagnosis of TB. The Xpert MTB/ RIF assay contributes to the rapid diagnosis of TB disease and rifampin drug resistance.
Prostatespecific antigen
Immunochromatographic A marker of seminal fluid indicating recent unprotected sexual intercourse. rapid test on vaginal or rectal swabs A marker for unprotected sex among females. The detection of YCS or lack thereof determines the presence or absence of male chromosomes in the vaginal swab sample. The assay has been shown to be 92–98% specific for YCS detection within 2–3 weeks of unprotected heterosexual intercourse (3). • Validate self-reported drug use; estimate proportion of population using drugs, by type. • Tests may use blood, urine or hair, with detection times (since last drug ingestion) varying by drug and specimen type. • Rapid tests exist that show reasonable accuracy and can test concurrently for multiple analytes
Y-chromosomal Qualitative PCR assay sequences (YCS)
Illicit drugs (e.g. amphetamines, cannabis, cocaine, opiates)
Anti-HBc, hepatitis B core antibody; Anti-HBs, hepatitis B surface antibody; HBsAg, hepatitis B surface antigen; IgM, immunoglobulin M; TP, T. pallidum
A-6.9 Selection of tests and testing algorithms Biomarkers are among the most important survey data collected, because they typically relate to a survey’s primary and secondary objectives. Table A-6.3 provides information about various STIs and their tests that can be used in selecting STIs for surveys. Appendix I-10 compares specimen types, advantages, limitations, cost and complexity of the two most commonly used HIV testing technologies, enzyme immunoassays (EIAs) and rapid tests. If participants are to receive HIV test results on the same day they
participate in the survey, a rapid HIV testing algorithm will probably be required. The first test performed should be highly sensitive (to detect all true positives), and should be followed by a second highly specific test to identify any true negative specimens falsely identified as positive (i.e. false positive) by the first test. Some national algorithms require a third or confirmatory test. Additional guidance on designing HIV test algorithms for different surveillance purposes and in different epidemiological contexts is provided in the Joint United Nations Programme on HIV/AIDS (UNAIDS)/WHO Guidelines for using HIV testing technologies in surveillance: selection, evaluation, and implementation – 2009 update (4). These guidelines also provide information on the qualification of HIV and other tests to be used for diagnosis (5).
54
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-6.10 Development of standard operating procedures Detailed SOP for sampling, transport, testing and data collection should be prepared for each test and type of specimen collected, ideally in consultation with laboratory technicians with expertise in the tests. Laboratory staff may adapt SOP from the package inserts of commercial test kits. SOP should include details about how specimens will be collected, processed, stored and transported, if applicable. Additionally, protocols and SOP should outline how and to whom test results will be returned. SOP also should address specimen disposal and adherence to biosafety guidelines. A sample SOP is provided in Appendix I-30. Training of survey staff should include exercises to identify and address unforeseen challenges.
Blood Blood – in the form of serum, plasma or dried blood spot (DBS) – is the preferred specimen for HIV testing because it has a higher concentration of HIV antibodies than oral fluid. It also allows for additional testing; for example, for viral load, recent infection, HIV type and subtype, and HIV drug resistance, as well as syphilis and hepatitis. Blood can be collected through finger stick or venipuncture. Generally, participants are resistant to having more than one blood draw or finger stick (pricking). In addition to the discomfort, double-pricking may also compromise confidentiality because it may indicate that an infected person requires a second test. A single finger stick often yields only 200 μL of blood or less, although some systems may facilitate collection of up to 500 μL. Collecting venous blood using evacuated (vacuum) blood tubes is therefore more productive. With this approach, relatively large volumes (5–10 mL or more) of blood can be collected, meaning that multiple tests can be performed. Collecting extra blood is useful in case tests have to be repeated for quality control, or to store specimens for future unspecified testing. The use of a needle and syringe and subsequent transfer to tubes is not encouraged because of biosafety issues and the potential for compromising the quality of the specimen (e.g. hemolysis). Evacuated tubes containing the anticoagulant ethylenediaminetetraacetic acid (EDTA) are especially useful because they yield both cellular and noncellular (plasma) fractions. Vacuum tubes without anticoagulant (i.e. untreated) should be used if serum is desired. Aliquots of plasma or serum should be prepared shortly after collection, then refrigerated, transported and frozen at the central laboratory; alternatively, they may be frozen on-site and transported on freezer packs to the central laboratory without thawing. Repeated freezing and thawing should be avoided because it may compromise specimen quality. Blood collected for CD4+ T-cell counting needs to be stored at ambient temperature. Special “long-life” blood tubes are available that stabilize the CD4+ T-cells up to 7 days, allowing for less stringent time requirements for specimen transportation and testing. DBS are easy to prepare and they provide a backup specimen in case plasma specimens are lost or spoiled. However, DBS are not the ideal specimen type for HIV serology using EIA. Some EIA kit manufacturers have protocols for DBS but, in general, protocols have to be adjusted (optimized) because of the high background optical density readings associated with serum eluted from DBS. Moreover, DBS may not be suitable for every biomarker (i.e. CD4+ T-cells).
A-6.11 Specimen type, collection, labelling, processing and transport Survey biomarkers are commonly based on blood specimens (e.g. HIV, syphilis and HSV). Depending on the test, biomarker and target population, other specimens collected may include urine, oral fluid, vaginal swabs, rectal swabs and oropharyngeal swabs. Less commonly collected specimens (e.g. urethral swabs or endocervical swabs) may require a clinician to collect them; such specimens may be useful in symptomatic populations or among key populations at high risk for STI. Biosafety and biosecurity precautions must be strictly adhered to during specimen collection, handling, testing, storage and waste disposal, even where surveys are conducted in informal settings such as truck stops, bars and nightclubs, or rented apartments.
A-6.11.1 Specimen type and collection Appendix I-10 provides a list of specimen types for HIV testing and the advantages and disadvantages of each. Table A-6.4 provides a list of optimal specimen types for the detection of Chlamydia trachomatis and Neisseria gonorrhoeae by nucleic acid amplification test (NAAT) type. Comprehensive information on laboratory methods for detecting Chlamydia trachomatis and Neisseria gonorrhoeae and other STIs can be found in Recommendations for the laboratory-based detection of Chlamydia trachomatis and Neisseria gonorrhoeae (6), and Strategies and laboratory methods for strengthening surveillance of sexually transmitted infections (7).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
55
Table A-6.4 Blood specimens – collection, fractions and use Blood specimen type 56 Venous blood Capillary blood *
HIV drug resistance
Viral hepatitis
HIV rapid test
HIV DNA PCR
Western blot
HIV recency
Other EIA
Viral load
Syphilis
HIV EIA
Collection container
CD4+
HSV
Blood fraction
Anticoagulated whole blood Vacuum, EDTA Plasma Dried blood spots* Vacuum, EDTA, preservative Vacuum, no anticoagulant Capillary or microcollection tube Anticoagulated whole blood Serum Anticoagulated whole blood Dried blood spots*
Prepared from whole blood on Whatman 903 filter paper cards EDTA, ethylenediaminetetraacetic acid; EIA, enzyme immunoassay; HSV, herpes simplex virus; PCR, polymerase chain reaction
Table A-6.4 outlines type of blood specimens, methods of collection, and use for measuring HIV biomarkers. Additional guidance on the selection of specimen types is provided in the UNAIDS/WHO Guidelines for using HIV testing technologies in surveillance: selection, evaluation, and implementation – 2009 update (4). For more specific information on specimen transportation and storage guidelines, refer to the WHO Guidelines for the safe transport of infectious substances and diagnostic specimens (8). A laboratory assessment should be performed before BBS implementation to ensure quality standards, followed by implementation of a quality assurance and monitoring plan to assure testing quality during survey implementation (see Section A-6.15). Oral fluid Oral fluid (saliva) testing is less invasive than venipuncture, and can be useful for survey situations where it may be difficult to collect blood specimens. However, there are limitations to this approach. For example, if more than one oral fluid test is used, specimen collection must occur at least 30 minutes apart; also, oral fluid can only be used with certain EIAs and rapid tests specifically designed for oral fluid specimens. Oral fluid rapid
HIV testing is sometimes performed in challenging survey environments where participants with reactive tests are counselled to seek additional testing for diagnostic confirmation elsewhere. Investigators may, however, choose to use the results for data analysis, without waiting for results of confirmatory tests. In general, oral fluid specimens may be stored at 4oC to 37oC up to 21 days. This temperature should be maintained during shipment. Specimens should be frozen (–20 oC or below) if stored for an extended time. Once thawed, specimens can be refrozen and thawed only once more. Urine Urine is the CDC-recommended specimen for use in NAATs to detect Chlamydia trachomatis and Neisseria gonorrhoeae in men (6). Because the bacteria in firstcatch urine specimens are remarkably similar to those found in paired urethral swab specimens, urine can be used when swabs are undesirable, such as in populationbased studies or where multiple sampling of participants is required (9). For BBS testing for Chlamydia trachomatis and Neisseria gonorrhoeae, survey investigators may choose to collect urine from men and from women;
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
however, because the sensitivity of vaginal swabs for detecting Chlamydia trachomatis and Neisseria gonorrhoeae among women is greater than that of urine, swabs are considered the optimal specimen type for women (6). Vaginal and endocervical swabs For NAAT detection of Chlamydia trachomatis and Neisseria gonorrhoeae in women, vaginal swabs are considered the optimal specimen because they have greater sensitivity than first-catch urine specimens, and are as sensitive and specific as cervical swab specimens (6). The sensitivity and specificity of self-collected vaginal swabs are equal to or better than swabs collected by a clinician.1 Some women may find self-collection more acceptable. Also, self-collection reduces staffing requirements for the survey. Formative assessment should include questions about willingness to provide specimens and which collection method is preferred. Vaginal swabs do not replace cervical exams for diagnosis of female urogenital infection; women may have cervicitis, urethritis, or urinary tract or vaginal infections due to causes other than Chlamydia trachomatis or Neisseria gonorrhoeae. Rectal swabs For MSM, rectal swabs should be collected for NAAT detection of Chlamydia trachomatis and Neisseria gonorrhoeae. Depending on the context, rectal swabs
may be collected among female SW (FSW) as well. These can be collected by a clinician or self-collected. As with vaginal swabs, formative assessment should probe the target population’s willingness to provide rectal swab specimens and their preferred method of collection. Oropharyngeal swabs NAATs are the recommended testing method for detection of Neisseria gonorrhoeae in oropharyngeal specimens. Nonharmful Neisseria species commonly found in the oropharynx might cause false positive reactions in some NAATs, and further testing, such as by polymerase chain reaction (PCR), might be required for accuracy. Chlamydia trachomatis is not considered to be a clinically important cause of pharyngitis. Urethral swabs Urethral swabs historically have been the specimen of choice for STI diagnosis in men. Urethral swab collection, however, causes considerable discomfort and has been identified as a disincentive for routine STI screening (4). Furthermore, urethral swab collection requires a clinician or trained health worker. Urine is the CDC-recommended sample type for NAAT-based diagnostics for Chlamydia trachomatis and Neisseria gonorrhoeae (6). Urethral swabs are still recommended for testing men for Trichomonas vaginalis. Table A-6.5 shows the ideal specimen method for each type of STI screening.
Table A-6.5 Optimal specimen types for BBS STI screening using nucleic acid amplification tests Organism Women
Optimal specimen
Alternative
Reduced sensitivity
Chlamydia trachomatis Neisseria gonorrhoeae Trichomonas vaginalis
Vaginal swab Vaginal swab Vaginal swab Endocervical swab First-catch urine
Endocervical swab Endocervical swab
First-catch urine First-catch urine
Men
Chlamydia trachomatis Neisseria gonorrhoeae Trichomonas vaginalis Chlamydia trachomatis Neisseria gonorrhoeae
First-catch urine First-catch urine Urethral swab
Urethral swab Urethral swab
SW, MSM, TG (specimens in addition to those above) Rectal swab Rectal swab Oropharyngeal swab
a
Adapted from ARUP Laboratories 2014 (10)
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
57
A-6.11.2 Specimen labelling All specimens (e.g. aliquots, DBS cards and rapid testing devices) must be labelled with the participant’s unique identification number (ID) or a laboratory ID that is linked with the participant ID. The label is required for linking biological and survey data, and for specimen storage, tracking and retrieval. Careful planning of labelling procedures is recommended: a single 7 mL venous blood specimen drawn into a plain or anticoagulant tube may result in one or two 1.5 mL aliquots of serum or plasma and a five-spot DBS card, each of which require a label. Labels should be preprinted before the survey begins. The number of labels per participant depends on the number of specimens and forms used. Extra labels should be printed for unanticipated needs. Ideally, labels are barcoded so that laboratory and survey staff can register specimens with a barcode reader to minimize manual data entry and associated errors. Preprinted labels that withstand humidity and freezing (down to at least –20 oC, and perhaps as low as –80 oC depending on where specimens will be stored) are the best option. Otherwise, a permanent marker should be used to label specimens. At a minimum, specimens should be labelled with the participant ID and date of collection. If not indicated on the specimens themselves, a record should be kept of which specimens may be stored for future testing and which must be discarded. Temporary storage on-site and processing Specimens may be stored temporarily in refrigerators or cooler boxes with cold or freezer packs. Specimen storage requirements vary by type of specimen. For example, DBS specimens should not be refrigerated immediately after collection, but held at room temperature and dried thoroughly before packaging. DBS should be packaged in a zip-lock bag containing a humidity indicator card and desiccant packets and stored in a refrigerator (4–8 oC) for up to a month or in a freezer (–20 oC) if available for longterm storage. If serum or plasma specimens will undergo any molecular testing (e.g. genotyping), they should be processed and frozen within 24 hours. Handling, storage and transport of specimens collected for NAAT detection of Chlamydia trachomatis and Neisseria gonorrhoeae testing varies by test kit manufacturer. Urine specimen storage requirements will vary by NAAT type and type of container (primary cup or transport tube). For example, urine is one type of NAAT that may be held in a primary cup at room temperature for up to 24 hours for female urine specimens and up to 3 days for male specimens; the holding time for other NAAT types ranges from 24 hours to 30 days at 2 oC to 30 oC (6). Thermometers should be used to monitor storage temperature daily. Thermometers that record the minimum and maximum temperature are useful.
A-6.11.3 Transport Specimens sent to the laboratory should be packaged and adequately stored during transportation (e.g. cooler boxes with or without ice packs, or liquid nitrogen), according to requirements of the assays to be performed. They should be accompanied by transportation or shipping forms, and chain-of-custody forms (Appendices I-13 and I-14). To ensure proper handling of specimens upon arrival, dates of shipping and receiving should be scheduled taking into account the receiving laboratory’s operating hours. When coordinating local transport, investigators should consider whether the receiving laboratory will be staffed to process specimens at the time of their arrival. Blood collected in tubes with anticoagulant needs to be separated by centrifugation and the plasma frozen within 24 hours of collection for any molecular testing. In addition to being trained on proper specimen handling procedures, staff transporting fresh specimens should be trained on relevant biosafety and biosecurity procedures. Specimen registration at the laboratory When specimens are received by the laboratory, they should be cross-checked against the shipping form, registered into the laboratory data-management system (or logbook), assigned a unique laboratory ID that is linked to the participant ID (if applicable), and stored appropriately until used for testing. The laboratory data-management system should ensure specimen tracking and linkage of laboratory and participant IDs. Any nonconformity should be documented; for example, too high or low temperature, hemorrhage, bacterial overgrowth, empty or insufficient specimen volume, unsealed bags or missing labels. Laboratory support BBS should be supported by a laboratory, ideally one accredited by an international accreditation agency. Staff at the laboratory should participate in all phases of the survey from design to results dissemination (e.g. final report and manuscript writing). Countries without accredited laboratories may have a national reference laboratory, a national public health laboratory, or a university or hospital laboratory that may support surveys. The survey laboratory should have in place robust quality management systems and the capacity to perform at least basic diagnostic testing (e.g. serology testing). The laboratory may receive, process, store and test specimens, either to provide primary results for the survey, or to conduct quality-control testing or future testing. It may also provide support for on-site laboratory procedures by providing training, supplies and supervision.
58
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-6.12 Laboratory data management Survey investigators should ensure that a system (preferably electronic) is in place for managing data; for example, cataloguing specimens, recording results, storing raw laboratory data and recording qualitycontrol results. Survey laboratory results (e.g. line-listed optical density values, viral RNA copies/mL and qualitycontrol test results) can be collected and merged with other survey data so that output for all the assays are maintained in the final database rather than just the final positive or negative interpretations. This will facilitate data review and troubleshooting, should testing issues arise.
To store leftover survey specimens for future unspecified testing, survey investigators must obtain informed consent from survey participants. Participants must also be informed as to whether test results from future testing will be returned to them. The survey protocol should specify the type (e.g. plasma or DBS) and amount (e.g. volume or number of spots) of specimens to be stored. If a specimen repository is important to the BBS objectives, then ideally, two sets of specimens should be prepared – one for immediate testing, the other for storage and future testing. Long-term specimen repositories must be continuously managed. For example, DBS specimens must be inspected every 6 months and their desiccants replaced as necessary. Use of a centralized laboratory information system may help with tracking and maintenance of long-term specimen collections. Temperature should be monitored.
A-6.13 Laboratory procedures Even when testing of specimens is not being done at the survey site, BBS usually require on-site procedures for specimen collection and processing at the very least. Quality-assurance measures need to be considered, including training, use of quality-control materials, periodic review of testing data, proficiency testing panels, and retesting of a subset of specimens at a designated reference laboratory. Traditionally, HIV and STI testing have been conducted within the confines of a standard laboratory; however, there is a clear trend towards conducting on-site rapid testing without the need for sophisticated equipment, and inclusion of point-ofcare testing (e.g. for CD4+ T-cells counts and viral load) with increasingly small and portable computerized equipment. A private room – or other designated location in the case of time-location sampling (TLS) – should be reserved for on-site laboratory procedures. On-site infrastructure may include refrigerators or cooler boxes (for temperature-sensitive reagents or specimen storage); sharps, biohazard and biowaste disposal containers; centrifuges; mobile phones; and tables and chairs. Appendix I-2 provides examples of various cost categories, including a list of laboratory commodities and equipment.
A-6.15 Quality assurance (assessment) and quality control A laboratory quality management system is crucial to ensure correct testing results. For BBS, some of the important elements of a quality management system include documentation of procedures and records management, SOPs, external quality assessment (assurance), use of quality-control specimens, and quality-control retesting of a sample of survey specimens. All laboratories should employ quality-assurance practices. Many laboratories participate in external quality assessment; that is, an evaluation of the testing process by an outside organization (e.g. a national reference laboratory). Such assessment can help to identify existing or potential problems, and activities include: • site visits with direct observations, and review of testing procedures; • quality control of survey test results by retesting a subset of specimens already tested (e.g. all positive and 5–10% of negative specimens retested by the national reference laboratory); and • proficiency testing of all survey staff performing HIV or other testing. Proficiency testing involves testing a panel of blinded samples sent to testers at testing sites and laboratories by a recognized proficiency testing programme provider. These samples are sent to the sites on a pre-defined frequency (2–3 times a year). Test results are submitted to the programme provider to be analysed for accuracy of testing. Poor performance is then addressed by retraining, increasing supervision or taking other corrective actions. Given the relatively short duration of
A-6.14 Specimen repository Survey protocols and consent forms often include text to allow for long-term storage of specimens for future unspecified needs. Protocols should specify a minimum and maximum storage period, and define who will control the use of the specimens. Investigators should consult relevant ethics review committees about the future use of specimens. Storage space is another important consideration. Many laboratories have limited space; thus, storage space needs to be secured in advance.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
59
most BBS, survey sites that conduct testing should be monitored closely. Proficiency testing samples should be distributed monthly or perhaps every 6 weeks for the duration of the survey.
A-6.16 Supply chain considerations The following considerations apply to supply management: • the number of tests and consumables that need to be ordered should be determined. • the amount of hardware equipment (e.g. testing instruments) needed per survey site should be determined. • when procuring commodities, expiration dates should be considered. Some commodities, especially testing reagents, have a short shelf life (a few months) so they should not be ordered too early or in too large a quantity. Partial deliveries may need to be arranged with vendors or suppliers. Extra stock (10–20%) should be ordered to account for training, repeat testing or losses. To minimize wastage of perishable materials, they should not be ordered until all required approvals (not just ethical approvals) have been obtained. • supplies that need to be ordered from abroad often have a long delivery time and may be subject to or delayed by customs. • all key equipment (e.g. pipettes, enzyme-linked immunosorbent assay [ELISA] machines, refrigerators or freezers and incubators) should be well maintained and in good working condition. • proper storage of supplies includes keeping them in a clean, organized locked shelf or cabinet and stored according to the manufacturer’s instructions in a wellventilated room away from direct sunlight. Supplies should be organized by expiration dates so that older supplies are used first (i.e. first expiry, first use).
• labels, printers and barcode readers should be procured, or the availability of preprinted labels ensured, before initiating the surveys. • the United States Agency for International Development (USAID) Laboratory Logistics Handbook, which outlines how a supply chain should be designed and managed to ensure the availability of commodities, may be a useful resource (11).
A-6.17 Biosafety and biosecurity Biosafety and biosecurity measures are intended to assure the safety of participants, health-care workers and laboratory workers, and should be applied at all stages of laboratory activities, from specimen collection to laboratory testing and specimen storage. “Laboratory biosafety” describes the containment principles, technologies and practices that are implemented to prevent unintentional exposure to pathogens and toxins, or their accidental release (12). “Laboratory biosecurity” describes the protection, control and accountability for valuable biological materials within laboratories that prevent their unauthorized access, loss, theft, misuse, diversion or intentional release (13). Laboratory managers and survey investigators are responsible for ensuring biosafety and biosecurity. The following should be available where specimens are handled (e.g. collected, processed, transported, tested and stored): protective gloves, laboratory coats, eye protection, waste bins for biohazardous materials, sharps containers, pipette tip disposal containers, regular waste containers, alcohol, bleach and other disinfectant solutions. Arrangements must be made for transport of biohazardous materials for subsequent incineration or autoclaving and re-use of suitable instruments. Access to post-exposure prophylaxis should be available in the event of accidental exposure to infectious agents (e.g. HIV) by survey staff.
A-6.18 References 1 Hauser RM, Weinstein M, Pool R, Cohen B. Conducting biosocial surveys: collecting, storing, accessing, and protecting biospecimens and biodata. National Academies Press. 2010 (http://www.nap.edu/catalog/12942.html, accessed 26 August 2016). Dicker LW, Mosure DJ, Levine WC, Black CM, Berman SM. Impact of switching laboratory tests on reported trends in Chlamydia trachomatis infections. Am J Epidemiol. 2000;151(4):430–435 (http://www.ncbi.nlm.nih.gov/ pubmed/10695602, accessed 26 August 2016). Ghanem KG, Melendez JH, McNeil-Solis C, Giles JA, Yuenger J, Smith TD et al. Condom use and vaginal Y-chromosome detection: the specificity of a potential biomarker. Sex Transm Dis. 2007;34(8):620–623 (http:// www.ncbi.nlm.nih.gov/pubmed/17308500, accessed 26 August 2016).
2
3
60
6. Biomarker considerations
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
4
UNAIDS/WHO Working Group on Global HIV/AIDS and STI Surveillance. Guidelines for using HIV testing technologies in surveillance – 2009 update. Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2009 (http://apps.who.int/iris/bitstream/10665/164358/1/9789241598057_eng.pdf?ua=1, accessed 30 August 2016). WHO. HIV assays operational characteristics: HIV rapid diagnostic tests (detection of HIV-1/2 antibodies): report 17. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/hiv/pub/vct/rapid-diagnosis-testsreport17/en/, accessed 10 August 2016). Centers for Disease Control and Prevention. Recommendations for the laboratory-based detection of Chlamydia trachomatis and Neisseria gonorrhoeae – 2014. MMWR Recomm Rep. 2014;63(RR-02):1–19 (http://www.ncbi. nlm.nih.gov/pubmed/24622331, accessed 30 August 2016). WHO. Strategies and laboratory methods for strengthening surveillance of sexually transmitted infections. Geneva: World Health Organization (WHO); 2011 (http://apps.who.int/iris/bitstream/10665/75729/1/9789241504478_ eng.pdf, accessed 30 August 2016). WHO. Guidelines for the safe transport of infectious substances and diagnostic specimens. Geneva: World Health Organization (WHO); 1997 (http://www.who.int/csr/emc97_3.pdf, accessed 30 August 2016). Dong Q, Nelson DE, Toh E, Diao L, Gao X, Fortenberry JD et al. The microbial communities in male first catch urine are highly similar to those in paired urethral swab specimens. PLoS One. 2011;6(5):e19709 (http://www.ncbi.nlm. nih.gov/pubmed/21603636, accessed 30 August 2016).
5
6
7
8 9
10 ARUP Laboratories. Sample collection for the diagnosis of STD using nucleic acid amplification tests: Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis. Salt Lake City: ARUP Laboratories; 2014 (https:// www.aruplab.com/Specimen-Handling/resources/pdf/SampleCollectionForSTD.pdf, accessed 30 August 2016). 11 United States Agency for International Development. Laboratory logistics handbook: A guide to designing and managing laboratory logistics systems. Arlington, VA: USAID; 2009. 12 WHO. Laboratory biosafety manual. World Health Organization (WHO). 2004. 13 WHO. Biorisk management: Laboratory biosecurity guidance. Geneva, World Health Organization (WHO). 2006 (http://www.who.int/csr/resources/publications/biosafety/WHO_CDS_EPR_2006_6.pdf, accessed 30 August 2016).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
6. Biomarker considerations
61
Section
A
7. Preparing biological data-collection instruments
This chapter focuses on the data instruments used for the collection, processing, testing, transport and storage of biological specimens. Procedures for each of these activities vary; for example, because of differences in survey protocols, choice of assays and algorithms, the site of specimen collection and testing, and whether specimens will be stored for future testing. Survey investigators should prepare a detailed flow chart to illustrate each step. Figure A-7.1 gives an example of a flow chart.
Figure A-7.1 Sample flow chart illustrating the process for rapid HIV testing at a survey site, ranging from obtaining consent to providing HIV test results Participant presents Participant Checklist Form 2 Laboratory technician: 1. Verifies informed consent. 2. Labels blood collection tube with PIN.
Key terms Algorithm: A step-by-step procedure. Biomarker: A biological characteristic measured through laboratory or other tests, generally using a biological specimen such as blood for testing; examples include HIV antibody or viral load. Calibrate: To compare and adjust readings of an instrument with those of a standard; adjust experimental results to take external factors into account or allow comparison with other data (source: adapted from Oxford Dictionary). Chlamydia Causative agent of chlamydia, a bacterial trachomatis: sexually transmitted infection. Molecular Testing for nucleic acid sequences (e.g. testing: DNA or RNA) as opposed to immunologic testing for antigens or antibodies. Neisseria Causative agent of gonorrhoea, a bacterial gonorrhoeae: sexually transmitted infection. Optical Intensity of colour as measured by a density: spectrophotometer. In some enzyme immunoassays (EIAs) that detect HIV antibodies, more intense colour (higher optical density) generally indicates more antibody, although EIAs are designed to be qualitative, providing an answer of reactive or nonreactive.
Laboratory technician:
Participant provides blood specimen
1. Provides pre-test counselling. 2. Collects blood specimen, places in rack, allows to separate at room temperature. 3. Informs participant that result will be ready before participant leaves the RDS survey site. 4. Ticks testing boxes on Participant Checklist Form 2. 5. Gives Form 2 back to participant. 6. Directs participant to the coupon manager.
Laboratory technician: 1. Labels microtube with PIN. 2. Prepares serum aliquots after serum and clot have separated. 3. Uses an aliquot to perform HIV rapid test(s) following IBBS testing algorithm. 4. Records results in log book and on test result form that will be filed in participant folder. 5. Places aliquots in cool box for transport to reference laboratory for quality control and future testing. 6. Disposes of tube containing clot in biohazard container.
Laboratory technician processes and tests blood specimen
Coupon manager directs participant to laboratory technician.
Participant receives HIV rapid test results
Laboratory technician: 1. Verifies that participant’s PIN matches PIN on test result form. 2. Provides HIV test result and post-test counselling. 3. Provides participant with referral information for HIV services. 4. If results are indeterminate, instructs participants where they should go for a repeat test in 4 weeks.
In some countries a tie-breaker test may be used, in others the result may be considered indeterminate
62
7. Preparing biological data-collection instruments
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-7.1 Specimen tracking forms Specimens are often collected at one site and transported to another for testing and storage. Paper or electronic forms are used to track specimens from the point at which they are collected through processing, testing, shipping, storage and final disposal. To describe specimen integrity, survey staff should maintain a record of the dates and times specimens were collected and the temperatures at which they are stored. This is especially important for tests that require specimens to be in good condition (e.g. not exposed to extreme temperatures or multiple freeze–thaw cycles). Testing standard operating procedures (SOPs) and test kit inserts include guidance for optimal specimen handling and storage processes. Specimen tracking forms should include the following information: • survey participant ID; • specimen ID (if different from participant ID); • date of specimen collection; • survey site location; • type of specimen; • date specimen stored in cool box, refrigerator or freezer at site; • storage temperature during transport – minimum and maximum – and transport conditions, (e.g. on dry ice, wet ice or cool packs, or at room temperature); • date specimen transported to another site (e.g. off-site or referral laboratory) for testing or storage; • name and location of second site (e.g. off-site or referral laboratory); • dates and times at which the laboratory received, processed, tested and stored the specimen; • number and volume of specimen aliquots or dried blood spot (DBS) cards or spots; • physical location and temperature of storage (e.g. refrigerator no. 2 or freezer no. 3) so that specimens can be easily retrieved; • initials or other form of identification for each person who handled the specimen at each step; and • other information locally determined to be needed. Appendix I-13 provides an example of a specimen tracking form.
Logs may include the following information: • initials or other form of identification of person (e.g. survey manager or laboratory technician) who handed specimens over and person (e.g. driver) who received them; • date and time specimens picked up by transport; • temperature of cool boxes (if used); and • date and time specimens delivered to laboratory or storage site. Some of these data (e.g. time and temperature) are critical for some specimens and assays (e.g. urine for Neisseria gonorrhoeae or Chlamydia trachomatis molecular testing). For consistency, and to allow for testing of specimens in the future, temperature monitoring should be performed for all specimens, even those that can be transported at room temperature. Appendix I-14 provides an example of a specimen transport driver log.
A-7.3 Refrigerator or freezer temperature logs Temperature monitoring should be performed at the start of each workday at any location (survey site or off-site referral laboratory) where specimens or temperaturesensitive reagents are stored. Temperature should be monitored using calibrated and certified laboratory thermometers. In spaces that experience larger temperature fluctuations (e.g. a cool box), temperatures should be monitored using special thermometers that record the minimum and maximum temperatures experienced in the container during the time specimens were being held. Appendix I-16 provides examples of specimen temperature monitoring log.
A-7.4 Test results form Just as survey questionnaires are used to record responses to questions, the test results form is used to record results of specimen testing for biomarkers. These testing data may be entered into a separate database from the survey interview database. The two files can be merged using the survey participant identification number (ID), to link biomarker test results with interview data.
A-7.2 Specimen transport logs Specimen transport logs are used to record the movement of specimens from one site (e.g. the collection site) to another (e.g. the off-site or referral laboratory). This form accompanies the specimens. A copy should be made and kept at the site where other survey information is maintained.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
7. Preparing biological data-collection instruments
63
Some laboratory testing machines automatically produce assay results and other information. This automation reduces transcription errors and allows survey staff to export testing data for merging with other data files (e.g. the interview data). The test results form (electronic or paper) should include the following: • survey participant ID number; • assay name, lot number and expiration date; • name or other form of identification of person performing each test; • date testing performed; • results from specimen testing for biomarkers; - optical density values, if applicable; - titres, if applicable; - individual assay and final test (algorithm) result; - test quality control and calibrator results, if applicable; • comments;
- notes about specimen integrity (e.g. gross hemolysis or bacterial overgrowth); - notes about specimen volume (e.g. amount of plasma received for testing, or quantity not sufficient); • date result reported to survey site or survey investigators, if applicable (e.g. testing performed offsite); and • date test result provided to participant, if applicable (e.g. on-site test results and same-day test result provision). Appendix I-17 provides an example of a test results log.
64
7. Preparing biological data-collection instruments
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms
8. Eligibility for participation
Participant eligibility for a BBS usually depends on age, recency of a risk behaviour, sex, and city of work or residence. Depending on the survey’s objectives, potential participants may have to meet other criteria as well. This chapter defines the target population and shows how screening for eligibility yields the survey sample. Figure A-8.1 The relationship between target population, survey population and eligible participants
Eligibility The description of all the requirements a criteria: potential survey participant must meet in order to participate. Eligibility Asking potential participants a series of screening: questions to determine survey eligibility.
Target population
A-8.1 Defining the population The target population is the group of people about which investigators are interested in drawing conclusions. This population is generally described broadly – for example, “men who have sex with men (MSM) in Mexico City” – and is not specific enough to provide guidance about who should be “allowed” to participate in the survey. For example, the above definition may include MSM who are not eligible to participate in the survey, including those that are not currently sexually active or who are too young to provide informed consent. The survey population is a subset of the target population narrowed to account for practical considerations related to sampling methodology and eligibility criteria. In a timelocation sampling (TLS) survey, this population might include men who congregate in venues where MSM are known to socialize and who are randomly selected by survey staff. In a respondent-driven sampling (RDS) survey, it would be men who are identified as MSM by other MSM, who are referred by a previous participant, and have access to the survey site. The survey sample is the subset of the survey population that actually participates in the biobehavioural survey (BBS).
Survey population (screened for eligibility)
Screened eligible participants
An individual must meet the eligibility criteria for a survey in order to participate. For example, a BBS conducted among female sex workers (FSW) may define eligibility as aged 15 years or over who have exchanged sex for money or goods at a brothel in the past 6 months. For MSM, the definition may include men aged 15 years or over who have had anal sex with other men in the past 6 months. For people who inject drugs (PWID), eligibility may be defined as aged 15 years or over and having injected drugs within the past 30 days or the past 6 months. Eligibility criteria ideally should be precise, time bound, measurable and reliable. “Precise” means that the information given by the potential participants can be used to accurately describe the survey participants. “Time bound” means that eligibility based on a certain behaviour is limited to a specific time frame, so that the potential participants can accurately recall and describe it.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
8. Eligibility for participation
65
“Measurable” means a criterion can be enumerated; for example, participant age in years. “Reliable” means the information provided by potential participants can be verified if needed. It can be difficult for eligibility criteria to fulfill these standards. Survey staff have to be well trained in order to spot potential participants who may not tell the truth.
ethical considerations of survey conduct and the human subjects review process. Sex BBS are typically restricted to one sex, with the exception of surveys of PWID, which often include both sexes. For MSM and transgender person (TG) surveys, investigators need to weigh the merit and costs of combining these populations into one sample or keeping them separate, which often results in one population not being surveyed. Male and female sex workers (SW) may represent distinct HIV epidemics and thus are better sampled in separate surveys. For surveys that include both sexes, eligibility screening should include the sex of the potential participants so investigators can determine if each sex is proportionally represented in the final sample. Risk behaviour To be eligible to participate, potential participants must engage in the risk behaviour being studied. For example, a survey of PWID should not include noninjecting drug users. A survey of women who sell sex for money should not include those who only engage in transactional sex. If investigators are interested in sampling those who engage in transactional sex as well, the eligibility criteria should be altered accordingly. Similarly, a survey among MSM should not include men who identify as gay but have not had sex with another man. During eligibility screening, potential participants should be asked questions that help verify their eligibility. For example, in a PWID survey, they may be asked to show track marks or to describe the injection process. Screening staff may probe MSM or SW for their knowledge of locations associated with the respective behaviour or terms used by the population; for example, venues they attend or terms they use to denote certain behaviours or groups of people. Time period Narrowing the participants to those who have engaged in specific behaviours within a specific time period allows investigators to focus on those who have engaged in risk behaviours most recently. Participants who have engaged in the risk behaviour recently will also be better able to describe the behaviour, and are less likely to be subject to recall bias. MSM who did not have sex with another man during the specific time period, or SW who did not exchange sex for money, goods or services during the specific time period, will not meet the eligibility criteria. Place of residence or work Most surveys have a defined sampling area, and therefore make residence or venue attendance (for selected venues) an eligibility criterion. For example, investigators may want to recruit only those who are residents of or work in the city/district/province. If
A-8.2 Verifying eligibility Eligibility is verified through an eligibility screening tool. Typically, participants are asked a series of questions to determine their eligibility. Other techniques may also be used; for example, PWID may be required to show injection track marks or demonstrate how to assemble a needle and syringe, and prepare drugs. Survey investigators should design a clear set of screening questions and a decision-algorithm so that the eligibility of potential participants is determined in a standardized fashion. In addition to excluding participants who do not meet the criteria, investigators may screen out potential participants who meet exclusion criteria; for example, those who are intoxicated (thus, unable to comprehend the questions or provide informed consent) or who have already participated in the survey.
A-8.3 Elements of eligibility The following sections discuss considerations for elements of eligibility: age, sex, risk behaviour, time period, place of residence or work, and language. Age Typically, members of the target population must fall within a certain age range to be eligible. The minimum or maximum age for participation depends on the influence of age on the risk behaviour in question, and the national policy regarding the minimum age at which someone can provide informed consent. All BBS have a minimum age for this reason; however, few have an upper age limit. It is important to follow the national policy regarding minimum age for participation in a survey. In some places, investigators may be legally obliged to report to authorities if participants are found to be minors. If, during formative assessment, a large or particularly at-risk proportion of the target population is found to be below the national required minimum age for participation, investigators may need to obtain special permission from their parents or guardians for them to participate. Sometimes it may be possible to sample underage people without parental consent if they are independent (i.e. they obtain their own livelihood) or if getting such consent puts them at risk. Local laws and guidelines, and rules and regulations set by the local institutional review board (IRB) should be consulted. Chapter A-3 provides additional information on
66
8. Eligibility for participation
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
individuals in the population migrate often, investigators may define residence as having lived in the area for a certain time. Residency information may also help inform local programming needs after the survey, or allow for estimations of the population within the survey area. While those who do not live or work in the sampling area may contribute to the HIV epidemic, they are unlikely to access prevention, care and treatment services and consequently may not reflect local knowledge, attitudes or practices. Language Survey investigators should take into consideration the languages spoken in the region of the survey. Individuals should be proficient in the language(s) in which the survey is conducted in order to participate.
A-8.4 Exclusion criteria Exclusion criteria are characteristics of potential participants that prohibit their enrolment. Such criteria may include being intoxicated (thus, unable to comprehend the questions or provide informed consent) or not presenting a valid coupon in an RDS survey.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
8. Eligibility for participation
67
Section
A Once the target population has been defined, investigators have the option of selecting every member of the group, or selecting a subset of members (i.e. a sample). A sampling strategy describes how the sample will be selected. It also determines whether and how “statistical inferences” can be made from the sample to the source population; that is, the extent to which the sample is representative of the source population. An appropriate sampling strategy can generate unbiased estimates that: • accurately describe the population; • can be compared with estimates for other populations; and • can be compared with estimates of the same population produced by other surveys or at different times. This chapter briefly describes concepts used in sampling, types of sampling, and factors involved in selecting an appropriate sampling strategy.
9. Sampling strategy
Sampling is the process of selecting individuals from a source population to participate in a survey. The purpose of this is to generate information about the population without having to study the entire population, thereby reducing time and cost.
Key terms Convenience The selection of individuals from a sampling: population based on accessibility and availability. Location or Places where the target population venue: spend time and may be accessed. Used interchangeably with “site”. Nonprobability A method of sampling where the sampling: probability of sampling is unknown. Reliability: The degree to which a technique or approach provides consistent results if repeated multiple times. Sampling: The process of selecting individuals from a source population to participate in a survey. Sampling bias: The situation in which some population members are systematically more or less likely to be sampled than others, leading to biased estimates. Sampling The source material or list from which a frame: sample is drawn in order to represent the population. Sampling Defined as N/n, where N is the population interval (SI): size and n is the sample size. Site: A place where survey sampling occurs. It may be a clinic, office or apartment for RDS surveys, or a public space (e.g. a street corner, bus stop, brothel or club) for TLS surveys. Statistical The process of using information from a inference: sample to make inferences, or conclusions, about the population from which the sample was taken. Validity: The degree to which an estimate is likely to be true and free of bias (systematic error).
A-9.1 Concepts used in sampling, and types of sampling A-9.1.1 Concepts used in sampling The main concepts used in sampling, discussed below, are: • statistical inference • sampling bias • validity and reliability • precision • stratified sampling • sampling domains Statistical inference Statistical inference is the process of using information from a sample to make inferences, or conclusions, about the population from which the sample was taken. If people are sampled in a biased manner, the sample may not be representative of the source population. Inferences or
68
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
conclusions made on the basis of observations from such a sample may not be valid or reliable. Sampling bias Bias refers to the difference between an estimate and the “truth”. Sampling bias occurs when some members of the population are more or less likely to be included in the sample than others. Without adjustments, the estimates produced by such a sample are not representative of the population. There are two types of sampling error: systematic and nonsystematic. Nonsystematic sampling error, also known as random sampling error, is the error that results from taking a sample rather than measuring every individual in the population. For example, 100 different samples of the same population will each produce different estimates. The degree to which the estimates differ will depend on the size of the sample relative to the total population. Nonsystematic sampling error can be decreased by increasing the sample size. For example, if two surveys of the same population differ only in their sample size, the survey with a larger sample size will have less random sampling error than the survey with a smaller sample size. Systematic sampling error is generally caused by either a poor sampling design or failure to follow sampling protocols. Unlike random (nonsystematic) sampling error, systematic sampling error cannot be predicted, calculated or accounted for. Examples of systematic sampling error include incomplete sampling frames, and samples that include people who do not meet the eligibility criteria. A sampling frame is the source from which a sample is drawn. It is a list of all those within the source population who can be selected. When using simple random sampling, the list contains all individuals. When using cluster sampling, the list will include groupings of individuals known as clusters. Examples of clusters for target populations might include venues where sex is sold, and venues where men meet male sexual partners. Some specific sampling (or selection) biases are especially relevant to biobehavioural surveys (BBS) and include the following: • self-selection bias – Some people may have more motivation to participate in a survey if they have an interest in the subject matter or if the reimbursement for participation is very high. • healthy worker bias – People who are healthier are more likely to be able to enrol in a survey than less healthy members of the population. • differential referral – In surveys that ask participants to recruit others, individuals may not recruit randomly from the population. In chain-referral surveys,
participants may recruit people who are more like them or who are more popular within their social network. Validity and reliability The term validity, as it is used in BBS, refers to the extent that the information collected in a survey answers the research question. Reliability refers to consistency; it is the extent to which a survey is likely to yield the same result when repeated. Whereas validity is concerned with the survey’s success at measuring what the researchers intended to measure, reliability is concerned with the consistency of the actual measurement instrument or procedure. There are two types of validity: internal validity and external validity. Internal validity refers to how valid the survey results are based on the survey design and any other factors that might influence accurate results. External validity refers to the populations, settings and variables to which this effect can be generalized. Precision Precision reflects the extent to which repeated measurements (e.g. surveys using the same methods) yield the same results. Precise results (i.e. estimates with small margins of error and narrow confidence intervals) may not necessarily be accurate (i.e. close to the truth). Instead, precision is an expression of variability, or the amount of imprecision. Stratified sampling Stratified sampling is the process of dividing members of the population into homogeneous subgroups before sampling. It is done to reduce variability or random sampling error, thereby improving the precision of the estimates produced by the sample. The strata should be mutually exclusive (i.e. every member of the population must be assigned to only one stratum) and exhaustive (i.e. every population member must be included). For example, if sex workers (SW) operate through three types of venues (brothel, street and residence), the sampling frame can be developed separately for each type of venue, and a predetermined proportion of respondents, reflecting the overall distribution of SW, can be drawn from each stratum. If the proportion from each stratum is not known, a similar effect is achieved by ordering the sampling frame by type of venue, unless there is an intent to oversample within some strata. Sampling domains When separate estimates are desired for different subsets of the population (e.g. different subtypes or different geographical locations), separate samples should be drawn from each. Aggregation of data from different sampling domains (e.g. to obtain national estimates from multiple local sampling domains) must be done with appropriate analytical techniques and weighting (Chapter C-1 provides further information on this).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
69
A-9.1.2 Types of sampling The two main approaches to sampling are nonprobability sampling and probability sampling. A nonprobability sample is one that produces estimates of unknown precision and reliability, and for which inferences should not be made or should be made with caution. A probability sample is one that produces estimates that can be inferred to the source population with a known level of precision and reliability. Nonprobability sampling Where populations are hidden because of stigma, discrimination criminalized behaviours or where sampling frames do not exist, nonprobability samples are often a necessity. Examples of nonprobability sampling methods include convenience, quota and snowball (a type of chain-referral) sampling. This section discusses where such types of sampling may be used. Further details of these methods are given in Section A-9.2.2. Probability sampling Probability sampling refers to sampling methods where every individual has a known (nonzero) chance of being selected, and the selection probability can be calculated. Examples of probability sampling methods include simple random sampling, systematic random sampling, stratified random sampling, cluster sampling, multistage sampling and respondent-driven sampling (RDS). Clusters (and cluster sampling) may be viewed in terms of: • location – conventional cluster sampling (CCS); for example, a survey among prisoners; • social ties – for example, RDS; or • time and space – time-location sampling (TLS); for example, sampling of mobile individuals at selected locations during selected time slots. Simple random sampling Simple random sampling is the most basic type of sampling. Each person is chosen entirely by chance, and each person has an equal chance of being selected. Because the sampling probability is equal for everyone, simple random samples are considered self-weighted and no sampling weights need to be calculated. However, if the sampling frame is large or the population is spread out geographically, survey implementation can be impractical. Because complete sampling frames are generally not available for hidden or hard-to-reach populations, simple random sampling is not feasible for most BBS. Systematic random sampling Systematic random sampling is the selection of individuals at regular intervals from an ordered list of individuals in the population, beginning at a randomly selected point. The list could be ordered either alphabetically or by other criteria such as address or size. The main advantage of using systematic sampling over
simple random sampling is its simplicity. For example, if the total population were 1000, a systematic random sample of 100 individuals from that population would involve observing every 10th person on a list. Stratified random sampling Stratified random sampling involves categorizing members of the population into mutually exclusive and collectively exhaustive groups, then choosing an independent simple random sample from each group, based on that group’s proportion to the total population. Investigators can also oversample in some strata to ensure that enough people are selected to obtain stratum-specific estimates. For example, if there is a relatively small number of respondents in a particular stratum (e.g. females aged 15–24 years), taking a random sample from the entire population would not produce a sample large enough to make inferences about that subgroup. Cluster sampling Cluster sampling can be used when it is impractical or impossible to list all individual population members, but it is possible to list locations or clusters where they gather. By randomly selecting a subset of clusters (e.g. venues or sites) and sampling population members at those locations only, the task can be made more manageable and also less costly, because of substantially reduced travel and labour costs. One limitation of cluster sampling is that it requires larger sample sizes than simple or systematic random samples. Multistage cluster sampling Multistage cluster sampling involves multiple stages of random selection of sampling units. The clusters selected at the first stage are called primary sampling units (PSU), and at subsequent stages are called secondary sampling units, tertiary sampling units, and so on. Multistage cluster sampling is typically used when the population or the sampling area is large. At any stage, clusters may be selected randomly or with probability proportional to size (PPS), until the last stage, when simple random sampling of individuals is performed within each of the selected clusters. Respondent-driven sampling RDS combines snowball sampling (in which participants recruit other participants) with a mathematical model to compensate for the fact that participation is not random. RDS is based on the principles of chain-referral sampling: it relies on respondents at each wave to select the next wave of participants. RDS expands upon chain-referral sampling by introducing several innovations that minimize sampling error and account for some types of selection bias. These innovations allow for probability-based selection of participants based on the network within the target population. RDS is one of the most popular methods to sample hard-to-reach populations, such as key populations (e.g. men who have sex with men [MSM], people who
70
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
inject drugs [PWID] and SW). However, it makes several assumptions in order to fulfill the requirements needed to make population-based estimates (see Section 9.7 for a discussion of the assumptions).
A-9.1.3 Selection of sampling strategies Several factors must be considered before selecting a sampling strategy; for example, the existence of a sampling frame and the degree to which the target population is mobile, is networked, or can be identified and accessed at physical sites or venues. Figure A-9.1 shows a flow chart that can be used to facilitate the selection of a sampling strategy. Factors to be considered in selecting a sampling strategy are as follows (2-6):
• does a sampling frame exist? If a sampling frame exists, survey investigators can opt for a probability-based sampling design, such as random or cluster sampling. • is the target population networked (socially connected)? If there is no sampling frame and it would not be possible to construct one, but formative research suggests that the target population is well networked, RDS could be used. For example, if PWID rely on other PWID to buy drugs, they may form a network, even if unintentionally. Similarly, networks may exist among SW and MSM. • is the target population identifiable and accessible at physical sites or venues? If formative research suggests that a large proportion of the target population can be found at venues, and access to these venues is possible, then a sampling frame can be constructed and cluster-based sampling designs including CCS or TLS may be appropriate.
Figure A-9.1 Criteria for selecting a probability-based sampling method
Formative Assessment
If the target population is too small or does not meet any of the three conditions below, then use convenience or other non-probability sampling method, or conduct a qualitative study.
No
Is the target population large enough to reach the required sample size?
Yes
Are members of the target population socially networked with one another?
No
Does a complete list of all potential participants exist?
No
Can members of the population be sampled in physical locations?
Yes Can individuals identify each other as part of the target population and willing to recruit others from the network?
Yes
Yes Are individuals associated with locations in a fixed manner?
Simple Random Sampling (SRS)
Yes
Yes
No
Respondent-Driven Sampling (RDS)
Conventional Cluster Sampling (CCS)
Time-Location Sampling (TLS)
If all the conditions are met, the decision as to which sampling method to use should be based on available financial and human resources, capacity of the survey team, the legal and political environment, and other key findings from the formative assessment.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
71
A-9.2 Nonprobability sampling methods Nonprobability sampling methods are flexible and easy to implement, but considerable care must be taken when using the information obtained from such samples because it is unlikely to truly represent the population being studied. Sampling bias is likely and, even if the direction of bias is known, the magnitude of bias is usually unknown. Usually, sampling bias also impedes comparison with other surveys. This section discusses when to use a nonprobability sampling method, types of these methods, their strengths and limitations, and ways to use them.
Key terms Chain-referral A sampling technique where enrolled sampling: survey participants refer or recruit potential participants from among their acquaintances (peers). Purposive A method of sampling used to select sampling: individuals with specific characteristics; for example, sex workers who take antiretroviral treatment. Quota A type of purposive sampling that specifies sampling: how many people with each characteristic will be sampled.
A-9.2.1 When to consider using a nonprobability sampling method Nonprobability sampling methods are most useful when conducting qualitative research or when the probability sampling is either unimportant or impossible. Nonprobability sampling may be used when: • the target population cannot be reached through a probability sampling method; • because of the small size of the target population, it is impossible to achieve the sample size required for a probability method; • for some other reason (e.g. resource constraints or low participation rates), it is impossible to attain the sample size required of a probability method, even if the target population is large; • funds are limited; • data must be obtained quickly; and • the research capacity among investigators or staff is limited. During survey planning, investigators should decide whether to use probability or nonprobability sampling methods. Probability-based methods are preferable but not always feasible. Sometimes, investigators begin with the intention of using probability sampling but are unable to generate a sampling frame and have no alternative but to adopt a nonprobability design. Nonprobability methods are vulnerable to bias; hence, it is important for investigators to identify possible biases in their BBS and assess how such biases affect the results.
(this information complements the discussion about types of sampling in Section A-9.1.2, above). Convenience sampling In convenience sampling, investigators select participants who are most easily available and accessible. This is fast and less resource intensive than other methods. Investigators may visit a number of locations and ask people to participate at each location. Selection probabilities are unknown, and not everyone who meets the eligibility criteria is guaranteed selection. Convenience sampling can be useful, particularly when surveying a population that is extremely small. It is best used in the exploratory stage of a research and surveillance activity. An example of convenience sampling is recruitment of participants based on availability (e.g. PWID who congregate near a train station or patients at a health clinic). Take-all approach The take-all approach is a kind of convenience sampling in which investigators attempt to sample every person that meets the eligibility criteria; it is most commonly adopted when the target population is small. Even under those circumstances, however, it is unlikely that 100% of the target population will be sampled. The take-all approach can be used at the final stage of a multistage cluster sample when identified clusters have few eligible people, or if there is limited access to many clusters. In these instances, it is more efficient to sample all eligible people who are approached for participation. Purposive sampling Purposive sampling is the process of sampling individuals with specific characteristics. Whereas a convenience sample might enrol all SW who are easily identified,
A-9.2.2 Types of nonprobability sampling methods There are many kinds of nonprobability sampling methods, the most common being convenience, purposive, quota and chain referral, as discussed below
72
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
a purposive sample might only select SW who have a regular partner. This method may be adopted when investigators only need information on a subset of a population, or are trying to ensure a diverse sample along certain specific dimensions. Quota sampling Quota sampling is a type of purposive sampling that specifies how many people with a given characteristic must be sampled; it is generally used to ensure that particular segments of the population are sampled. For example, in a survey of PWID, investigators could decide that they want to interview 80 PWID, comprising 20 female and 20 male homeless PWID, 20 people who also sell sex, and 20 additional PWID of any characteristics. This approach does not establish statistical criteria for selecting the individuals with these characteristics; hence, their selection is not probability based and the sample is not representative of the entire population of PWID. Chain-referral sampling Chain-referral sampling encompasses a group of sampling methods in which individuals recommend others to investigators or directly recruit others to participate in the survey. The sample grows through peer referral. Snowball sampling and indirect sampling are types of chain-referral sampling. Indirect sampling is discussed in Section A-12.4. Snowball sampling Snowball sampling gets its name from the image of a snowball that gets larger as it rolls down a hill, collecting more snow as it goes. Snowball sampling is a type of chain-referral sampling in which a few well-informed initial participants are selected purposively and then help to enlarge the group. Investigators ask them to provide names of other individuals who are well informed on the survey topic. These people are then invited to participate in the survey, and the process continues (7). In another version of snowball sampling, the initial wellinformed individuals are asked to recruit others who are eligible for the survey. The recruits, in turn, are asked to
recruit others, who then recruit others of their own, until the sample size is reached. As in the other nonprobability methods described above, the probability of selection of any given participant is unknown.
A-9.2.3 Strengths and limitations of nonprobability sampling methods As with all sampling methods, nonprobability sampling methods have strengths and limitations, as indicated in Table A-9.1.
A-9.2.4 How to use a nonprobability sampling method Surveys using nonprobability sampling methods are relatively easy to implement. In all other regards, they face the same requirements for implementation as probability sampling methods. For example, nonprobability sampling methods also require standard operating procedures (SOPs), informed consent and standardized questionnaires. Surveys that use nonprobability sampling methods can be implemented in many different ways. For example, investigators can sample people and ask them to come to a survey site, or the survey can be administered wherever survey staff finds potential participants. The appropriate method may depend on the social visibility of the target population (8). If investigators forgo the selection of a survey site, it is important to consider how to test people for HIV and other sexually transmitted infections (STIs) in public settings. This scenario is described in more detail in Section A-9.5 on CCS. If investigators adopt snowball sampling, they may need to create coupons to assist participants in recruiting peers and directing those peers to the survey site. Considerations for preparing coupons and peer recruitment can be found in Section A-9.7 on RDS.
Table A-9.1 Strengths and limitations of nonprobability sampling methods Strengths • Relatively inexpensive • Relatively easy to implement • Can be conducted with any sample size • Can be used when there is no sampling frame
Limitations • Sample is not representative of the underlying population because the selection probability of participants is unknown. • Results from multiple surveys are not comparable. Nonprobability methods have great potential for sampling bias, so it is impossible to determine if observed differences in results are due to changes over time in behaviour, or merely in the sample. • No basis for assessing the precision or reliability of results.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
73
A-9.3 Cluster sampling overview The purpose of this section is to give an overview of cluster sampling methods used in BBS. It covers characteristics common to all cluster sampling techniques and the most important differences among the methods commonly used to survey key populations. Cluster sampling is the method of choice in many surveys because it can provide quick, relatively inexpensive and representative data on populations. It has been adapted for use with key populations over the past 15–20 years. Cluster sampling can be used when it is impractical or impossible to list all individual population members, but when it is possible to list locations (clusters) where target population members gather, and where they can be identified and approached for survey participation. This is often the case for key populations. The sampling frame is an exhaustive list of clusters. Clusters are places where population members live, work, socialize or conduct other activities related to their defining characteristics. Cluster sampling, in its simplest form, involves a two-stage process. First, a subset of clusters are selected from an exhaustive list of all possible clusters; then individuals are selected from within each selected cluster. Surveys carried out in very large areas sometimes have additional stages beginning with selection of a subset of geographical areas at the first stage (e.g. districts, wards or census enumeration blocks of a large city), followed by selection of clusters within the selected geographical areas at the second stage, and the actual sampling of individuals within the selected clusters at the third stage. frame. The survey team instead samples a limited number of (randomly) selected clusters. By selecting clusters, and listing and sampling population members in those clusters only, the sampling task becomes more manageable and less costly compared to the alternative of simple random sampling.
A-9.3.3 Limitations of cluster sampling People who gather together in clusters may be more likely to have certain traits in common with each other than they do with people in other clusters. For example, young SW might prefer to solicit clients on corners with other young SW whereas older SW may choose to solicit clients elsewhere. Furthermore, people who inject heroin do not necessarily interact with people who inject methamphetamine. This social or behavioural clustering (also referred to as design effect) can skew survey results if not accounted for in analyses (see Chapter C-1). Accounting for the design effect in analysis usually has the effect of increasing the standard error, so results from cluster surveys tend to have less statistical precision than results from surveys that use simple random sampling and an equal sample size. When cluster sampling is venue based, which is typically the case, it has the additional disadvantage of only capturing the subset of the population who frequent venues, and who are identifiable and accessible at those venues. So, population members who do not frequent venues are excluded, as are population members who frequent only inaccessible or unidentifiable sites. To the extent that these population members differ from those who are accessible at venues, venue-based cluster sampling may introduce bias. Examples of key population members who may not be reached through venue-based sampling include: • SW whose clients contact them through brokers or via mobile phone, and who do not solicit clients from physical locations; • MSM who only gather in private locations such as homes or who find sexual partners only on the internet; and • PWID who inject in private locations (e.g. homes), or who get their drugs and injecting equipment from friends or family rather than directly from dealers.
A-9.3.1 Types of clusters These guidelines distinguish between two main types of clusters: physical clusters and time-location clusters. Physical or fixed clusters are places where a fixed number of people is present (e.g. a brothel). Surveys selecting such clusters are using CCS. Time-location clusters are places where the number of people may change over time (e.g. people attending a bar or club). Surveys selecting such clusters are using TLS.
A-9.3.2 Strengths of cluster sampling Cluster sampling can provide a statistically valid representative sample of a population spread over a large geographical area, without requiring a complete list of eligible population members for its sampling
74
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-9.3.4 Planning for cluster sampling A number of decisions should be made early on in the survey planning process to determine how the sampling frame will be constructed and what information will be required to complete it. Ensure a venue-based cluster sample will be appropriate for the population Be sure the definition of the target population is compatible with a survey design that uses cluster sampling. For example, a survey of brothel-based SW will not include SW who find their clients via phone or internet. If formative research or existing information indicates that a significant proportion of the target population will not be reached by cluster sampling, consider another approach. Define the geographical catchment area Typically, a survey will define a sampling area through political or geographical boundaries (e.g. a city’s political boundaries). If it is believed that substantial urban and rural differences exist, then this may necessitate including enough respondents in both types of geographical areas and stratifying the analysis by urban and rural areas (see below). Consider stratification Stratification is the process of separating members of the target population into more homogeneous subgroups (or strata) before sampling. Stratification may become necessary if two or more subgroups of the target population differ so much from each other (e.g. in their risk behaviours, their burden of disease, or ethnicity) that it would be better to view them as separate populations. Common examples are brothel-based and street-based SW or MSM and transgender women, who may be so different from each other that investigators need to sample each subpopulation (stratum) independently. Prior knowledge about the cluster types and relative sizes of the subpopulations from formative assessments are used to assign every sampling unit (venue, time-location slot) in the sampling frame to one, and only one, stratum. Usually, stratification requires over-sampling (an increase in sample size) in some or all strata to allow for separate estimates for each stratum. If stratification is not feasible because of budget or time constraints, it may be necessary to redefine the target population or venue types so that the survey is restricted to one subpopulation only. Build design effect into sample size calculation Simple random samples have a design effect of 1, but for cluster surveys, the design effect is usually greater than
1. The expected design effect can be used to guide how much to increase the sample size to compensate for the increased variance. Design effects are sometimes available from estimates of the primary outcome from previous surveys. If they are not, some speculation on the variance for important variables including the primary outcome may be required. Statistical software packages, including Epi Info, SPSS, SAS and Stata, can calculate the design effect for each variable in a given survey (see Chapter A-10).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
75
A-9.4 Procedures for two-stage and multistage sampling A-9.4.1 Two-stage cluster sampling Most BBS using cluster sampling will use a two-stage sample design. PSUs or clusters are chosen at the first stage of sample selection, and individual respondents are chosen from within each of the selected PSUs at the second stage. This sampling scheme, adapted to meet the needs of the different subpopulations and local conditions, will likely satisfy the needs of most BBS efforts.
Key terms Cumulative The cumulative sum (across clusters) of measure of the estimated number of individuals who size: meet the approach criteria. Estimated The number of individuals estimated to measure of be present and eligible to be recruited size: during formative assessment or sampling frame development. This number helps investigators plan sampling strategies, and could be used to carry out selection of clusters using probability proportional to size (PPS). Multiplicity: The chance that any one sampling unit (person) could be captured in multiple clusters (i.e. an individual could be recruited in multiple venues). Primary The unit used for the first stage of sample sampling unit selection (e.g. a brothel) – in a two-stage (PSU): cluster sample, the PSU and the cluster are the same. Probability With PPS, a sampling unit (e.g. a cluster) proportional gets selected according to its size (relative to size (PPS): to that of others); hence, larger clusters have a larger probability of selection. Random start A randomly selected number between 1 (RS): and the sampling interval. Sampling The time period when field staff are at the event: venue identifying, selecting, recruiting and interviewing respondents. Sampling unit: In a statistical analysis, refers to one member of a set of entities being studied (e.g. bars or respondents). Sampling The population to which the results can be universe: inferred.
A-9.4.2 Multistage cluster sampling If the sampling universe is very large, meaning the sampling area covers a large geographical area, additional sampling stages, such as province, district or segment of a city, can be added. For example, in a provincial survey of SW, investigators may include towns as the first sampling stage, venues as the second stage, and individual SW as the third stage.
A-9.4.3 Selecting primary sampling units This section describes procedures for selecting sampling units at each stage of cluster sampling. Clusters selected at the first stage of sampling are called primary sampling units (PSUs), and clusters selected at subsequent stages are called secondary sampling units and tertiary sampling units. When two-stage cluster sampling is used, the PSU and the ultimate clusters from which individual respondents will be selected are the same. Probability sampling requires that the selection probability of every respondent is known. With simple random sampling (SRS), every respondent has an equal probability of selection. With two-stage or multistage cluster sampling, the selection probabilities are the combined product of the sampling probabilities at each stage of selection. Equal probability versus probability proportional to size Sampling units (clusters) must be selected at every sampling stage. There are two ways to select clusters: by probability proportional to size (PPS) or by equal probability. When clusters are selected by PPS, it means they are selected proportional to their size, so larger clusters have a larger probability of selection. When clusters are selected by equal probability, it means they are selected without regard to size. So, large clusters have the same probability of selection as small clusters.
To use PPS selection procedures, the number of individuals associated with each PSU must be known in advance. This number, known as a measure of size (MoS), does not need to be an exact count. A rough approximation for each site may be good enough, and because it is rough, it is referred to as an estimated measure of size (EMoS). Selecting clusters by PPS improves sampling efficiency by increasing the likelihood that larger clusters are selected. However, the larger the size of the cluster, the smaller the probability of each individual within that cluster being selected. To compensate for this, a weighting process is used in the analysis, which works by assigning
76
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
a weight to individuals from every cluster that is inversely proportional to the size of the cluster. Self-weighted designs When the combination of sampling probabilities at each stage produces a selection probability that is the same for individuals from every cluster, this is known as a self-weighted design. Combinations that will produce self-weighted samples include: • PPS at the first stage and a fixed (equal) number at the second stage; and • equal probability sampling (EPS) at the first stage and select all, or select fixed proportion (e.g. 20% from each cluster) at the second stage. Statistically, the first of these (PPS with equal number at second stage) is the more efficient two-stage sampling design. However, it does require accurate measures of size prior to selecting clusters. The second (EPS with select all or fixed proportion at second stage) can be challenging to implement if very large clusters are selected. The final sample size also becomes unpredictable. As an alternative, it is possible to select fewer respondents (or an equal number of respondents) from clusters selected by EPS. However, the final sample may not be self-weighted. Self-weighted samples are easier to analyse, but if the EMoS are very different from the actual measures of size (AMOS), then the sample may no longer be selfweighted. Therefore, it is always best to collect and record all information needed to conduct a weighted analysis, in the event that it becomes necessary to perform a weighted analysis. Note that even for selfweighted survey datasets the design variables are still needed in data analysis to correctly calculate standard errors and p-values. Decision criteria for EPS and PPS Ideally, investigators should obtain a current list of clusters with their EMoS, to facilitate PPS sampling. Choose EPS Figure A-9.2 Decision criteria for EPS and PPS No Information on EMoS available?
only if information on EMoS is not available. If information on EMoS is available and clusters have little or no variation in EMoS, investigators may choose EPS as well as it would equal a self-weighted design. However, in most cases EMoS will vary substantially across clusters, and will warrant a PPS design (Figure. A-9.2). If investigators choose to generate a self-weighted sample, and EMoS is available, PPS is the method of choice. For discussion on how to approach PPS vs EPS for TLS designs, see Section A-9.6).
A-9.4.4 Procedures for selecting sampling units Determine how many clusters are needed and what the ideal cluster size should be The number of clusters to be selected must first be determined before selecting clusters by PPS or EPS. The number of clusters will be a function of the desired cluster size (number of respondents to be sampled from each cluster), and the overall sample size for the survey. More information on how to sample clusters using EPS and PPS can be found in Appendix I-18. Desired cluster size Sampled clusters should be small enough to make a listing of individuals feasible, but large enough to support a minimum number of sampled individuals per cluster. Generally speaking, a sample with smaller clusters will result in a smaller design effect than a sample with fewer larger clusters. But more clusters may also mean higher travel/transport costs. Another factor to consider is the degree of expected homogeneity (similarity) within clusters as opposed to between clusters, with respect to the outcomes of interest (e.g. behavioural variables, HIV status). If homogeneity is expected to be low, then fewer clusters will be needed. Number of clusters to select Survey statisticians often recommend that there be no fewer than 30 clusters (see also Chapter A-10). A common approach to deciding how many clusters to select is to divide the total sample size for the survey by the desired cluster size. The desired cluster size should be smaller
No Yes Large variation in EMoS?
Yes EPS EMoS, estimated measure of size; EPS, equal probability sampling; PPS, probability proportional to size
PPS
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
77
than the (average) EMoS; that is, the expected number of eligible people in a given cluster (which should be available from the sampling frame). It is possible to sample a fraction of all eligible persons at a given cluster; however, cluster sizes should still be as large as or larger than the average EMoS in order to avoid shortfalls in obtaining the desired sample. Nonresponse (refusal to participate) should also be factored into the design. Some designs deal with nonresponse by inflating the initial sample size, others by doing replacement (see Chapter A-10).
Example of calculating number of clusters 1 - calculate the average EMoS from the sampling frame. For this example, assume EMoS = 10. 2 - only aim to sample a fraction of the average EMoS; for example, 30% to 50%. It is unrealistic to assume everyone in a given cluster can be sampled. In this example, investigators aim at a cluster size of no more than 3 or 5 (30–50% of 10). This improves efficiency of the cluster sampling (reducing design effect) and minimizes the risk of shortfall due to nonresponse. 3 - the required number of clusters, therefore, should be 600 (target sample size) divided by 3, 4 or 5; that is, 200, 150 or 120 clusters.
Collecting information to calculate sampling weights Regardless of whether EPS or PPS is used, it is critical to document the information that will be needed to calculate sampling weights. This involves the following steps: • preserve the sampling frames with information about each cluster, including cluster name, cluster identification number (ID), MoS (if applicable) and selected units at each sampling stage. • record sampling event information: - actual measure of size - # approached and eligible - # consented to survey - # nonresponse (e.g. refusals, incomplete interviews or incomplete blood samples). • record individual information and answers to the multiplicity question (if applicable). Multiplicity refers to the possibility that some people may be present at different clusters at different times and so may get approached or sampled more than once. Maintain the original sampling frame for reference, or in case a second round of cluster selection is required. Document the sampling information in accordance with the steps above; this information is required to calculate sampling probabilities.
78
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-9.5 Conventional cluster sampling versus time-location cluster sampling A-9.5.1 Features of CCS and TLS There are two major types of venue-based multistage cluster sampling typically used with key populations: Table A-9.2 Features of CCS and TLS Conventional cluster sampling CCS should be used when there is a relatively fixed population associated with fixed venues (sites) such that: • individuals are associated with only one venue • the same individuals are associated with the same venue no matter when it is visited • a complete list of individuals who are associated with the venue may be obtained or created Examples of situations where CCS might be appropriate include sampling of stationary: • brothel-based SW • prisoners • military personnel • PWID in residential treatment facilities.
conventional cluster sampling (CCS) and time-location sampling (TLS). Table A-9.2 summarizes the basic features of CCS and TLS.
Time-location sampling TLS should be used when the target population is mobile, meaning different people may be present at a given venue at different times. It may be used when individuals come and go freely from the venue (site), and when their composition may differ depending on the time the venue is visited. In TLS: • individuals may be associated with more than one venue • it may be possible to create a list of attendees at a venue, but only of those who are present during the sampling event (time period) PSU in TLS contains both a time and a location element. The same physical location may be included in the sampling frame multiple times, to account for the variation in target population members frequenting the site at different times, hence the name “time-location” or “venue-time” (or “venue-date-time”) cluster sampling. Examples of situations where TLS might be appropriate include: • SW at street-based sites, bars or nightclubs • MSM at bars or MSM cruising sites • PWIDs at shooting galleries or other injecting venues
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
79
A-9.5.2 Types of venues used in cluster sampling for key populations Table A-9.3 on the next page lists the types of venues typically used for key populations (SW, MSM and PWID). Target population Sex workers (SW)
Conventional Cluster Sampling
Time-location Sampling
Brothels,a massage parlours,a hair salons Brothels,a massage parlours,a bars, restaurants, discos, city blocks, streets, public parks Bars, nightclubs, streets, parks, restaurants, bathhouses, railway platforms, beaches, social organizations Community of hijras linked to a particular guru (in India)b Drug treatment centres Bars, nightclubs, streets, parks, restaurants, bathhouses, railway platforms, social organizations Street locations, shooting galleries, injection parlours, syringe exchange centres Workplaces, locations where young people gather Truck stops, depots, border checkpoints
Men who have sex with men (MSM)
Transgender persons (TG)
People who inject drugs (PWID)
Young people Long-distance truck drivers Prisoners Students Military personnel Hospital patients Miners
Households, schools Transport companies (with rosters of employees) Prisons, jails, detention facilities Schools Barracks, camps Hospitals Mining camps
CCS, conventional cluster sampling; TLS, time-location sampling Note: Some of the examples are listed in both categories because the nature of the relationship between the cluster and the population may vary depending on the context. a b
Brothels and massage parlours can be sampled using CCS or TLS. If the SW live or always work at the same brothel, CCS may be appropriate. Hijras is a term used for transsexual or transgender people in India who sometimes live in communities led by a guru.
A-9.5.3 Conventional cluster sampling This section describes conventional cluster sampling (CCS) and the general steps required to prepare for using this method. CCS is used when population members are fixed in relation to the venues from which they will be sampled.
When to use CCS Use CCS when there is a relatively fixed population associated with a site or cluster such that: • individuals are associated with only one site or cluster; • the same individuals are associated with the site or cluster no matter when it is visited; and • a complete list of individuals who are associated with the site or cluster can be obtained or created. The PSU in a two-stage cluster sample using CCS is the site or cluster. Examples of situations where CCS might be appropriate include sampling of students in schools, detainees in prisons, military personnel in barracks, PWID in treatment facilities and SW living in brothels.
A-9.5.4 Overview of CCS design CCS is a cluster sampling method. Typically there are two stages of sampling, but there can be more. The first stage of sampling consists of selection of PSUs, which can be drawn with equal probability or with PPS. At the second stage, the target population members in the selected clusters are randomly or systematically selected from the complete list of individuals in the PSU.
80
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Strengths of CCS • CCS needs only an exhaustive list of clusters where eligible participants can be accessed and recruited. • CCS is a robust sampling technique that has been commonly and successfully used in the past. • under ideal circumstances, it is possible to create a self-weighted sample. Limitations of CCS • if the design effect is greater than 1, a larger sample size will be required to obtain a comparable level of precision to that of a simple random sample. • obtaining self-weighted samples may be difficult if accurate measures of cluster size are not available in advance of the survey. • although listing eligible respondents at venues is theoretically possible, it may be difficult to obtain a complete list. • it may be necessary for the team to return multiple times to the site to find all the people who have been selected. If the team takes shortcuts and replaces respondents who are unavailable during the first visit, with those who are more easily available, this can introduce bias. • though the members of the community are theoretically available at any time, the lists may change frequently. Prisoners, PWID in treatment facilities and military personnel, for example, may transition in and out of that location at any time. CCS plan A number of preparations specific to sampling in a site with a fixed population are required before the first sampling event can take place (see Appendix I-1). • sample size – Determine the sample size using the methods described in Chapter A-10. • method for sampling PSU – Two approaches exist for sampling clusters: - equal probability should be used when there is minimal variation in EMoS between clusters or if there is little information on the EMoS of each cluster; and - PPS should be used when there are fairly good EMoS for all the clusters in the sampling frame and high variation in MoS between clusters. Individuals within clusters are sampled randomly or systematically. Two options are available for sample selection from each selected cluster (see Section A-9.4.4 on desired cluster size): • option 1 – The sample size within each cluster may be equal (i.e. the same number of individuals are selected from each cluster). Choosing this option after selecting PPS at the first stage will produce a self-
weighted sample, provided that the EMoS is accurate. This option also has the advantage of providing control over the final sample size. • option 2 – The sample size within each cluster may be a fixed proportion of individuals from the total in each cluster. This option has the advantages of not requiring EMoS ahead of time, and it will produce a self-weighted sample. However, the disadvantage is lack of control over the final sample size. Sampling key populations Some special considerations are warranted for sampling key populations such as SW, MSM and PWID: • community advisory board – Establish a community advisory board (CAB), made up of target population members. The board will provide a liaison to the community to facilitate trust and safeguard the community’s interests. A terms of reference for the CAB can be helpful. • team members – It is advisable to include key population members on the survey team, either as community consultants, or in some cases as interviewers or field team leaders. This can help with establishing trust and rapport with the community, and also with ensuring the identity of key population members. • management permission – Locations or venue owners or managers, where relevant, may be approached for permission to conduct sampling at the site. If they refuse, some information about the location or venue should be collected (numbers and types of target population members) to provide some measure of possible sampling bias.
A9.5.5 Steps for using CCS The steps for using a CCS are described here and can be found in the flow chart at Figure A-9.3. 1. eligibility criteria: Define the population being surveyed clearly, and also the types of venues where they will be accessed (i.e. inclusion and exclusion criteria). For example, in a survey of SW, the eligibility criteria might be women who work in brothels and are aged 15 years or over. This may be appropriate if there is a clear programmatic decision to focus on brothel SW. But it may exclude many women who work outside of brothels (e.g. those who are street based or bar based, or operate only through phone or social media sites). 2. catchment area: Define the geographical area that will be represented by the survey, and where sampling will occur. This might be an area defined by a political boundary like city limits (e.g. the district of Kampala) or a geographical boundary like “east side of the river”.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
81
3. sampling frame: The sampling frame is a comprehensive list of locations where the population can be found, and from which the sample will be selected. Sampling frame development can be easy or complex, depending on whether up-to-date lists of venues exists, or whether they need to be created. If existing lists are used, they must be updated to verify that they are current. That means locations where population members no longer gather must be removed from the list and new locations must be added. Typical types of locations that serve as clusters are brothels, lodges, nightclubs and street corners (for SW), schools or classrooms (for in-school young people), and barracks (for uniformed services). However, ultimately the clusters must match with whatever is specified in the eligibility criteria. If EMoS are available for clusters, they can be used to facilitate the sampling scheme. Figure A-9.3 Conventional cluster sampling steps
4. sampling of clusters: If EMoS are available, clusters can be selected by PPS. When PPS sampling is used, larger clusters have a higher probability of selection than smaller clusters. This can make sampling more efficient; however, adjustments must then be made to account for the lower selection probability of people in smaller clusters. If EMoS are not available, clusters can be selected with equal probability. The number of PSUs selected will be a function of the sample size and the desired cluster size. This is described in more detail in Section A-9.4.3. 5. inclusion of sampled clusters: All selected PSUs must be visited and contribute to the sample even if the target sample size has already been reached. 6. sampling of individuals: Each PSU is expected to contribute survey respondents, according to the chosen option.
Two stages
Two stages or more?
More than two stages
Create sampling frame of PSUs* (clusters) and select clusters by EPS or PPS, as appropriate
Conduct the following steps at the final sampling stage
Create sampling frame and select units at each stage by EPS or PPS, as appropriate
Schedule sampling events, taking geography and transportation of field teams into account
Conduct sampling event, selecting eligible respondents using equal size, proportional data or take all as appropriate (see survey implementation at Chapter B)
*
Depending on the target population and the type of cluster, this may be as simple as obtaining a list of locations from a minor partner (e.g. in case of prisons) or as complicated mapping location such as brothels or sites where sex workers solicit clients (e.g. in the case of sex workers)
82
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-9.6 Time-location sampling Key terms Focus group: Semistructured group conversations conducted with usually 6–8 members of the target population under the guidance of a facilitator using an interview guide. Focus groups are convened to confirm or refute observations made in key informant interviews, and to gather additional information to facilitate survey preparation and implementation. Key informant People who have first-hand knowledge of (interview): the target population or subject matter. Time-location A cluster sampling technique that not only sampling (TLS): considers the venue (location) but also the time of sampling. TLS is used when the number of people attending a venue may vary over time or day. Examples include sampling of patrons at bars, clubs or bathhouses. Venue The entirety of all venues from which (location) clusters or time-location clusters are being universe: drawn.
An example of a time-location cluster might be a specific gay bar between 10 pm and 2 am on a Friday night. Men who are found at this time-location cluster may differ from men found at this same location at another time (e.g. 2 pm to 6 pm on Mondays), not only as individuals, but in terms of their levels and types of risk behaviours. For example, men frequenting sites on weekdays may be more likely to be unemployed and therefore have less disposable income. For this reason, it is important to be able to select venues during different time periods. In its simplest form, TLS is a two-stage cluster sampling design. The PSUs are the time-location clusters. TLS clusters are selected by equal probability or PPS (in the same manner described in Section A-9.4.3 on selecting PSUs). The second stage comprises the sampling of population members at the selected time-location clusters. Because individuals can come and go from the venues, they must be sampled during defined sampling events. This is in contrast to CCS, where respondents are “stationary” with respect to the venue, and so can be listed and sampled at any time. To use PPS sampling with TLS, the “expected” EMoS at different times must be known in advance, and it is understood that by definition, the MoS will vary, so the estimated sizes can only serve as expected values.
A-9.6.1 Overview of time-location sampling design This chapter describes the steps involved in sampling participants using time-location sampling (TLS). TLS is used when participants are mobile with respect to the venues from which they will be sampled. For example, street-based SW or MSM attending clubs or bars may move around between venues, and will not be associated with any one venue in a “fixed” or stable manner. This is in contrast to CCS where sampling happens at conventional clusters or “sites” with “fixed” or stable populations (e.g. children in school or inmates in prisons). To implement TLS as a cluster sampling method, a sampling frame is needed that allows all members of the population to have a nonzero chance of selection into the sample, and a method for calculating selection probabilities. The sampling frame should be constructed in such a way that it can capture the different types of people who may frequent venues at different times of the day and on different days of the week. This is accomplished by constructing a sampling frame that is composed not only of venues frequented by target population members, but also of time slots when the venues are operational. These combined venue or time clusters are known as time-location clusters.
A-9.6.2 Strengths of time-location sampling TLS provides a means of sampling mobile populations in a representative and probabilistic manner.
A-9.6.3 Limitations of time-location sampling • if the design effect is greater than 1, a larger sample size will be required to obtain a comparable level of precision to that of a simple random sample. • the yield from individual sampling events can be unpredictable because of changeable MoS. • likewise, obtaining self-weighted samples may be difficult if accurate measures of cluster size are not available in advance of the survey. • in theory, recruitment of respondents should be evenly distributed throughout the sampling time interval. This can be difficult to manage logistically.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
83
Venue universe versus time-location sampling frame The location or “venue universe” is a list of all of the potential locations or venues in a given geographical area, including those that may not be appropriate for the survey. It contains information about the location or venue type and each of the potential day-time periods, EMoS of each time-location slot, any stratification information, and any information that informs the investigator if the venue and daytime period is appropriate to include in the sampling frame. The “time-location sampling frame” is an exhaustive list containing all the relevant time-location clusters to be sampled from. The variables included are location or venue ID, location or venue name, time period and EMoS. Additional variables may need to be included, depending on factors such as target population and cultural context.
the site in the sampling frame multiple times because of the risk of duplicate sampling. But if there is high turnover at a site, then the site can be included in the sampling frame multiple times. Some sampling frame development exercises record on the site information sheet peak times and times with few people at the site, and list maximum and minimum numbers of target population members expected at different days and times. At the end of the sampling frame development exercise, investigators can develop a comprehensive sampling frame consisting of all time-location clusters. The sampling frame should exclude time-location clusters that are likely to be too small (i.e. have too few attendees) to warrant a sampling event; venues and locations for which venue management refuses to grant permission for sampling events; and venues that are deemed too dangerous for survey staff. 5. select the clusters – Using the complete sampling frame, time-location clusters are selected (first sampling stage). These clusters may be selected either through EPS or PPS (see Section A-9.4.4 on procedures for selecting sampling units). The selected time-location clusters are then added to the “sampling event calendar” (i.e. a calendar table with cells for each day). Care must be taken to avoid days when sampling will be difficult or impossible, such as when staff may not work or certain venues may be closed or low in attendee numbers (e.g. public holidays or gay pride festivals). Once the calendar is completed, preparations to conduct the field work can begin and field recruitment and enrolment initiated.
A-9.6.4 Time-Location Sampling Plan These steps (see also Figure A-9.4) can be followed for developing a sampling plan for a TLS survey (9): 1. define eligibility criteria – Define survey eligibility criteria, keeping in mind that eligibility will have to be established quickly in the field and from visual assessments. 2. establish a CAB – Establish a CAB made up of target population members. The board will provide a liaison to the community to facilitate trust and to safeguard the community’s interests. A terms of reference for the CAB can be helpful. 3. plan the team composition – Plan to include some key population members on the team as community consultants (liaisons) or interviewers/field team leaders to help with trust and rapport. 4. develop a sampling frame – Develop a comprehensive list of potential time-location clusters (sampling frame). This involves visiting venues and completing site information sheets that document information such as name of the site, name of contact person at site, physical location, characteristics and boundaries of site, type of site (e.g. for FSW this might include brothels, bars or street corners), estimated number of target population members expected to be found at the sites on different days and at different times, and frequency of target population members visiting sites. This last item is to provide some indication about how often venues should be included in the sampling frame. If target population members frequent sites daily, it would not be desirable to list
84
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure A-9.4 Time-location sampling steps
Two stages
Two stages or more?
More than two stages
Construct venue universe
Conduct the following steps only in selected clusters
Create sampling frame and sample at each stage
Create TLS frame from information in venue universe
Decide whether EPS or PPS in appropriate
Sample time-location clusters from sampling frame using EPS or PPS
Schedule sampling events, taking into account geography and transportation of field terms
Conduct sampling events (see survey implementation at Section B) EPS, equal probability sampling; PPS, probability proportional to size; TLS, time-location sampling
Time-location sampling considerations This section presents considerations relevant for development of a protocol and SOPs. A protocol checklist (see Appendix I-1) and the TLS section of the survey design checklist (see Appendix I-4) can facilitate planning. Formative assessment By the end of formative assessment, investigators should have a good knowledge of the types of locations or venues where the target population can be accessed. The formative assessment may reveal location or venue types that investigators had not anticipated. It is important that all location or venue types be represented in the venue sampling frame. General sampling event recruitment planning • because public locations or venues are often relatively unstructured and each one is different, it is best to create strategies for recruitment in a variety of situations so that the field team can approach and
recruit people in a systematic and efficient manner rather than according to a strict set of procedures that may not always fit well. • geographical boundaries for each location or venue should be fixed and agreed upon by the team when the locations or venues are mapped. • a standard text for introducing recruiters and the survey should be developed and rehearsed for each sampling event. • recruitment should be done in small teams of no fewer than two people. The approach should ensure safety of the team while facilitating adherence to the SOPs. Documentation of sampling procedures All information relevant to each sampling stage should be documented and retained. This includes sampling frames at each stage (including time-location units at the final sampling stage), procedures for selecting sampling units at each stage, cluster sizes, number sampled in each cluster and nonresponse. Reasons for nonresponse may
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
85
include refusal of venue managers to allow for a sampling event to take place, and refusal or failure of an individual respondent to give consent or to successfully complete all stages of data collection. Completeness of sampling frame Create clear SOPs to collect an exhaustive list of locations and venues where individuals could be accessed for an interview and then develop a comprehensive sampling frame consisting of all the relevant time-location clusters to ensure that all locations or venues and day-time periods are listed. Ideally, the same methods can be repeated in subsequent surveys. It is not sufficient to simply copy the list of venues used in a previous survey. Sources of venue information Sources of venue information include key informant interviews and focus group members, websites, newspapers and field observations. The location or venue data-collection form or site information sheet is the data instrument to record all information about each location or venue. The data collected through this form informs the location or venue universe and the time-location cluster sampling frame. The information collected is used to decide if the location or venue is appropriate for the survey, what day-time periods the target population is present and which venue-day-time periods should be listed on the sampling frame, as well as logistical considerations. Potential information to be collected about each venue needed to construct the time-location sampling frame is listed in Table A-9.4, and examples of forms to capture the information collected are in Appendices I-19 and I-20. The data needed to execute a TLS survey differs by target population, culture and location, and may be guided by the formative assessment findings or observing some locations or venues themselves. Missing location or venue types The formative assessment may reveal that there are locations or venues where the target population gathers, but where it is impossible or unsafe to approach, recruit and interview them. For example, it may be unsafe (e.g. abandoned railway platforms) or impossible (e.g. brothel owner refuses to allow entry of survey staff) to conduct surveys at specific locations. Investigators should explore what subset of the population might be missed at these locations or venues and the potential sampling bias that may be introduced because these locations or venues are not in the sampling frame. This information should be reviewed when writing survey results to better understand survey limitations. Location or venue suitability This refers to criteria that a location or venue must meet in order to be included in the sampling frame. A location or venue is generally suitable if there are day-time periods with potential participants; if interviewing and HIV testing
Location or venue attendance patterns The SOP for construction of the location or venue universe and location or venue sampling plan should outline how to determine if attendance patterns are different at different time periods. If the attendance pattern differs significantly over different times of the day or different days of the week, or if the characteristics of the venue attendees belonging to target population change over time, the same location or venue may be included in the sampling frame multiple times based on changes in the attendance patterns over time/day. Multiplicity Multiplicity refers to the possibility that individuals may attend multiple locations or venues during the sampling period and therefore have a higher sampling probability. The most practical way to account for multiplicity is to ask each survey participant how often they attend other venues. This information can be used to calculate weights at the individual level to account for the frequency with which individuals attend venues. Approach criteria Approach criteria are used by the survey staff to decide who should be approached (and counted) for recruitment into the survey. These criteria are not identical with survey eligibility criteria; approach criteria allow for a decision about whether to count and approach an individual based on observation only. For example, in a survey of FSW, staff should approach only women who appear to meet the minimum age for participation. All survey recruiters need to use the same approach criteria, to avoid bias. The approach criteria are used to quickly gauge who is likely to meet survey eligibility criteria since both target population members and noneligible individuals may be present. The approach criteria should be informed by the formative assessment and venue observation (see Section B-1.3 for more information on approach criteria). Method for selecting potential participants to approach at a sampling event There are several ways to systematically sample potential participants at the selected time-location cluster (see Section B-1.3). Selecting a method beforehand can help reduce sampling bias because some potential participants may appear more approachable than others. Rehearse the method before using it in the survey. It is possible to have more than one approach in a survey period, but it should not change during a sampling event. Community liaison It may be helpful to enlist the help of a community liaison to identify the potential participants who could be approached and recruited at a selected conventional
is logistically feasible at or near the venue; and where safety for staff and participants can be assured.
86
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
or time-location cluster. This person can be a member of the community being surveyed and who could help in establishing rapport with the community. Length of sampling events Sampling events can vary to between 2 and 4 hours. Sampling events typically are of equal length, but may differ in order to allow for more recruitment at low volume clusters, or less recruitment at high volume clusters. If a venue is active for a long time, for instance 6 pm to 2 am, it is better to break the time into two separate periods, or venue-day-time events, to avoid survey teams having to spend excessive amounts of time at a venue once the required number of people have been sampled. Management permission Location or venue owners or managers, where relevant, may be approached for permission to conduct sampling events in the future. If they refuse, some information about the location or venue should be collected to provide some measure of possible sampling bias. This information includes numbers and types of target population members. Method for obtaining the actual measure of size At each sampling event, the individuals who are eligible to be screened or recruitment (i.e. who meet the approach criteria) are counted (i.e. enumerated) and this number is recorded by a “counter”. For any timelocation cluster, the counter will be present for the entire time duration (even if interviews are completed before the end of the time slot) and will count each and every eligible and attending individual. Plan for interviewing or testing outside the location or venue Sometimes it is not possible to find an interview place or site at the selected location or venue that is appropriate to conduct interviews and testing. The protocol should account for this by either allowing for alternate-location or other-day interviews. It is fairly common to approach, screen and recruit people at a location or venue and then ask them to come to a different location for interview and biological testing. • alternate-location interviews: When selected locations or venues cannot provide privacy, a tent or van nearby may be used, and participants escorted from the venue to the nearby survey location (it should be within walking distance). • other-day interviews: Appointments on other days can be made if it is impossible to conduct interviews and testing at the time of enrolment. In this case, recruiters provide a survey ID and an appointment card with a time, date, and location for the interview. Use of this approach should be limited
because some recruited persons may not come to the interview, leading to higher nonresponse. It may also complicate confidentiality. Still, it may be preferable to accommodate deferred data collection for some recruits or venues rather than losing data or having biased sampling. Steps to determine the venue universe Determining the venue universe essentially means to create a sampling frame that will then be used to select the clusters for inclusion in the survey. The following steps are discussed below: • develop an SOP on how to create the sampling frame (describing the “venue universe”); • identify and describe the locations or venues; • observe the locations or venues; and • review and finalize the venue universe description. Step 1: Develop an SOP for assessing the location or venue universe • data elements: Decide on the information required for each location or venue. - number of attendees and time slots: What information is needed to decide on the time periods to include on the sampling frame? - venue attendance pattern: What information is required to determine if the venue is attended by people with different levels of risk at different times? - multiplicity: What information is required to determine if individuals at the venue also attend other venues? - safety and logistics: Where should data collection and testing take place? • interviewees: Decide whom observers should talk to and what topics they should address. • update the venue observation form as appropriate (Appendix I-20). Step 2: Identify and describe the locations or venues • data sources: Identify key informants or form focus groups. Key informants (single interviewees) should be knowledgeable about potential venues; they may be identified from among venue patrons, outreach service providers, venue managers or anyone else with good knowledge about the relevant venue “scene”. After some key informant interviews, focus group discussions may be conducted to confirm or refute observations made in key informant interviews and gather additional information. - begin by asking key informants and focus group participants to identify locations or venues where the target population gathers and add them to the location or venue universe.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
87
- find additional venues on the internet, in publications geared towards the target population, from outreach organizations and from survey staff. • the location or venue universe should include all venues. Locations or venues that are not deemed “suitable” will be eliminated from the location or venue sampling frames.Determine venue type: Assign a code to each venue type that will be part of the venue code (e.g. bars = B, street corner = S). • assign a venue ID: Use the location or venue type designation in combination with a number to create a unique venue ID (e.g. S3). Consider using a standardized code for province or geography if the survey is being conducted in multiple areas. • add the venue’s name and geographical information (address or street corner location) so that staff can visit the venue for observation. Step 3: Observe the locations or venues Start with locations or venues that are less understood and move to those that are more familiar. Use the
venue observation form (at Appendix I-20) to document observations at each venue. • establish rapport with the venue owner or manager; talk to venue staff and attendees. • determine venue suitability using the venue suitability criteria. • identify day-time periods appropriate for sampling events. • gather other logistical information useful for field staff. Step 4: Review and finalize the venue universe information Review the gathered venue universe information for completeness and accuracy. Determine venues and day-time periods suitable for inclusion in the timelocation cluster sampling frame (See Chapter B for more information about construction of the sampling frame).
Table A-9.4 Potential information to be collected about each location or venue (see Appendix I-20: Venue observation forms) Venue information • Venue type • Venue name • Venue code • Venue site contact info • Address of venue • Comments/description
Time periods • Minimum and maximum number of individuals affiliated with a particular location or venue time slot • Operational days of the venue • Peak and nonpeak days (days of the week when the maximum/minimum number of individuals are likely to be found) and EMoS on those days • Peak and nonpeak times (times of the day when the maximum/minimum number of individuals are found) and EMoS at those times
Target population observation • Potential language barriers • Stability of location or venue • Visibility of location or venue • Observed behaviours • Observations about demographic characteristics • Seasonal variations • Mobility within as well as across location(s) • Detailed sketch of venue layout • Venue boundaries • Indicate intersections, landmarks to highlight venue boundaries • Location where key informants were spoken to • Locations where target population are grouped (if applicable) • If location or venue is a room in a building, be specific about how to find the room • If location or venue is a stretch of a street, show the side(s) of the street • Participant flow: use arrows to indicate the direction of entry to the site
Venue logistics • Safety concerns? • Potential interview/ testing location outside venue needed? • Barriers to recruiting? • Barriers to interviewing? • Barriers to testing? • Opened/closed since initial venue listing creation? • Parking for staff or van? • Comments • Space for interviewing, testing specimens at or around this site • Name and address of place • Nearest landmark • Contact person • Number of rooms available • Condition of rooms • Toilet available to collect urine samples? • Running water available?
The items on this list are provided as an example. Investigators should modify the information needs as appropriate. The location or venue information is confidential data, and should be treated as such. Access to these records should be kept to a minimum number of people, records should be kept in a locked cabinet or encrypted files, and nonessential information should be destroyed when the data collection is over. EMoS, estimated measure of size
88
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-9.6.5 First stage sampling: time-location This section describes the sampling of time-location slots, including creation of the sampling frame from the venue universe and scheduling recruitment events. Sampling frame construction and time-location cluster selection Step 1: Create the sampling frame in database form from the information in the venue universe. The “time-location cluster sampling frame” is a database with one line (row, record) for each time-location to be sampled from. The time periods should be chosen so that they are long (or short) enough to yield a sufficient (but manageable) number of sampled participants. Within this time slot, the patrons frequenting the venue should not change (e.g. patrons frequenting a bar early evening may differ from those later in the evening). The EMoS (i.e. the likely or anticipated number of eligible venue patrons or attendees) can be derived as part of the formative assessment or through venue observation, and should be generated with as much care as possible. The variables included are location or venue ID, location/venue name, time period and EMoS. Figure A-9.5 Sampling frame Order # Location (name, address) 1 2 3 4 5 ... 18 Venue 1 Venue 1 Venue 1 Venue 2 Venue 2 Venue ... Venue 8
Step 2: Decide whether EPS or PPS is appropriate. Regardless of the approach chosen, save the selected clusters with their unique cluster IDs on a separate worksheet for planning fieldwork. Maintain the original sampling frame for reference and in case a second round of cluster selection is required. Document the sampling information in accordance with the steps above: this information is required to calculate sampling probabilities and nonresponse. Step 3: Select clusters with EPS or PPS. Step 4: Create a calendar (see Figure A-9.6) for each field team. • step 4.1: Ascertain staff availability; determine the dates and times the field staff will NOT be available to conduct recruitment events because of holidays, vacations or other planned absences. These dates and times should be blocked off the calendar. • step 4.2: Block off other dates and times on the calendar when events cannot occur. Also consider staff burnout; in some settings, field teams are given a week off after 4 weeks of field work to return to their families.
Time 9 pm–11 pm 11 pm–midnight Midnight–2 am 8 pm–10 pm 10 pm–11 pm ... 8 pm–10 pm
Estimated measure of size 8 15 10 5 15 ... 14
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
89
Figure A-9.6 Calendar example with dates and times when sampling cannot take place blocked out January 2017 Sunday Monday Tuesday Wednesday Thursday Friday Saturday
1 2 8–10 PM: S1 7–9 PM: BR2 11 PM–1 AM: B2 8 9
3 4 5 6 7 8–10 PM: BR1 10 PM–Midnight: 7–9 PM: B1 6–8 PM: BR4 7–10 PM: S4 11 PM–1 AM: S2 S3 10 PM–1 AM: BR3 9 PM–Midnight: 11 PM–1 AM: B3 B3 10 11 12 13 14
Step 5: Plan and schedule sampling events. Schedule the sampling event for each of the selected time-location clusters. The specifics on how to do this should be determined in the local context and consider factors such as field team size, geography and infrastructure/roads. It may be helpful to order the selected venues geographically; for example, from north to south or near a main highway.
A-9.6.6 Eligibility and response rates in conventional and time-location cluster sampling: summary of numerical characteristics Table A-9.5 highlights the information needed to facilitate the calculation of sampling probabilities at each stage and the response rate. There are a number of factors that influence the selection probabilities in cluster sampling, especially time-location sampling.
Table A-9.5 Data requirements for calculating sampling probabilities and response rate Row Characteristic Conventional Cluster Sampling Time-location Sampling
Approach A # Selected individuals Randomly or systematically selected from a list of eligible individuals Systematically selected and approached from individuals who meet the approach criteria in a selected time- location cluster The selected person refuses approach n/a The selected person agrees to hear about the survey and be screened for eligibility
B C D
# Refused approach # Not present # Accepted approach (A-B)
The selected person refuses approach The selected person is not present when the survey team is therea The selected person agrees to hear about the survey
E
Acceptance rate (#Accepted/#Approached, D/A)
90
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Row
Characteristic
Conventional Cluster Sampling
Time-location Sampling
Eligibility screening F G H I J K L M # Age or language ineligible # Who live outside geographical catchment area # Previous participantsc # Who were not the correct sex/ gender # Who had not engaged in risk behaviour during the time period # Ineligible using any one of the eligibility criteriac # Eligible for participation (E-K) % of those approached who were eligible for participation (L/D) n/a The list provided to or created by the survey team ideally includes only eligible participantsb
Participation N O P Q # Who refused to participate after eligibility screening or consent # Who consented to participation in survey # Who consented to provide a specimen for HIV testing Survey participation rate (O/D) Normally, in CCS, only eligible individuals will or should be in the sampling frame and therefore offered participation, whereas in TLS that is not always possible. Individuals who are ineligible should not be considered when computing the participation rate.
R S T U V a b
HIV testing participation rate (P/D) # Survey records lostd # Incomplete survey recordse # Complete records available for analysis Enrolment rate
Schedule up to two return trips to interview those who are not present, if possible. If a large proportion of individuals interviewed appear to be ineligible, review the eligibility criteria with site stakeholder, re-create the list and resample. c Approached individual may tell the recruiter they have already participated, or they may be identified by another one of the field staff. This may happen when the reimbursement for participation is high. d Occasionally, paper survey forms are inadvertently destroyed or the electronic record is lost during transfer. e Some analysts may choose to include incomplete records or records of those who consent to the survey but not the HIV testing in their analysis. This number is meant to represent the number of records with complete information.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
91
A-9.7 Respondent-driven sampling A-9.7.1 Overview of respondent-driven sampling Respondent-driven sampling (RDS) is a peer-driven chainreferral sampling method that was first implemented in 1994 by Douglas Heckathorn (10). This probability-based sampling methodology is useful for sampling populations that lack a sampling frame. It relies on participants’ ability and willingness to identify and recruit other members of the target population into the survey. In RDS, recruitment relies on peers recruiting their peers rather than survey investigators selecting the sample. RDS surveys start with a small number participants called “seeds” who are recruited by investigators or survey staff. These seeds participate in the survey and are then asked to invite peers within their social network to participate in the survey. A social network can include friends, colleagues and other personal contacts who know one another and who share characteristics that define them as members of the target population. Thus, for a survey among PWID, investigators enrol a small number of PWID seeds who then invite PWID they know to join the survey. These participants then recruit their peers, who are also invited to recruit other PWID who have not already participated in the survey. This process continues until the sample size is reached. Differences between RDS and other chain-referral methods Chain-referral sampling comprises a group of sampling methods that begin with a convenience sample of initial subjects (seeds) through which wave 1 subjects are recruited; these wave 1 subjects in turn recruit wave 2 subjects, and so on (11). Snowball sampling is perhaps the best known chain-referral sampling method. RDS is similar to snowball sampling in that the target population must be socially networked so that participants can invite their peers into the survey. Although snowball sampling is useful for rapidly identifying potential participants, the methodology is biased. For instance, participants may recruit peers with characteristics similar to their own, producing a sample that is not representative of the target population. Additionally, if participants are allowed to refer an unlimited number of peers, people with larger social networks will dominate the sample. These limitations prevent investigators from drawing inferences from the sample and generalizing them to the entire target population (12). In snowball sampling, investigators neither need to record who referred whom or estimate participants’ personal network sizes, whereas in RDS they do.
Key terms Convergence: The point at which the sample characteristics no longer change, no matter how many more individuals enter into the sample. Convergence is an indication of seed dependence, like equilibrium, but is based on the population estimate for a given variable. Whereas equilibrium is based on the wave, convergence is based on the order of enrolment into the survey. Coupon: An invitation – by participants to peers – to enrol in an RDS survey. Equilibrium: The point at which the distribution of participant characteristics is similar between waves Page 128 of 296 (1). Equilibrium, like convergence, is an indication of seed dependence with respect to RDS, but is calculated based on the sample. Personal Number of personal contacts or peers network size who are part of the target population. (or degree): Such contacts may include friends, acquaintances, partners, coworkers and relatives. Primary Also called primary incentive, this is compensation: money or an item given after the first visit is completed to compensate for time and transportation. Recruitment All participants sampled through the same chain: seed and connected through the resulting waves constitute a recruitment chain. Secondary Also called secondary incentive, this is compensation: money or an item given at the second visit for each recruited peer that participates in the survey. Compensation for transportation may also be provided at the second visit. Seed: A participant that is recruited by investigators or survey staff. All RDS studies begin with the selection of at least one seed. Social A social structure made up of a network network: of personal contacts who share characteristics that define them as members of the target population. Wave: The set of participants in a given number of recruitments from a seed. Individuals in wave 1 are those recruited directly by the seed. Individuals in wave 2 are recruited by those in wave 1.
92
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
RDS minimizes biases associated with chain-referral sampling by limiting the number of peers that each participant can recruit, which prevents people with larger networks from dominating the sample. It requires longer recruitment chains (i.e. more waves), but these should eventually result in a sample that is no longer biased by the characteristics of the purposefully selected seeds. RDS is both a sampling method and an analytical method. RDS analysis considers that participants have different probabilities of being recruited and uses statistical weights based on the participant’s personal network size or degree, and their recruitment patterns to estimate the prevalence of variables of interest. A participant’s personal network size reflects the number of peers that could have recruited the person into the survey. When to use RDS RDS can only be used if the target population is socially networked and its members can recognize and recruit one another. All subpopulations and subgroups within the network must be connected. For example, to conduct an RDS survey among FSW, the subpopulation who sell sex in brothels should have at least some ties to the subpopulation who sell sex on the street. If this is not the case, investigators may want to conduct two separate RDS surveys: one among brothel-based SW and another among street-based SW. Social networks may also differ by other things such as neighbourhoods, age, ethnicity or nationality. Because RDS recruitment is peer driven, it is often used when members of the target population do not congregate in large numbers in public venues. It is also employed when stigma and discrimination towards the target population make other sampling methods difficult. However, RDS is not only used among hard-to-find populations. Groups recruited using RDS include university students (13), migrant populations (14), and heterosexual men (15) and women at high risk of HIV (16). Strengths and limitations of RDS Investigators choose RDS as their sampling method for numerous reasons; for example: • it allows for sampling of less visible segments of the target population; • the target population conducts recruitment; • the sample can include individuals who attend venues and those who do not; • it maintains or increases the privacy of the target population; • it facilitates field operations and minimizes logistical needs; • it minimizes the number of additional questions needed to create sampling weights; and • it allows for computing population-based estimates.
On the other hand, limitations of RDS include the following: • RDS relies on several assumptions that must be met in order to produce valid results, and it can be difficult to measure some of these assumptions, especially the assumption of random recruitment within a network; • some social network components may be separated from others by factors such as geography, language or age, which results in a bottleneck in which recruitment occurs within one part of the network that is not linked to other parts; • it is difficult or impossible to confirm the validity of RDS-based survey estimates; • analysis of RDS-based data is challenging; and • it is difficult to measure nonresponse in recruitment (e.g. when peers are offered a coupon but choose not to join the survey). RDS theory: functional and analytical assumptions Before implementing an RDS survey, investigators should understand what makes RDS a probability-based sampling methodology, and the assumptions upon which it relies. RDS involves two types of assumptions: functional and analytical (17). Functional assumptions Functional assumptions of RDS are that: • respondents know one another as members of the target population; • respondents are linked by a network composed of a single component; and • the target population is large enough to allow sampling with replacement. An RDS survey will fail or be compromised if these functional assumptions are not fulfilled or are violated substantially. These assumptions, which should be examined during the formative assessment before RDS protocol development, are discussed below. Respondents know one another as members of the target population RDS assumes that members of the target population can recruit other eligible members to be in the survey. Therefore, RDS is feasible only when individuals can recognize one another as part of the target population. For example, MSM may recognize one another because they have sex with and socialize with other MSM. However, RDS may not be successful in situations where most SW work in isolation (e.g. by finding clients on the internet or working in private homes). Respondents are linked by a network composed of a single component Members of the target population must be part of a larger social network where each individual is directly
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
93
or indirectly socially connected to all other individuals through a series of social connections. These connections must be close enough to facilitate recruitment; in other words, individuals must interact with one another often enough so that a sufficient number of people can be recruited in the time available. Participants should know multiple people who meet the survey’s eligibility criteria. Participants should be able to recruit both people with whom they have weaker relationships, such as acquaintances, and people with whom they have stronger ones, such as friends and family. Bottlenecks occur when recruitment becomes “trapped” within a certain subgroup of the survey population. Members of the subgroup are socially linked to other members of the survey population; however, recruitment becomes bottlenecked because they disproportionately recruit other members of the same subgroup. Bottlenecks reduce the precision of RDS estimates because the bottlenecked group is more homogenous than members of the larger survey population (18). The target population is large enough to allow sampling with replacement Sampling with replacement means that each participant can be sampled more than once. With a fixed population, a person’s chance of being selected for participation would be the same at any stage of the sampling process. In reality, sampling in RDS occurs without replacement (19). Assuming that sampling is with replacement allows investigators to model a participant’s probability of selection as being proportional to the participant’s personal network size. A relatively new RDS analysis estimator (the successive sampling estimator) assumes sampling without replacement and instead models each participant’s probability of selection as a function of the participant’s personal network size relative to all the personal network sizes remaining in the population after people already in the sample have been removed (20). If the sampling fraction (the target sample size divided by the number of people in the target population) is less than 10% or 20%, the successive sampling estimator may perform similar to the Volz-Heckathorn estimator (RDS II). Analytical assumptions Two analytical assumptions of RDS need to be met to make inferences from the sample to the target population: • respondents can accurately report their personal network size: and • peer recruitment yields random selection from the recruiter’s network. If these assumptions are not met, the estimators may be biased and thus will not be representative of the target population.
Respondents can accurately report their personal network size The size of a participant’s personal network is the number of personal contacts or peers who are members of the target population. Contacts may include friends, acquaintances, sex partners, coworkers and relatives, as long as they meet the eligibility criteria. To determine a participant’s personal network size, survey staff ask a series of questions about the number of eligible people known to the participant. An estimate of personal network size is crucial for calculating sampling weights for RDS, because it determines the probability that a participant will be recruited into the survey. Specific network questions are described further in this chapter under “Additional questions about personal network size”. Peer recruitment yields random selection from the recruiter’s network The assumption of random selection implies that every person in the participant’s network has an equal probability of being recruited. Survey staff should not give participants direction about whom to recruit except to say other members of the target population. Random selection may be influenced by factors such as how often peers see each other, how far they live or work from each other, or how well they know each other. Participants may recruit disproportionately from their own subgroup (e.g. they may be more likely to recruit people from their own age group). The assumption of random selection would be violated if one subgroup was more likely to receive or accept coupons than another. However, if the variable associated with biased recruitment (e.g. age group) is not correlated with any of the primary outcome variables (e.g. HIV status), the RDS estimates for those outcomes may not be biased. The random recruitment assumption is plausible only if members of the survey population have reasonably easy and comfortable access to the survey site, an appropriate time frame is used for the network-size question, and appropriate compensation is provided. However, nonrandom recruitment, if it occurs, will not necessarily bias the RDS estimate if recruitment is not correlated with a variable important for estimation; for example, with the survey’s main outcome or degree (21). Preparing for RDS survey implementation For an RDS survey to gather a probability-based sample, it must have a way to track recruitment (usually through the use of numbered coupons) and the personal network size of each participant. Most participants, with the exception of those at the very end of the survey, are given uniquely numbered coupons that they can use to recruit peers. The coupon serves as an invitation to join the survey. The coupon numbers
94
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
allow investigators to map the recruitment process and document the chain of recruits stemming from each participant. A participant’s personal network size, or “degree”, approximates how many peers the participant can choose to recruit from. Investigators estimate this number by asking each participant how many peers they know who also know them; the investigators can then use this number to adjust for bias and, together with the recruitment information, to estimate the probability of an individual being selected into the survey. All estimates adjusting for RDS assume that these relationships are “reciprocal”; that is, the participants know the peers and the peers know the participants. With these two measures, investigators can make the necessary analytical adjustments to yield a probabilitybased sample. If these measures were not factored into the analysis, an RDS sample would still be biased, and considered a convenience sample. It is not essential to offer respondents compensation for participating in an RDS survey, but it is strongly recommended because participants use their time, effort and money to recruit peers. RDS uses a dual compensation system to encourage participation and
recruitment. Primary compensation is given at the end of the first visit. Secondary compensation for the successful recruitment of peers is given during the second visit. Steps in recruitment The steps in RDS recruitment are as follows: 1. investigators select and enrol a few individuals from the target population to serve as seeds. Investigators administer questionnaires and biological tests to each seed, and give them a predetermined number of coupons with which to recruit their peers. 2. the seeds use the coupons to invite their peers to join the survey (further information on how many coupons to give each participant is provided below). 3. individuals who received a coupon from a seed report 4. individuals recruited by a seed who then participate in the survey constitute the first wave. In turn, their recruits who then participate in the survey form the second wave (Figure A-9.7). 5. the recruiting process continues until the survey achieves the calculated sample size or higher until the sample reaches equilibrium or converges for the main outcome and salient variables.
Figure A-9.7 Recruitment chain of eight waves generated from one seed Seed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
95
Ideally, the RDS recruitment process generates long recruitment chains, which are needed to reach equilibrium or convergence (22). When equilibrium is reached for a given variable, the sample’s values for that variable are stable and no longer influenced by the characteristics of the seeds, which were chosen by investigators, whereas when convergence is reached, the estimators for that variable are stable and are no longer influenced by the characteristics of the seeds. Although it is important to reach a stable composition with respect to key variables, recruitment should continue until the predetermined sample size has been reached. Not all seeds may be successful in recruiting peers or generating long recruitment chains. It is often the case that only one or two seeds achieve recruitment chains long enough to attain equilibrium and those individuals account for most of the achieved sample size. However, this is not a problem as long as the social network fulfills the assumption of one complete social network (e.g. in Figure A-9.8, that people in the blue chain could have recruited people in the green chain). Figure A-9.8 shows an example of an RDS sample where each successive node and arrow from a seed represents one wave of recruitment. One seed (the grey seed in the figure) produced only one wave of recruitment whereas another (the green seed) produced 13 waves. Because these are all part of one complete network component, the maximum number of waves for any one chain represents the maximum for the entire sample. Figure A-9.8 Recruitment by 10 seeds
Preparations for implementation Formative assessment results play a key role in informing investigators about how to prepare for implementation. Coupon design and distribution Coupon design. Coupons are unique survey invitations that participants give to their peers. The coupons are usually paper and can be passed between people. In web-based RDS surveys (23) they may also be electronic (Appendix I-21 provides more information about unique participant codes). Regardless of the format, coupons should provide key information about the survey to potential participants, including: • coupon start and expiry date; • coupon ID number; • survey name; • survey purpose – for surveys in which it is necessary to protect the safety and confidentiality of the participants, this may need to be generic (e.g. “men’s health survey” instead of “survey of men who have sex with men”); • survey location, phone or email address; • hours and days of operation; and • other important information regarding the survey (e.g. amount of compensation). Start and expiry dates. Investigators may choose to indicate the start and expiry dates on the coupon. These
Seed
Seed
96
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
dates inform recruits of when they can visit the survey site. The duration of coupon validity is defined by the investigators and can vary. A future start date (e.g. 1–3 days after a participant’s enrolment) may be used to prevent coupon trading and an underground economy (i.e. where peers are waiting outside the survey site and participants immediately recruit them to redeem their secondary compensation). An expiry date (e.g. 2 weeks after a participant’s enrolment) may motivate peers to come to the survey office sooner and may better ensure that a peer’s enrolment occurs before the recruiter’s second visit during which the recruiter receives compensation for peer recruitment. An expiry date also allows staff to calculate how many valid and expired coupons still circulate “on the streets”. Investigators may decide to allow enrolment of persons presenting with expired coupons as long as they meet the eligibility criteria. Figure A-9.9 Coupon design example Front
Coupon numbering. Each coupon must include a unique ID and be linked to the recruiter. This allows investigators to determine from which recruiter the coupon was received. Coupons should be numbered serially (e.g. 101, 102, 103, 104 and so on). Recruitment should be tracked using a coupon-tracking system such as RespondentDriven Sampling Coupon Manager (RDSCM). To avoid errors and duplicate coupons, coupons should be numbered and printed before implementation. Coupon number as survey ID. Ideally in an RDS survey, the coupon code serves as the survey ID, which links interview, biomarker and other data collected from the participant. The survey ID should also be linked to any laboratory code (as applicable) and a unique participant code (UPC) created to confirm participant identity at the second visit (described below). Figure A-9.9 shows an example of a coupon design.
Address: 123 Main St. (across the street from Store A)
MEN’S HEALTH SURVEY COUPON NUMBER:
Open Monday through Friday 10 a.m. – 7 p.m. Call for more information: XXX-XXX-XXXX Valid between / / and / /
Back Bring this coupon to join! If you are eligible, you can: • Check your health • Receive free treatment • Receive compensation for your time Coupon is not transferrable or valid after expiry date.
Map MAIN STREET SITE
STORE A
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
97
Number of coupons. In most RDS surveys, participants are given up to three coupons (24). If participants are given too many coupons, recruitment chains may be wide rather than long and the sample size may be reached before equilibrium or convergence, the point at which the sample is independent from the seeds (24). Therefore, giving more than three coupons per participant should be considered a last option. If recruitment is poor, first consider speaking with participants and others in the community to determine reasons for poor recruitment, undertaking outreach to promote the survey among the target population, improving recruiter training or adding more seeds. The number of coupons distributed should be reduced only when recruitment is robust. If participants are given too few coupons, the recruitment chains may die out. Survey site selection RDS surveys are usually implemented at one or more fixed locations, or survey sites. Survey investigators should choose a survey site that is easy to get to via public transportation, is secure, offers privacy for participants, is large enough for all survey procedures and is acceptable to the target population. Formative work can help determine how to improve confidentiality and safety of the site. Participants may feel more comfortable if the site is in a residential area; in other contexts, they may prefer a commercial area or a clinic where other people in the area may assume they are patients. Accessibility is a matter of location as well as the days and times the survey site is open. If some potential participants have jobs during the day, a survey site that is only open during normal business hours is likely to exclude them from the sample. Some surveys allocate specific days for first visits and second visits. Survey investigators should decide whether setting up appointments or having drop-in times, or some combination of the two, is most appropriate for the target population. If the recruitment area is large and participants need to travel far to reach any given location, investigators should consider having more than one survey site, and consider the following if selecting multiple survey sites: • participants may try to participate more than once if the survey operates in multiple sites. In this case, investigators may consider having one survey team that alternates between the sites so that only one site is open per day. Staff may then recognize participants who try to enrol again. • coupon management requires greater attention to avoid giving out duplicate coupon numbers. One possibility is for each site to provide unique identifiers in the coupon number itself. For example, all “Site A” coupons could begin with the letter A (e.g. A101, A102,
A103 and so on) and “Site B” coupons begin with the letter B (e.g. B101, B102, B103 and so on). • cross recruitment should be monitored to confirm that some recruiters who participated in the survey at “Site A” recruit peers who participate at “Site B”. Mobile sites. Investigators may consider using a mobile site if the survey population is geographically dispersed, or is in an area with a poor or limited public transportation system, or might not visit a fixed site because of security concerns. Mobile sites should have the same characteristics as a fixed site; that is, they should be safe and comfortable, and have a dedicated area for interviewing and specimen collection. In general, a mobile site should be inconspicuous and should not attract attention. Investigators can operate the mobile site in several locations with assigned hours. Alternatively, they can schedule appointments to meet participants at a convenient public location. The coupons should provide clear instructions about how to find the mobile site. Everyone with a coupon must be given equal opportunity to participate in the survey. Preference must not be given to individuals who live in areas that are more convenient to travel to or visit. All individuals who call to make an appointment should be offered participation if eligible. Site layout and staff Figure A-9.10 shows an example of the ideal layout for an RDS survey site. A survey site should have a reception area where individuals are greeted and can wait between survey procedures. Initial screening for a valid coupon may occur in the same area. Another room or separate space is needed for the coupon manager to screen individuals for eligibility and obtain informed consent. Later, the coupon manager (i.e. an RDS survey staff member responsible for the coupon management system that tracks receipt and distribution of coupons) will explain the recruitment process and give coupons to the participant. An ideal site has more than one interview room, and has space for a counsellor to conduct pre-test and post-test counselling, for specimen collection and for running any tests. A bathroom should be available for participants as a convenience and is also needed for the collection of urine specimens. A kitchen is useful for preparing refreshments for participants waiting in the reception area. It can also serve as a break room for survey staff. Depending on the biomarkers included in the survey, space for a small laboratory may be needed as well. Details of what occurs at the first and second visit for each participant are described in Section B-1.4. For more information on survey sites (25) and details regarding staffing positions required for an RDS survey, see the information on staff selection in Chapter A-14.
98
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure A-9.10 Layout of an RDS site
SCREENING INTERVIEW ROOM #1 INTERVIEW ROOM #2 COUNSELLING ROOM
COUPON MANAGER
ENTRANCE
KITCHEN
BATHROOM
SPECIMEN COLLECTION ROOM
RECEPTION/WAITING AREA
Seed selection Seeds are target population members who have been invited to join the survey by survey investigators, to begin the RDS recruitment process. Factors that should be considered when selecting seeds include the number of seeds, the diversity in characteristics of seeds and the likelihood that the seed will recruit others. Number of seeds. If investigators select too many seeds, the survey site may be overwhelmed with more participants than it can accommodate, and the target sample size may be reached before reaching equilibrium. If investigators select too few seeds, it may take more time to reach the desired sample size. It could also result in some social networks being underrepresented in the sample; for example, people from a certain part of a city. The number of seeds selected may be influenced by the following factors: • calculated sample size – Achieving a larger sample size may require more seeds. • compensation – The higher the level of compensation offered, the easier it may be to motivate participation and recruitment; however, this could also lead to ineligible people trying to enrol. • degree of stigma or hiddenness of the target population – Populations that are subject to more stigma may be more difficult to recruit, in which case investigators may select more seeds if recruitment is not as effective.
• number of connections among survey population members – If the population has many connections (large social networks) this is a good indication that recruitment will be easily sustained. In this situation, fewer seeds may be needed. Seeds should roughly represent the diversity of the target population, including characteristics such as geography, key outcome measure (i.e. HIV positive and negative), socioeconomic status, age and gender. Although diversity is not required, the more representative the seeds are of the target population, the faster the sample may converge towards equilibrium (10, 26). It is also ideal to select seeds that represent a diversity of socially salient variables, which may affect the way each seed recruits. For example, sexual identification may be a factor that influences the way MSM would recruit other MSM and should be considered in seed selection by ensuring that, for example, gay and non-gay identified men are seeds. Ideal seeds have a large social network and, more importantly, are well connected to their peers, trusted and well liked, and communicate well orally to promote survey participation. Seeds should be well informed about the survey and be enthusiastic about recruiting others to participate. Groups or organizations that work closely with the target population are well placed to help identify seeds. Ideal seeds will be able to recruit across subpopulations to ensure that no bottlenecks occur (i.e. that the sample comprises one complete network component) (25).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
99
Compensation Although it is not required to offer compensation for participating in an RDS survey, it is strongly recommended. The amount and type of compensation should be based on findings from the formative assessment, and the protocol should include the method for determining the compensation amount. Investigators should choose an amount of compensation that is sufficient to compensate individuals for their time and effort, but not so much that they participate purely for the compensation, that coupons are bought and sold within the target population, or that people attempt to participate more than once in the survey. Compensation for RDS is best looked at as a remuneration that shows respect for participants’ time and effort in recruitment, and their judgement to use remuneration for personal needs (27). A key component in RDS is peer pressure – a mutual like and respect between peers is a strong incentive for participating. Additional questions about personal network size Personal network size is the number of people a participant knows who are likely to be eligible for survey participation. It is crucial to assess the size of each participant’s personal network in order to weight the data. This information is usually collected through a series of questions that go from broad (e.g. “How many people do you know and they know you who have injected drugs in the past 6 months?”) to specific (e.g. “How many [PWID in this city who are at least 18 years old] have you seen in the past 2 weeks?”) (24). The data from the last question is used as the personal network size. Because these questions produce the weights used for data analysis, they should be asked face-to-face, even in surveys that otherwise use audio computerassisted self-interview (ACASI). This makes it easier for participants to ask clarifying questions.
Examples of network size questions for an RDS survey are: • how many people do you know and they know you who have sold sex in the past 6 months? • of these, how many live in this city? • of these, how many of them are aged 15 years or over? • of these, how many have you seen in the past 14 days? • of these, how many would you consider giving a coupon to? Unique participant code Eligible participants should be assigned a UPC (see Appendix I-21). This code is used at the second visit to verify that the person is the same person who participated in the first visit and was given a specific survey ID. Once a participant’s identity has been confirmed, survey staff can compensate the person for recruitment efforts, provide any outstanding test results, and conduct the second visit interview about recruitment efforts. Investigators can create the UPC by using the answers to a series of questions that only the participant knows the answer to. As such, only the participant will be able to recreate the resulting code (i.e. one that cannot be used to identify the participant). The following pieces of information can be used to create the UPC: • the first two letters of the participant’s first name; • the first two letters of the participant’s mother’s name; • the participant’s age in years at the time of the initial interview; and • the first two letters of mother’s place of birth. In this example, a person named JOHN DOE – whose mother’s name is MARY, who is aged 18 years and whose mother was born in RDSVILLE – would have the following UPC.
JOHN DOE UPC = JOMA18RD
MARY
18
RD SVILLE
The code can be created during the eligibility screening and verified by the receptionist or screener to prevent duplicate participation. Alternatively, some surveys use fingerprint scanners that create a unique alphanumeric code without storing the actual fingerprint image, thus retaining anonymity. This code cannot be used to reconstruct the fingerprint. Upon re-scanning the same finger at the second visit, the same code will be generated and the participant’s record can be retrieved. There are guidelines for the reporting of RDS data (28). Also, for guidance on analysis of RDS, WHO has produced a guidance on RDS analysis (29). The University of California in San Francisco has an operations manual for RDS studies, as well as other resources for RDS (30).
100
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-9.8 References 1 2 3 McCreesh N, Frost S, Seeley J, Katongole J, Tarsh MN, Ndunguse R et al. Evaluation of respondent-driven sampling. Epidemiology. 2012;23(1):138. Foreman EK. Survey sampling principles. New York, Marcel Dekker. 1991. Magnani R, Saidel T, Rehle T. Sampling strategies for monitoring HIV risk behaviors. In: Rehle T, Saidel T, Mills S & Magnani R (eds.), Evaluating programs for HIV/AIDS prevention and care in developing countries: a handbook for program managers and decision makers, Arlington, VA, Family Health International. 2001. Munro HL, Pradeep BS, Jayachandran AA, Lowndes CM, Mahapatra B, Ramesh BM et al. Prevalence and determinants of HIV and sexually transmitted infections in a general population-based sample in Mysore district, Karnataka state, southern India. AIDS. 2008;22:S117–S125. Som RK. Practical sampling techniques. Boca Raton, FL, CRC Press. 1996. UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines for second generation HIV surveillance: an update: know your epidemic. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/ World Health Organization (WHO); 2013 (http://apps.who.int/iris/bitstream/10665/85511/1/9789241505826_eng. pdf, accessed 3 August 2016). Patton M. Qualitative evaluation and research methods. Beverly Hills, Sage publications. 1990. Magnani R, Sabin K, Saidel T, Heckathorn D. Review of sampling hard-to-reach and hidden populations for HIV surveillance. AIDS. 2005;19 Suppl 2:S67–72 (http://www.ncbi.nlm.nih.gov/pubmed/15930843, accessed 30 August 2016). CDC. National HIV behavioral surveillance system: men who have sex with men – Round 4: operations manual. Atlanta, GA: Centers for Disease Control and Prevention (CDC); 2014 (http://www.cdc.gov/hiv/pdf/statistics/ systems/nhbs/nhbs-msm4-operations-manual--version-date-05-16-2014.pdf, accessed 4 August 2016).
4
5 6
7 8
9
10 Heckathorn DD. Respondent-driven sampling:a new approach to the survey of hidden populations. Soc Probl. 1997;44(2):174–199. 11 Heckathorn DD. Snowball versus respondent-driven sampling. Sociol Methodol. 2011;41(1):355–366. 12 Van Meter KM. Methodological and design issues: techniques for assessing the representatives of snowball samples. In: Lambert EY (ed), The collection and interpretation of data from hidden populations. NIDA Research Monograph 98, Rockville, MD, National Institute on Drug Abuse. 1990:31–43 (https://archives.drugabuse.gov/pdf/ monographs/98.pdf, accessed 8 August 2016). 13 Rutherford GW, Anglemyer A, Bagenda D, Muyonga M, Lindan CP, Barker JL et al. University students and the risk of HIV and other sexually transmitted infections in Uganda: the Crane survey. Int J Adolesc Med Health. 2014;26(2):209–215. 14 Montealegre JR, Risser JM, Selwyn BJ, McCurdy SA, Sabin K. Prevalence of HIV risk behaviors among undocumented Central American immigrant women in Houston, Texas. AIDS Behav. 2012;16(6):1641–1648. 15 Townsend L, Johnston LG, Flisher AJ, Mathews C, Zembe Y. Effectiveness of respondent-driven sampling to recruit high-risk heterosexual men who have multiple female sexual partners: differences in HIV prevalence and sexual risk behaviours measured at two time points. AIDS Behav. 2010;14(6):1330–1339 (http://www.ncbi.nlm.nih.gov/ pubmed/20625926, accessed 8 August 2016). 16 Townsend L, Zembe Y, Mathews C, Mason-Jones AJ. Estimating HIV prevalence and HIV-related risk behaviors among heterosexual women who have multiple sex partners using respondent-driven sampling in a high-risk community in South Africa. J Acquir Immune Defic Syndr. 2013;62(4):457–464.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
9. Sampling strategy
101
17 Heckathorn DD. Extensions of respondent‐driven sampling: analyzing continuous variables and controlling for differential recruitment. Sociol Methodol. 2007:151–207. 18 Goel S, Salganik MJ. Assessing respondent-driven sampling. Proc Natl Acad Sci U S A. 2010;107(15):6743–6747 (http://www.ncbi.nlm.nih.gov/pubmed/20351258, accessed. 19 Gile KJ, Handcock MS. Respondent-driven sampling: an assessment of current methodology. Sociol Methodol. 2010;40(1):285–327 (http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3437336&tool=pmcentrez&rend ertype=abstract, accessed 3 August 2016). 20 Gile KJ. Improved inference for respondent-driven sampling data with application to HIV prevalence estimation. J Am Stat Assoc. 2010;106(493):135–146 (https://arxiv.org/pdf/1006.4837.pdf, accessed 8 August 2016). 21 Volz E, Heckathorn DD. Probability based estimation theory for respondent driven sampling. Journal of Official Statistics. 2008;24(1):79. 22 Gile KJ, Johnston LG, Salganik MJ. Diagnostics for respondent-driven sampling. J R Stat Soc Ser A Stat Soc. 2015;178(1):241–269 (http://www.ncbi.nlm.nih.gov/pubmed/27226702, accessed 8 August 2016). 23 Wejnert C, Heckathorn DD. Web-based network sampling: efficiency and efficacy of respondent-driven sampling for online research. Sociol Methods Res. 2008:105–134 (http://www.respondentdrivensampling.org/reports/web_ rds1.pdf, accessed 8 August 2016). 24 Johnston LG, Malekinejad M, Kendall C, Iuppa IM, Rutherford GW. Implementation challenges to using respondentdriven sampling methodology for HIV biological and behavioral surveillance: field experiences in international settings. AIDS Behav. 2008;12(4):S131–S141 (http://www.ncbi.nlm.nih.gov/pubmed/18535901, accessed 3 August 2016). 25 WHO Regional Office for the Eastern Mediterranean. Introduction to HIV/AIDS and sexually transmitted infection surveillance: Module 4: Supplement. A guide to using RDS Analyst and NetDraw. World Health Organization (WHO) Regional Office for the Eastern Mediterranean/Joint United Nations Programme on HIV/AIDS (UNAIDS); 2013 (http://applications.emro.who.int/dsaf/EMRPUB_2014_EN_1686.pdf?ua=1&ua=1, accessed 8 August 2016). 26 Heckathorn DD. Respondent-driven sampling II: deriving valid population estimates from chain-referral samples of hidden populations. Soc Probl. 2002;49(1):11–34. 27 Semaan S, Heckathorn DD, Des Jarlais DC, Garfein RS. Ethical considerations in surveys employing respondentdriven sampling. Am J Public Health. 2010;100(4):582. 28 White RG, Hakim AJ, Salganik MJ, Spiller MW, Johnston LG, Kerr L et al. Strengthening the reporting of observational studies in epidemiology for respondent-driven sampling studies: “STROBE-RDS” statement. J Clin Epidemiol. 2015;68(12):1463–1471. 29 WHO Regional Office for the Eastern Mediterranean. Introduction to HIV/AIDS and sexually transmitted infection surveillance: Module 4: Introduction to respondent-driven sampling [WHO-EM/STD/134]. Eaton, L. A.: World Health Organization (WHO) Regional Office for the Eastern Mediterranean; 2013 (http://applications.emro.who. int/dsaf/EMRPUB_2013_EN_1539.pdf, accessed 8 August 2016). 30 Global Health Sciences. Toolbox for conducting integrated HIV bio-behavioral surveillance (IBBS) in key populations. San Francisco: University of California; 2016 (http://globalhealthsciences.ucsf.edu/pphg/gsi/epidemiologicsurveillance/ibbs-toolbox, accessed 30 August 2016).
102
9. Sampling strategy
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms
10. Sample size requirements
This chapter describes how to inform and compute sample size requirements for probability-based surveys, including one-time surveys and surveys repeated over time. Although many software packages and online calculators for sample size calculations are available, it is advisable to consult a statistician for such calculations, especially for complex sample designs or the planned use of test statistics in data analysis.
Confidence The estimated range of values which likely interval: includes the true, unknown population parameter. Design effect A factor expressing how much larger a (DEFF): sample size for a complex survey design must be compared to simple random sampling. It is the ratio of the observed variance over the (expected) variance if it were based on a simple random sample. Power: The probability of not making a type II error (i.e. accepting a null hypothesis that is false). Precision: How close an estimate is to other estimates made using the same methodology. The higher the precision, the narrower the confidence interval. Variance: A value that indicates the dispersion (scattering) of a variable’s values around its average (mean). Z-score: Represents the distance between the raw score and the population mean in units of the standard deviation.
A-10.1 Considerations for determining sample size Biobehavioural survey (BBS) sample size requirements are usually calculated with one of the following goals in mind: • to estimate the prevalence of a certain characteristic at a single point in time with some specified level of precision (e.g. HIV prevalence among men who have sex with men [MSM] with a defined margin of error); or • to detect differences between two groups or changes in a certain characteristic over time (e.g. whether condom use at last sex with a client changed among male sex workers (SW) between survey rounds, or whether street-based SW have a lower or higher HIV prevalence than brothel-based SW). Before calculating a sample size, investigators should consider the types of estimates required, and the variable(s) on which to calculate sample size. Types of estimates required The sample size required to detect changes in estimates over time or differences between subgroups is typically larger than that required to estimate the prevalence of a certain characteristic at a single point in time. The purpose of the survey and whether subsequent rounds are expected should be considered when calculating sample size, to ensure sufficiently large sample sizes. Most BBS base their sample size requirements on certain population attributes of interest measured as proportions. Examples include the proportion of participants who test HIV positive, or the proportion of participants who shared injecting equipment the last time they injected drugs. Given the importance of viral load suppression in combating the HIV epidemic, investigators are increasingly basing sample size
In practice, target and actual sample sizes are often influenced both by statistical and resource considerations. Actual sample size achieved may vary depending on the sampling speed, sampling duration or challenges in implementation. This chapter applies only to surveys using probability samples. For surveys using nonprobability (i.e., convenience) samples, conventional tests of statistical significance do not apply. Instead, practical considerations such as budget and access to the target population will determine the achievable sample size.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
10. Sample size requirements
103
calculations around this variable. As a general rule, the smaller the proportion, the bigger the sample size needed (e.g. a “rare event” requires investigators to sample more people in the population to get a precise estimate). Selecting a variable for sample size calculation Investigators use a variable that represents the primary objective of the survey to determine the required sample size. If the survey has several primary objectives, investigators determine the sample size needed for each, then select the largest size required. This approach ensures that the target sample size is large enough to satisfy all primary objectives. Viral load suppression in a population is often the most important variable for sample size calculation. Calculating sample size to estimate prevalence at one point in time The formula below shows the components of a sample size calculation for a given point estimate. Although we can easily calculate sample sizes with software programs, it is helpful to be familiar with the elements of the formula:
Confidence level The confidence level refers to the percentage of all possible samples that can be expected to include the true population parameter. For example, a 95% confidence level implies that 95% of the confidence intervals include the true population parameter.1 In general, the narrower the confidence interval chosen, the larger the required sample size. The confidence level is often calculated based on the z-score distribution, which indicates how many standard deviations a value is from the mean. A 95% confidence level (implying an alpha of 0.05, meaning there is a 5% likelihood that the true population parameter lies outside the 95% confidence interval) corresponds to a z-score of 1.96. The z-score for a 90% confidence level is 1.645, whereas that for a 99% confidence level is 2.326. Precision or margin of error Because a sample only provides an estimate of the true population parameter, the precision of a sample reflects the reproducibility of the survey’s measurement using the same methods. One measure of precision is the standard error, with precision being inversely related to the standard error.2 The higher the precision, the narrower the confidence interval. For example, for a survey with an assumed prevalence of 50% and a confidence level of 95%, a precision of 0.10 (10%) means that 95% of the time the survey (with a given sample size) will yield an estimate falling within 40–60% (the confidence interval); that is, 50% ±10%. When calculating the sample size, the precision needs to be used as an absolute value (i.e. 50% ±10% in the above example) and not as a relative value (i.e. 20% in the above example). Design effect A design effect is included in the sample size formula when a probability survey other than a simple random sampling survey is used. It quantifies how much larger the sample size needs to be to provide the same precision as a simple random sample, and is a multiplier used to inflate the sample size by a certain value. A design effect of 2, for example, means that a sample size of 400 needs to be doubled to 800 in order to achieve the same precision as a simple random sample. A design effect inflates the necessary sample size compared to a simple random sample because the “variance” in such samples is larger than in simple random samples. Variance indicates how far a set of numbers (values) is spread out in the population. The design effect reflects the ratio of the actual (observed) variance to the expected variance if a simple random sample were used. Probability-based nonrandom
n = DEFF*Z21-α/2*P*(1-P)/d2 where n = minimum sample size required DEFF = design effect Z1-α/2 = z-score for the desired confidence level (usually 1.96 for 95% confidence) P = expected proportion d = precision Components of the sample size calculation to estimate prevalence at a point in time The components of the sample size calculation are expected proportion (P), confidence level (Z1-α/2) precision or margin of error, and design effect, as discussed below. Expected proportion Estimates (e.g. HIV prevalence) from previous surveys or other sources can inform the value of P. If no previous surveys are available, estimates from nearby areas or countries may be used. Alternatively, investigators may assume (e.g. through a literature review) a certain prevalence ratio for their target population compared to the general population. For example, if no HIV prevalence estimate for male SW is available, investigators may multiply the (known) HIV prevalence among the general male population in the same area by a factor of 2 or more, based on data from other studies.
1 2
Adapted from: http://stattrek.com/statistics/dictionary.aspx?definition=confidence_level Adapted from: http://stattrek.com/m/statistics/dictionary.aspx?definition=precision
104
10. Sample size requirements
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
sampling designs (e.g. cluster, time-location or RDS) produce less varied samples, leading to wider confidence limits. In other words, less varied sampling designs provide less precision than a simple random sample; therefore, such nonrandom sampling designs lead to smaller “effective” sample sizes. Importantly, the design effect varies by the variable of interest (participant characteristic); for example, the design effect to measure HIV prevalence with a certain precision may be different from the design effect to measure ever having been tested for HIV. Investigators need to be careful not to confuse the design effect with the design factor (DEFT), which is the square root of the design effect. Sometimes, survey reports publish DEFT rather than the design effect. In these cases, the DEFT value (e.g. 1.5) needs to be squared in order to derive the design effect (in this example, 1.5 × 1.5 = 2.25) and so obtain the factor by which the effective sample size needs to be inflated. How to calculate the design effect The design effect cannot be predicted with accuracy before a survey, and therefore needs to be estimated based on recommendations or data from previous surveys. After a survey, an estimate of the design effect can be calculated by dividing the observed variance by the variance based on a simple random sample. Calculating the design effect after the survey informs future surveys and gives a sense of whether the assumed pre-survey design effect value was accurate. Design effect in respondent-driven sampling (RDS) surveys In RDS surveys, design effect values not only vary substantially across surveys and within surveys by indicator, they may also be high. In the past, a design effect of 2 was commonly anticipated for RDS surveys. However, recent theoretical work suggests that a larger design effect should be expected (e.g. 3 or 4) (1, 2). Although such a large design effect can dramatically reduce the effective sample size (i.e. the sample size corresponding to that of a simple random sample), investigators may have few alternatives available other than accepting a lower precision, striving for the largest sample size possible or opting for a different sampling method. The two sources of uncertainty in RDS estimates are bottlenecks in recruitment (i.e. recruitment across distinct and poorly connected networks leading to segmentation of the target population) and nonrandom recruitment within recruiters’ networks. Investigators should refer to the current literature for theoretical examinations of expected design effect values as well as observed design effect values in comparable RDS surveys.
Design effect for cluster-based surveys In conventional cluster sampling (CCS) or time-location sampling (TLS) a design effect around 2 should be used. Results within clusters are typically expected to be correlated, which requires that the sample size be increased when compared to that needed under simple random sampling. Example sample size calculation for a single proportion In an RDS survey designed to estimate HIV prevalence among people who inject drugs (PWID), the expected HIV prevalence (P, informed by a previous survey) is 20% (P = 0.2), if the design effect is set at 3, the confidence level at 95% (Z1-α/2 = 1.96) and precision at 5% (d = 0.05): Design effect = 3 Z1-α/2 = 1.96; (1.96)2 = 3.84 P = 0.20 d = 0.05; (0.05)2 = 0.0025 Using these data, the final sample size is 738 (rounded up): 3 × (3.84 × 0.2 × (1–0.2))/0.0025 = 738 Other elements to inform sample size calculation Sometimes, investigators are interested in a particular subgroup of the population – such as users of a particular drug, SW who only work outside brothels or TG – which may make up only part of the survey sample. If a certain precision is required for an estimate in a subgroup, then the total target sample size needs to be larger. Investigators need to estimate the relative size of such a subgroup (e.g. 50%), calculate the sample size for the intended precision, and then “inflate” the total sample size by that factor (e.g. a subgroup comprising 50% of the target population would imply a doubling of the sample size). Limitations of the presented method for sample size calculation are discussed below. Possible underestimation of required sample size The presented method is somewhat non-conservative (i.e. may suggest lower sample sizes than warranted), especially where assumed prevalence values (proportions) are low. In such situations, other approaches (e.g. Wilson score or asymptotic bounds with continuity correction) are preferable and would typically give larger sample sizes. As mentioned earlier, a statistician should be consulted regarding the choice and appropriateness of the sample size formula. Missing data and the effect on sample size Even when surveys meet their target sample size, the resulting datasets may be incomplete due to accidental
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
10. Sample size requirements
105
data loss (missing data), erroneous values, refusal to provide an answer (nonresponse), or loss of specimens or refusal to provide a biological specimen. If the possibility of incomplete or erroneous data or missing specimens cannot be excluded during survey planning, a corresponding increase in sample size is recommended to offset their effect. To do this, the original sample size is divided by (1-x/1), where x represents the proportional extent of nonresponse. For example, if investigators fear that 10% (= 0.1) of the target 1000 specimens will not yield biomarker data due to factors such as loss, then the corrected sample size is derived through 1000/((1-0.1 )/1), yielding 1111. Lack of finite population correction The formula shown above for sample size calculation assumes that the sample size is much smaller than the population size. This means the total population is considered “infinite”; that is, much larger than the sample size. If this is not the case, a finite population correction should be used. The final population correction (fpc) is computed as
Selecting P1 and P2 P1 should be determined as described above (see “Components of the sample size calculation to estimate prevalence at a point in time”). P2 is set at the target proportion (20% or 0.2 in this example). In practice, it is usually set at the smallest meaningful change investigators expect might have occurred. For example, a decline in HIV prevalence from 30% to 20% may be considered meaningful, but a decrease from 30% to 28% may not. The larger the difference between the two proportions, the smaller the sample size needed to determine a statistically significant difference. Investigators should consider P1 and P2 carefully. Programmatically, it is often easier to achieve a drop in HIV prevalence from 30% to 20% (a drop of one third) than from 15% to 5% (a drop of two thirds), even although the absolute difference is the same. Consider also the time interval and type of outcome measurements. It is difficult or impossible to detect substantial changes in HIV prevalence in a short period of time (e.g. a drop in HIV prevalence from 30% to 20% over 2 years is unlikely to occur), whereas changes in behaviours are much likely to be detected in the short term (e.g. uptake of HIV testing or condom use). In addition, the closer P1 and P2 are to 50%, the larger the sample size needed to meet a desired precision. The sample size required for each survey for the measurement of change between two survey rounds for a given indicator is a function of four factors: • initial prevalence of the characteristic; • size of the difference to be detected; • level of significance, or the degree of certainty that a change of that magnitude would not have occurred by chance; and • power, or the degree of certainty that a change of that magnitude can be observed if it actually does occur. The formula for calculating the sample size for each survey round (n) is given by:
f pc = √((N − n)/(N − 1))), where N denotes the total population size, and n denotes the sample size. The computed fpc is then used to refine the expected standard error, which determines the confidence level. For example, if the total population is 10, 000 and the sample size is 600 the finite population correction would be: √((10, 000−600)/(10 000−1) = 0.9695≈0.97 To use a finite population correction, the total population size must be known, but often it is not. Hence, total population size is often not considered when determining sample size. In all cases, investigators should be confident that the computed sample size is achievable given the total population size, meaning that the total population size should be substantially larger than the sample size. Once data collection has started, investigators should examine the incoming data early on to see whether the assumptions are likely to be met; this may allow investigators to redo the sample size calculations and alter the target sample size in time. Calculating sample size to detect changes over time An alternative framework for calculating sample size is based on looking for a change in prevalence over two survey periods rather than a desired precision in the estimate for one survey. Detecting changes in proportions For a variable of interest, P1 and P2 are the proportions expected at Time 1 and Time 2. For example, investigators may wish to detect a change in HIV prevalence among SW from 30% (P1) to 20% (P2) or less some time later.
n = DEFF × (Z1-α/2 × 2√2p(1−p)+Z1-β × 2
√p1(1−p1)+p2(1−p2))2 / (P2 – P1)2 where DEFF = design effect P1 = the estimated proportion at the time of the first survey P2 = the target proportion at some future date (second survey), so that (P2 – P1) is the size of the difference that is to be detectable
106
10. Sample size requirements
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
P = (p1 + p2 )/2 Z1-α/2 = the z-score corresponding to desired level of significance, usually 1.96 for 95%, and α is the type I error rate (reflects the probability of the test suggesting that there was a change in the proportions when there was not one). Z1-β = the z-score corresponding to the desired level of power, usually 0.83 for 80%, and β is the type-II error rate (reflects the probability of the test suggesting that there was no change in the proportions when there was one). As indicated by the z-score, this formula is for a twotailed test; that is, the required sample size will enable the detection of a difference in P in either direction (down or upwards). Statistical power The statistical power score (Z1-β) corresponds to the power required to detect change over time. Usually, 80% power is selected, meaning there is an 80% chance that a survey will detect a change in a certain proportion over time or between groups, if there actually is a change. The ability to detect a difference will depend on the sample size, the magnitude of the difference and the variance. The more statistical power needed, the larger the resulting sample size. Examples of sample size calculations to detect a change in proportions Example 1. Suppose investigators are planning a survey of female sex workers (FSW). The goal is to measure whether condom use at last sexual intercourse with a client will increase from 20% in the baseline survey to 30% or more in the next survey round. They set the design effect at 2 (DEFF = 2), confidence at 95% (Z1-α/2 = 1.96, for a two-sided test3) and power at 80% (Z1-β = 0.84). Using these parameters, the final sample size is 587: 2 × (1.96 × √(2 × 0.25(1–0.25))+0.84 × √(0.20 × (1–0.20)+0.30 × (1–0.30)))2 / (0.30–0.20) = 587 In this example, investigators would need to conduct two surveys with a sample size of at least 587 participants in each survey.
Example 2. Suppose investigators want to detect a decrease of 15 (absolute) percentage points in the proportion of transgender male vocational students who had unprotected sex in the past 12 months. A level of significance of 95% and a power of 80% is desired. On the basis of earlier survey data, it is thought that the appropriate “baseline” value on the indicator would be 55%. Set P1 = 0.55 and P2 = 0.40, and use z-score values of Z1-α = 1.645 (95% significance level for a one-sided test) and Z1-β = 0.84 (corresponding to 80% power) and obtain: n = 2 [1.645 √2(0.475)(0.525) + 0.84 √(0.4)(0.6) + (0.55)(0.45)]2/(0.40–0.55)2 = 2 × [(1.1617+0.5865)2/0.0225] = 272 vocational students in each survey round. Note: Sample design assumes a design effect of 2. Appendix I-22 shows examples of sample size requirements for a range of different scenarios. The same limitations and considerations as shown for one-time surveys above apply.
A-10.2 Additional considerations Considerations for web-based surveys Many web-based surveys are considered convenience samples, so considerations such as statistical power and precision may not apply. Web-based surveys attempting a sampling design to yield some form of probability sample (e.g. a web-based RDS, representative of MSM accessing the internet) should follow the methods described in this chapter to calculate sample size (3, 4). Time and budget Sampling speed may have a substantial influence on achieving the sample size, because many survey cost elements are time dependent (e.g. staff time, rent, utilities and transport) rather than a function of the total target sample size. Hence, the more participants a survey can enrol each day or week, the more cost effective it generally is and the higher the resulting sample size. For example, an RDS survey that starts recruiting slowly and fails to increase the speed of recruitment will yield a correspondingly smaller total sample size by the time the budget is exhausted than a survey that enrols far more respondents each week.
3
A two-sided test it is one that can detect a change in either direction (i.e. either an increase or a decrease in this example). A one-sided test is one that is only powered to detect a change in one direction. Unless it is clear that a change in the outcome of interest can only be in one direction, a two-sided test should be used. The second example uses a one-sided test because investigators are only interested in detecting a decrease, rather than a change in either direction. The sample size required for a one-sided test is smaller than for a two-sided test.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
10. Sample size requirements
107
Calculating the size of a second BBS Often the decision to conduct a survey to detect a change compared to a baseline value comes after the first survey has been conducted. In such a situation, investigators should use the first survey’s results to inform the second survey, including the baseline value (proportion) and surrounding confidence limits, as well as the design effect. This allows investigators to have more confidence
that their second survey’s target sample size will be able to detect the anticipated difference. Sample size calculators Many statistical software packages include tools for sample size calculations, such as EpiInfo.4 A simple Excelbased calculator for cluster-based surveys can be found on the WHO website.5 Various sample size calculators are included in Appendix I-22 and online.6
A-10.3 References 1 Johnston LG, Chen Y-H, Silva-Santisteban A, Raymond HF, DiNenno E. An empirical examination of respondentdriven sampling design effects among HIV risk groups from studies conducted around the world. AIDS Behav. 2013;17(6):2202–2210. Wejnert C, Pham H, Krishna N, Le B, DiNenno E. Estimating design effect and calculating sample size for respondent-driven sampling studies of injection drug users in the United States. AIDS Behav. 2012;16(4):797–806 (http://www.ncbi.nlm.nih.gov/pubmed/22350828, accessed 8 August 2016). Bauermeister JA, Zimmerman MA, Johns MM, Glowacki P, Stoddard S, Volz E. Innovative recruitment using online networks: lessons learned from an online study of alcohol and other drug use utilizing a web-based, respondentdriven sampling (webRDS) strategy. J Stud Alcohol Drugs. 2012;73(5):834–838 (http://www.ncbi.nlm.nih.gov/ pubmed/22846248, accessed. Bengtsson L, Lu X, Nguyen QC, Camitz M, Hoang NL, Nguyen TA et al. Implementation of web-based respondentdriven sampling among men who have sex with men in Vietnam. PLoS One. 2012;7(11):e49417 (http://www.ncbi. nlm.nih.gov/pubmed/23152902, accessed 8 August 2016).
2
3
4
4 5 6
http://www.cdc.gov/epiinfo/ http://www.who.int/chp/steps/resources/sampling/en/ http://statpages.org/
108
10. Sample size requirements
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A
11. Population size estimation methods used with surveys
In many countries, the HIV epidemic was first identified among key populations. The key populations covered in these guidelines are often “hard-to-reach” or “hidden” populations because their behaviours are illegal or stigmatized in some settings; hence, the size of these populations is usually unknown. This chapter discusses methods used to estimate population size, because knowing the size of a key population can help investigators to understand the scale of the response needed to ensure the population gets services to prevent HIV infection.
Purpose or objective of population size estimation Estimating the size of key populations at risk for HIV is important in a number of ways. For example, size estimation data can be used to: • inform policy and advocacy; • provide denominator data for indicators related to reaching coverage for key populations; • provide critical information for models used to estimate and project HIV impact; • inform HIV response planning, target setting and resource allocation (e.g. funding or budget); and • inform service delivery and facilitate programme monitoring and evaluation (e.g. programme coverage) (1, 2). A biobehavioural survey (BBS), in conjunction with other studies, can provide an opportunity to estimate the size of the surveyed population. The Joint United Nations Programme on HIV/AIDS (UNAIDS) global, regional and national HIV estimates for all but generalized HIV epidemic settings use key population size estimates to estimate the number of new and prevalent infections, the number of people eligible for care and treatment, and the number of HIV-related deaths. Population size estimates have also been used to: • mobilize political support and commitment for the HIV response; • direct funding by characterizing the extent of the epidemic; • plan HIV programmes for key populations nationally, regionally and locally; and • monitor and evaluate programmes in terms of coverage, quality and effectiveness.
Results of a size estimation study can garner muchneeded media attention to demystify often-held notions that the actual number of certain key populations is negligible and hence not worthy of public health action. Population size estimates can be used as denominators for reporting on international monitoring indicators, and for grant applications. Population size estimation methods currently in use A number of size estimation methods are available for use. No gold standard exists. All methods have their strengths and limitations, and where possible, multiple estimates should be produced to derive a consensus estimate. Table A-11.1 presents an overview of different size estimation methods. The available methods can broadly be categorized into direct methods (census and enumeration) and indirect methods. Direct methods count members of the population directly, whereas indirect methods use data from different sources to estimate their size. In addition, these methods can be classified through their data source: data collected from at-risk populations, and data collected from the general population.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
109
Table A-11.1 Summary of methods for estimating the size of key populations
Method Census
Description Count all members of the population
Strengths • Real count, not an estimate or sample • Can produce credible lower limit • Can be used to inform other methods • Can produce credible lower limit
Assumptions/limitations • At-risk populations are often hidden; misses some or many members of the population • Stigma may prevent members from identifying themselves • Time consuming and expensive • At-risk populations are often hidden; misses some or many members of the population • Stigma may prevent members from identifying themselves Relies on four conditions that are hard to meet: • the two captures must be independent and not correlated • each population member should have equal chance of selection • each member must be correctly identified as “capture” or “recapture” • no major in or out migration
Enumeration
Develop a sampling frame and count all members of the population at the selected locations
Capturerecapture
Size estimate is based on two independent captures (samples): • capture 1: “tag” and count number tagged • capture 2: “tag” and count number “retagged” (i.e. who had also been “tagged” the first time)
• Relatively straightforward to do with access to population • Does not require much data
Multiplier method
Apply a multiplier (e.g. number receiving particular service or having membership, or number receiving a unique object distributed before a survey) to survey estimate (proportion of survey sample sharing same characteristic)
• Uses data sources already available • Flexible in terms of sampling methods; first source need not be random, but second source should be representative of population
• The data sources must be independent • The data sources must define population in the same way • Time periods, age range and geographical areas must be aligned • Data collected from existing sources may be inaccurate
General populationbased survey
Ask respondents if they engage in the behaviour of interest (e.g. male–male sex, money for sex, inject drugs) Generates proportional estimates (% of the general population who engages in a certain high-risk behaviour)
• National surveys are common and familiar • Easy to implement if a survey is underway • Straightforward to analyse • Sampling is easy to defend scientifically (“gold standard”)
• Low precision when the behaviours are rare • Respondents may be reluctant to admit to stigmatized behaviours • Only reaches people residing in households (mobility) • Privacy and confidentiality risk to subjects
110
11. Population size estimation methods used with surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Method Network scaleup method (NSUM)
Description The general concept is that an individual’s social network is representative of the whole population (i.e. one person’s group of friends somehow reflects the characteristics of the whole community); thus, members of the general population are asked whether their acquaintances have highrisk behaviours (e.g. buying or selling sex, having anal sex between men or injecting drugs). The average proportion of the respondents’ acquaintances who have these behaviours is used to estimate the proportion of the adult population with these behaviours.
Strengths • Can generate estimates from general population rather than hard-to-reach populations • Does not require survey respondent to disclose stigmatizing behaviours • Clear formulation and calculations have been developed and tested • It is possible to get confidence intervals
Assumptions/limitations • Average personal network size is difficult to estimate • Some subgroups may not associate with members of the general population • Transmission error: respondents may be unaware someone in their network engages in the behaviour of interest • Reporting bias (i.e. social desirability) may arise • May work well with some groups and in some contexts but not in others
• Does not require collection of Reverse tracking Compares the “observed size” method with the “estimated size” for additional information outside a selected site, calculates an of a TLS survey adjustment/correction factor and modifies the total estimated size in the sampling frame accordingly
• Requires an exhaustive list of sites • Would not be able to estimate the hidden part of the population (i.e. that portion that never accesses these sites) • Could result in overestimation if the same individuals are found in many sites
Successive sampling population size estimation
• Does not require collection of • Estimate’s validity depends A probability model is created as a function of the observed additional information outside heavily on representativeness personal network sizes in the RDS of the BBS of the sample sample, investigators’ knowledge • Can be conducted with multiple • Prior information on the about the population size, and BBS sampling methods population size may be the unobserved network sizes including TLS and RDS unavailable or poorly estimated • Does not require collection of Requires sampling from each participant’s network connections additional information outside and matching those connections of an RDS survey against the other participants in the sample and the list of their respective contacts Truncated Poisson estimates for sparse data rely on certain assumptions, some of which may be hard to meet: • a closed population (no entry or exit); • behaviorally homogeneous population members; and • a constant encounter rate, with no behavioral response to an encounter that reduces or increases the rate for future encounters
One-sample capturerecapture method
Sources: Abdul-Quader et al 2014, UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance, 2010 (2, 3)
For a detailed description of each of these methods, please refer to the recent UNAIDS/WHO guidelines
document on size estimation (2) and other relevant literature in the references.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
111
A-11.1 Population size estimation based on surveys Because these guidelines focus on integrated BBS, among the methods listed above, two methods can be integrated within a BBS: the multiplier method and the capture-recapture method (CRC). The remainder of this chapter covers key assumptions and strengths and limitations of these two methods as well as emerging methods, and considerations for selecting a method.
The variance can be computed as:
Var(N) = [B*C2*(B–m) × (C2–m)]/m3 The 95% confidence intervals can be computed as:
95%CI = N ±1.96 × √Var(N) Example of service multiplier method A representative survey of 800 FSW was conducted in a city. Participants were asked whether they had visited a certain key population health centre in the past six months. The survey found that 11.0% (88/800) of participants had visited the health center in the past six months. The health centre reported having 395 individual FSW patients in the past six months. Thus, the estimated number of FSW in the city is: B=395 m=88/800=11.0% N = B/m = 395/0.11 = 3,591 Object multiplier Another version of the multiplier method involves distribution of a unique object (i.e. one that cannot be bought easily or otherwise obtained) to the target population before the survey. Investigators often discuss a suitable object with representatives of the survey population (key informants) to ensure the object is acceptable and has intrinsic value (utility); that is, the object is useful to keep, but does not have commercial value that may facilitate its sale or trade for other things. Bottle openers, key chains or nail clippers are such objects. Once identified, investigators will often order a set number of custom-made objects. The object is distributed widely in the sampling area and only to those individuals likely to meet the survey’s eligibility criteria. Recipients of the object are told that they should not lose it or give it to someone else. When the survey participants are recruited and interviewed, they are asked whether they received the object that was previously distributed. If any survey participants are not able to show the object that was distributed before survey recruitment, they are shown a number of similar objects and asked to identify the correct one. The total number of the survey participants who correctly show or identify the object they had received and the proportion of them recruited to participate are used to estimate the size using the formula presented above. Investigators should aim to distribute at least twice as many objects as the target sample size. Given that this may not always be feasible, investigators should distribute as many objects as possible. The more objects distributed, the more precision can be expected in the estimate. Appendix I-23 includes a calculator to determine the number of unique objects required to obtain a given level of precision.
A-11.1.1 Multiplier method The multiplier method compares two independent sources of data to estimate the total number in a population. The first source is a count or listing from programme data including only the population whose size is being estimated, and the second source is a representative survey of the same population (1). Service multiplier • data source 1 – the count: Review the programme interventions and identify specific programme services. From programme data, count the total number (B) of the population (whose size is being estimated) who received that particular service within a specified geography and given time period. • data source 2 – the multiplier: Estimate the proportion (m) of the population (whose size is being measured) receiving that particular service during the same reference period through a representative survey among the target population within the same geographical region. This is known as “service multiplier”. • multiply the count (B, data source 1) by the inverse of the proportion of the population (1/m) who say they received services over the same period using the following simple mathematical formula:
N = B/m where
N = estimate of total population size B = total number of the population who received a m = proportion of the population receiving a C2 = sample size of the second capture (i.e., the survey) Programme data can be from sources such as outreach HIV testing services, needle or syringe exchange programmes, or sexually transmitted infection (STI) clinics. particular service (survey data). particular service (programme data)
112
11. Population size estimation methods used with surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Event multiplier Another version of the multiplier method involves conducting an event for the target population before the implementation survey. When the survey participants are recruited and interviewed, they are asked if they had attended the specific event that was held on a certain day and time at a certain location. The total number of survey participants who attended the event and the proportion of them recruited to participate are used to estimate the size using the above formula. Assumptions There are four key assumptions to be fulfilled for the multiplier method: • the data are for unique individuals (i.e. no double counting); • both data sources are using the same definition for the population; • data from two sources are available for the same time period; and • both data sources have the same catchment area (geographical coverage). Strengths and limitations Multiplier methods are relatively straightforward to use and perhaps the most widely used of size estimation methods (1). However, they are subject to the quality of the service (event or object) data. Poor service data or programme data that are challenging to match with the survey’s eligibility criteria in terms of geographical coverage, risk behaviour and age often present a limitation to the quality of the resulting size estimates. Further, investigators need to make sure that the service data can provide numbers on unique persons reached (i.e. exclude persons who show up in the service data multiple times). The survey questions related to the
size estimation need to be very clear (i.e. sensitive and specific). Questions should ask about the specific service that is being used to estimate the size. Unique object multipliers may be the choice where service data is nonexistent or of poor quality. The greatest difficulty in using multiplier methods is finding data for institutions and populations that correspond with one another in terms of the definition of population, time reference period and, particularly, catchment area. Because of the catchment area issue, multiplier methods based on service use are most commonly used at the local level. The multiplier method based on unique object distribution could be used at a larger level or in an area without good institutional data; however, it relies on access to members of the key populations in which a particular unique object is distributed as widely as possible. Multipliers based on unique events rely on having the survey population aware of the event and attending the event without the fear of stigma and discrimination. Whatever multiplier is chosen, the key assumptions should be met and both data sources (multiplier and survey) should be representative of the survey population, otherwise, the resulting estimates are subject to bias.
A-11.1.2 Capture-recapture method CRC originated in biology (counting fish in a pond), and until recently it was used primarily to estimate wildlife populations (where it was known as the Lincoln-Peterson estimator). Other terms sometimes used include “mark and capture” and “capture and release” (4-10). Figure A-11.1 illustrates CRC.
Figure A-11.1 Capture-recapture method
Captured in list 1 and list 2
Captured in list 2 only Not captured Not captured
Captured in list 1 only Source: UNAIDS/WHO 2010 (2)
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
113
How capture-recapture works • prepare a list – as exhaustive as possible – of the sites where the population can be found, based on information collected during mapping or sampling frame development exercises of BBS. Visit all the sites on a given day within a certain time period, and tag all members of the population found at the site (give them some unique identifier, such as a card or memorable gift). Keep a count of the persons tagged (C1). • revisit all the sites 1 or 2 weeks later within the same time period, and tag all members of the population encountered at the second visit. Ask all those encountered (approached) whether they were tagged (received the unique identifier) in the first visit (1–2 weeks earlier) by someone from the team. Keep a count of the persons tagged at the second visit (C2). • identify and count the persons who were tagged twice, both in the first and second visit (R). The three key counts mentioned above: total number of persons captured during the first visit (C1), total number of persons captured on the second visit (C2), and total number of persons captured twice (R) are used to estimate the size using the following simple mathematical formula:
• capture in the second sample is independent of capture in the first (i.e. the two samples are independent); that is, either key population members are tagged at all sites, or investigators draw a random sample of sites on each occasion (investigators using a sample of sites should not go to the exact same sample of sites both times). CRC can also be conducted by sampling a selected number of sites (locations) when the size estimation area is much larger and the total number of sites (locations) are numerous. In this instance, a random sample of sites (locations) are selected for the first capture. For a recapture exercise, another sample of sites (locations) are selected randomly. This helps with drawing independent samples for R and C. Example of the use of capture-recapture method The following example summarizes the steps involved in a population size estimation activity conducted to estimate the size of the population of street-based sex workers (SW) in Dhaka, Bangladesh. Step 1: Existing data were reviewed and analysed to understand the nature of street-based sex work, including characteristics, estimated size, locations, accessibility and operating hours. Step 2: The population was defined; in this case, FSW who negotiated with their clients on the street.
N = (C1 × C2)/R where
N = estimate of total population size C1 = total number of persons captured on the C2 = total number of persons captured on the R = total number of persons captured on both the visits. The variance can be computed as: second visit first visit
Step 3: The geographical area was defined, to be included for the estimation. Step 4: The defined geographical area was divided into a number of zones and field staff were assigned to each zone. Step 5: Ethnographic observation and mapping of each of the zones were conducted to assess the following: • prevalence of street-based SW within that zone; • locations or venues within that zone where the street-based SW were likely to congregate and where they could be approached for counting during the estimation exercise; • when they can be found and counted (day of the week and time of the day, e.g. early evening or late evening); and • estimated number of SW that congregated at each of the identified locations within the zone. In addition to conducting ethnographic observation, the field staff conducted brief informal interviews with local
Var(N)=[C1*C2*(C1-R)*(C2-R)]/R3 The 95% confidence intervals can be computed as:
95%CI=N±1.96*√Var(N)
Assumptions There are four key assumptions that need to be fulfilled for applying CRC: • the population is closed (i.e. no significant in or out migration); • individuals captured on both occasions can be matched (i.e. no loss or misclassification of marks); • for each sample, each individual has the same chance of being included (i.e. same catchability); that is, people in the first sample are not more or less likely to be included in the second sample than people who were not included in the first sample; and 11. Population size estimation methods used with surveys
114
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
business owners to verify information collected based on observation and mapping, including estimated number of street-based SW within that zone. Step 6: SW were recruited as volunteer field staff from known sources (e.g. from existing STI clinics). Step 7: Orientation was provided to the SW about the purpose of the project activity. Step 8: Field teams were formed for each zone, with each team including at least one project field staff and two to three SW recruited as volunteers. Step 9: Field teams were trained on the size estimation method (CRC) including the principles and operations of the method. Step 10: Based on information and secondary data collected, it was decided that the first count (first capture) would be conducted on a given day between the times of 7:00 pm and 3:00 am. Step 11: About 8000 small cards were printed for distribution during the first count, which was based on existing information on street-based SW in the city. The cards were sequentially numbered from 0001 to 8000. In addition, the cards had information about one of the STI clinics (address, telephone number and services available) that provided STI services to SW. Each field team received a set number of cards, and records were kept. The field team was instructed to keep track of the number of cards distributed and those not distributed and returned to the project office. This was done to determine the total number of cards that were distributed in the first count. Step 12: On the day that was selected for the first count (first capture), field teams were instructed to go to their respective zones, previously assigned. The team in each zone walked around various locations. When they encountered a SW (verified by the volunteers) she was approached by the volunteer and asked whether someone had approached her and given her a card (she was also shown the card that the team carried). If the SW mentioned that she did not receive any card, she was given a card. This continued until about 3:00 am. At the end of this first count, the total number of cards distributed was determined for each zone (total number of street-based SW captured – first capture).
Step 13: The same process was repeated a week later. The field teams went to their respective zones at the same time and stayed from 7:00 pm until 3:00 am. However, this time the team carried a different card. The new cards had a different colour with different information and a different sequence of numbers printed on them. Each field team was given a set number of these new cards and instructions on how to maintain records of new capture and recapture. When they encountered a SW (verified by the volunteers) she was approached by one of the volunteers and asked whether she had received a card. If she responded that she had, she was shown the new card to determine whether she received the new card or had received a card during the first count but not during this second count. If she mentioned that she received a card (different from the new one) during the previous week, she was given the new card and recorded as a recapture. If she mentioned that she never received any card, then she was given the new card and recorded as a capture during the second count. This process continued until 3:00 am. Step 14: Based on the total number of cards distributed during the first count (first capture – NC1), total number of cards distributed during the second count (second capture – NC2), and the total number of SW who received cards during the first count and also during the second count (recapture), calculation was conducted to estimate the total size of the street-based SW who operated primarily in the evening between 7:00 pm and 3:00 am in the selected geographical area. Strengths and limitations A simple two-sample CRC is relatively easy to use. It also does not require statistical expertise. However, the four assumptions listed above can sometimes be difficult to meet, and such limitations should be documented and later presented together with the results of the size estimation. Steps can be taken to avoid violating the assumptions and minimize bias that may result when assumptions are not fully met (2), including by adding a third source and effectively conducting a capturerecapture-recapture. Three-source capture-recapture lets investigators relax the assumption of independent sources by using interaction terms between dependent sources in a log-linear model (11).
A-11.1.3 Network scale-up method The network scale-up method (NSUM) is used as part of a general population survey. NSUM facilitates the estimation of a participant’s personal network size (i.e. the number of people personally known to the participant – the denominator) and probes the number of key or priority population members personally known to
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
115
the participant (the numerator). Dividing the numerator by the denominator yields the key or priority population prevalence within the respondent’s personal network. Averaging these prevalence estimates across all survey respondents and weighting them will yield an overall estimate of prevalence in the key or priority population (12-14). NSUM can also provide regional or city-level size estimates if the sample size is large enough at these levels. Strengths and limitations Few extra resources are required for each additional population added to an NSUM exercise, the primary resource being staff time for the additional 15–30 questions asked of each participant. When added to an already planned general population survey, the costs to produce NSUM-based estimates are small, although this process does require preparatory work. NSUM may underestimate the population size if few people know that individuals in their social network are key population members (i.e. transmission error) or when participants do not reveal that they know key population members. Transmission error potentially has the biggest impact on results. This error can be adjusted using the game of contacts (GoC) technique (15), which comprises a set number of questions that are used in interviews with key population members as part of a key population BBS. It facilitates a quantitative estimate of that population’s social visibility, ranging as a proportion between 0 (total invisibility) and 1 (total visibility). For example, if men who have sex with men (MSM) on average have a visibility of 0.5 (i.e. half of their network members know that a given man is an MSM), then the NSUM population size estimate needs to be doubled to account for this bias.
based on key informant interviews during mapping or sampling frame development exercises. • a proportion of all the sites from the sampling frame for the actual survey is selected, with the sites chosen in such a way that they are representative of all the sites in the sampling frame. For example, 30% of all the sites in the sampling frame is selected. • during the BBS, the survey team visits the selected sites and interviews (and collects biological specimens from) a random sample of the population group at each selected site. At each selected site, information is counted and recorded; for example, the total number of completed interviews, not completed interviews, refusals and total population attached to the site (“observed size”). This information is collected through cluster information sheets for weighting purposes. • this provides, for 30% of the sites, the “estimated size” from the sampling frame and “observed size” from the cluster information sheets of the BBS. • the method compares the “observed size” with the “estimated size” for each selected site, calculates an adjustment or correction factor, and modifies the total estimated size in the sampling frame accordingly. The simple mathematical formula is (17):
where
A-11.2 Emerging methods In addition to the two methods discussed above, there are a number of emerging methods to use in conjunction with BBS among key population surveys.
N = estimate of total population size n = total number of selected sites for BBS Yi = “observed value” at i-th site Mi = “estimated value” at i-th site M = total estimated size in the sampling frame; i.e. sum of Mi for all the sites in sampling frame. In the formula, the total estimated size in the sampling frame (i.e. M) is adjusted or corrected by
A-11.2.1 Reverse tracking method The reverse tracking method (16) is a new population size estimation method that can be integrated within a BBS. It is an indirect size estimation method that modifies or corrects the estimated total size in the sampling frame based on actual observations. The method is based on a simple mathematical principle and does not require collection of additional information. How reverse tracking works • the BBS sampling frame for any at-risk population group may include an exhaustive list of the sites where members of the population can be accessed for an interview and an “estimated size” for each site,
and it is the average variance of the observed values from the estimated values across the selected sites. Assumptions There are three key assumptions that need to be fulfilled for the reverse tracking method: • the sampling frame should include an exhaustive list of sites; B. Survey implementation & quality assurance
116
11. Population size estimation methods used with surveys
A. Survey preparation
C. Data analysis and use
• each site in the sampling frame should have an estimated size based on key informant interviews; and • the observed size should be recorded and be available for each selected site for the BBS. Strengths and limitations The strength of this method is that it does not involve any additional data collection. It uses already collected data during a BBS to generate another estimated size. The estimates from this method would be as good as the sampling frame of the BBS (in terms of its coverage of the sites). This method would not be able to estimate the hidden part of the population (i.e. that portion that never accesses these sites). Accuracy of the “observed size” at site level during the survey is also a critical factor. In many countries where the laws are extremely harmful to target populations, it will be difficult to find sites. Another limitation is that this method could give an inflated size for a “floating population” due to presence of the same set of individuals across different time-location sampling (TLS) surveys for a particular site, and a further round of correction may be needed to address this particular issue.
prior distributions that may incorporate previous or concomitant information about the population size. Handcock, Fellows and Gile (17) developed RDS analyst software to implement this approach. The strength of this analysis approach is that when RDS surveys are conducted, network size information is also collected and used in the estimation procedure. However, the method depends on prior information on the population size or sample proportion, and this information may be unavailable or poorly estimated, resulting in large interval estimates. In addition, the method is relatively new and will benefit from validation studies, comparison with other approaches and possible refinement. Lastly, if the survey is based on sampling designs other than RDS, then this method will not be suitable. This new approach is still being validated and may need further refinement.
A-11.2.3 One-sample capture-recapture method Recently, Dombrowski et al. (19) innovatively used CRC entirely within a BBS. They used a network-based variant of CRC, whereby participants maintained anonymity. It required sampling from each participant’s network connections, and matching those connections against the other participants in the sample and also the list of their respective contacts. RDS was used to recruit the first capture using the internet (Craigslist), and the second capture was based on respondents’ friends in their personal networks. Recapture sampling was based on links to other participants derived from demographic and “telefunken” matching procedures – the latter being an anonymized version of telephone numbers. This technique made it possible to estimate the population size without physically recruiting a second sample.
A-11.2.2 Successive sampling population size estimation Handcock, Gile and Mar (18) proposed a Bayesian approach to estimate population size of target populations using data from a single respondent-driven sampling (RDS) survey. In the Bayesian approach, a probability model is defined for the observed data given the population size, where it is assumed that we already have some knowledge about the current population size. The probability model or likelihood for the data is a function of the observed personal network sizes (i.e. degrees) in the RDS sample, as well as investigators’ knowledge about the population size and the unobserved network sizes, both of which are assumed to be generated from a super-population model based on some unknown distribution. The likelihood is formulated as an approximate RDS sample obtained by successive sampling from the full population with degrees dependent on a super-population parameter. Through use of Bayes’ rule, the likelihood is multiplied by an assumed prior distribution for the population size and assumed prior distribution for the super-population parameter to arrive at a posterior distribution for population size, from which it is possible to calculate a posterior mean (i.e. population size estimate) and 95% probability interval for the population size. Because prior knowledge about the size of the population might be incorrect, it is possible to perform sensitivity analyses for unknown population sizes through use of different
A-11.3 Selecting a method A-11.3.1 Factors to consider – single method versus multiple methods All methods have limitations. Estimates are uncertain and different methods are likely to give different results. Factors that can make estimates substantially uncertain include variability in quality of service data (multiplier method), assumptions that are hard to meet (CRC) and transmission error (NSUM). There is no reason to use only a single method for estimating the size of the population. Multiple methods make for checks and balances and cross-validation. Different methods may produce different estimates. However, when different methods are used and produce estimates in the same overall range, they are likely to inspire more confidence than a single estimate produced by a single method. There are several methods
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
117
that could be integrated within any BBS. The choice of method and the number of methods used is often determined by the resources available. Those planning and implementing size estimation activities should look into the possibility of using more than one method when planning BBS or when other surveys are being planned or implemented. Formative assessment should be conducted before implementing a BBS. Formative assessment may include mapping (social and geographical) and ethnographic observation. During mapping and ethnographic observations, attempts can be made to estimate the size of the population accessible at the sites (see Chapter A-4). If a household survey is planned, the investigators should look into the possibility of using NSUM for estimating the size of multiple key populations. The most important step is better planning and preparation before integrating any particular method. When planning a BBS, it is important to inventory capacities as well as resources available to see which method can be integrated into the BBS. If using a service multiplier, it is important to assess what type of service data are available, the quality of the data, access to data, and any legal or other constraints to using the data. A formative assessment that includes members of the key population conducted before designing the survey clarifies the social and geographical distribution of the target population and the likelihood that the survey staff can reach them. Working with target population members informs the selection of the most appropriate methods. Piloting the method(s) in a subset of the population and validating the method(s) in a known setting (e.g. a university), where possible, are important. The estimates are limited to the local area where the BBS was implemented. Local estimates cannot be considered as the national estimates. Considerations should also be given to sampling methods that are used for conducting the BBS; a multiplier method may be suitable to use with an RDS survey but CRC may not be. Similarly, TLS may require use of different size estimation methods to those that can be used with RDS surveys. Estimating population size is important and useful; however, the calculated estimates should be accompanied by the corresponding degrees of uncertainty. Collected data have various limitations. Because population is an estimated number, a range is acceptable. A full description of how to analyse data from population size estimation activities is beyond the scope of these guidelines. Investigators planning to conduct a population size estimation should consult with experts and review the literature for more information.
A-11.3.2 Selecting the method for key population size estimation This section provides examples of considerations about the choice of method to estimate key population sizes, focusing primarily on local level size estimation and on methods within the context of BBS. Method choice considerations Review of past survey and size estimations and key informant interviews Planning a BBS may involve conducting formative assessment, which includes review of available data as well as focus groups and key informant interviews with the stakeholders. Any available data should be reviewed to understand whether and where target population members usually congregate and how mobile they are, perhaps due to police activity, and whether they access services. This brief preliminary information can be obtained from reviews of existing information, talking with other “stakeholders”, and conducting a brief “ethnographic” assessment. The information collected through this initial “assessment” can help in understanding the feasibility of reaching the population directly – how accessible they are in general and at select locations. A formative assessment may also indicate that a BBS is not warranted at this time but that size estimation is needed. Social visibility and mobility of key population If the target population is mostly visible and can be accessed in certain locations within the city, then census or enumeration may be appropriate methods. Both may provide information for the creation of a sampling frame for surveys using cluster sampling. Target populations that are highly mobile may make census and enumeration more difficult and thus tend to violate CRC assumptions. Geographical level for which the size estimation is needed Size estimation in conjunction with a BBS can provide only local size estimates; for national level estimates, data triangulation, projection or other methods such as NSUM are needed. Whenever a household survey is being planned to be conducted, attempts should be made to use NSUM to estimate the key population sizes for the national and subnational level. NSUM may also be used within local general population surveys, especially when other methods such as CRC or multiplier may be difficult to implement. Financial and technical resources Census and enumeration are usually resource intensive, depending on the size of the town and the number of venues. In contrast, using Bayesian statistics on already collected RDS data usually only requires minimal
118
11. Population size estimation methods used with surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
additional time during data analysis, but does require a technical understanding of the applied data analysis. CRC may require relatively more resources compared to census and enumeration and technical capacity of the staff. Service provision Where services tailored to the target population are available or where service providers can reliably count target population members among their clientele, this facilitates the use of the multiplier method. Method-specific considerations Capture-recapture method CRC may require relatively more resources than census and enumeration, and greater technical capacity of the staff. CRC relies on several assumptions being met if it is to produce valid results. It also works best where the target population is accessible and visible (to those
conducting CRC). It can be conducted as part of a survey (the survey can be used as the recapture stage) or independent from a survey. Multiplier method Use the multiplier method when reliable service data are available; that is, where service providers can provide accurate counts of target population members served and where the characteristics of the clientele served reflects the eligibility criteria in the BBS. Census and enumeration Census and enumeration should generally be seen as providing minimum key population size estimates (i.e. they often lead to underestimates). Because these methods are typically venue based they may work reasonably well only when all or most of the target population can be counted at accessible venues.
Table A-11.2 Factors to consider when choosing a key population size estimation method Capture-recapture Target population social visibility and mobility High mobility will bias CRC size estimate. Target population members need to be visible to capturing staff. Outreach may be more resource intensive than enumeration. It may be technically demanding as assumptions for CRC need to be met. Not applicable.
Multiplier Less affected mobility if service data stem from same time as sampling period. Relatively easy. Not resource intensive.
Census and enumeration Target population members need to be visible to enumerators. High mobility may affect estimates. Relatively easy. Outreach may be resource intensive, especially for census.
Financial and technical resources
Service provision
Accurate service provision data essential (when using service multiplier). Warrants probability-based survey sample.
Not applicable.
Sampling design
May inform sampling frame for both cluster sampling surveys and RDS compatible with CRC.
Both may inform sampling frame for cluster sampling and RDS compatible with CRC.
CRC, capture-recapture method; RDS, respondent-driven sampling
A-11.4 References 1 UNAIDS/WHO Working Group on HIV/AIDS/STI Surveillance. Estimating the size of populations at risk for HIV: issues and methods. Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2003 (http://data.unaids.org/publications/external-documents/estimatingpopsizes_en.pdf accessed 4 August 2016). UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines on estimating the size of populations most at risk to HIV. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO); 2010 (http://www.unaids.org/sites/default/files/media_asset/2011_Estimating_Populations_ en_0.pdf, accessed 4 August 2016).
2
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
11. Population size estimation methods used with surveys
119
3 4 5 6
Abdul-Quader AS, Baughman AL, Hladik W. Estimating the size of key populations: current status and future possibilities. Curr Opin HIV AIDS. 2014;9(2):107–114. Black JF, McLarty D, Mtasiwa DE. Capture-recapture techniques. Difficult to use in developing countries. BMJ. 1994;308(6927):531. Larson A, Stevens A, Wardlaw G. Indirect estimates of 'hidden' populations: capture-recapture methods to estimate the numbers of heroin users in the Australian Capital Territory. Soc Sci Med. 1994;39(6):823–831. Mastro TD, Kitayaporn D, Weniger BG, Vanichseni S, Laosunthorn V, Uneklabh T et al. Estimating the number of HIV-infected injection drug users in Bangkok: a capture–recapture method. Am J Public Health. 1994;84(7):1094– 1099. Neugebauer R, Wittes J. Voluntary and involuntary capture-recapture samples – problems in the estimation of hidden and elusive populations. Am J Public Health. 1994;84(7):1068–1069. Sarkar S, Durandin F, Quader AA. Estimation of number of street-based female commerical sex workers (CSWS) and their HIV situation in the city of Dhaka, Proceedings of the 4th International Congress on AIDS in Asia and the Pacific, Manila, Philippines. 1997:184. Stephen C. Capture–recapture methods in epidemiological studies. Infect Control Hosp Epidemiol. 1996;17(4):262– 266.
7 8
9
10 Weir S, Wilson D, Smith PJ, Schoenbach VJ, Thomas JC, Lamptey PR et al. Assessment of a capture-recapture method for estimating the size of the female sex worker population in Bulawayo, Zimbabwe. Chapel Hill: Carolina Population Center, University of North Carolina; 2003 (http://pdf.usaid.gov/pdf_docs/PNADA379.pdf, accessed 4 August 2016). 11 Hook EB, Regal RR. Capture-recapture methods in epidemiology: methods and limitations. Epidemiol Rev. 1995;17(2):243–264. 12 Salganik MJ, Fazito D, Bertoni N, Abdo AH, Mello MB, Bastos FI. Assessing network scale-up estimates for groups most at risk of HIV/AIDS: evidence from a multiple-method study of heavy drug users in Curitiba, Brazil. Am J Epidemiol. 2011;174(10):1190–1196. 13 Feehan DM, Umubyeyi A, Mahy M, Hladik W, Salganik MJ. Quantity versus quality: a survey experiment to improve the network scale-up method. Am J Epidemiol. 2016;183(8):747–757. 14 Feehan DM, Salganik MJ. Generalizing the network scale-up method: a new estimator for the size of hidden populations. Stat.AP (arXiv:1404.4009v3). 2015;(https://arxiv.org/pdf/1404.4009.pdf, accessed 8 August 2016). 15 Salganik MJ, Mello MB, Abdo AH, Bertoni N, Fazito D, Bastos FI. The game of contacts: estimating the social visibility of groups. Soc Networks. 2011;33(1):70–78. 16 Vadivoo S, Gupte MD, Adhikary R, Kohli A, Kangusamy B, Joshua V et al. Appropriateness and execution challenges of three formal size estimation methods for high-risk populations in India. AIDS. 2008;22:S137–S148. 17 Handcock MS, Fellows IE, Gile KJ. RDS Analyst: software for the analysis of respondent-driven sampling data, Version 0.42, 2013. Available from: http://hpmrg.org. 18 Handcock MS, Gile KJ, Mar CM. Estimating the size of populations at high risk for HIV using respondent-driven sampling data. Biometrics. 2015;71(1):258–266. 19 Dombrowski K, Khan B, Wendel T, McLean K, Misshula E, Curtis R. Estimating the size of the methamphetamineusing population in New York City using network sampling techniques. Adv Appl Sociol. 2012;2(4):245–252.
120
11. Population size estimation methods used with surveys
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms
12. Supplemental studies
Investigators and programme implementers often need more information than that which a standard biobehavioural survey (BBS) provides. Supplemental studies can provide additional critical information to explore the individual, socioeconomic and cultural context that determines who continues to be vulnerable to HIV infection and why.
Cohort study: An epidemiological study that follows a group of people over time to observe which and how many experience certain outcomes (e.g. HIV seroconversion). Also known as longitudinal study. Data In qualitative research, the point at which saturation: no new useful information emerges. Exposure: The condition of being (potentially) in contact with something, such as an infectious agent or an intervention. Homogeneity: In respondent-driven sampling surveys, the degree to which people in a population have similar or identical characteristics. Outcome: The measure of interest or endpoint in clinical trials or in data analysis, or the result or consequence of a public health intervention (e.g. condom use or HIV status). Partner Use of one population to recruit another survey: related population; for example, when SW in a survey are asked to refer their clients for survey participation. Primary The survey that directly samples the survey: target population. Probe: Request(s) by the interviewer for more information, or a method used during an interview to help a participant recall information. Semistructured Using an interview guide with themes interviews: or open-ended questions (grouped by topic areas) to conduct an in-depth, open-ended conversation with the respondent(s).
The BBS itself may also raise additional questions that need further investigation, including about the sex partners of the target population. The next three sections describe three data-collection approaches that can be paired with a BBS: qualitative studies, cohort studies and partner surveys.
A-12.1 Qualitative studies Although a BBS provides information on the behaviours of participants, a qualitative study helps investigators understand the behaviours people engage in and the meanings attached to them (1). This section discusses the role of qualitative studies in a BBS and introduces qualitative research methods. Specifically, it discusses two kinds of qualitative studies: formative assessment and supplemental qualitative studies. A formative assessment is conducted before a BBS and provides investigators with the information needed to plan a BBS. A supplemental qualitative study can be conducted during or after a BBS. More information on formative assessment can be found in Chapter A-4. While quantitative research focuses on the what, where and when, qualitative research focuses on the why and how. Together, quantitative and qualitative methods can provide a deeper and more nuanced understanding of knowledge, behaviours and practices, including how individuals and communities understand HIV, sex and drug use, and how best to design interventions. Table A-12.1 provides examples of different types of information that can be obtained from men who have sex with men (MSM) and people who inject drugs (PWID) from both quantitative and qualitative interviews.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
121
Table A-12.1 Example of information provided in quantitative and qualitative interviews Quantitative Question For FSW: For how many years have you been selling sex for money? Answer 3
Qualitative Question Answer
For FSW: “I first got paid to have sex 3 years ago. A guy at a bar offered me money. I didn’t feel good about it and said For how long have you been selling sex? I’d never do it again. About a year later, my sister was in an accident and we needed money for her medical bills. Selling sex was the fastest way to make money. I stopped once we had enough money but after a few months decided to quit my normal job and do this instead because I could make so much money.” For MSM: In what circumstances do you not wear a condom during anal sex? For TG: Tell me about your gender identity. “My boyfriend and I do not use condoms. We have been together for 2 years and get tested for HIV together. We agreed to be monogamous. I do not think he has sex with other men but it’s possible. I am always the insertive partner though so my risk of HIV is less. Because of this I have less to worry about if he cheats so I am okay not using a condom with him.” “I’m a woman. Some people say I’m trans and others say I’m a transwoman. But look at me, I’m a woman. I’ve always been a woman. It’s just that I started out with male genitals. So how can I be a transwoman if I’ve always been a woman? It’s what I am now. A woman.” “I used to use clean needles but one of the places I got them from closed. I don’t have a steady job and cannot afford transport to the other place. It takes too long to get there too. Sometimes I reuse my own needles. Sometimes I share with others. It depends on the situation. If I’m out at a party and I’m able to get heroin, I’ll use whatever equipment is there. I don’t carry that stuff around with me all the time. But if I’m at home by myself, that’s where I have my stash. I don’t want to share it with others. I want to keep it clean.”
For MSM: Do you always use a condom when having anal sex?
No
For TG: Do you consider yourself as male, female, transgender or other? For PWID: In the past 6 months when you injected, how often did you use a new, sterile needle?
Female
25% of the time
For PWID: Why do you not always use a sterile needle?
Strengths and limitations of qualitative research methods Like quantitative methods, qualitative methods have strengths and limitations for both the participant and study investigator, as shown in Table A-12.2.
122
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table A-12.2 Strengths and limitations of qualitative studies Strengths for participants • Offers an opportunity to tell their story • Provides option to disclose as much or as little information as they like • Allows them to emphasize what is important to them • Can be liberating to discuss openly with someone issues that are not openly discussed in the community • Usually allows them control over the date, time and place of the interview
Strengths for researchers • Easily administered and requires very few resources to conduct • Can provide insights into quantitative data • Offers freedom to change the order of questions • Provides opportunity to prompt for more in-depth information • Provides opportunity to record spontaneous answers or responses (e.g. crying) • Provides opportunities to clarify questions • Allows spontaneous questions to be asked to address emerging issues • Provides control over who participates in interviews
Limitations for participants
Limitations for researchers
• Less anonymity • Time consuming to analyse • A participant may not feel comfortable with the interviewer • Possible influence of interviewer on participants’ responses or the questions • Relies on the skill of the interviewer to elicit a good • Time consuming interview (e.g. building rapport, providing good follow-up questions on the spot, avoiding leading questions) • Emotionally exhausting • Emotionally and physically exhausting to conduct interviews, particularly those that address sensitive issues such as HIV, sexuality and violence • Usually small, nonrepresentative sample; cannot apply statistical methods in analysis
Qualitative data-collection methods As described in Chapter A-4, three qualitative datacollection methods are commonly used with BBS: • semistructured interviews; • focus groups; and • ethnographic mapping. These methods are discussed below. Semistructured interviews Semistructured interviews are conducted with a series of open-ended questions. Having questions that are open-ended rather than more structured (e.g. multiple choice questions) encourages two-way communication between the interviewer and the participant. It allows the interviewer to modify the course of the interview
based on the participant’s responses and ask follow-up questions about ideas needing further elaboration. Also, semistructured interviews provide an opportunity for investigators to delve deeply into a topic and understand the reasoning and motivation behind behaviours. Key informants Semistructured interviews are often used to illicit information from key informants – people who have important information that may help in designing the BBS. They include members of the target population and those with whom the members associate. Potential nontarget population key informants are listed in Table A-12.3. They can be allies in working with the target population and maneuvering local politics.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
123
Table A-12.3 Target population and relevant key informants Target population Sex workers Men who have sex with men Transgender persons People who inject drugs
Nontarget population key informants Venue owners or staff, pimps, taxi drivers, HIV service providers, ministry of health officials, clients, NGO staff Venue owners or staff, HIV service providers, ministry of health officials, NGO staff Venue owners or staff, HIV service providers, ministry of health officials, NGO staff Drug treatment providers, harm-reduction service providers, HIV service providers, drug dealers, law enforcement, ministry of health officials, NGO staff
NGO, nongovernmental organization
Key informants should be diverse to capture a broad range of viewpoints. For example, for a BBS among MSM, consider including MSM who are different in terms of age, neighbourhood of residence, socioeconomic status, education, HIV status, marital status, sexual identity, or the extent to which they hide the fact that they have sex with men. Many countries have multiple terms for different types of MSM. Be sure that these typologies are also represented among qualitative study participants. Similarly, non-MSM key informants should include individuals from diverse venues or service providers. If the formative assessment, BBS and supplemental studies are to be successful it is important to build trust among the population through key contacts and by visiting venues the population frequents. This also helps to mobilize the target population and introduce the survey team to them. An interview guide for use with gatekeepers can be found in Appendix I-6. Focus groups Focus groups are semistructured group conversations that are usually conducted with about 6–8 members of the target population under the guidance of a facilitator using an interview guide. They are useful for collecting many viewpoints quickly, particularly about topics that do not require sharing private information. Key informant interviews tend to elicit more thoughtful and personal responses, whereas focus groups tend to elicit the “public” response, and to generate group discussion and debate. For instance, participants might be more likely to discuss their risk behaviours in an individual interview than in a focus group, but might have no preference as to whether to discuss access to health services in an interview or a focus group. Using semistructured key informant interviews and focus groups together speeds data collection and enables investigators to compare responses from both settings. Focus groups also allow participants to interact with each other to spur ideas and confirm or contradict information shared by others, and collect data faster than individual interviews. It is
often good to stagger interviews and focus groups so that different kinds of data from different sources can be analysed throughout the formative assessment or supplemental qualitative study. Whereas interviews are often conducted with members and non-members of the target population alike, focus groups are usually conducted with only the target population. Separate focus groups may be needed for subgroups of a population to understand their unique perspectives. Possible subgroups might include brothelbased and street-based sex workers (SW), younger and older MSM, or methamphetamine and heroin injectors. A brief demographic questionnaire to characterize formative assessment participants can be found in Appendix I-5. A sample interview guide for semistructured interviews and focus groups can be found in Appendix I-7. Considerations when conducting a semistructured interview or focus group Type of interviewer In some settings it is useful to include members of the target population in the survey team because they can facilitate buy-in and linkages to the population. These people may be seen as more trustworthy and may provide insight during data analysis. Conversely, it is possible that participants may not be as honest with their peers, especially with a person already known to them. There may also be political sensitivities to including individuals from one part of the population on the survey team and not another. Interviewers should be trained to collect and manage data, and should use a consistent approach with a clear set of instructions. Limiting the number of interviewers results in more consistent data collection. Interviewers for the qualitative study should be given a clear introduction to the objectives of the BBS and the qualitative study, as well as a thorough review of the data-collection items and instruments.
124
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
It is important to consider the cultural and gender aspects of interviewer selection. For example, in an MSM survey, it may be useful to ask local experts (e.g. members of an MSM service organization) what kinds of people MSM will be comfortable speaking to. It may be appropriate in some settings to have peer-matched or gender-matched interviewers to put participants at ease. In other settings, participants may not accept peers as interviewers. Audio recording interviews and focus groups It can be difficult and distracting for an interviewer to take notes and conduct an interview at the same time because semistructured interviews are open ended, allowing for detailed and in-depth discussions of issues. It is thus recommended to audio record the interview and to have two people on each data-collection team: an interviewer and a note taker. Notes can be used to formulate follow-up questions and probes. Many ethical review committees require that consent be obtained for audio recording interviews and focus groups, and that consent be audio recorded. Audio recordings are considered personal identifiable information and require special security safeguards. The protocol should explain how digital recordings will be kept secure and destroyed. Example consent forms for qualitative interviews during the formative phase and the BBS phase can be found in Appendices I-25 and I-26. Interview guide An interview guide for a qualitative study may focus on a few specific topics or cover a range of topics. A qualitative interview produces the richest results when it is conducted in a conversational rather than a strict question–answer format. This allows for flexibility in the ordering of questions, the prompting of participants and, if necessary, changes in thematic emphasis based on the interests and direction of the participant and investigator. The narrative story elicited through this process results in a vivid description of the context in which participants exist and highlights the greatest influences on their environment. Administering a qualitative interview is more challenging than a quantitative interview because it requires flexibility and readiness to adapt the discussion based on the respondents’ answers. Staff skilled in qualitative methods are required to conduct the qualitative interviews. Additional considerations for conducting qualitative interviews can be found in Chapter A-4. Appendix I-24 includes an interview guide to help survey implementers design a data-collection tool for qualitative studies. It is important to remember that the interview guide is simply a guide. Survey implementers should consider the following when developing an interview guide: • stay consistent with the goals of the qualitative survey. • focus on a few key topics.
• keep the interview to no more than 1 hour to avoid participant fatigue. • establish trust with the participant before the interview begins. Participants should know that they have something important to tell the interviewer, and that the interview is confidential and safe. • use open-ended questions and limit the use of questions that can be answered with one word. • keep questions neutral to avoid asking leading or judgemental questions. For example, ask “What do you think about HIV testing?” rather than “Don’t you think you should get an HIV test?” • include probes – requests by the interviewer for more information – on key questions, as respondents may not always provide enough information when answering a question. More information on probing is included later in this section. • use the same terminology as participants. For example, if a FSW describes herself and other sex workers as “ladies” rather than “sex workers”, the interviewer should similarly refer to them as “ladies”. Types of qualitative questions Many types of questions may be used in a qualitative study (2). The most common and important questions that can be used as part of a qualitative study include values and opinions, knowledge and information, behaviours and experiences, and feelings and meanings. Table A-12.4 provides examples of how these question types may be used to obtain information on a topic (using condoms as the topic). Additional examples can be found in Appendices I-6 and I-7.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
125
Table A-12.4 Question types and examples about condoms Type of question Values and opinions: These questions are aimed at understanding personal, cultural and religious values and their impact on how people think about an issue, experience or event; for example, MSM can be asked how stigma against homosexuality affects their lives and their social and sexual networks. Knowledge and information: These questions aim to understand what people think is factual; for example, if antiretroviral therapy is effective for treating HIV.
Examples of questions • Describe how effective you think condoms are. • Tell me about what young people think about condoms, especially those who use them and those who do not. • Describe how prisoners access condoms. • What do you think about prisoners having access to condoms? • Describe the places where you can access free condoms in your community. • I’m interested in learning your perspective on the relationship between condoms and HIV prevention. Do you think that condoms can prevent HIV? Why or why not? • Describe the range in cost of condoms in this town. What do condoms cost here? • Tell me about what religious groups say about condoms. • Can you tell me about how you first learned to use condoms? Who taught you? Describe for me how to use a condom. What questions do you still have? • How do you negotiate the use of condoms with casual partners? Tell me about the last time you negotiated condom use with a casual partner. How did it go? Is that how you usually approach it? • Who do you use condoms with and why? Tell me about your partners who you use condoms with and your partners who you don’t use condoms with. How do you make the decision about when to use a condom? • When are you most likely to use a condom? • When are you least likely to use a condom? • Describe your feelings when you ask your sex partner to use a condom. How does your sex partner usually react? What feelings come to you when your sex partner reacts this way? • How do you feel when your sex partner refuses to use a condom? • Could you tell me what it’s like when your sex partner asks to use a condom? How do you respond? • What is it like going to buy condoms at a kiosk or pharmacy or requesting them from a health provider?
Behaviours and experiences: These questions are used to elicit people’s descriptions of events, people, experiences and places; for example, the experience of an MSM telling a health service provider that he has sex with men.
Feelings and meanings: These questions are used to understand the emotional response of a participant (or others) in response to an issue or an event; for example, how a person felt after being diagnosed with HIV.
Probing Beyond knowing the themes of an interview and the type of questions to ask, a good interviewer should know how to probe for more information. Probes can ask for elaboration, a definition, a comparison or context (3). The ability to probe is one of the most important skills an interviewer can have. The interviewer must ensure that probes are respectful and nonjudgemental, and draw the participants into a conversation. The most interesting stories or information may sometimes be about events or opinions that the participant is uncomfortable talking about. In these cases, the interviewer needs to have already established trust with the participant, and
may need to offer a reminder of the importance of the participant’s story at that moment. The interviewer must also know when to probe. To do this, the interviewer must understand the goals of the qualitative study and be able to recognize immediately when the participant has mentioned something that could be an important finding. Six useful probes are as follows (4): 1. Elaboration (to obtain more detail): a. Can you tell me a little more about that? b. What else did she say to you?
126
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
2 Continuation (to encourage the participant to keep talking): a. Please go on. b. What happened then? 3. Clarification: a. I am not sure I understand what you mean by that. b. What do you mean by that? c. Do you mean that you saw her do that? 4. Attention: a. That is really interesting. Can you please tell me more? b. I see. 5. Completion: a. You said you spoke to him. What happened then? 6. Evidence (to identify how sure a person is of their interpretation): a. How certain are you that things happened in that order? b. How likely is it that you might change your opinion on that? Avoid asking “why” when probing because it can feel confrontational. Instead, use probes such as “describe”, “how come” or “tell me more about …”. Participants can provide information in a qualitative study through more than just words; for example, they can draw a timeline of their sex partners in the past 6 months and indicate how they met them and which partners know each other. Ethnographic mapping Ethnographic mapping uses simple graphics or maps to convey information about the environment or context in which a survey is being conducted. It helps investigators to understand where the target population congregates and the activities they engage in. Maps can show where risk behaviours occur (e.g. where sex is sold or drugs are used), and locations where the target population gathers (including expected numbers by location, time and subgroup) or accesses health services; thus, maps can show where the target population can be approached. Ethnographic mapping can indicate whether time-location sampling (TLS) is the most appropriate or feasible sampling method, based on whether populations can be readily mapped and accessed. While ethnographic mapping may help with creating a list of venues for cluster-based methods, it may not include all venues and it may include items that are not relevant to a sampling frame. Mapping of resources, including HIV services, where condoms or lubricants can be obtained or where needles can be exchanged, can facilitate identification of referral facilities for the BBS and the location of the survey site.
Maps should also identify barriers that prevent people from accessing a location (5, 6). For example, a religious institution or law enforcement agency near a needle and syringe exchange centre or a survey site may discourage PWID from visiting the centre or participating in the survey. Maps have the added benefit of providing important information that can be used immediately to improve or increase HIV services even before the BBS is completed. Maps must be kept secure and shared with only certain stakeholders, including service providers and community members, to protect the target population. The map should not be too detailed about specific locations. Table A-12.5 provides examples of physical structures, geographical areas and behaviours that may be included in a map.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
127
Table A-12.5 Examples of content of ethnographic maps Physical structures • Service providers or outreach areas • Brothels • Bars, clubs • Pharmacies • Parks • Hotels • Schools • Transport hubs • Community centres • Churches, temples or mosques • Cafes • Health clinics or hospitals • Police stations • Military barracks • Homeless shelters • NGOs • Safe houses for those escaping domestic violence • Needle and syringe exchange centres • HIV testing and counselling sites NGO, nongovernmental organization
Geographical areas • Ethnicity of population • Bus routes • Street intersections • Low-income areas
Social behaviours Places where: • sex is sold or traded • sexual partners are found • drugs are purchased • drugs are used • target population members socialize together • violence occurs • outreach services occur
Ethnographic mapping can take different forms. Key informants or focus group participants can either draw maps themselves or add important information to existing maps. They can also be asked to verify the existence of pre-identified items on a map and make any changes. Maps from different groups or individuals can then be compared and verified. As places on a map may differ by the number or type of people who go to the location, as well as the day or time they may be there, it is useful to also solicit this information. An interview guide for collecting mapping information as part of a key informant interview or focus group can be found in Appendix I-8. For more information on mapping, see Tools together now! 100 participatory tools to mobilize communities for HIV/AIDS (7). Maps vary depending on the participants who create them. For example, an older MSM who is married with children may create a very different map of where he goes for sex or health services than a young MSM who is open about his sexuality. Their barriers to health services are also likely to differ. Neither map is right or wrong; each provides details to investigators on MSM activities in that setting. When diverse participants contribute to ethnographic mapping efforts, investigators get the most complete map. Many mapping exercises include visits to places on the map to observe the site and confirm information provided by participants about locations and behaviours (8, 9). Efforts should be made to visit “high density” and
“low density” locations, because those who frequent them may have different characteristics, behaviours and attitudes. Investigators should note whether there is homogeneity (i.e. people there are similar to one another) within a venue, and across venues. This can facilitate sampling if conventional cluster sampling (CCS) or TLS are used. Observing and engaging the community can bring insight to both the development and conduct of the survey, as well as the interpretation of the findings. While this is relevant for all sampling methods in BBS, it is most relevant to cluster-based methods (described more in Chapter A-9). Qualitative data analysis After the interview or focus group, the interviewer and the note taker should note their impressions about the session, its main themes, and comments and reactions of the participants. This should be followed by a debriefing with the entire survey team. The debriefing should compare the newest data with data from previous interviews and focus groups, to develop an understanding of the context and opinions of the target population. Investigators should frequently reflect on the objectives of the formative assessment and the BBS during data analysis to maintain the focus of the analysis. Because formative assessments usually focus on a few specific questions with only a few participants, data analysis software is not necessary. However, if the number of
128
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
participants and amount of data in a supplemental qualitative study are large, the use of electronic data analysis software is recommended. Qualitative software packages allow the user to transcribe and code interviews, making it easier to identify themes and ideas. It is good practice for each interview to be coded by two people to ensure consistency and quality. One common software application is Dedoose1; others are NVivo2 and ATLAS.ti3. If multiple people conducted the interviews, they should debrief regularly about themes and questions emerging from their interviews, to confirm common themes and explore issues in more depth in future interviews. For example, an interviewer may learn that some transgender persons (TG) indicate that they do not sell or exchange sex; however, upon further exploration the participant may admit to seeking out sex partners who will give things. The identification of this subcategory of transactional sex should be shared among the study team so that it can be explored in future interviews by the entire team. In another example, previous interviews may indicate that PWID had few problems with police or health-care workers; then one interview participant reveals there are actually many problems. New or different information such as this should be explored in additional interviews to find out why this person’s experience is so different. It could be related to where the person spends time or which other activities the person is engaged in (e.g. theft or selling sex). Consent to participate in a qualitative study is separate from consent for the main survey and should be obtained before the start of the qualitative interview. Many institutional review boards (IRBs) consider a person’s voice to be identifiable information. Investigators should check in advance what relevant ethics committees will allow investigators to collect. Audio recordings of interviews should be kept in a secure place and destroyed as soon as the interview has been transcribed and verified. Transcriptions should be verbatim (i.e. an exact reflection of the entire conversation). The transcriber should not abbreviate or paraphrase any text. Noteworthy appearances and undertones (e.g. laughing or crying) should be captured and clearly noted as remarks or comments by the transcriber. During the course of a qualitative interview, participants may provide identifying information about themselves or other members of the target population. This information should be altered to ensure anonymity. Real names should never be transcribed or used in data analysis.
Timing of supplemental studies Supplemental qualitative studies can be implemented during or after the BBS. In a BBS that does not collect identifiable information from participants, it is best to implement the qualitative study during the survey, so that individuals can easily be invited to participate. The qualitative interview should be conducted on another day, so that the participant does not become overwhelmed by the length of the interview process. Individuals are more likely to be available and interested in participating in the qualitative study if they are invited before the BBS ends. Participants should be provided with additional compensation for their transport to the survey site and time spent during the qualitative study. Sampling Participants in supplemental qualitative studies are often sampled purposively; that is, they are selected based on their responses during the quantitative interview. Common criteria for selecting participants include demographics, HIV status, treatment status, history of exchanging sex (for non-SW), and history of violence or discrimination. Other examples include MSM who have female sex partners and PWID sampled in an SW survey. In addition, it is important to select participants who are talkative and seem as though they have information to share. Alternatively, survey implementers may sample every nth participant in the BBS. It is important to note that doing this does not result in representative qualitative data. Sample size As described in Chapter A-4 on formative assessment, sample size is not set a priori for qualitative research. Sample size in qualitative research is based on data saturation (also known as redundancy); that is, the point at which no new useful information emerges from the interviews. It is possible to interview 100 participants who have little to share and obtain very little information. It is also possible to interview 20 participants and obtain all the information needed. The depth of information shared by participants and the analytical abilities of investigators are more important than the sample size (2). The protocol should explain the rationale for the selected sample size (10). A supplemental qualitative study will probably require a larger sample size than a formative assessment due to the greater diversity and depth in responses. Depending on the questions of interest, it may be necessary to sample more participants of a certain type (e.g. individuals who always use condoms). The exact number of semistructured interviews to conduct depends on whether focus group discussions are conducted with members of the target
1 2 3
www.dedoose.com http://www.qsrinternational.com/product http://atlasti.com
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
129
population. Often, the more focus groups conducted, the fewer semistructured interviews needed. It is common to conduct approximately 6–9 semistructured interviews with diverse key informants who are not part of the target population, and 12–15 members of the target population. During data analysis of the qualitative interviews, investigators may decide to stop coding (analysing) interviews once the stage of saturation or redundancy has been reached.
Sample size and sampling method A BBS generating a representative sample is a good way to start building a cohort. The calculated sample size for the BBS should be used as a starting point to determine the sample size for the cohort study. The following additional factors should be considered to determine whether the sample size should be increased due to the addition of the cohort study: • the frequency of the event of interest: - This is the most important factor. If survey implementers are interested in observing HIV incidence, a relatively rare event, they will either need to follow participants for a long time and risk loss to follow-up, or follow more people for a shorter period. They should consider the cost and logistical trade-offs. • cohort-specific enrolment criteria: - At time of enrolment into the cohort, participants should lack the main outcome of interest. For example, if the objective is to examine HIV incidence, only HIVnegative participants may be offered enrolment. • unwillingness of BBS participants to participate in the cohort study: - Not all participants in the BBS will be willing to participate in the cohort study. Willingness to participate should be assessed during the formative assessment. • loss of participants: - Participants may initially be willing to participate in the cohort study but eventually withdraw or be lost to follow-up due to a lack of time or interest in the survey. This will reduce the power to detect statistical significance as the sample size decreases; it also opens the study to bias if the participants who drop out are different from those who remain in relation to the variables of interest. Participation in the BBS counts as the first observation (baseline) for each participant in the cohort study, which continues after the BBS is completed. Data-collection instruments Depending on the goal of the cohort study, investigators may need to develop additional tools and forms that are not used during the BBS. Additional tools may include appointment reminder cards, and personal identification tracking forms to contact participants before future visits. If adapting a questionnaire that was made for a crosssectional survey, it is useful to consider how the variables in the questionnaire track changes over time.
A-12.2 Cohort studies BBS are generally cross-sectional, meaning that information on each participant is collected at only one point in time. Cross-sectional surveys cannot show causation, or measure the effect of exposures on outcomes over time, or time to an event (e.g. death or initiation of treatment). Cohort studies, in contrast, can show causality because they follow the same participants over time (11).1 Cohort studies enable survey implementers to explore many things, including: • directly observed HIV incidence; • behaviour changes; • linkages to HIV care and treatment; and • retention in HIV care and treatment. Despite the important information that cohort studies can provide, they are not very common due to the financial costs and logistical challenges of following the same people for an extended time. They also require substantial planning to implement. However, cohort studies can be conducted when resources allow and when it is important to demonstrate causal relationships. In these cases, a BBS can provide the initial sample for a cohort study if it is planned well in advance. Participants in the BBS can then be offered enrolment in the subsequent cohort study. This chapter provides some guidelines for conducting a cohort study as part of BBS planning. Timing The decision to implement a cohort study together with a BBS should be made early in the planning stages. This will allow survey investigators to assess the feasibility of a cohort study during the formative assessment (see Chapter A-4). Implementers can use that assessment to determine the willingness of target population members to: • participate in a cohort study; • return to the survey site for follow-up visits for additional interviews or to provide biological specimens; and • be contacted for reminders about scheduled survey appointments.
130
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Data analysis Data from a BBS and supplemental cohort study can be analysed separately or together, using the BBS data as the cohort’s first observation point (12, 13). BBS data should be analysed using appropriate analysis techniques for cross-sectional surveys. These data should be analysed while the cohort study is being implemented to make results available faster. Other analysis methods, namely those using person-time to measure exposures and outcomes, are used in cohort studies (11). Operational and cost considerations Cohort studies can be expensive because of the amount of time they take to implement and the costs associated with tracking people over time. Collaborations with service providers may reduce costs but may also risk discouraging participation by individuals who do not want to be associated with these providers. Survey implementers should ensure that budgets can sustain staff salaries and other recurring costs during the cohort’s duration. Conversely, extra coordination will be necessary if participants start returning for their second observation while recruitment (first-observation visits) is still going on. This would require survey staff to manage two different procedures and possibly high participant load at the same time. Participants are unlikely to forget about a follow-up survey visit if it is scheduled shortly after the previous visit, but too many visits may result in participant fatigue and dropout. The number of survey visits per participant and the amount of time between visits depend on the goals of the cohort study and study operations. For example, the standard indicator time frame (see Section 3) is to measure linkage to care within 6 months. Thus, asking HIV-infected participants if they have gone for care 1 month after being diagnosed does not facilitate reporting on this indicator. Similarly, if seeking to measure directly observed HIV incidence and compare it to results from an assay measure, the time between survey visits will depend strongly on the kind of HIV test being used and its window period. Human subjects considerations If a cohort study is planned on top of a BBS, the consent language used for the BBS will need to reference that survey staff will also approach the potential participant about enrolment in the cohort. A separate protocol and review by an IRB will be required.
A-12.3 Partner surveys A-12.3.1 What are partner surveys? In a partner survey, participants in a primary survey recruit another related population to produce a second sample. Participants in the primary survey may be sampled using any method, whereas participants in the partner survey are identified using referral sampling. For example, a primary survey of SW could be coupled with a partner survey of their clients. In this case, SW who participated in the primary survey would be asked to recruit some of their clients for the partner survey. Unlike snowball sampling, the eligibility criteria for the partner survey is different from that of the primary survey, and partner survey participants do not recruit others. Investigators may consider using a partner survey when direct sampling methods such as snowball sampling, CCS, TLS or respondent-driven sampling (RDS) are unable to successfully reach the target population. Direct sampling is difficult or impossible when the target population members do not interact with one another (e.g. clients of SW), do not frequent venues (e.g. sex partners of PWID), or are unaware of their membership in the population (e.g. female partners of MSM, regular partners of SW, or individuals whose partners are in multiple concurrent partnerships). This method can be employed to recruit female sex partners of MSM, clients of SW or sex partners of TG. The probability of being recruited into a partner survey is usually unknown. Researchers are working to identify methods for obtaining probability-based results using this method, and partner surveys using sampling weights have been published (14). Currently, there are three options for analysis: • analysing the data as though it came from an unweighted nonprobability sample; • applying the same weights as those used that were applied to their recruiter in the primary survey; and • deducing a weight for the participants; for example, by asking MSM how many female sex partners they have, or asking SW how many clients they have had in the past 2 weeks.
A-12.3.2 Preparing for an indirect sampling survey Adding a partner survey to a BBS requires additional preparations: • planning how to indirectly sample participants; • determining the target sample size;
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
131
• protecting the confidentiality of primary survey participants from their recruits and vice versa; and • identifying an appropriate survey site. Planning how to indirectly sample participants The most common means for recruiting participants for a partner survey is to give participants of the primary survey (e.g. SW) a fixed number of coupons to distribute to another target population (e.g. clients of SW). The coupons used for each population should be distinct in colour and survey identification numbers (IDs), so that participants in the partner survey are not recruited into the primary survey or vice versa. More information on the use of coupons for recruitment can be found in Section A-9.7 on RDS. Although it is not yet clear whether RDS analysis methods can be extrapolated to partner survey data, investigators should still track who recruits whom to facilitate different analyses. The number of coupons given to participants of the primary survey to recruit participants in the partner survey will depend on the characteristics of this second population. For example, SW could be given three coupons to recruit clients because they have multiple paying clients, but only one coupon for regular nonpaying partners. It is preferable to recruit one client each from many SW rather than multiple clients from just a few SW, because the clients of any one SW may have more in common with each other than with clients of other SW. The number of coupons per secondary population may also depend on the survey budget. Determining the target sample size It is difficult to anticipate the sample size for the partner survey because it is often unknown how many partner survey participants a primary survey participant may recruit on average. This depends on the number of “partners” any one primary survey participant may have, the willingness of the primary survey participant to actually recruit (hand out a coupon to) a “partner”, and the willingness of such partners to enrol. For the purposes of planning and writing the protocol, investigators can estimate the sample size relative to the number of participants in the primary survey. For instance, in a primary survey of MSM and a partner survey of their female sex partners, investigators may anticipate that not all MSM will have female sex partners, so the sample size of the partner survey will be smaller than the size of the primary survey. In contrast, a partner survey of clients of SW may approach that of the primary survey of SW, because all SW have clients and most have more than one. More information on sample size can be found in Chapter A-10. Protecting the confidentiality of primary survey participants from their recruits The ethics and confidentiality of partner surveys are complicated by the necessity to protect the
confidentiality of the primary survey participants. For example, investigators should not tell female sex partners of MSM that they are eligible for the survey because their recruiter (sex partner) is an MSM. In this case, providing the participants with information about the person who recruited them is a violation of confidentiality. Similarly, participants in the partner survey should not be informed of the eligibility criteria for either survey population (primary and partner survey). For example, in a survey examining sexual networks and multiple concurrent partnerships, partners of survey participants should not learn of the existence of concurrent sex partners or even why they are eligible to participate. Preventive measures could include scheduling appointments for different participants on different days. Survey staff should be trained to withhold information from participants of the partner survey about the true nature of the primary survey population. Recruiters may need to be trained to provide an alternate explanation to their recruits about the purpose of the survey and the eligibility criteria that still allows them to be truthful. One useful way of explaining to partner survey participants why they did not receive coupons to recruit others is that a lottery system was used to select who was asked to do this. Identifying an appropriate survey site Although the primary survey may be implemented at one or more fixed sites or at venues, a partner survey is best implemented at a fixed survey site that participants can easily visit according to their schedule. It is not necessary for primary survey participants to visit this survey site. It may also be helpful to pick a site where a diverse population is present, so that no population stands out to the participants. For example, in an RDS survey of MSM who also recruit their female sex partners, the site would be frequented by many gay men but few women, which could cause suspicion among the female partners. If instead many groups use the site, MSM and their female sex partners will blend in with the other populations. Alternatively, separate days may be dedicated to enrol either primary or partner survey participants. More information on how to identify and set up survey sites can be found in Section A-9.7 on RDS.
A-12.3.3 Implementing a partner survey The survey procedures for partner survey participants may be similar or identical to those of the primary survey. If the partner survey aims to collect biobehavioural data, then these participants may undergo similar survey activities to that of their recruits (e.g. responding to questionnaires and providing biological specimen). But survey activities may also differ. Investigators may wish, for example, to conduct qualitative rather than quantitative interviews in the partner survey, especially if investigators suspect that the achievable sample size is small.
132
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Below are several considerations for survey procedures in the primary and partner surveys. For participants in the primary survey: • recruitment - participants in the primary survey should be asked to recruit individuals for the partner survey only after completing the survey questionnaire. This prevents participants from confusing the activities of the two surveys. It also helps investigators to assess whether the primary survey participant may recruit “partners” and how many they may recruit. For instance, only MSM who indicated in the survey that they have female sex partners should be asked to recruit them. - In RDS surveys, primary survey participants should first be trained in recruiting peers for the primary survey. They may then be trained to recruit for the partner survey. Investigators should use different coupons (using different colours or other visual cues) to help to keep the two processes distinct. For example, green coupons could be used for recruiting peers (primary survey) whereas orange coupons could be used for the partner survey. - If the primary survey sampling method is not RDS, survey staff should introduce participants to the concept of peer recruitment. • biomarkers - because partner surveys must generally be conducted in a survey office, primary survey participants surveyed at a venue, as in TLS, can go to the survey office to collect any laboratory-based test results. For participants in the partner survey: • eligibility - the recruitment coupon and eligibility should be verified.
• recruitment - partner survey participants may not know the real reason they were recruited. In some cases it may be possible to ask partner survey participants to recruit others, but doing so may further complicate data analysis. • biomarkers - because a survey site is used, participants can return to the office to collect any laboratory-based test results. Where feasible, it is best to use rapid tests to ensure that participants receive the results of their tests in real time. The research team should track the link between recruits and recruiter through referral management software,4 Microsoft Excel or some other system to allow for possible weighting of the survey data. More information on tracking sampling can be found in Section A-9.7.
A-12.3.4 Coordination Surveying more than one population at a time requires substantial coordination. Separate eligibility screening, coupon systems, consent forms and data instruments may need to be used. If the protocol indicates that participants will not be provided with complete information about their recruiter or the reasons for their eligibility, survey staff should be trained to ensure that such information is not inadvertently disclosed; for example, that the female sex partner of an MSM is not accidentally given a consent form for MSM participants or coupons to recruit others. Survey staff should use different participant ID numbering systems, different coloured paper for tracking, or other defined procedures for each target population to keep them distinct.
A-12.4 References 1 2 3 4 Ozawa S, Pongpirul K. 10 best resources on ... mixed methods research in health systems. Health Policy Plan. 2014;29(3):323–327 (http://www.ncbi.nlm.nih.gov/pubmed/23564372, accessed 30 August 2016). Patton M. Qualitative evaluation and research methods. Beverly Hills, Sage publications. 1990. Mariampolski H. Qualitative market research. Sage. 2001 (http://dx.doi.org/10.4135/9781412985529.n312, accessed 30 August 2016). Rubin H. Qualitative interviewing (2nd ed.): the art of hearing data. Thousand Oaks, California, 2005.
4
www.respondentdrivensampling.org
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
12. Supplemental studies
133
5
Emmanuel F, Blanchard J, Zaheer HA, Reza T, Holte-McKenzie M, team H. The HIV/AIDS Surveillance Project mapping approach: an innovative approach for mapping and size estimation for groups at a higher risk of HIV in Pakistan. AIDS. 2010;24 Suppl 2:S77–84 (http://www.ncbi.nlm.nih.gov/pubmed/20610953, accessed 30 August 2016). UNAIDS. Issues brief: local epidemics. Joint United Nations Programme on HIV/AIDS (UNAIDS); 2014 (http://www. unaids.org/en/resources/documents/2014/20140707_JC2559_local-epidemics, accessed 16 November 2016).
6
7 K4Health. Tools together now: 100 participatory tools to mobilize communities for HIV/AIDS. 2016 (https://www. k4health.org/toolkits/pc-hivaids/tools-together-now-100-participatory-tools-mobilize-communities-hivaids, accessed 4 August 2016). 8 9 Aronson RE, Wallis AB, O’Campo PJ, Whitehead TL, Schafer P. Ethnographically informed community evaluation: a framework and approach for evaluating community-based initiatives. Matern Child Health J. 2007;11(2):97–109. Spradley JB. Participant observation. Belmont, CA, Wadsworth Cengage Learning. 1980.
10 Guest G, Bunce A, Johnson L. How many interviews are enough? An experiment with data saturation and variability. Field Methods. 2006;18(1):59–82. 11 Grimes DA, Schulz KF. Cohort studies: marching towards outcomes. Lancet. 2002;359(9303):341–345. 12 Horyniak D, Higgs P, Jenkinson R, Degenhardt L, Stoove M, Kerr T et al. Establishing the Melbourne Injecting Drug User Cohort Study (MIX): rationale, methods, and baseline and twelve-month follow-up results. Harm Reduction Journal. 2013;10(1):11. 13 Yang H, Hao C, Huan X, Yan H, Guan W, Xu X et al. HIV incidence and associated factors in a cohort of men who have sex with men in Nanjing, China. Sex Transm Dis. 2010;37(4):208–213. 14 Shah NS, Shiraishi RW, Subhachaturas W, Anand A, Whitehead SJ, Tanpradech S et al. Bridging populations—sexual risk behaviors and hiv prevalence in clients and partners of female sex workers, Bangkok, Thailand 2007. J Urban Health. 2011;88(3):533–544.
134
12. Supplemental studies
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms Data The protection of data against confidentiality: unintended or unauthorized access. Data security: The requirements (technical and physical) that keep data secure and confidential.
13. Data management
The purpose of this chapter is to describe data-management planning to facilitate the collection of high-quality data. The survey team must have a clear plan for compiling and processing data, and for ensuring the confidentiality of survey participants in the earliest stages of survey planning. Investing in data management before data collection saves time and allows effective use of resources.
Examples of survey-level data documentation: • description of the database: narrative of the objectives and methods of the survey to provide context for the database; • detailed chronology of events and activities through the life of the survey (e.g. starting and ending dates of data collection, and number of participants); • compilation of all data-collection instruments (e.g. questionnaires, consent forms, specimen tracking sheets and screening tools); • list and descriptions of all materials used (e.g. interview schedules, logbooks, standard operating procedures [SOPs], training materials and confidentiality agreements); • information on data validation and quality assurance procedures performed; • information on data confidentiality, access and conditions for sharing or use; • SOPs on how to make changes to the data (e.g. cleaning, error checking and creation of variables); and • description of changes made to the data since its collection (e.g. cleaning, error checking and creation of variables).
Creating a data-management plan A data-management plan helps investigators to design and implement the survey; monitor data collection; prepare for data entry, data cleaning and data analysis; and guide data use and sharing. A data-management plan generally includes the following components: • data documentation • data dictionary • unique participant identifier • data-quality checks • skip patterns • data entry • data confidentiality • data access and use • data backup • data archiving and version control • data security and storage
A-13.2 Data dictionary A data dictionary or codebook describes the content of a dataset. Many data analysis or data-capture systems can automatically generate a data dictionary or codebook. An example of a data dictionary is given in Appendix I-28. A data dictionary typically contains the following information: • variable names, labels, descriptions and values; • acceptable values; • variable type (e.g. text or string, numeric, data or time, or Boolean); • codes and reasons for missing data; • checks or restrictions (e.g. valid ranges, logic checks or skip patterns); 13. Data management
A-13.1 Data documentation Data documentation is a detailed record of how the data are collected, defined, organized, processed and analysed. Both paper-based and electronic data should be documented at each step of the data-management process to ensure all data can be traced accurately from collection to analysis.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
135
• derived data or variables created using code, algorithm or command file; and • weighting variables created, if needed (for probability-based sampling methods).
A-13.3 Unique participant identifier Each participant should be assigned a unique participant identifier or survey identification number (ID). All data collected for a participant must be linked back to this survey ID. The data-management plan should include specific quality-control measures to prevent reuse of the survey ID. Data entry or transcription errors can be reduced by using barcodes with the survey ID on preprinted labels that can be placed on forms and specimens. This requires the use of barcode scanners or other devices (e.g. tablets) with digital cameras and with the capacity to scan coded data. There should also be processes to account for unused labels to prevent accidental reuse. If barcode scanning is not feasible, preprinted labels should still be considered to reduce transcription errors and the reuse of survey IDs. Most laboratories assign their own unique ID to each specimen. Survey investigators should determine Table A-13.1 Validation rules Validation rules Application Acceptable values
whether they need additional data-management tools to link the survey ID and the specimen ID. Linking the main survey ID with the specimen ID can help detect coding errors (or in respondent-driven sampling [RDS], the recruit or coupon ID). Also, the specimen ID can serve as a backup unique participant identifier.
A-13.4 Data-quality checks Assessing data quality is an integral part of data collection, data entry, data cleaning and data analysis. Quality-control measures help prevent data loss and minimize data entry errors, ultimately saving time and resources. Examples of data-quality checks include creating validation checks and skip patterns. Ideally, validation rules for a questionnaire should be developed before data collection for both electronic and paper-based administration. In electronic data collection, data validation rules can be programmed into the software to automatically detect invalid responses. For paper-based data collection, interviewers or their supervisors in the field will have to detect the errors themselves. Therefore, interview staff must be properly trained to apply the validation rules. Table A-13.1 shows sample validation rules for a survey.
Example The question “Have you ever had sex?” could have the following values: • Yes • No • Don’t know • Refuse to answer The values “1”, “2”, “8” and “9” are the only acceptable values for responses to this question.
Assign values that are acceptable for a categorical variable, including values that represent “Other”, “Don’t know” and “Refuse to answer”
Valid range
Assign a minimum and maximum value based on allowable responses
The question “What month were you born?” has the following minimum and maximum values based on the number of months in the year. • Minimum: 1 • Maximum: 12 Values outside of this range are not allowed. If a participant responds he is 22 years old, all agerelated responses cannot be larger than 22. For example, the participant cannot respond “27” to the question “When did you first have sex?”. Example 1: Survey ID Example 2: The participant’s sex is needed because it will affect the type of questions asked in the questionnaire.
Logic or consistency checks Mandatory fields
Ensure data collected at different points throughout the interview are internally consistent
Require that a response be provided. For some fields, responses such as “Don’t know” and “Refuse to answer” may not be relevant as response options; for example, questions determining eligibility. In general, however, it is preferable not to leave blanks and to record a response for each field even if it is “Don’t know” or “Refuse to answer”.
ID, identification number
136
13. Data management
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
A-13.5 Skip patterns Skip patterns differ from validation rules. They ensure that only relevant questions are asked to each participant, because questions are only asked based on responses to previous questions within a questionnaire. For example, if a participant responded that he is male, a question regarding pregnancy will be skipped. Incorporating skip patterns can improve the flow of administrating a questionnaire. It is important to pilot test a questionnaire to make sure that skip patterns are applied correctly.
on any relevant laws that exist in the country and the penalties for breaching confidentiality. The original agreement should be held in the staff member’s personnel file and a copy given to the staff member. Staff should review confidentiality and data security procedures, and sign the agreement on an established periodic basis (e.g. annually). An example of a data-use agreement is in Appendix I-37. Any personal identifying information collected (e.g. name, ID and address) should be removed from electronic forms before data entry. Personal identifying information should only be collected if it is required (e.g. returning laboratory results). If collected electronically, this information should be encrypted. If collected by paper, the information should be stored separately from other collected data. Special consideration is needed for relational data where connections between variables in related datasets can disclose participants’ identities. Sensitive words – for example, men who have sex with men (MSM), people who inject drugs (PWID) and female sex workers (FSW) – should not be used on any datacollection tools or other educational materials (e.g. educational pamphlets). The same is true for geospatial data that may reveal sensitive information. Documents such as signed consent forms should be stored separately from the data files.
A-13.6 Data entry Data entry errors are common and can have a measurable impact on data analysis. Double data entry, especially for paper-based questionnaires, is strongly recommended to produce a valid dataset. Double data entry is when two different people enter the same data and compare the data for errors. Although this process costs more than single data entry, it reduces the number of data entry errors. There are two ways to clean data after double entry: • compare the two data entry files to find discrepant observations. This can usually be done automatically using statistical software (e.g. Epi Info, SPSS, SAS or Stata). • have the statistical software flag inconsistencies as data are entered the second time. When an inconsistency is identified, the person responsible for entering the data the second time determines whether the original entry is correct. Cleaning of data entry errors is further described in Chapter B-3.
A-13.8 Data access and use Guidelines for data access and use should be an integral part of the data-management plan. This includes determining: • data access: - who has access to the data; - who has access to computers, tablets and servers that contain data; - who has the authority to give access to the data; • data ownership: - who owns the data; - whether or how the data will be shared; • data use: - how the data will be used; and - who will manage the data. Defining access should not be limited to datasets but also data collected on paper such as logbooks, questionnaires, checklists and laboratory results. Access to data with personal identifiers should be limited to staff who are authorized to access those data for operations (e.g. the data manager checking for duplicates or a counsellor giving a participant test results). Data analysis should be conducted only on a de-identified dataset.
A-13.7 Data confidentiality Data confidentiality is the protection of data against unintended or unauthorized access, especially participants’ identities. Biobehavioural surveys (BBS) should be done anonymously; if this is not possible, investigators must implement measures to assure data confidentiality and security. Data confidentiality must be maintained during data entry and processing, and staff should be trained in ethical issues important for surveys. All staff with access to the data should sign a confidentiality agreement before the start of the survey. This agreement outlines how data must be kept confidential and secure, and it may include information
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
13. Data management
137
A-13.9 Data backup It is imperative that all data are protected against loss. Electronic data can be lost due to hardware or software failure, viruses, power failure or human error. Paper forms can be lost during collection or transfer, destroyed by accident, ruined by physical means (e.g. beverage spills), or misfiled and never located again. These and other topics are described in an example data-management SOP for surveys using a questionnaire development system (QDS) in Appendix I-27. Investigators should back up data daily during data collection. Systems that automate data backups are ideal because they do not require conscious effort to safeguard data. The backup plan should include strategies for all systems where the data are stored, including tablets and network-based and non-network-based computers. Backup files should be verified and validated regularly by comparing them with the originals and checking them for completeness. The data backup plan should include all the steps for data backup and validation of the backups, and should be an appendix to the data-management plan. Access to backed-up data should be restricted, similar to access to the original data file locations.
A-13.11 Data security and storage Storing backups solely on-site does not provide adequate protection, because the survey office’s computers are usually part of the same network, or the office may be burgled, or raided by law enforcement. Furthermore, safeguards must be put in place to limit risks to participants through accidental or malicious disclosure of data by theft of forms or data-collection devices. These safeguards can include use of passwords, encryption and limiting the data stored at survey sites, especially on portable devices. To secure paper-based data: • keep all hardcopies (paper questionnaires and other survey forms if applicable) in a locked cabinet with controlled access; • plan secure transport systems of paper forms to prevent loss of data; • store personal identifying information separately from other survey data; and • keep a log of the survey ID of all paper-based forms and where they are located; if documents are transported from one survey site or office to another, the log should indicate the survey ID of all forms in the shipment. To secure electronic data:
A-13.10 Data archiving and version control Every time a dataset is backed up a new version is created. All versions of the database must be archived and good version control practices should be implemented, to protect against loss of information. Versions of the files and backup files can be identified by dates recorded in the file name or version numbering in the file name (e.g. v1, v2, v3, and so on). For paper-based data-collection tools, the version number can be inserted as a header or footer on the document. Maintenance of the master files of each dataset should be assigned to specific team members who are responsible for updating specific components of the datasets or database, as needed. Any changes should be managed through direct supervision according to the data-management plan. The process of freezing the database (i.e. when the data files are archived) should also be included when describing version control. Freezing the database involves making copies at specific intervals for archiving and data use. This ensures that any analysis from the database can be reproduced and provides a structured time frame for incorporating changes.
• use access controls (password protection) on devices such as computers and tablets. Strong passwords have at least eight characters, including at least one lower case letter, one upper case letter, one digit and one symbol; a password should be required every time an electronic device is turned on; • encrypt devices or the data stored on them; • restrict access to the server to those who need it, and limit user permission to prevent unintended and undocumented changes to the data (e.g. view-only rights); • control internet access and installment of new software to minimize the risk of malware installation; • avoid storing computer data files with personal information on portable devices (laptops, tablets or flash drives) that can physically be lost; • minimize transfer and use of portable devices (e.g. flash drives) to prevent physical loss of data and transfer of viruses or malware; • ensure that portable devices that contain data are kept by field staff at all times when in the field, and are shut down when not in use; • purge computers and other data-capture devices of data after the last interview of each day by uploading the collected interviews to a main, secure database, to minimize both the amount of data carried on the handheld device and the number of records lost or compromised if the handheld device is lost or stolen; and • use an external source to back up data.
138
13. Data management
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
To secure electronic data during transmission: • do not transmit any data over an unsecured Wi-Fi or internet connection; • use a virtual private network (VPN) for connecting between sites and the server; • avoid using email to transfer electronic files; if files must be emailed, they should be password protected or encrypted; and • schedule regular antivirus software updates, software patch installs and virus scans. When a computer used for data collection is taken out of service, any hard drives that may have once contained data must be wiped clean and reformatted before being used for another purpose. Software that repeatedly overwrite data on a disk with random patterns have been developed for this purpose and should be used. Other removable computer hardware (e.g. compact disks or
DVDs used to store data backups) that are no longer needed should be destroyed and not used for another purpose. Upon conclusion of the survey, all documents containing sensitive information must be shredded before disposal. Similarly, all electronic files with sensitive information that are not part of the master dataset should be wiped clean and overwritten. Breaches in the data security procedures should be reported immediately to survey staff leadership and investigated by staff to assess the causes and implement remedies. Confirmed breaches resulting in the release of sensitive information should be reported to the principal investigators, who must report breaches and any remedial action taken to all institutional review boards.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
13. Data management
139
Section
A A-14.1 Considerations for field staff selection Staffing decisions are influenced by the survey’s goals, objectives and activities. It is important to consider the number of field staff needed, their level of experience, their personal characteristics, and any accreditation they may need to collect and test biomarker specimens. Number of field staff The number of field staff needed at a survey site may depend on various factors, including the: • complexity of the survey process (because a survey with more steps may require additional staff) and the sampling method used; for example, respondent-driven sampling (RDS) versus time-location sampling (TLS): - RDS is a passive sampling method, meaning that investigators have little control over how many potential participants to expect on any given day; - in TLS, survey staff spend a lot of time traveling to and setting up the survey site at the various locations; • sample size (because the larger the sample size, the more interviewers will be needed to conduct the survey in the predetermined time frame); • data-collection method – for example, audio computerassisted self-interview (ACASI) or computer-assisted personal interview (CAPI) – and whether data are collected on paper or electronically; Paper-based data collection requires additional staff to double enter and clean the data; • inclusion of a second-visit activity (e.g. follow-up questionnaire, provision of biological test results); • security and transportation needs (which may necessitate security guards or additional staff, such as drivers for field staff); and
14. Staff selection
This chapter discusses assembling field staff for biobehavioural surveys (BBS), including their number, roles, and how survey activities are distributed between them. The roles and responsibilities of non-field-based staff (e.g. principal investigator and statistician) are not discussed here because these positions have usually been developed and filled before the decision is made to conduct a BBS.
• method and type of biomarker specimen collection and processing (e.g. testing on-site or off-site). Experience of field staff The level of experience required from each staff member depends on the demands of their position. At a minimum, field staff should have some experience with the target population, and they should also have: • previous experience working on other health surveys; • knowledge of the topics and issues related to the survey; • experience using computers or tablets if data are collected electronically; and • experience in specimen testing and counselling if on-site testing is part of the survey design. Characteristics of field staff The personalities of staff members should be considered. Field staff will be more successful if they reflect some of the following characteristics: • friendly and able to build rapport; able to make participants feel welcome and comfortable; • responsible and reliable (e.g. arrive on time when the survey site opens, and take appropriate action when unexpected situations arise); • mature and professional (e.g. able to interact with others in a professional manner); • nonjudgemental and non-discriminatory toward the target population; • able to work in a team; • able to speak the language(s) spoken by the target population; • committed to working for the duration of the survey; and • available to work nontraditional hours (if required).
140
14. Staff selection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Consideration for inclusion of target population members as staff Investigators may consider filling positions with members of the target population. This approach should be determined based on whether the formative assessment indicates a widespread acceptance of having community members serve as staff, because it has implications for survey participants’ sense of confidentiality and safety. On one hand, community members may build rapport with survey participants more easily. On the other hand, survey participants may be concerned that staff members who are part of the community and their social network could compromise their privacy.
A-14.2 Field staff roles and responsibilities: non-method specific Investigators must define staff roles and build teams within the context of the survey’s objectives, budget and activities. As such, investigators may need to merge positions or divide them to make the best use of available resources, time and effort. They should adapt descriptions of the positions as appropriate for the country in which it is being conducted, the sampling methodology, and the survey site layout. For example, nurses may be nurse counsellors and also return results. In some settings and with adequate training, nurses may also collect blood and perform rapid testing. Table A-14.1 lists staff positions that are needed for all BBS, regardless of the recruitment method.
Table A-14.1 General BBS roles and responsibilities Roles Field staff coordinator Interviewer
Responsibilities Supervise daily activities at site Provide support to staff Communicate on regular basis with principal investigator Know the questionnaire and where sensitive issues may arise Build rapport with survey participants Obtain informed consent from eligible survey participants Conduct interview or oversee the interview process Record responses Ensure that information collected (including paper questionnaires and electronic data) is securely saved or stored after the interview Provide pre- and post-test HIV counselling based on local regulations Collect biological specimens Complete relevant forms and processes, and package specimens for testing or transport Conduct on-site rapid testing as specified in protocol Manage data entry and database development Oversee data editing and quality control Capture data and enter it into database Analyse and interpret data Test specimens Record test results Transfer results data to main survey database Meet with community members Share survey progress with members Encourage participation in survey Transport staff and supplies Protect field staff and participants Maintain survey site
HIV counsellor/nurse or laboratory technician
Data manager Data entry staff Data analyst Laboratory staff
Community liaison/mobilizer
Driver Support staff
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
14. Staff selection
141
A-14.3 Field staff roles and responsibilities: method specific Some BBS methods require specific additional positions to be filled. Tables A-14.2 and A-14.3 list positions specific to TLS and RDS. Because TLS and RDS survey sites are not investigator-controlled environments, it is important for the survey investigators and field staff to meet and work together before survey implementation.
Survey coordinators should research pay scales and payment schedules of comparable survey projects in the respective country or region. Payments should be comparable to the salaries of similar local jobs, and decisions about salary rates, payment schedules and working hours should be made by someone who is knowledgeable about the local context. Investigators should also consider roles for staff that are remunerated in-kind rather than paid. Local laws and employer policy will govern contractual arrangements, sick and annual leave, and termination of contract.
A-14.4 Considerations for staff payment Funding for paying staff salaries should be accounted for in the budget as early as the survey-planning phase.
Table A-14.2 TLS-specific roles and activities TLS specific Roles Enumerator or counter Recruiter Responsibilities Collect quantitative data used to determine inclusion of venue in survey Count venue attendees and direct recruiters to approach selected attendees Approach venue attendees Explain survey to potential participant, describing: • the purpose • interview procedures • privacy protection • compensation
Table A-14.3 RDS-specific roles and activities RDS specific Roles Receptionist Responsibilities Greet people entering the survey site Check that people entering the survey site have a valid coupon Inform people how the coupon process works when they arrive without a coupon Give information about the survey over the phone or at the survey site Track and manage the flow of participants Inform participants of how long they may have to wait for an interview Screen participants to see whether they fulfil the eligibility and inclusion criteria Ask specific questions about behaviours related to the target population to identify and disqualify imposters Initiate checklist form Establish a unique identifier to link a recruit’s follow-up visit to initial visit Verify coupons Develop coupon-tracking system (if not already existing) Manage the coupon-tracking system Explain the survey to the participant Conduct and record consent Give instructions on peer recruitment Issue referral coupons Provide primary or secondary compensation, as warranted
Eligibility screener
Coupon manager
142
14. Staff selection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
A Key terms
15. Standard operating procedures
This chapter provides guidance for creating and using standard operating procedures (SOP) for a biobehavioural survey (BBS). An SOP is a set of written instructions for distinct survey activities. Developing and using SOP helps to ensure that staff uniformly conduct the survey as recommended by the investigators and as outlined in the protocol. SOP help to maintain quality and consistency, particularly when staff change.
The first, or title, page should list: • the staff concerned; • a summary of content, with purpose and field application; • any related SOP; • the name of the author; and • the name of the person authorizing the SOP (optional). SOP review and approval The draft SOP should be reviewed by other individuals who have experience with the activity being described. Survey implementers should field-test complicated SOP, and investigators should approve the final version of the SOP. SOP revisions SOP need to remain current in order to be useful. When procedures change, SOP should be updated and reapproved. An SOP change log can be used to track these changes and the versions approved. SOP should be reviewed periodically during the survey to ensure that procedures are still appropriate. The survey coordinator is responsible for maintaining and updating SOP as needed. After an update, the survey coordinator should provide the revised SOP and associated training to all relevant staff, and ensure that staff are following the revised SOP.
Standard A set of written instructions for distinct operating survey activities. procedure:
An example SOP that can serve as a template is provided in Appendix I-30.
A-15.1 How to write a standard operating procedure Style SOP are written in concise, clear, simple language (active voice and present tense) using a step-by-step approach. Flow charts can help illustrate steps for staff to follow. SOP preparation Survey investigators determine which activities or processes warrant an SOP. An SOP should be written by individuals knowledgeable about the survey design and activities (i.e. principal investigator, survey coordinator, data or laboratory manager), and should contain enough detail for staff to carry out the procedure without supervision. Each page of the SOP should have a header or footer, or both, with: • a unique title; and • an approval date, a version number, or both.
A-15.2 BBS activities that require an SOP Many survey activities and procedures can benefit from a written SOP; for example: • selecting and training field staff (e.g. training in sampling; see Table A-15.1 for examples); • recruiting staff and eligibility screening; • obtaining consent;
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
15. Standard operating procedures
143
• ensuring confidentiality; • standardizing field site operations (participant flow, hours of operation and appointment system); • collecting data; • monitoring data collection; • developing survey identification numbers (ID); • providing test results to participants; • providing HIV counselling; • providing or referring for treatment; • treating sexually transmitted infections (STI) or vaccinating for hepatitis B; • reimbursing funds; • ensuring safety of participants and staff; • developing personal evacuation plans; • dealing with adverse events; • dealing with repeat participants or imposters; • dealing with survivors of violence; • undertaking laboratory activities: - specimen collection and long-term storage; - specimen transfer to the laboratory; - algorithms for specimen testing ; - laboratory quality-control procedures; - communication of laboratory results to survey site; • managing data (including data entry and data cleaning); and • ending data collection or closing survey site.
Appendix I-27 includes an example SOP on datamanagement for RDS surveys using Questionnaire Development System (QDS).
A-15.3 Sampling-specific SOP Some sampling methods include activities that require specific SOP (see Table A-15.1).
A-15.4 SOP for checklists, logbooks labels and forms Various types of survey-specific instruments (e.g. checklists, forms, logbooks and labels) should have an SOP that provides detailed instructions for making use of the instrument. Some SOPsinclude checklists for staff to use when conducting a BBS; such SOP help staff to identify important survey tasks during BBS implementation. Table A-15.2 gives examples of these instruments and the corresponding SOP.
Table A-15.1 Activities associated with sampling methods that require standard operating procedures Sampling method Conventional cluster sampling, time-location sampling Sampling frame preparation (list of venues) and observation of venues Enumeration, cluster information collection Participant selection, listing of people Participant recruitment Logistics of prepping for each day of collection Respondent-driven sampling Seed selection Coupon distribution and management Peer recruitment training Reimbursement management Managing multiple survey sites (e.g. multiple sets of staff or one set that moves among sites) Welcoming to survey site. Waiting room activities Visiting procedure (e.g. what happens at first and second visits)
144
15. Standard operating procedures
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table A-15.2 Survey-specific instruments that require an SOP Instrument Checklist Logbook
Examples • Survey procedure checklist • Field supervision monitoring checklist • Survey ID log • Appointment log • Specimen collection log • Specimen shipping log • Specimen registration log (at laboratory) • HIV test results log • Log for other biological tests • Supply lists • Unique ID • Specimen label • Participant tracking form • Cluster information sheet
Corresponding SOP • Flow planning of survey procedure • Specimen collection and transfer
Label Form
• Specimen collection and transfer • Flow planning of survey procedure • Sampling
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
15. Standard operating procedures
145
Section
A
16. Training and training methods
This chapter provides guidance on staff training and the information that should be covered in the training materials. Training should familiarize staff with the goals and purpose of biobehavioural surveys (BBS) and provide background information on HIV and treatment.
At the end of the training sessions, staff should be aware of: • steps in the survey implementation process and related tasks; • roles and responsibilities of each staff member involved in the survey; • sampling methodology; • laboratory procedures; • data collection and management; • ethics and confidentiality; and • safety and security procedures.
confidentiality. They should also be trained in professional behaviour and demeanor towards fellow staff and survey participants. Staff should not “date”, have sexual intercourse with or share drugs with participants. Staff should also complete cultural competency and sensitivity training in order to learn how to work with highly stigmatized groups, and should be trained in positive body language and good interpersonal skills. Interviewing The attitude of interviewers can influence participants’ answers in face-to-face interviews, especially in interviews that ask about illegal or stigmatized behaviours. To encourage honest responses, interviewers must be thoroughly trained in open and nonjudgemental questioning techniques and accurate recording of responses. The amount and type of training required will vary depending on who is conducting the interview. For example, if the peers of those in the participant group are selected as interviewers, they may be less likely than professional staff to appear judgemental. Conversely, they may be biased towards recording or coding responses in a way that reflects their own opinions or behaviour. If the questionnaire is electronic and self-administered, staff should be trained in how to use the equipment, how to assist a participant to complete the questionnaire, and how to troubleshoot problems with computer equipment and the electronic instrument interface. Biological specimen collection, handling and transportation Detailed information about how to collect, handle and transport biological specimens to the laboratory should be included in SOP. Survey staff responsible for biological specimen procedures should be thoroughly trained in these procedures. This training should include a practice run (“dry run”) to detect and address issues that may arise.
A-16.1 Training Training should also increase staff comfort in interacting with participants and discussing sensitive topics in the BBS questionnaire. Trainers should instruct on all BBS standard operating procedures (SOP), and should make these and all other training materials (e.g. copies of data instruments, protocols and role-play scenarios) available to staff members to serve as a reference. The cost, required hours and location of training should be estimated during the design phase and included in the budget (see Appendix I-2). In addition to survey procedures and tasks, roles and responsibilities, and SOPs, other important topics include staff behaviour, how to conduct interviews, specimen handling, and safety and security. Professional conduct Staff should conduct survey activities in a manner that adheres to ethical research principles, by respecting and protecting the privacy, confidentiality and autonomy of participants; also they should sign a confidentiality agreement (see example in Appendix I-29). Survey staff should remember that survey participation is voluntary, and should be trained in obtaining informed consent and in the consequences of breaching participant
146
16. Training and training methods
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Safety and security The safety and security of the survey team and participants should be a high priority for investigators. Survey staff should be trained in safety procedures and should practise how to respond to potential security threats, including requests for information or entry to a survey site by police or other authorities. Appendix I-9 has examples of security SOP.
Role play Role playing allows staff to practise scenarios and demonstrate their knowledge. Staff should participate in role-play exercises after completing training sessions on qualitative and quantitative interviews, consent administration, eligibility screening, reception of recruits (for RDS), counselling and testing, and returning results (if applicable). Survey procedure simulation Once staff have been trained in their respective tasks, a practice session of survey activities, using a mock survey in a target population similar to the intended survey population, is a good exercise for understanding the survey event process. The practice should follow the prescribed flow of events and be as realistic as possible. After completing the exercise, the trainer should discuss staff knowledge of roles and performance of activities.
A-16.2 Training methods An interactive practicum-style setting helps with learning and allows trainees to rehearse survey procedures and tasks. Role-playing and simulating (i.e. practising) survey events from beginning to end are two methods for training staff.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
16. Training and training methods
147
B. Survey implementation and quality assurance
1. Stakeholders 2. Planning considerations before starting a biobehavioural survey 3. Ethical consideration in planning and conducting surveys 4. Formative assessment 5. Questionnaire development 6. Biomarker considerations 7. Preparing biological data-collection instruments 8. Eligibility for participation 9. Sampling strategy 10. Sample size requirements 11. Population size estimation methods used with surveys 12. Supplemental studies 13. Data management 14. Staff selection 15. Standard operating procedures 16. Training and training methods
16 19 23 30 34 43 60 63 66 101 107 121 135 140 143 146
A Survey preparation
Data analysis and use
C
1. Data analysis 2. Use and dissemination of survey findings 3. Evaluating survey implementation 4. Data sharing for public use
186 200 206 208
1. Data collection 149 2. Survey monitoring 173 3. Preparing data for analysis 176
B A. Survey preparation
148
Survey implementation & quality assurance B. Survey implementation & quality assurance C. Data analysis and use
Stakeholders
Section
B
1. Data collection
This chapter discusses the implementation and monitoring of nonprobability surveys. As mentioned in Section A (Survey Preparation), nonprobability sampling is the method of choice when more rigorous sampling designs are not possible due to cost, complexity or relative inaccessibility of the target population.
B-1.1 Surveys using nonprobability sampling methods Key terms Community- An organization largely comprised of based members from a given population organization: (e.g. men who have sex with men), often working in advocacy or providing services. Community A person acting as a link to a target liaison: population. May be a representative of a population. Drop-in centre: A safe venue that provides health, educational or counselling services to a particular population (e.g. people who inject drugs). Gatekeeper: A person who controls access to people or a place.
In most cases, investigators choose the sampling method before developing the protocol. Sometimes, however, a probability-based survey may fail to produce a representative sample. Possible reasons include differential recruitment, refusal of venues to provide access, and sparse social networks. The survey then becomes, in effect, a nonprobability-based survey and the data should be analysed as such. Documenting survey procedures and results that prompt the decision to treat the data as nonrepresentative helps investigators to adjust for these conditions and better address biases. It also helps readers to assess the validity of results.
B-1.1.1 Convenience sampling Convenience sampling is perhaps the simplest nonprobability sampling method because individuals are recruited for participation based on the ease of finding them. For example, a survey may consecutively enrol people who inject drugs (PWID) who frequent a community-based organization or drop-in centre. Convenience sampling may take many forms, all of which generally facilitate sampling for the investigator. Although selecting the sample through multiple means (e.g. referral, at an STI clinic or brothel, and through advertisements at bars or clubs) may produce more diversity in the sample, it cannot be regarded as probability-based. Take all approach A very simple version of convenience sampling is the “take all” approach, in which all eligible persons identified are offered enrolment. Examples of this approach include sampling all eligible SW in a brothel, all transgender women (TG) in a bar or all PWID frequenting a drop-in centre. If the number of venues is too large, the aim should be to include as many as possible with
Because the probability of selection for a given participant is unknown in nonprobability surveys, sampling bias poses the largest risk to the representativeness of such surveys. For example, if a survey is designed to sample sex workers (SW) at clinics treating sexually transmitted infections (STIs), a substantially higher HIV and STI prevalence would be expected among this sample compared to a survey that samples SW at venues where SW solicit clients. Although such nonprobability samples do not allow inference to the larger target population, survey investigators should still strive for a varied sample to potentially reduce selection bias (e.g. by sampling from as many STI clinics as possible, and from other venues). Before planning a nonprobability sampling survey, investigators should consider the limits of nonrepresentative survey results and the allocation of time and money to the survey.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
149
as much variety as possible to help diversify the sample. First, all known sampling locations should be listed, then a feasible number selected, and all eligible persons encountered at these venues enrolled.
Snowball sampling differs from RDS in that, in snowball sampling: • there is no limit on the number of coupons issued per person; • the recruiter–recruit relationship does not need to be tracked; • recruiter and recruit do not need to know one another; and • survey staff may make initial contact with potential participants.
B-1.1.2 Purposive sampling Sometimes a survey targets a particular subset of a population such as men who have sex with men (MSM) who are also married to women, or HIV-infected members of the target population. The implementation of a survey using purposive sampling depends on the desired sample characteristics. For example, a survey targeting MSM who are married to women may require the use of peer referral, whereas a survey targeting HIV-infected members of the target population may sample people at an antiretroviral therapy (ART) clinic. Purposive sampling may also be used to select particular attendees at an ART clinic for participation. For example, if investigators find that most people come from only one part of the city, they may preferentially recruit those from other parts so that the sample is more reflective of the whole population.
B-1.1.5 Recruitment strategies Potential survey participants can be recruited from a target population in various ways, including at commercial venues or public spaces, and through service providers, community-based organizations or the internet, as discussed below. Commercial venues and public spaces Sampling at venues and other locations may be preferred if many or most members of the target population frequent these locations. Venues may include bars, clubs, parks, hotels, brothels and selected streets. Usually, the more kinds of locations sampled, the less biased the sample. Survey investigators may need to ask venue owners for permission to sample at the venue. See the section on time-location sampling (Section A-9.6) for discussion of the practicalities of sampling at venues. Service providers and community-based organizations Service providers can help identify target population members. Service providers may include individuals working for STI clinics, community-based organizations that provide services, drop-in centres, nongovernmental organizations (NGOs) or other groups that distribute prevention materials (e.g. condoms, lubricants, needles and syringes). Survey investigators may offer enrolment either to a portion or to all service recipients or community-based organization members or attendees. If feasible, potential participants should be sampled directly at the service delivery site or office of the communitybased organization to facilitate the speed of recruitment. If direct sampling is not possible, service providers and staff of community-based organizations can distribute vouchers or flyers inviting individuals to participate in the survey at another location. Survey investigators may also enlist the help of outreach workers who know where to find or contact members of the target population; these workers may enjoy more trust and have better rapport with the target population than do survey staff. Internet-based surveys Websites of interest to a target population (e.g. gayfriendly websites) can also be used for recruitment. Investigators may place a link on a website that redirects
B-1.1.3 Quota sampling Quota sampling is similar to purposive sampling, except that it aims to sample a given number of people with specific characteristics. This form of sampling may be used to ensure inclusion of a sufficient number of people with characteristics of interest, or to ensure that the sample is more reflective of the whole population. For instance, investigators may choose to sample 100 PWID who know their HIV status and 50 PWID who do not. PWID of known HIV status may be found by collaborating with care providers or social gatekeepers. PWID of unknown status may be sampled (selectively) through chain-referral sampling (described below) or at drop-in centres. Other examples include having quotas for the number of street-based and brothel-based SW, or for SW born domestically and internationally.
B-1.1.4 Chain-referral sampling or snowball sampling Snowball sampling may be a good option when target members know one another. Snowball sampling begins by purposively sampling some initial participants who are then asked to help recruit others. The purposively sampled participants may share with the staff how and where to locate their peers; alternatively, the staff give the participants vouchers and information about the survey and ask them to either recruit or spread awareness of the survey to their peers. Additional details on using vouchers (coupons) for sampling are available in Section A-9.7 on respondent-driven sampling (RDS).
150
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
the user to the survey webpage. Before the actual self-administered interview, participants undergo selfadministered eligibility screening and consent. Some internet surveys also ask participants to refer their peers; for example, by providing an email address that is used to automatically send an invitation for survey participation. Participants’ internet protocol (IP) addresses should be considered as personal identifiers; survey investigators must ensure that the addresses are protected if they are collected or stored, or that participants can remain anonymous.
B-1.1.6 Preparation for sampling Each sampling method and recruitment strategy has unique operational requirements, but the considerations below apply to all methods in some way: • contact gatekeepers and describe the survey objectives and confidentiality. • set expectations for the survey activities with the gatekeeper. Explain (orally or in writing, as appropriate) the requirements for listing individuals at the location; the physical space needed for interviewing, collecting specimens or providing test results; the time needed for each interview; and any testing procedures. • identify a private area appropriate for conducting interviews and collecting biological specimens. • train survey staff on protections for participants, including ensuring that participants understand and sign confidentiality agreements. • schedule participant sampling or interviews.
to do so, they may need to carry backpacks, cooler boxes, flashlights, clipboards and tablet computers with them. If feasible and affordable, a van or bus can provide the mobility needed for sampling at venues, facilitate data and specimen collection, and increase security and privacy. Regardless of the type of datacollection activity, relevant staff require certain items. These include copies of the letter of support from the ministry of health, and institutional review board (IRB) approvals; sufficient paper copies of the data collection forms; sufficient and appropriate supplies for specimen collection and transport; information or brochures about HIV and STI prevention, care, treatment and services; and reimbursements for participants (e.g. money in a cash box, condoms, lubricants and clean injection kits).
B-1.1.8 Monitoring of survey progress Adherence to standard operating procedures (SOPs) is equally important for surveys that use nonprobability sampling and those that use probability sampling. Documenting and monitoring survey progress and sample composition is required to gauge both the extent of various biases and data quality. SOPs should be developed and changes documented. As far as possible, investigators should document how sampling was conducted, including the sampling universe (i.e. the number and location of sampling sites, and the proportion of each included in the survey). They should also record data that inform the potential denominator (i.e. the probable size of the target population accessible through the planned sampling design). During the sampling process, staff should document the number of people screened for eligibility, the number who are eligible, the number who consent, and the reasons for refusal. They should also document the number of participants who agree to specimen collection and the number of people who receive their results, as well as reasons for refusal. If vouchers or coupons are used, the numbers issued and redeemed should be tracked. In addition to documentation of the actual sampling procedures, attention to data and specimen quality is important. Staff should note the potential for non-sampling error, such as reporting bias in the interview process or testing quality in a mobile or outdoor environment.
B-1.1.7 Enrolment and data and specimen collection Every effort should be made to ensure data quality and minimize errors. Measures to ensure security for staff and participants should also be considered. Most surveys, including those that employ nonprobability sampling and those with more rigorous sampling designs (see Section B-1.2) use the same procedures (e.g. consent, confidentiality measures, interviewing, specimen collection and testing). Investigators should follow the survey protocol and flow chart of survey activities. Enrolment may take place at the location of sampling (e.g. at a venue or in a van, a room in a clinic or a park bench). Other surveys may refer or transport potential survey participants to a fixed survey office with better security. Sampling and enrolment may need to occur in the evening or late at night when the target population may be more accessible. Depending on where data and specimen collection take place, survey staff should be prepared to administer surveys and collect and test specimens in the field;
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
151
B-1.2 Conventional cluster sampling This section provides guidance on using conventional cluster sampling (CCS). Procedures explained here follow the activities described in Chapter A-9 about creating the sampling frame and selecting clusters. This section also describes preparations for participant sampling, special confidentiality and anonymity procedures, field staff arrival and set-up, methods for sampling individuals from a list, recruitment techniques and other survey procedures. In CCS, the target population is associated with clusters in a “fixed” manner. That is, individuals are associated with only one cluster location, the same individuals can reliably be found at the same cluster regardless of when it is visited, and a complete list of individuals who are associated with a cluster can be obtained or created. Appendix I-20 provides additional guidance on mapping cluster locations and creating lists of potential participants. Examples of survey situations in which CCS might be appropriate include SW working in brothels or homes, detainees in prison, military personnel in barracks, and PWID in treatment facilities.
Key terms Actual The actual number of individuals who measure of meet the approach criteria during a size: sampling event. Approach Criteria for who, among all the venue criteria: attendees, staff should approach for participation in the survey. Community A person acting as a link to a target liaison: population. May be a representative of a population. Counter: Device used by enumerator to count people who cross into the recruitment area; or a person who conducts the enumeration. Enumeration: Counting all people in a recruitment area. Gatekeeper: A person who controls access to people or a place. Intercept: Approaching or engaging a potential participant with the expectation of screening the person for survey eligibility. Recruitment Location where recruiters approach area: potential participants to invite them to participate in the survey. Sampling The period in which field staff are in event: a venue recruiting and interviewing respondents.
B-1.2.1 Preparation for sampling The steps listed below take place during the survey planning phase after the cluster locations have been selected from the sampling frame. These steps relate to the logistics of participant sampling at each selected cluster location. • contact a gatekeeper at each selected cluster location and describe the survey objectives. • visit each selected cluster location to build rapport, understand potential field issues, answer questions and address concerns. • observe the physical layout of the location and \ work with the gatekeeper to identify a private area appropriate for conducting interviews and collecting biological specimens. • discuss confidentiality with key people in the selected clusters. • set expectations for the survey activities with the gatekeeper. Explain (orally or in writing, as appropriate) the requirements for listing individuals at the cluster location; the physical space needed for interviewing, collecting specimens or providing test results; the time needed for each interview; and any testing procedures. • request a complete list of eligible individuals who are connected to the cluster location. If the plan is to
recruit a stratified sample of participants (e.g. by gender or age group), request a list that includes the information needed to stratify the sample. • schedule participant sampling date. Special confidentiality considerations Several special confidentiality concerns should be considered with CCS. First, gatekeepers do not have the right to see information collected from survey participants. Second, the information on the list of eligible individuals should be specific enough that any one individual can be identified among all individuals in the cluster. To protect the confidentiality of individuals on the list, investigators must follow these procedures: • after sampling is complete (i.e. after all selected individuals in the cluster have been asked to participate), destroy the list. • gatekeepers may try to claim the right to see the information collected from survey participants, including HIV test results. Human subjects protections dictate that this information not be shared. If a
152
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
gatekeeper insists on access to the survey data in exchange for sharing a list of eligible individuals, do not include this cluster location in the survey. • train survey staff on protections for participants and ensure they understand and sign confidentiality agreements. The following subsections describe set-up, screening and interviewing, and closeout of sampling events via specific steps that should be taken after the team arrives at the selected cluster location. Preparing for sampling events Step 1: Meet with gatekeeper Ensure that the gatekeeper knows the purpose of the survey and the activities that will be taking place. If a community liaison is part of the team, ensure that person is part of this meeting. Step 2: Obtain or create a list of individuals in the cluster If a list of individuals associated with the cluster location is not available, ask the gatekeeper to create one. The investigator or other responsible supervisor can support the gatekeeper and other key informants at the cluster location to create the list. Characteristics of a good list of individuals in a cluster include: • each individual associated with the cluster is on the list. • no individual is listed more than once. • only eligible individuals are listed.1 • each individual can be identified based on the information on the list. • if stratification is being used at this stage, the elements needed to stratify the sample should be included on the list (e.g. if stratifying by age, the list should specify the age of each potential participant). • the list includes space to specify which selected individuals are absent or unavailable on the day the survey takes place. Setting up at sampling events Step 1: Set up the space for conducting interviews and HIV testing The physical space for conducting interviews and HIV testing and counselling should be out of view and earshot of others. Control the space to prevent onlookers or others from invading the privacy of participants.
Random selection from a list of participants Participant selection should be a quick and simple process. The objective is to give an equal chance of selection to all of the eligible individuals on the list; that is, to randomly select people within a cluster. These are two of the many ways to randomly select individuals: • Systematic random sampling: Divide the total number of people on the list (N) by the number needed to participate (n). Select each n/Nth person on the list. Note: It is best practice to roll a dice or to use some other random-number generator to identify a starting point on the list. For example, if a rolled dice shows a “4”, start counting every n/Nth individual with the fourth person on the list. • Random sampling: Create N (the total number of people on the list) pieces of paper with sequential numbers from 1 to N and put them in a container. Without looking, select papers until the number needed to participate (n) is selected. Match the number on the paper to the ordered number of people on the list. For example, if the number “5” is chosen, recruit the fifth person on the list. This method is more appropriate when a smaller number of eligible participants is needed.
Step 2: Select individuals for participation from the list Select the required number of individuals (including those who are eligible but not present at the time of selection) according to the procedures laid out in the protocol (see Chapter A-9). To avoid bias, sampling is done solely by the supervisor, not by the community liaison. Note: After the initial sample of individuals has been approached, it may be necessary to re-sample individuals if the number of refusals is higher than expected – this step is only necessary when the number of refusals prevents the team from enrolling the target number of participants.
1
If eligibility includes behaviours, the method for determining eligibility for individuals who have a fixed association with a site should be described in the protocol and operations manual. If it is not possible to obtain a list of only those individuals who are eligible, the survey instrument should include an eligibility screener to determine the individuals who may be interviewed.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
153
Example of recruitment talking points Thank you for your interest in our survey. I would like to tell you about the survey: • We will describe the survey procedures in detail before we collect any information from you. • Being part of the survey is voluntary and confidential. • The survey includes an interview and HIV testing. The interview will take up to 60 minutes to complete. • We will ask you questions about your background, and about behaviours surrounding sex and substances you may use. • We would like to take a blood sample for an HIV test. • There are minimal risks to being part of the survey. • By being part of the survey, you will help us to plan better HIV services for your community. • You will receive condoms, lubricants and referrals to services, if you wish. • We will answer any questions you have about the survey before you start.
Sampling individuals from a list Investigators should select from the complete list of individuals associated with the cluster location. There may be situations in which a number of eligible individuals on the list are not available to participate when the survey is conducted. In this case, do the following: • Schedule an alternative time to survey those who are not available. This alternative time should be as close as possible to the initial survey date. Staff should make at least three attempts to complete the survey with those who are unavailable before seeking a replacement participant. • Document selected participants who are not available to participate as “nonresponse”. • Document the “measure of size” or the total number of eligible individuals associated with the cluster.
The field team may have fewer interviewers than the number of selected participants in a cluster location. For example, the team may have two interviewers but need to interview five individuals from a cluster. In this case, the team should first approach two of the potential participants to conduct the interviews, and when an interviewer is free, then approach the third participant. This technique helps to ensure that participants do not have to wait for an interview to begin. If selected potential participants are not present, schedule alternative times to conduct the interviews with them. Step 2: Escort individual to interview area Escort potential participants to the physical location where interviews and HIV testing will take place. Step 3: Assign a survey identification number The specific procedures for assigning survey identification numbers (ID) and deciding which staff member is responsible for this step be included in an SOP. Survey ID assignment is commonly done by one staff member to avoid duplication of IDs. If more than one person is assigning survey ID, allocate different survey ID numbers to each relevant staff member. Step 4: Screen for eligibility Ensure the selected potential participant is eligible by screening the person using a standardized tool. Document eligibility for each individual approached. If someone is ineligible, inform the person and offer thanks for their time. Note: Supervisors should monitor the number of ineligible individuals logged by each interviewer for quality improvement purposes. Interviewers who log an unexpectedly high or low number of ineligible persons may require retraining or mentoring to ensure proper screening procedures.
Recruiting, screening, consenting, interviewing, counselling and testing Step 1: Approach and recruit selected individuals This step may be combined with or designed to complement Step 5 in below. The supervisor and, if appropriate, the community liaison speak with each person selected from the list. Rather than talking to all potential participants as a group, aim for one-on-one conversations to ensure an individual is free to decide about participation without pressure from others. Build rapport with the potential participant; discuss the survey, including its purpose and benefits; and explain the random selection process and confidentiality protections. During this step, it may be helpful to use recruitment “talking points” to briefly describe survey procedures, risks and benefits, and other elements of the consent process.
154
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Step 5: Obtain informed consent Obtain informed consent according to procedures described in the SOP and document the outcome of the consent process (see Chapter A-3 for more information). Step 6: Interview the participant Administer a face-to-face or self-administered interview using an electronic questionnaire. If both methods are offered, document the interview mode used. Step 7: Conduct counselling and testing The interviewer escorts the participant to the area designated for counselling, and specimen collection and testing. Provide pre-test counselling and collect biological specimens. Conduct post-test counselling according to the protocol and SOP. Ensure that the result of each test in the testing algorithm is recorded. Return test results if conducting rapid testing, or schedule a date to return test results. Step 8: Provide referrals as appropriate Participants newly found to be HIV positive and those who are not in care for HIV infection must be referred for medical care and treatment as described in the SOP. All individuals testing negative for HIV should be given information about where they can be tested in the future. Participants in need of other health or social services should also receive referrals to the appropriate service providers in the community. Step 9: Provide treatment for treatable STIs Conduct this step if it is part of the protocol, according to the SOP for STI treatment. Step 10: Provide prevention materials Provide participants with HIV-infection prevention materials and supplies (e.g. informational pamphlets, condoms, lubricant and clean injection kit). Step 11: Provide compensation as appropriate Provide participants with compensation if this is part of the protocol and SOP. Closeout Debriefing At the end of data collection at the cluster location, a debriefing for field staff is useful in order to discuss any issues or problems that occurred. This debriefing can be either a meeting convened with all field staff or a series of one-on-one conversations. Possible questions for discussion include: • In general, how did data collection go at the selected cluster? • Were there any cluster-related issues that affected survey activities?
• Were there specific concerns related to selection of the participants, consent process, and so on? • Were there any barriers to recruitment? What strategies were successful in overcoming barriers? • Were there any unusual or adverse events (e.g. a participant ended the survey early, or a participant initially consented to an HIV test but changed their mind later)? • Were there any problems with the electronic devices? • Were there any possible errors within the survey data? • Were there any problems with specimen collection or test kits? If necessary, the field supervisor can record this information for future reference. The field supervisor should notify the principal investigator or other staff if an SOP or the protocol requires amending. Field notes The field supervisor should document any problems, barriers or challenges at the cluster location where the data were collected. This information is important when investigators are reviewing recruitment statistics and can help to adjust field operations in future clusters, if necessary. Collect and review forms and logs At the end of the data-collection activities at a selected cluster, the field supervisor should collect all materials used by field staff, review them for accuracy, make corrections as necessary, and tabulate recruitment outcomes (e.g. number of individuals selected, number approached, number accepting participation, number of refusals and reasons for refusal).
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
155
B-1.3 Implementation of time-location sampling This section describes the implementation of surveys using time-location sampling (TLS), including the core activities of preparation for the sampling event: event setup, enumeration, sampling, enrolment and other survey procedures. As described in Section A-9.6, TLS is used to survey populations who come and go freely from physical venues (locations) but may not visit the venues regularly. Although it would be possible to create a list of all people at a venue, the list would vary depending on the day and time the location is visited by the survey team. In TLS, investigators create a sampling frame comprising all possible venues and times that the target population may visit. Investigators then sample from a list of venue– day–time combinations to obtain a representative sample of the people attending venues. Survey staff then visit the venues to collect data during “sampling events”. Individuals present at a venue during a sampling event are eligible to be sampled in the survey. Most of the people at the venue during the sampling event should be part of the survey’s target population. TLS may be appropriate for surveying SW waiting for clients at bars, nightclubs, street locations or hotels; MSM or TG attending bars, clubs or bathhouses; and PWID attending drug-buying or drug-using venues or other similar open spaces. include bar or brothel owners, pimps, nongovernmental organization (NGO) staff, or drug dealers. In some cases, there may be not be an obvious gatekeeper. Together with the community liaison, the field team should meet with the venue gatekeeper to inform them about the objectives of the survey and the sampling and survey procedures. A community liaison is a person acting as a representative of or a link to a population to facilitate survey planning, implementation, interpretation and dissemination of results. For a park or street-corner venue with no obvious gatekeeper, the field team should work with the community liaison to identify any key people to meet. Because the gatekeeper may differ each time the field team visits a venue, be prepared at every meeting at the venue to meet the gatekeeper present and do the following: • Explain the goal of the survey, including that it will benefit the target population and that their privacy and confidentiality will be protected. • Set expectations for the survey activities. • Explain requirements for enumerating and intercepting individuals at the venue, including that data collection may occur inside the venue, and the need for privacy for interviewing and testing. • Emphasize that the survey team will do its work with minimal possible interference with “normal business” at the venue. Remind the gatekeeper that all survey data are confidential and cannot be shared with the gatekeeper or venue staff, if they exist. Step 2: Prepare materials for the sampling event Ensure that all supplies are taken to the sampling event. Below is a non-exhaustive list of important items to be taken to every sampling event: • letters of support from the ministry of health and institutional review board (IRB) approvals: take at least two copies of each document to show potential participants, gatekeepers, or authorities, if needed. • information and brochures about prevention, care, treatment and services for HIV and sexually transmitted infections (STIs); compensation for participants (e.g. money in a cash box), and condoms, lubricants, clean injection kits. • copy of the venue observation form for reference (see Appendix I-20). • data-collection instruments and forms: approach forms, consent forms, questionnaires, test-result forms,
B-1.3.1 Survey staff For each sampling event, the survey team should include a combination of the following staff positions: supervisor, counter, recruiter, interviewer, counsellor and laboratory technician. Ways in which these positions can be combined are described below. For details regarding the staff involved in a TLS survey, see Chapter A-14.
B-1.3.2 Preparation for the sampling event The following steps take place after the venue day–times have been selected (see Section A-9.6). The steps relate to the logistics of participant sampling and recruitment at each site. Step 1: Meet with the gatekeeper of each selected venue and, if appropriate, the community liaison Gatekeepers should be engaged whenever possible to facilitate access to the target population and survey planning. As survey activities may interfere with the normal routine of a venue, a gatekeeper’s acceptance of or support for the survey may be needed. Gatekeepers may
156
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
referral forms, sampling event forms, incident report forms, and counting device. Bring paper backups of all electronic documents. • electronic hardware (if applicable): tablets, cell phones, internet modems. • laboratory supplies: specimen-collection devices (e.g. needles and syringes, lancets, tourniquets, tubes, sharps container, and swabs), testing devices (i.e. rapid tests), equipment for specimen storage (e.g., cooler box) and transport. • other supplies: tables, chairs, tents, lights, pencils and pens, paper, folders, waste bin. Step 3: Identify the recruitment area and interview and testing area(s), and setup area(s) for conducting interviews and HIV testing A good recruitment area should have the following characteristics: • clearly defined. The recruitment area should be a well-defined space where potential participants who enter or cross the space are approached for recruitment. Another option is a “moving line”. For indoor venues, such as brothels, bars and some shooting galleries, the recruitment area may be inside or outside the venue. • specific to the venue. The area should be defined to ensure that only people attending the venue are recruited to participate in the survey. • easy to manage. The recruitment area should be selected to facilitate the survey team’s ability to manage recruitment. In larger venues, avoid areas where venue attendees may cross several times, such as near the toilet or bar. These locations can make it difficult to ensure individuals are counted only once. • of an appropriate size. The recruitment area should be large enough to ensure that the target sample size for each sampling event is reached, but not so large that the survey team becomes overwhelmed. - using a small recruitment area for large venues with many people can be helpful. - at small venues or those with few people, it helps to use a large recruitment area to recruit a larger proportion of venue attendees. In such cases the entire venue can act as the recruitment area. - when there are many people in the recruitment area (e.g. a busy street corner or the entrance to a bar), it is useful to restrict recruitment to only those individuals who enter or cross the recruitment area from a single direction. - during an event, the size of the recruitment area can be adjusted to match changing numbers of attendees. Be sure that doing so does not exclude some people from the opportunity to be sampled. • accessible to the rest of the survey implementers. The location of the recruitment area should allow the field supervisor (or other survey staff) to effectively direct
recruitment. It should also be located close to where interviewing and testing are conducted. • private. The space for conducting interviews and HIV testing and counselling should be out of view and hearing of others at the venue. The space around this area should be controlled to prevent others from invading the privacy of participants, and to limit distractions. Dividers, curtains or tents can be used to create a temporary space for these activities. Venues without suitable space for these activities should not be included in the sampling frame. Step 4: Determine the measure of size for the event (counting) As described in Section A-9.6, sampling the target population during a sampling event is similar to secondstage cluster sampling. By this point, investigators have identified the sampling events (similar to first-stage cluster sampling) and now are selecting individuals from each cluster or sampling event. In cluster-based sampling, investigators generally know the number of eligible people in each cluster, but in TLS they do not know beforehand. Therefore, investigators must determine the measure of size (MoS) for an event, that is, the number of eligible people at a venue during a sampling event. The MoS is used to calculate the probability of a person being approached and recruited into the survey. During survey planning, investigators determined the number of venue–day–times from which they can sample. Now, they count the number of people to sample from at a specific venue–day–time. During each sampling event, a member of the survey staff should be the counter and count venue attendees with a counting device. The same staff member should be the counter for an entire recruitment event to ensure accurate tracking of venue attendees and to avoid counting individuals more than once. There are many ways to approach counting. One of the most common methods for determining the MoS is to count the number of people at a venue before the sampling event begins, and then count the number that enter the venue during the event. Whichever method is chosen, survey staff should implement it accurately and consistently at all venues in the sampling frame. Step 5: Hold pre-event staff meeting Survey teams should arrive at a venue before it becomes crowded in order to set up for the sampling event. This strategy facilitates the pre-event count, planning and set up for the event. Before the sampling event begins, hold a meeting to: • discuss roles and responsibilities; • distribute supplies to team members; • review survey ID, venue code and event number information;
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
157
• identify the recruitment area; • identify the counting area; • identify the interviewing and HIV testing areas; • discuss observations about the venue and venue attendees; • build enthusiasm and raise the energy level of the staff; • discuss potential safety concerns; and • conduct other activities as appropriate.
Counting in a venue with an entrance Two counts are obtained for recruitment events conducted at venues with an entrance: the pre-event count and the entry count. For the pre-event count, the counter counts all members of the target population who are already inside the venue immediately before the sampling event begins. For the entry count, the counter counts all members of the target population who enter the venue for the first time during the recruitment event. Together, these two counts represent the number of target population members who attended the venue during the sampling event (the MoS). The two counts should be recorded separately and totaled at the end of the event. Pre-event count The pre-event count inside the venue should be conducted even if recruitment, interviewing or HIV testing occurs outside the venue. The counter must personally conduct all pre-event counts; pre-event counts should not be obtained from venue gatekeepers, managers or staff. Survey teams should arrive at a venue before it becomes crowded in order to set up for the sampling event before many people arrive. This approach makes the pre-event count and planning and set up for the event easier. To obtain the pre-event count, the counter should start counting target population members at the point farthest away from the primary entrance that most attendees use to enter the venue. Starting from this farthest point, the counter should count as they move towards the primary entrance, ending the count at the primary entrance. For example, the counter could begin in the back of the bar and then count while walking towards the doorway at the front of the bar (see Figure B-1.1). By starting at the farthest point from the primary entrance, the counter will be able to capture any individuals who enter the venue during the pre-event count. This technique is useful even in smaller one-room venues. The following examples illustrate how to conduct the count, starting with smaller venues (see Figure B-1.1), venues with obstacles (see Figure B-1.2), and progressing to venues with multiple rooms (see Figure B-1.3).
B-1.3.3 Counting Where to count The location of counting, or counting area, and the method of counting depend on whether a venue has a dedicated entrance (e.g. a doorway, gate or similar structural entry); counting is conducted differently if the venue does not have a dedicated entrance (as discussed below). People the counter should not count include: • Individuals who do not meet the approach criteria. The approach criteria describe who, among all the venue attendees, staff should approach for participation in the survey. These criteria are most useful in venues where not all those present are part of the target population. For example, if a survey’s eligibility criteria include MSM aged 15 years and over, males who look much younger than 15 years should not be approached. • Venue employees or others workers who are not “attending” the venue, and thus do not meet the approach criteria. For example, the male bartender at an MSM bar should not be recruited into an MSM survey, even if he is an MSM. This criterion does not apply for brothels or other venues where the employees are the target population. Other people who may have to enter the recruiting area for work include police officers and street cleaners. • Individuals who have already been counted. • Self-referrals, who are people who deliberately enter or cross the counting area trying to enrol in the survey or obtain the incentive. They may have learned about the survey from another venue attendee, or they may be attracted by the activity generated by the survey team. Some individuals will ask if they can participate. They should not be allowed to participate nor should they be counted. Figure B-1.1 Venues with one room a) The counter should start at the point farthest away from the primary entrance and count as they move toward the primary entrance. Primary Entrance
b) To count all the individuals in a crowded venue, it may be necessary to move crisscross through the crowd
Primary Entrance
158
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure B-1.2 Venues with obstacles Flow of Attendees: BAR Direction of counting: If there is an obstacle in a room, such as a large bar, that prevents the counter from moving directly across the room toward the primary entrance, the counter should count while moving around the obstacle and against the flow of venue attendees. By moving against the flow of venue attendees, the counter will minimize the number of individuals who are missed and not counted.2
Source: CDC 2014 (2)
If there are multiple rooms in a venue, the counter should begin counting in the room farthest away from the room with the primary entrance (see Figure B-1.3). If there are multiple levels and the primary entrance is on the lowest level, the counter should begin counting on the level farthest away from the level with the primary entrance and then count through to the entrance.
Counting in a venue with a complex layout When the rooms or levels in a venue are not arranged in a sequence that ends with the room with the primary entrance, the counter should decide where to start counting based on the flow of people in and out of each room, or on and off each level. The counter should begin counting in the room or on the level with the lowest flow of people and move to the area with the highest flow, and then end in the room with the primary entrance (see Figures B-1.4 and B-1.5). If there is no clear pattern to the flow in the venue, the counter should decide where to start counting based on the number of the target population in each room or on each level. People in the room with the primary entrance should always be counted last.
Figure B-1.3 Venues with multiple rooms 2nd 1st If there are multiple rooms in a venue, the counter should begin counting in the room farthest away from the room with the primary entrance and end counting in the room with the primary entrance.
Primary Entrance
Figure B-1.4 Venues with multiple rooms not in sequence Room with lowest flow (or smallest number) 1st 3rd 2nd
Primary Entrance
Room with highest flow (or largest number)
2
Source; CDC 2014: DHAP NHBS Operations Manual.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
159
Figure B-1.5 Venues with multiple floors not in sequence Floor with lowest flow (or smallest number) 1st 5th Floor with highest flow (or largest number) 3rd
2nd
Primary Entrance
4th
4th
Basement
Ground (First) Floor
Second Floor
Figure B-1.6 Crowded venues that have equally distributed attendees If venue is so crowded that it would be difficult to obtain an accurate count, the counter may divide the venue into equal sections, count all the individuals in one of the sections, and then multiply this count by the number of sections to estimate the total count. When the individuals attending the venue are equally distributed throughout the venue, the counter just needs to ensure that the sections selected are the same size. In this example, the counter divides the venue into quadrants, counts the individuals in one quadrant, and then multiplies this count by four to estimate the total count for the venue.
Primary Entrance
Primary Entrance
Figure B-1.7 Crowded venues that do not have equally distributed attendees STAGE STAGE If the counter decides to divide a crowded venue into sections, but the individuals attending the venue are not equally distributed throughout the venue, the counter must take this into consideration. In addition to ensuring that the sections the venue is divided into are the same size, the counter must also ensure that the distribution of individuals is the same in each section. The counter therefore divides the venue into four sections that are perpendicular to the stage to ensure that the distribution of individuals is the same in each section. The counter counts the individuals in one of the sections and then multiplies this count by four to estimate the total count for the venue. Note that the counter moves toward the primary entrance along the line that divides the section. This allows the counter to keep track of the boundary of the section and makes it easier for them to count because they have to count only the individuals between themselves and the wall.
Primary Entrance
Primary Entrance
Counting in a venue with a large number of people In some cases, the number of people attending a venue may be so large at the start of a sampling event that it is difficult to obtain the pre-event count. If this occurs, the counter may divide the venue into equal sections, count all the members of the target population in one of the sections, and then multiply this count by the number
of sections to estimate the total pre-event count (see Figure B-1.6 and Figure B-1.7). For example, the counter could divide a busy bar into quadrants, count the people in one quadrant, and then multiply this count by four to estimate the pre-event count for the entire bar. This method should only be used when the entire venue is extremely crowded and it would otherwise be impossible to obtain the pre-event count.
160
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Venues with low attendance or few target population members Some venues may have very low attendance and hence few members of the target population enter the recruitment area. In this case, an alternative approach plan would be to consecutively approach people seated at a bar, standing along a wall, or already in the recruitment area. If survey staff establish an alternative approach plan for a venue, all staff must follow the plan and not arbitrarily recruit people.
Entry count The entry count is the number of potentially eligible members of the target population who enter the venue during the sampling event. This count should be obtained at the primary entrance to the venue. The same counter who obtained the pre-event count should obtain the entry count. When the survey staff are ready to begin recruitment, the counter should clear the tally counter to zero and start counting target population members who enter the venue. The counter should stop counting when the last person is approached for recruitment or the field supervisor decides to end the event, usually at the end of the venue–day–time period. Counting should be uninterrupted between these starting and ending points. The counter should continue to count even when all the interviewers are busy with participants. Some venues may have multiple entrances. If the counter cannot monitor multiple entrances simultaneously, the counter should only count people entering at the primary entrance. The field supervisor should document if a venue has additional entrances where people were not counted. Counting in a venue without an entrance Venues without a dedicated entrance such as a doorway or gate include street blocks or parks. If these venues are small enough to conduct a pre-event count, the counting should follow the same steps as those for venues with an entrance. Many venues without an entrance are so big that it may be difficult to count everyone inside before the sampling event begins. In this case, only collect the entry count. The counter should count individuals entering the recruitment area defined by the survey team. The methods for counting at venues without a dedicated entrance are similar to those with an entrance. When the survey staff are ready to begin recruitment, the counter should start counting members of the target population who enter the recruitment area. The counter should
not count those who are already in the recruitment area when counting begins. If a person is in the recruitment area when counting begins, leaves the recruitment area, and re-enters later during the recruitment event, the person should be counted upon re-entry. The field supervisor should also note that the pre-event count was not obtained because the venue did not have a dedicated entrance. If there are many places for individuals to enter the venue, count people who enter only from a specific place. The counter should stop counting when the last person is approached for recruitment or, if enrolment is slow, the field supervisor decides to end the event. Counting should be uninterrupted between these starting and ending points. The field supervisor should document when a venue has additional entrances where people were not counted. Another method is to divide the venue into equal segments of manageable size and conduct the pre-event and entry counts for only one segment. At the end of the event, the survey team should multiply the counts by the number of segments to determine the MoS for the entire venue. For example, if a venue is split into four segments, the MoS of one segment should be multiplied by four to produce the count for the entire venue. This is similar to what is shown in Figure B-1.6. This approach is only applicable if all segments would have approximately equal counts. If one segment of a park has many more people than another, multiplying the count of a given segment would not produce a valid count for the entire venue.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
161
Approach criteria and eligibility screening Not all venue attendees are part of the target population. The approach criteria describe who the staff should approach to determine whether the person is eligible to participate. Approach criteria should be broad enough to not exclude eligible individuals. However, venues may have multiple populations present. Immediately asking a person if they sell sex may cause problems for the survey team if the person does not sell sex and is unaware of the survey. Thus recruiters should not directly reveal the survey’s target population. For example, in a survey of female sex workers (FSW), when approaching a potential participant, the recruiter should say that she is doing a survey of women who go to the venue. Then the recruiter can tell the person how she will be compensated for her time and ask if she would be interested in joining. To determine eligibility while also being discreet in case the person is not an FSW, the eligibility screener should include questions that ascertain eligibility as well as a few questions that hide the purpose of the survey. Such questions might include: • are you married? • do you have children? • where do you meet men? • how do you earn money? • have you had sex in the past 6 months?
B-1.3.4 Recruitment and survey steps Start the entry count after the set-up and the pre-event count have been completed and all field staff are in place and ready to start. Then begin recruiting, interviewing, counselling, collecting and testing specimens as outlined below. Steps 7–9 below may occur in a different order. Provided recruiters and interviewers are available, recruiters should approach venue attendees consecutively. That is, once the first attendee has been counted and approached, the second attendee should be counted and approached. Recruitment continues while interviewers are available. If there are no interviewers available, counting continues but approaching and recruiting temporarily stop until an interviewer becomes available, at which point they resume. The counter must continue counting during this time to enable investigators to calculate the selection probability of an individual. Step 1: Approach and recruit When recruitment begins, the counter or field supervisor directs a recruiter to approach and recruit a specific venue attendee who meets the approach criteria. This process is called the “approach” and can be done by either a recruiter or an interviewer who carries an intercept form (Figure B-1.8). Each recruiter should carry one form. To ensure that recruiters do not preferentially select friends or a certain type of person for participation, the counter or supervisor can identify who should be approached. Recruiters may select individuals themselves provided all have an equal chance of being approached. The counter can direct recruiters at venues without an entrance. At venues with an entrance, the counter can direct recruiters only when the recruitment area is near to the counter, that is, the primary entrance or near the primary entrance the primary entrance to the venue.
The approach begins with the recruiter engaging the targeted venue attendee in conversation. If the person ignores the recruiter, the recruiter should mark a tick next to “No” on Line 1 of the intercept form (see Figure B-1.8 and text box). If the person is willing to talk, the recruiter should briefly describe the survey to the prospective participant and determine whether they have previously participated in this survey round. If the person has already participated, the recruiter should offer thanks for the person’s time and mark a tick next to “Yes” on Line 2 of the form before returning to the counter to receive a new person to approach. If the person has not participated in the current survey round, the recruiter should ask to screen the person for eligibility. If the person is not willing to be screened for eligibility, mark a tick next to “No” on Line 3 of the form. If the person is willing to be screened for eligibility, mark a tick next to “Yes” on Line 4 of the form. Each person intercepted should be represented by only one tick. The number of ticks on Line 4 should equal the number of eligibility screening records conducted by the recruiter during the sample event. The number of ticks on Line 5 should equal the sum of all the ticks on the form, which in turn equals the number of people the recruiter approached during the sampling event. See Appendix I-19 for an example of the TLS intercept form and how to summarize the information from these forms after the sampling event is completed. Figure B-1.8 shows an example of an approach form. Each recruiter should use one approach form for a given sampling event. The form should be used to record recruitment efforts for all individuals approached.
162
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure B-1.8 Intercept form Event Data Date: Event #: Venue ID: Recruiter ID:
Line
Intercept data Accepted Intercept?
Sum
A
No: Previous Participant?
B
Yes: Recruited to be Screened for Eligibility?
C D E
No: Yes: Total:
If five refusals to approach or to participate occur in succession, the team should stop counting, analyse the situation together, determine the cause of the problem, and develop a plan to resolve the issue. The field supervisor should also evaluate the recruiter’s performance and provide recommendations. Potential solutions include seeking advice and assistance from the gatekeeper and trying different recruitment approaches (e.g. more aggressive, less aggressive, using a recruiter of a different gender or ethnicity). After a solution has been identified, counting and recruitment can resume. If the event remains unsuccessful, the team may consider terminating the event. However, all possible solutions to the problem should be tried first. The number of interviews conducted during each sampling event will depend primarily on attendance at the venue and the number of field staff. The number of interviews conducted during a sampling event at a high-attendance venue will usually be much higher than the number of interviews conducted at a low-attendance venue. To prevent the total sample from being dominated by attendees at a few well-attended venues, investigators should set a maximum number of persons that can be interviewed at one event. Counting and recruitment should continue until the end of the sampling event unless this maximum is reached.
If the count remains low by the end of the sampling event, determine the reason for the low count (e.g. TG do not go to the bar that night or there is a temporary curfew due to civil unrest). If multiple venues have low counts, consider alternative venues and sampling with replacement. The results from the intercept form are used to calculate the level of response. The MoS is used to weight the data to produce population estimates of results. These topics are discussed in more detail in Chapter C-1.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
163
Sample completed approach form In this example, 17 people were approached at the sampling event (equals the total number of ticks) by Recruiter 4. Three people refused to speak to the recruiter and one person had already participated in the survey. Two people refused to be screened for eligibility and 11 agreed to be screened.
Approach form Event data Date Jan 14, 2017 Line Participant data Accepted approach 1 No Previous participant 2 Yes Recruited to be screened for eligibility 3 4 5 No Yes TOTAL PEOPLE ││ ││││ 17 ││││ │ │ │││ Event # 3 Venue ID B001 Recruiter ID 4
Step 2: Assign a survey ID All those who agree to be screened for eligibility (those on Line 4 of the approach form) should be assigned a preprinted survey ID. Step 3: Screen for eligibility Screen participants in a private area (see Section A-8 for more information about eligibility criteria). Either the recruiter or the interviewer can screen for eligibility. If the prospective participant is not eligible for participation, the recruiter should offer thanks for the person’s time, escort the person from the survey area, and return to the counter to be assigned another person to approach. Step 4: Obtain informed consent After a person has been identified as eligible, the interviewer should obtain informed consent according to the protocol and SOP. Step 5: Interview the participant The interviewer should interview the participant or supervise a self-administered interview.
Step 6: Counselling and testing Following the interview, the interviewer should take the participant to the laboratory area to receive pre-test counselling and provide biological specimens. Perform biological testing and relevant counselling according to the protocol, local testing policies and SOP. Ensure that the result of every test (assay) in the algorithm is recorded. Step 7: Provide reimbursement and prevention material Field staff can use the time while rapid tests are running to compensate participants and provide them with material on HIV and STI prevention (e.g. informational pamphlets, condoms, lubricants, clean injection kits). Step 8: Provide test results, post-test counselling, and treatment If rapid tests are used to test for STI, participants who test positive for treatable STI should be provided with treatment, where possible. If some tests are performed or confirmed off-site, follow protocols to ensure the participant receives test results and treatment later.
164
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Step 9: Provide referrals as appropriate Refer newly diagnosed and out-of-care participants infected with HIV for care and treatment. All individuals testing negative for HIV should be given information about where they can be tested in the future. Individuals in need of other health or social services should also receive referrals to the appropriate service providers in the community. Figure B-1.9 illustrates these key steps in survey implementation.
can be either a meeting convened with all field staff or one-on-one conversations. Some potential questions for discussion: • in general, how did the sampling event go? • were there any site-related issues that affected project activities? • were there any barriers to recruitment? What strategies were successful? • were there any unusual events (e.g. a participant ended the survey early, or a participant initially consented to HIV testing but had a change of mind)? • were there any problems with the electronic devices? • were there any errors with the survey data? • were there any problems with HIV test specimen collection or test kits? • were there difficulties with HIV testing and counselling? If necessary, the field supervisor can record this information in the forms in Appendix I-20, and notify the principal investigator or other staff if the SOP or protocol need to be amended. Sampling event notes The field supervisor should record notes about the sampling event. These notes document any barriers to survey activities and changes at the venue during the sampling event. This information is important when interpreting recruitment statistics and provides a framework for improving operations. Field supervisors should consider collecting the following information: • description of the counting, intercepts, recruitment, interviewing and testing areas at the venue; • barriers to survey activities or safety issues at the venue; • significant changes in the demographic characteristics or eligibility of venue attendees since the initial venue observation was conducted; • new venues or day–time periods that were suggested during the sampling event; • reasons for removing the venue from the sampling frame; and • information about parking or transportation. Review of forms and logs At the end of the sampling event, the field supervisor should collect all materials used in the sampling event from field staff, review them for accuracy, make corrections when necessary, and tabulate recruitment outcomes. Be sure to thank any gatekeepers or others who made sampling possible at the venue.
B-1.3.5 Supervision in venues Strong supervision is crucial during each sampling event. Knowing what occurs during intercepts, monitoring trends in recruitment (refusals, successes), and working with each field-staff member's strengths and weaknesses are critical components of good field supervision. Investigators should observe field staff in a systematic way and provide them feedback on their performance during both training and survey implementation. Other supervisory activities during the sampling event include: • Ensure the survey is implemented according to the protocol and SOP. • Check in with recruiters and suggest ways to improve recruitment techniques (especially responses to recruitment barriers), increase participation and troubleshoot difficult intercepts. • Monitor recruitment and enrolment throughout the event to determine individual and team performance. Make changes to the counting area, recruitment techniques, team operations or other procedures if necessary. • Maximize the strengths of survey team members. Determine the best recruiters by observing the quality of the recruitment. Determine who works best at which venues, and with which populations. • Build team morale. Recognize a job well done and encourage the team to help and support one another.
B-1.3.6 Sampling event closeout The sampling event finishes either at the end of the venue–day–time period or when the maximum number of participants has been enrolled. Post-event debriefing At the end of each sampling event, a post-event debriefing allows field staff to discuss any issues or special circumstances that occurred. This debriefing
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
165
Figure B-1.9 Biobehavioural survey venue recruitment and survey participation steps Counter counts and directs recruiter/interviewer to intercept Intercept Accept intercept?
No
End
Yes Recruit Successful recruit?
No
End
Yes Screen
Eligible?
No
End
Yes Consent (survey, testing) Able/willing to consent?
No
End
Yes Survey Complete interview? End interview Insert comments
No
Yes
Collect biologicql specimen(s) – (after pre-test counselling)
Provide incentive
Return rapid HIV (and other) test results
Provide referrals as appropriate
Provide treatment for treatable conditions
166
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
B-1.4 Respondent-driven sampling Participant recruitment, enrolment, and participation in respondent-driven sampling (RDS) surveys is a complex and multistep process. This section outlines these and other key steps in the implementation and monitoring of RDS surveys. Participation in an RDS survey usually requires two visits to a survey site. The first visit includes coupon verification, eligibility screening, informed consent, interview administration, biological specimen collection and testing, recruitment training, and compensation for time and transport. The second visit includes return of any remaining test results, recruitment effort interview, and compensation for successful peer referral.
B-1.4.1 RDS survey staff and process The survey team should include a combination of the following staff positions: field staff coordinator, receptionist, eligibility screener, coupon manager, interviewer, counselor and laboratory technician. For details regarding the staff involved in an RDS survey, see Chapter A-14. First visit in an RDS survey Because the first and second visits for an RDS survey involve multiple steps, checklists are recommended to ensure that all steps have been performed for each participant (see Appendix I-32). Figure B-1.9 shows survey procedures for the first visit.
Figure B-1.9 Procedures for the first visit for an RDS survey Person arrives at site
Assess Coupon valid?
No
Yes Screen Eligible?
No
Ineligible
Yes Consent Able/willing to consent?
No
Yes
Administer Interview
Collect biological specimen(s) (after pre-test counselling)
Issue coupons and train on recruitment. Give primary reimbursement
Person leaves site
Return results and provide post-test counselling (for rapid tests)
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
167
Coupon verification Potential participants are greeted by a receptionist upon entering the survey site. The receptionist verifies that each participant: • has a valid coupon (except for seeds, for whom a coupon is not required); and • is able to provide informed consent (i.e. is not drunk or under the influence of drugs, and has no mental or emotional disabilities that may preclude informed consent). A valid coupon is a coupon that has not been forged or photocopied. If a start date, expiry date, or both are included, the potential participant must generally participate during the specified period (see Section A-9.7), although the investigators may choose to relax the rules to encourage participation. Doing so would not violate any RDS assumptions. The participant should be given an information sheet describing the survey while waiting for the next step in the survey process: eligibility screening. A sole receptionist is often the first person who can identify repeat participants: because the receptionist will see everyone who enters the survey site. Individuals who have already participated in the survey must be refused participation. Their coupons should be marked “VOID” and kept by the survey team. Survey staff should treat all participants in a courteous manner; participants who feel dissatisfied or offended may refuse to recruit peers. Also, participants who have a bad experience may tell others about their experience, which can further reduce participation. Ensure that the waiting area is a pleasant place with, for example, a television or reading materials (including health-related materials) to keep visitors occupied. Population-specific materials should be avoided to protect participants in the event of unexpected visitors. Investigators should decide whether family members (e.g. children of SW) or friends are permitted in the waiting area. Although some participants may feel uncomfortable if people who are not members of the target group are in the waiting area, allowing others to be present might also help reduce barriers to participation. Formative assessment can help identify the best approach to this decision. Eligibility screening The first step in eligibility screening is to verify the coupon’s validity in the survey database, which is often done by the coupon manager. If the coupon is valid, the manager should stamp it “USED”. If the coupon is invalid (i.e. copied, edited or already used), the manager should stamp the coupon “VOID” and file it for record-keeping. All coupons should be kept for record-keeping and to prevent reuse. Individuals bringing invalid coupons are
not eligible to participate in the survey and must be refused participation. Their coupons should similarly be collected, stamped “VOID” and filed. Eligibility screening often involves asking sensitive questions and should occur in a private room or area. A key component of the screening is to determine whether the potential participant is a member of the target population. If possible, the screener (i.e. the person responsible for the eligibility screening, often the coupon manager) should have experience working with the target population, or be a member of the target population. Screeners can also verify membership through further probing. For example, PWID usually have visible injection marks. If a potential participant has no visible injection marks, the screener can ask how drugs are prepared and injected, or the cost of the drugs. Only individuals who actually inject drugs will be able to answer these questions correctly. Experts on the local context can help tailor questions to identify the target population. Identification of members is harder with SW and MSM, because they can be part of the target population but not know the main areas where sex is sold or where men meet each other. Membership verification questions for each key population are included in the questionnaire module for eligibility in Section II of the supplementary materials. Sometimes, potential participants try to enrol by pretending to be a member of the target population, especially if compensation for participation is high. Ineligible individuals should be asked to leave the survey site, especially if they are ineligible because they are not part of the target population. Staff must still be polite to these individuals. If using an electronic data-collection tool, consider saying “I’m sorry, but the computer indicates you are not eligible”, which reduces the chance of protest. Informed consent After verifying eligibility and before conducting any other survey procedures, survey staff must obtain informed consent. A member of the survey staff reviews the consent form with the participant and answers any questions. Depending on the protocol, the participant can either sign the informed consent form (written informed consent) or verbally agree (verbal informed consent). The staff member should sign the document after informed consent is obtained. Many institutional review boards (IRBs) allow the use of verbal informed consent for surveys of key populations in order to facilitate participant anonymity. See Chapter A-3 for more information on informed consent. After informed consent is obtained, the participant is enrolled in the survey and the interview to determine the size of their personal network can start. If informed consent is not provided,
168
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
if appropriate, ask the person the reason for declining to participate in the survey. Document the reason and provide the person with transportation compensation before escorting them from the survey site. Size of the personal network Questions on the size of the personal network are essential for calculating weights for RDS population estimates. These questions should be administered face-to-face by the coupon manager or interviewer to ensure participants understand the questions correctly. Size of the personal network is determined by a series of questions that lead to an estimated total number of people the participant knows who are probably eligible for participation and have been seen during a defined period. See Section A-9.7 for examples of these questions. Interview The questionnaire should be administered to all enrolled participants, including seeds. After the interview, the participant proceeds to another area for counselling and to provide biological specimens for testing. Specimen collection Participants should receive counselling before they provide biological specimens. Participants must provide consent for the interview and collection of biological specimens. Consent for these two distinct activities may be obtained together or separately. Participants may consent to the interview but not to testing and specimen collection. Reasons for refusing to provide biological specimens may include initial fear of or unfamiliarity with the survey and the process. Some participants may provide informed consent for specimen collection but later change their mind. Any change in the informed consent should be noted and respected. Asking such participants why they changed their minds is acceptable in order to respond to their concerns and improve procedures for future participants. For example, participants may decline an HIV test if they already know they are infected; the staff member can explain to that testing will help the survey to determine the needs of everyone in the population. The survey may also offer additional tests for those living with HIV, including CD4 and viral load. Coupon issuance, peer-recruitment training and primary compensation After the interview and collection of biological specimens, the participant is given coupons to recruit peers into the survey. Each coupon has a unique code. The participant’s own coupon is linked to the coupons they give to peers. The link between the coupons can be tracked by RDS Coupon Manager or by using a logbook (see Section A-9.7 for more information on RDS coupon design and distribution).
At this step in the process, the role of coupons is explained to participants. Survey staff train participants on how to recruit peers using the following instructions: • safeguard the coupons. • give coupons only to people you know who are part of the target population. • give only one coupon per person. • inform peers of the purpose of the survey and what the survey entails, for example, completing a questionnaire and providing blood for tests for HIV and STI testing. • return to the survey site for the second visit to be compensated for successful recruitment, that is, for all of your recruits who completed the interview (as a minimum). To ensure consistency in training recruiters, provide a written script for staff members administering the training (see Appendix I-31). A script minimizes selection bias from inconsistent recruitment training. The unique participant code (UPC) may also be created at this point (see Chapter A-9) to verify participant identity at the second visit. Participants should be encouraged to collect their recruitment compensation during the second visit, when they may also receive results for any nonpoint-of-care tests. A second visit should be scheduled, usually 2 weeks after the first visit. If coupons include an expiry date, the second visit should be scheduled after this date to encourage recruits to return before the recruiter’s second visit. The last activity of this step is to provide compensation for participation. Primary compensation should be provided only to those who complete at least the interview, and should be recorded in a database or logbook. Also, provide the participant with risk-reduction materials such as condoms, lubricants, syringes and needles, as well as educational materials on HIV and other STIs.
Issues with compensation Participants may contact survey staff to learn which of their recruits have participated in the survey, in order to encourage others to participate and thus gain more compensation. Others may ask at their second visit. Giving the coupon numbers of those who participated to the recruiter may jeopardize participant confidentiality and is not recommended. Further, recruiters may place undue pressure on those who do not participate, and responding to such requests can add logistical challenges for survey staff because the person contacted usually does not have recruitment information.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
169
Figure B-1.10 Second visit in an RDS survey Person arrives at site
UPC verified?
No
Person ineligible to receive secondary reimbursement
Yes
Complete peer-recruitment interview or nonresponse questionnaire
Return test results (as needed)
Refer or treat if positive STI test
Person leaves site
Pay reimbursement for peer recruitment
Verify peer recruitment
Return of test results If rapid testing is part of the survey, results should be returned to the participant towards the end of the first visit. Post-test counselling and treatment, if feasible and appropriate, should also be provided. Link participants who test positive for HIV and those testing positive for other STIs to appropriate care and treatment. If tests require processing at an off-site laboratory, participants will need to return for their test results. Results should be returned during the participant’s second visit. Second visit in an RDS survey The purpose of the second visit is to compensate a participant for recruiting peers, to conduct a short interview about peer-recruitment efforts, and to return test results that were processed off-site. Figure B-1.10 shows the steps in a participant’s second visit.
Verification of participants’ identity The second visit begins with verification of participants’ identity, for which participants are asked to show their appointment reminder card with their survey ID. If a UPC has been created (see Section A-9.7), the series of questions is asked again to recreate and confirm the personalized code. This procedure can also be used to find participants’ survey ID if they lose it. Fingerprint scanning can substitute for a UPC. Secondary compensation Verify whether any of the participant’s coupons have been returned by their recruits. Compensation should be provided only for recruits who joined the survey and completed the interview. Peer recruitment or nonresponse interview In most other sampling methods, investigators can easily estimate the nonresponse rate because survey staff directly sample and enrol participants. In RDS, however, measuring nonresponse is difficult because participants
170
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
do the sampling on behalf of the investigators. There are three reasons for unsuccessful recruitment: • the participant lost the coupons or did not offer coupons to peers. • the participant offered coupons to peers but the coupons were rejected. • the coupons were accepted by peers who did not attend the survey site. The peer recruitment questionnaire (see Section II-40 for an example) can be used to determine patterns in coupon distribution, acceptance and refusal. Because not all participants will return for the second visit, results from this questionnaire should be interpreted with caution. Return of test results If testing was conducted off-site, test results should be returned during the second visit, with STI counselling and treatment provided if appropriate and feasible. If treatment is not available on-site, participants should be referred to target population–friendly clinics for treatment.
maximum number of coupons issued can be increased either temporarily or permanently in order to boost recruitment (e.g. from 3 to 5 coupons). The number of coupons given to each participant and the date of this procedural change should be documented. The increase in coupons issued per participant should ideally be temporary, lasting only long enough to boost recruitment. Successful RDS recruitment includes long recruitment chains. If too many coupons are issued to each participant, shorter chains with more branches may result. Add more seeds The number of seeds investigators can add is unlimited. Adding more seeds can boost recruitment, especially if new seeds can reach untapped social networks. However, adding more seeds can also reduce the chances of producing the long recruitment chains required in RDS. Increase compensation Increasing compensation can further encourage the target population to participate in the survey and recruit peers. However, a large increase may lead participants to sell their coupons or to lie about their eligibility or their membership in the target population in order to receive the compensation. Excessive compensation may also be coercive and unethical. Conduct continued formative assessment Investigators can solicit feedback on how to improve sampling from participants and key informants. For example, through informal conversations with participants and key informants, investigators may learn that the site location is not accessible to the entire target population and therefore an additional site is needed; or that perhaps site hours need to be changed. An anonymous suggestion box can also be placed in the waiting or reception room to encourage feedback. Other aspects In addition to monitoring the sample size, other aspects of sampling require monitoring, including the survey response rate, sample composition, number of recruitment waves, convergence, bottlenecks, and cross recruitment between sites. A coupon-management system is required to track who recruited whom, ensure that compensation is paid to the correct person, and track the number of waves reached. If on-site computers are available, coupons can be tracked in an electronic database. A commonly used database is RDS Coupon Manager2 (see Section A-9.7).
B-1.4.2 Recruitment monitoring Because the participants conduct sampling in RDS, predicting when the target sample size will be reached is difficult. In addition to sample size, the analytical and functional assumptions of RDS (see Section A-9.7) must be met in order for the sample to produce valid population estimates. Recruitment monitoring throughout the survey process helps investigators identify and address issues in recruitment. Sample size The time taken to reach the target sample size depends mostly on the actual sample size chosen, the number of coupons given to each participant, the number of seeds used, the amount of compensation, the density of the survey population’s networks, and the willingness of individuals to join the survey. Although the investigators cannot control the size of the target population, other factors can be controlled by them, including accessibility of the survey site, the days and hours of operation, the number of participants the survey team can enrol per day, seed quality, and the quality of participants’ experience. Investigators have a number of tools available to increase recruitment. Increase the number of coupons The number of coupons issued per participant can be modified throughout the recruitment period. The
2
www.respondentdrivensampling.org
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data collection
171
Survey response rate The nonresponse rate in an RDS survey cannot be accurately measured because investigators depend on participants to refer potential participants. The investigators have little knowledge of those who were offered a coupon but chose not to participate, and how these people differ from those who came to the survey office. Nevertheless, nonresponse can be approximated by monitoring the following: • number of coupons issued and redeemed. Track how many coupons have been issued to survey participants. A coupon is considered “redeemed” when a person with a valid coupon is screened for survey eligibility. • peer recruitment interview. The peer recruitment questionnaire (see Section II-40) is administered when participants return for a second visit. This questionnaire can determine how many of each participants’ coupons were offered, accepted and refused. Monitor the proportion of second visits to determine whether those who return for a second visit are similar to those who do not return. Monitoring sample composition Because the external validity (generalizability) of RDS surveys is difficult to determine, it is good practice to compare sample characteristics with formative assessment findings. Such characteristics include age, sex, ethnicity, neighbourhood of residence and HIV status. If the sample composition during recruitment does not approximate that of the target population, additional seeds with the missing characteristics can be included in an attempt to increase their representation. For example, if a formative assessment revealed that both male and female individuals who use drugs were found in the survey area, the sample composition should comprise both males and females. If only males were recruited, female seeds should be added to encourage recruitment among females. Number of recruitment waves, equilibrium and convergence The number of recruitment waves by seed should be monitored weekly. Equilibrium and convergence should also be assessed throughout recruitment. The two indicate whether the sample is independent of the characteristics of the seeds. While equilibrium is based on the sample, convergence is based on the population estimate. Having diverse seeds can help achieve equilibrium and convergence faster. Variation in both sample and population characteristics will be limited after they are reached. Continually perform, for example, bottleneck and convergence plots for key variables (these
plots can be created in software such as RDS Analyst3) to determine whether equilibrium and convergence have been reached. If they have not been reached for key variables by the time the sample size is reached, continue enrolment if funding and the protocol allow. Seek ethical approval if the protocol needs to be revised for a larger sample size. Multiple survey sites If the survey has multiple survey sites, separate networks can form around each survey site. This violates the RDS assumption that the survey population is made up of a single network. Cross recruitment between sites must occur (i.e. participants from one site must recruit people who go to another site) in order to prevent violation. Cross recruitment is easily monitored if coupon numbers identify where the participant was recruited (see Section A-9.7). To encourage cross recruitment, investigators should inform participants that their recruits can go to any survey site.
B-1.4.3 Ending an RDS survey As the survey approaches the target sample size, investigators need to end sampling in a controlled manner. The number of coupons issued can be gradually reduced or stopped altogether. The goal is to ensure that the sample size is reached while limiting the number of people who may seek participation after this point. To the extent possible, participants presenting with valid coupons should continue to be accepted and enrolled. If additional people can no longer participate (e.g. the IRB-approved sample size has been reached or the budget has been exhausted) survey staff should politely explain the situation to those seeking participation. A script should be developed to guide survey staff in communicating this message to potential participants. During recruitment training, survey staff can also advise participants that their recruits should come to the survey site quickly before the sample size is reached. As survey sites will remain open for scheduled second visits, eligible participants with valid coupons presenting themselves during this time should be reimbursed for their transportation, even if not enrolled.
3
http://hpmrg.org/software/
172
1. Data collection
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
B
2. Survey monitoring
This chapter describes considerations for survey monitoring. Monitoring is the real-time assessment of survey implementation to ensure that the protocol and standard operating procedures (SOPs) are followed, and to identify challenges while the survey is being implemented so that immediate corrective action can be taken. Monitoring is not a one-time event but an ongoing activity throughout the survey. Monitoring key aspects of survey implementation can contribute to improved data quality and should be done by survey staff and investigators.
Table B-2.1 lists key areas for monitoring. It may be useful to create a simple checklist based on Table B-2.1 to systematically document the monitoring of each activity, including what was monitored, by whom and when; any problems or challenges identified; and how each problem Table B-2.1 Monitoring topics and areas of focus Topic Staff Training and training material Sampling Recruitment
was addressed. Documenting monitoring activities can be useful not only for improving the existing survey, but also for future survey development and for when an evaluation is conducted after the survey (see Chapter C-3).
Areas of focus • Adequate number and type of survey staff (e.g. interviewers, nurses or counsellors, data managers, field staff supervisors) • Field staff trained on SOPs to ensure protocol adherence • Sampling conducted properly according to protocol and SOP • Verification of duplicate enrollees • People posing as members of target population • Fake/invalid coupons, coupon uptake (RDS) • Inappropriate reimbursement, selling of coupons (RDS) • Participant consent administered properly (see Supervision section below) • Any respondents enrolled that were actually ineligible • Ineligible people received coupons (RDS) • Respondent return rates for follow-up activities including return of test results and necessary referrals • Adequate spaces for interview administration • Missing interviews • Revision of questions which: - are difficult to understand - lead to refusals - lead to extreme values • If the questionnaire is modified for any reason, field staff should conduct mock interviews to validate any changes in electronic data entry, including verifying new skip patterns, and valid ranges for new variables
Enrolment
Data collection
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
2. Survey monitoring
173
Topic Data management Data quality
Areas of focus • Adherence to data-management SOPs including data storage, security, backup (see Section C-3) • Data-quality checks stated in protocol and SOPs conducted to identify and address the following: - missing data - invalid data values - outliers - skip pattern errors - refusal rates
Laboratory
• Adequate space for specimen collection • Adequate equipment • Supervision of biological specimen collection and processing for transportation to the testing laboratory • Record-keeping (e.g. specimen quality form or specimen tracking form) consistent with protocol and SOP for both survey laboratory and/or referral laboratory • Trained laboratory staff perform testing • Temperature logs completed and maintained • Test kits, reagents, and specimens not expired and being stored at appropriate temperatures • Verification of the number of specimens at laboratory and the number of completed interviews with the number of participants who provided consent for interview and/or specimen collection • Proportion of test results returned, delayed • Missing specimens, poor quality specimens • Poor testing quality • Insufficient supply of test kits • Testing errors • Inclusion of external quality assessment to validate results (see Chapter A-6) • Conduct of HIV counselling and testing • Provision of relevant referrals and reimbursement • Breaches in confidentiality • Informed consent procedures followed • Time spent for survey participation • Aborted interviews or survey procedures • Actual data-collection costs compared with estimated budgeted costs
HIV counselling, referral and reimbursement Human subjects Participant burden and acceptance Data-collection costs
B-2.1 Observing data collection Field supervisors should observe about 10% of the total number of interview sessions. Participants must give verbal permission before the session observation. Field supervisors should ensure that privacy is maintained during the interview. More frequent observations (about 25% of the observed interviews) may occur towards the beginning of the survey to identify any potential problems or issues in the administration of the survey. The field supervisor may use a site supervision checklist to identify potential issues. Problems should be addressed as soon as possible. For qualitative interviews, field supervisors should verify that interviewers have sufficiently probed key themes. If qualitative interviews are conducted with audio recording, the interviewers should also listen to a sample
of audio recorded interviews to ensure that the recording quality is adequate for future transcription. Feedback from survey participants may be obtained either immediately following survey participation by randomly approaching participants, or using a locked suggestion box where participants may leave anonymous notes about concerns they may have.
B-2.2 Laboratory monitoring As with interviews, field supervisors should select and observe about 10% of specimen collections. More frequent observations (at least 25% of specimen collections) may occur at the beginning of the survey. Field supervisors should conduct daily or more frequent
174
2. Survey monitoring
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
checks of specimens to identify and correct potential problems with specimen quality, including: • haemolyzed blood specimens • improperly dried blood spots • illegible specimen labels • missing labels • inadequate specimen storage For quality control and assurance, the laboratory may be enrolled in a proficiency programme relevant to assays performed there. In addition, a proportion of specimens may be sent to a reference laboratory or external laboratory for retesting to assess consistency and quality of results. All assay runs should include testing of valid controls and calibrators, if applicable. Any invalid runs should be repeated. Concordance between initial and confirmatory runs should be assessed as an additional measure of quality.
Participant follow-up and referral Field supervisors should check that test results are available for post-test counselling and referral. For sites that do not provide same-day test results, supervisors should monitor logs to assess the proportion of participants who return for their results. If the proportion of participants who do not return for their results is high (e.g. greater than 20%), survey investigators should actively encourage participants to return for test results during pre-test counselling sessions. Please refer to Chapter A-6 on biomarker considerations for further information.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
2. Survey monitoring
175
Section
B
3. Preparing data for analysis
This chapter provides guidance on data cleaning, preparing data for analysis, and assessing data quality. It discusses the different types of errors commonly found in surveys, and provides information on how to assess the quality of survey data, address identified errors, append and merge datasets, and document any modifications to the data.
There are several steps in preparing a dataset for analysis. Data cleaning (Steps 1–3) should ideally occur throughout data collection. Merging and preparing datasets for analysis (Steps 4–5) should occur after all data have been collected and cleaned.
Data errors can originate from various sources, including data collection, data entry (e.g. transcription errors, creating duplicate data) or sampling. Using a systematic process to identify and address errors can improve the quality of survey data. The following are examples of common data errors and concerns: • transcription errors • duplicate records • measurement error • lack of internal consistency • out-of-range values • outliers • other errors These errors are described in Table B-3.1.
B-3.1 Identifying and correcting data errors Data cleaning is the process of detecting and addressing data errors in a dataset. Data cleaning can be minimized if survey investigators take steps to assure data quality throughout data collection (see Chapter A-13). If electronic data collection is used, errors due to data entry can be minimized by proper coding, correct skip patterns, and correct data-validation rules. In addition, the survey instrument should be thoroughly tested before implementation and data collection should be monitored closely, to find and correct repeated mistakes early. Data managers should maintain a log showing corrections that are made. Some software programs automatically track corrections. If corrections are made on paper before data entry, then notes and the initials of the person making the changes should be placed on the paper form.
176
3. Preparing data for analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table B-3.1 Common data errors and how to prevent them Common data errors Transcription errors Description of error: Transcription errors are mistakes made when a person enters data from one form of documentation to another, for example, when paper-based data are entered into an electronic database. How to prevent error: Double-data entry can minimize errors of transcription. Two survey staff members each enter the data once and examine values that do not match within the two versions of the dataset. Many programs used for data entry will include functions and internal checks that can identify such discrepancies. How to resolve error: Each error identified in the dataset should be verified with the original response from the paperbased questionnaire. After data verification, all transcription errors should be corrected and any changes or corrections documented for future reference. Duplicate records Description of error: Duplicate records occur when one participant has more than one record in a dataset. The two records are the same for all variables for this participant and can result from transcription errors or errors in data merging. How to identify error: Duplicate records can be identified by running frequencies of the survey ID variables to identify duplication, during either the data-entry process or data management, including uploading of surveys and merging with other data sources. How to resolve error: For truly duplicate records (all variables for two or more records are identical, including the survey ID), delete the extra records. Measurement error Description of error: Measurement error occurs when the response provided differs from the true value. It can be caused by: Questionnaire design: The wording of questions can affect the responses provided by the participants, for example, if questions are unclear, ambiguous or suggestive. Interview administration: The mode of administration – face-to-face, computer-assisted personal interview (CAPI), audio computer-assisted self-interview (ACASI). For example, participants may not answer questions of a sensitive nature in front of an interviewer but may be more comfortable responding through a computer. Alternatively, they may not fully understand a question asked in ACASI and may not ask survey staff for clarification. Interview: The interviewer can introduce error in survey responses by not reading the question or response options as intended, by probing inappropriately, or by adding other information that deviates from the question or confuses the participant. Participant: Participants can interpret the meaning of, or respond to, questions differently. Errors can also be due to recall (i.e. the participant cannot accurately remember the true answer). How to prevent error: Measurement error can be reduced through piloting of the data-collection instruments before survey implementation and obtaining feedback from participants on their understanding of the questions and response options (i.e. cognitive interviewing). If measurement errors are found during survey implementation, the data collection instrument can be modified (ensuring that the version change of the instrument is documented). Additionally, if face-to-face or CAPI is used for data collection, interviewers should be trained thoroughly in order to not influence participants. Uniformity in interview administration is essential. How to resolve error: Measurement errors can be difficult to detect, particularly if the responses are not out of range or are consistent with responses to other variables in the dataset. If measurement errors are detected, keeping the original values and documenting the source of the error may be the best option. Lack of internal consistency Description of error: Internal consistency means answers to different questions do not contradict each other. Answers that do contradict each other are called logic errors and indicate a lack of internal consistency. Internal inconsistency can result from poor questionnaire design, incorrectly defined skip patterns, incorrectly programmed skip patterns, or a combination of these factors. For example, a participant states being male and later in the interview states being pregnant. A skip should have been programmed to prevent males from responding to any questions directed towards females, such as being pregnant. Or participants may incorrectly respond to a question; for example, reporting always using a condom in the past 6 months but indicating in a different question that they did not use a condom at last sex. Sometimes logic checks can prevent such errors; for example, a logic check can be placed in an electronic questionnaire to require that age at marriage is no more than current age.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Preparing data for analysis
177
Common data errors How to prevent error: Most errors in logic are preventable when validation rules and skip patterns are programmed correctly (see Chapter A-13). Investigators should pilot data-collection instruments before implementing the survey to ensure the tools are programmed correctly. If errors are found during data collection, the tool should be revised or reprogrammed as appropriate, with changes documented. How to resolve error: In some cases, inconsistent responses can be coded as missing. However, some errors due to incomplete data or faulty skip patterns cannot be resolved. In this event, document the error and be aware of its implications for data analysis or presentation of results. Out-of-range values Description of error: Out-of-range values are values that do not make logical sense. An example of an out-of-range value is a participant response or record of “15” for month of birth when the range for the variable should represent the number of the month (i.e. 1–12). An invalid value would include values outside the given response choices. For example, if a question requires a response of “1” for YES or “2” for NO and the response value of “3” is given, the error must be corrected and set to “Missing” because it is an invalid response. How to prevent error: Programming minimum and maximum values for ranges and valid values into an electronic data collection tool can minimize out-of-range and invalid values. How to resolve error: Errors due to out-of-range and invalid values can be resolved in three ways. First, set the value to “Missing”. Second, for continuous variables, use either the mean or the median value for the entire population, depending on the distribution of the data; for data that are not normally distributed, using the median is preferable because it minimizes the impact of outliers. Third, infer the correct value through other responses or data sources, if possible. Each change and the reason for the change should be documented. Outliers Description of error: Outliers, or extreme values, are not always errors. Outliers can be defined statistically. Formal statistical tests can identify outliers. These tests are designed to identify values that may influence statistical analyses practically. For an outlier to be considered an outlier, the value is illogical. For example, to the question “How many sexual partners have you had in the past week?” a response of “100 partners” may be possible, but is illogical. Typing errors may be suspected of producing some outlier values; for example, a response of “33” partners in the past 6 months could have resulted from typing the number 3 twice, when the participant actually responded “3”. How to prevent error: Because outliers or extreme values are not always errors, there are very few ways to prevent them. Questions can be worded carefully to elicit more consistent responses from participants. Interviewers can be instructed, or ACASI/CASI can be programmed, to ask the question again in a nonjudgemental manner to ensure the participant understood the question. Additionally, outliers or extreme values that do not make practical sense can be minimized by programming ranges into the data-collection tools. How to resolve error: Some experts recommend that these values remain “as is” in the analysis because they are not necessarily errors (3). However, the investigator may want to consider the influence of these extreme values on the analysis before deciding whether to include the data in the final analysis. The analysis can be repeated with and without the outlier data to assess the impact of the outlier on the analysis. After observing the outlier’s effect, and if these values are set to “Missing”, they should be reported as excluded from the analysis. The effect of outliers and extreme values can sometimes be minimized through categorization of response values. For example, a question probing the number of noncommercial sex partners in the past 6 months may yield a few outliers such as “250”, when all other values are below, for instance, 50. Placing all the continuous values in categories denoting 0, 1, 2 or 3+ noncommercial sex partners would minimize the effect of outliers. Other errors Other errors that require decision rules for correction include missing forms or records, programming errors, and values or records that have been incorrectly changed or deleted during data handling or cleaning. Although these errors cannot be corrected, they must be documented and noted in any reporting of the analysis.
Step 1: Assessing data quality The first step is to assess the quality of the survey data by examining the data for the common errors mentioned above. Screening data can help identify missing data, errors such as out-of-range and invalid values, outliers and unusual patterns due to measurement errors or
lack of internal consistency. Methods for data-quality assessment can be observational and not necessarily statistical. Invalid values, ranges and logic checks should be determined before beginning the data-quality analysis. Some commonly used methods for assessing data quality are described below (see Table B-3.2).
178
3. Preparing data for analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table B-3.2 Methods of screening data for errors Method Frequency distribution
Description Visually check for errors to identify missing data, out-ofrange values, invalid values and other potential errors.
Pros and cons Pros: • high level of skill not required for observation Cons: • may not identify all errors due to subjective review • rules defined by data-cleaning plan may not be consistently applied across all versions of a dataset. Pros: • high level of skill not required for observation • can verify internal consistency or logic checks Cons: • some analysis skill required to create 2 × 2 tables • may not identify all errors due to subjective review • rules defined by data-cleaning plan may not be consistently applied across all versions of a dataset.
Level of difficulty Easy
Bivariate display Bivariate displays or crossor 2 × 2 table tabulations look at expected relationships among variables. Particularly helpful to find logic errors.
Medium
An automated programme can be used to flag errors. The programme should be created using the rules defined in the data-cleaning plan, including logic checks, valid ranges and values, and missing values. Such programmes objectively screen all data and can be consistently and repeatedly applied to all versions of a dataset. However, their creation requires knowledge of syntax for statistical software and, often, significant time, due to the level of detail needed if the programme is based on the defined rules. Examples of using the data-quality assessment methods described in Table B-3.2 are provided below. Example 1: Transcription error and missing data Observing the frequency distribution of a variable can help to identify missing or duplicated data. For example, by monitoring the frequencies of the survey IDs, the number of participants can be confirmed. If the survey has collected information from 15 participants, but there are only 14 participants according to the frequency distribution (Figure B-3.1), investigators should determine why a record is missing and which record is missing. The data could be truly missing, or, as shown in Figure B-3.1, one individual may have mistakenly been given the survey ID of another participant.
A frequency distribution of survey IDs can also help determine duplicates. A survey ID should be unique; therefore, the frequency of every survey ID should not be greater than 1 (see survey ID 8 in the table below). If the data is duplicated, go back through the original data and forms if possible to determine the correct value for the variable. Consider whether a variable has too many missing values. Generally, if more than 10% of the values of a variable are missing (and the missing values are not due to skip patterns), the variable may have too many missing values to be interpreted.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Preparing data for analysis
179
Figure B-3.1 Example of data table with transcription error Survey ID 1 2 3 4 5 6 7 8 9 10 11 12 13 Total
Frequency 1 1 1 1 1 1 1 2 1 1 1 1 1 14
Percentage 7.14 7.14 7.14 7.14 7.14 7.14 7.14 14.29 7.14 7.14 7.14 7.14 7.14 100.00
Cumulative percentage 7.14 14.28 21.42 28.56 35.70 42.84 49.98 64.26 71.40 78.54 85.68 92.82 100.00
Example 2: Out-of-range values and invalid values Out-of-range values and invalid values can be prevented if valid ranges and values are assigned and programmed correctly during the creation of data-collection tools. During data collection, out-of-range and invalid values can be identified using frequency distributions.
For example, for month of birth, the range of valid values should be 1 through 12 to represent each month in the year. In Figure B-3.2, the value 13 is out-of-range and should be corrected based on the rules in the datacleaning plan. The electronic programme for the datacollection tool should also be modified to prevent future out-of-range values for this variable.
180
3. Preparing data for analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure B-3.2 Example of data table with out-of-range value Month of birth 1 2 3 4 5 6 7 8 9 10 11 12 13 Total
Frequency 5 2 2 1 3 2 3 4 7 3 4 3 1 40
Percentage 12.50 5.00 5.00 2.50 7.50 5.00 7.50 10.00 17.50 7.50 10.00 7.50 2.50 100.00
Cumulative percentage 12.50 17.50 22.50 25.00 32.50 37.50 45.00 55.00 72.50 80.00 90.00 97.50 100.00
Figure B-3.3 Example of data table with other potential error HIV test results Negative Positive Indeterminate Total
Frequency 122 26 2 150
Percentage 81.33 17.33 1.33 100.00
Cumulative percentage 81.33 98.66 100.00
Example 3: Other potential errors Frequency distributions can be used to identify other potential errors. For example, if the laboratory identified 27 HIV-positive specimens but the frequency distribution shows only 26 HIV-positive observations (Figure B-3.3), then the correct number requires verification through further investigation.
Example 4: Bivariate displays Bivariate displays or cross-tabulations can also indicate errors in data collection such as missing values, outof-range or invalid values, and other potential errors. Bivariate displays can verify internal consistency and perform logic checks in the data. For example, a crosstabulation of the variables “ever had sex” and “had sex in the past 6 months” should show zero responses in the cell for those who responded “no” to ever having sex and “yes” to having sex in the past 6 months (not “3” as in Figure B-3.4). Changes to the data should be based on the rules in the data-cleaning plan.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Preparing data for analysis
181
Figure B-3.4 Example of bivariate display Had sex in the past 6 months Ever had sex No No Yes Total 21 3 24 Yes 54 177 231 Total 75 180 255
Step 2: Address identified errors If errors are detected, the next step is to address the identified errors and missing values based on the rules defined in the data-cleaning plan. As part of this process, the person responsible for resolving the possible error and the period in which the resolution will occur should be specified. Having these rules available to survey staff involved in data collection and management is recommended. Correcting errors may involve returning to the original source of information (i.e. questionnaire or initial dataset). Three solutions exist if the error is in the original data: • correct the value: - use value in the original source of information (for paper-based collection); - if continuous, set the value to the mean or median; - if possible, locate information from another response or source; - let the value stand; or • set the value to “Missing”. Any changes made to the dataset must be documented in a data-audit log that describes the screened errors and the resolution of the errors. Recording the reasons for missing values Missing data can have various effects on the data-set. They can decrease the effective sample size, making estimates less precise (i.e. wider confidence intervals [CIs]) and statistical tests less powerful in detecting any significant differences. Also, missing data may not be randomly missing – that is, certain characteristics may be associated with missing data, which can affect the results of analysis. Some groups of participants may have different proportions of missing data. Therefore, as far as possible, investigators should prevent or minimize missing data. If there are no patterns in the missing data among the different groups, estimates will not be biased by the missing data. If missing data are ignored, those who did not respond to a certain question are assumed to be, on average, similar to those who did respond to the question.
For missing values, imputation can also be considered. Imputation is a statistical method in which values are prescribed for the missing data based on other data provided by the participant. There are various ways to impute data, including multiple imputation, but these methods are complex and require consultation with a statistician. Variables can have a value of “No response” for any of several reasons, which may be useful to distinguish. A value of “No response” could be due to the following: • the participant did not know or remember a response to a question; • the participant refused to answer a question; • the question was skipped based on a previous question; or • the participant did not input a response and left a blank space for the question. Some situations require distinguishing between the missing response types. For certain questions, a response category may mimic the default category “Don’t know”. For example, a question such as “Can a healthylooking person have HIV?” has the response options “Yes”, “No”, “Don’t know” and “Refuse to answer”. In this example, the response option “Don’t know” is a legitimate response and should not be set to “Missing”. In the questions below in Figure B-3.5, the first example demonstrates when “Don’t know” should be considered a response option. The second example demonstrates when “Don’t know” should be set to “Missing.” Ideally, make the numeric code for “Refuse to answer” or “Don’t know” the same for each question. In the example below, “8” is the code for “Refuse to answer” each time. If using the same value consistently, ensure that this value cannot be a potentially valid response in one or more questions. For example, do not use “8” for “Refuse to answer” if a question that needs a “Refuse to answer” response option asks for the number of years the individual has lived in the location, because “8” could be a valid response to this question. “888” may be a better choice for “Missing” in this case.
182
3. Preparing data for analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure B-3.5 Example of the different meaning of “Don’t Know” Example 1. “Don’t know” is a legitimate response.
Can a healthy person have HIV?
YES 1 NO 2 DON’T KNOW 3 REFUSE TO ANSWER 8
Example 2. “Don’t know” is not a legitimate response.
Are you currently living with a sexual partner? If applicable, any identified errors that are due to the questionnaire design or programming of the electronic data-collection tool should be corrected. These changes should be documented and a new version of the datacollection tool should be saved. Step 3: Append and merge datasets After the data from each dataset have been cleaned and the errors have been corrected, the next step is to append the data (if there are multiple sources) and merge them, as discussed below.
YES 1 NO 2 DON’T KNOW 7 REFUSE TO ANSWER 8
Appending datasets If the same data are collected through multiple sources (e.g. multiple sites), creating a database by appending the datasets from those sources is highly recommended. Appending a dataset means to directly add the observations from one dataset to another, usually because each dataset is derived from different participants. When investigators append, they are usually adding two datasets measuring the same variables in different groups together, essentially stacking one dataset on top of another. The variables and variable names should be consistent throughout the data sources. In Figure B-3.6, datasets 1 and 2 are appended, resulting in one dataset with four records.
Figure B-3.6 Example of appending datasets Dataset 1 Variable 1 Survey ID 1 Survey ID 2
Variable 2
Dataset 2 Variable 1 Survey ID 3 Survey ID 4
Variable 2
Appended dataset Variable Survey ID 1 Survey ID 2 Survey ID 3 Survey ID 4
Variable 2
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Preparing data for analysis
183
Figure B-3.7 Example of merging datasets Dataset 1 Variable 1 Survey ID 1 Survey ID 2
Variable 2
Dataset 2 Variable 3 Survey ID 1 Survey ID 2
Variable 4
Merged dataset Variable 1 Survey ID 1 Survey ID 2
Variable 2
Variable 3
Variable 4
Merging datasets If different data are collected from the same participant (e.g. behavioural questionnaire and laboratory results), the data should be merged (Figure B-3.7). Merging must be performed using the key variable in common (usually survey ID), and checked to detect unmatched records from each data source. After the datasets are merged, a selection of records from the merged file should be compared to the original data sources to ensure the data were correctly merged. The data manager should verify that records from each dataset are not missing from the merged set. If there are missing records, the source of the error must be identified and corrected. If a correction is not possible, the team must document the decision on how the missing records were handled. Data managers should ideally merge datasets multiple times throughout the data collection, instead of waiting until all data are collected, in order to ensure that any problems are discovered early and thus can be rectified. Step 4: Documentation A good data-management practice is proper documentation of all procedures, including any changes made to the data as part of the data-quality assessment. Documentation is important for several reasons. It enables each survey staff member who is working on a dataset to systematically follow all the steps and actions that were taken, such as records deleted, variables recoded, new variables created, and categories combined or collapsed. Further, if the dataset were shared internally or with external partners, others could readily duplicate and verify the analysis if all data preparation
steps were properly documented. Documentation is also important if the survey results are to be disseminated and published, especially in a peer-reviewed journal. The following are suggestions for good documentation practice: • keep the original dataset separate and save all programs leading to the creation of the final dataset, to allow recreation of any intermediate dataset if needed; • document in a separate log any modifications made due to data entry; • document computer programmes used to modify the original dataset through assigning a unique identifier that includes a descriptive title and the date the programme was created. if possible, describe the changes and modifications within the programme itself; • adopt a system of naming conventions for datasets when they are merged or modified; • maintain a log of programmes run on the dataset, including when the programme was run, the dataset used, and a brief description of what was run; • when creating variables through recodification, select new and meaningful names to preserve the original variables; • if surveys are conducted at multiple sites, variables should be uniformly named and coded across sites and surveys; • maintain the source files as originals before applying corrections; and • ensure all variables and categorical values have descriptive labels and are formatted appropriately.
184
3. Preparing data for analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
C. Data analysis and use
1. Stakeholders 2. Planning considerations before starting a biobehavioural survey 3. Ethical consideration in planning and conducting surveys 4. Formative assessment 5. Questionnaire development 6. Biomarker considerations 7. Preparing biological data-collection instruments 8. Eligibility for participation 9. Sampling strategy 10. Sample size requirements 11. Population size estimation methods used with surveys 12. Supplemental studies 13. Data management 14. Staff selection 15. Standard operating procedures 16. Training and training methods
16 19 23 30 34 43 60 63 66 101 107 121 135 140 143 146
A Survey preparation
Data analysis and use
C
1. Data analysis 2. Use and dissemination of survey findings 3. Evaluating survey implementation 4. Data sharing for public use
186 200 206 208
1. Data collection 149 2. Survey monitoring 173 3. Preparing data for analysis 176
B B. Survey implementation & quality assurance A. Survey preparation
C. Data analysis and use
Survey implementation & quality assurance Stakeholders
185
Section
C
1. Data analysis
This chapter provides an overview of how to analyse data from biobehavioural surveys (BBS). This includes using the data analysis plan (developed with the protocol) as a guide in preparing a dataset for analysis, creating sampling weights for BBS data, and reviewing analytical methods for BBS data. The chapter includes only basic analytical methods, such as descriptive statistics that summarize BBS data for survey reports. More advanced statistical methods such as regression analysis are beyond the scope of these guidelines.
The BBS data-analysis strategies presented in this chapter focus on probability-based sampling designs and inferences (see Chapter A-9). Data analysis that fails to account for the complex survey design of BBS may produce inaccurate estimates, confidence intervals (CIs) and P-values. Accounting for sampling design during analysis ensures that the CIs (the measure of uncertainty around the estimates) include both random sampling error (the measure of precision of the estimates) and other sources of error, such as refusals or selection bias. The standard errors of complex sampling designs (e.g. cluster sampling) are different from those of simple random sampling (SRS). Briefly, the survey design should aim to maximize the precision of the estimates and minimize the standard error; however, there are many factors that influence decisions on survey design, including budgets. For example, cluster sampling may improve the ease and reduce the costs of surveying the sample, but tends to reduce the precision of the estimates and increase the standard error because participants sampled within clusters are often more similar to one another than to participants from other clusters (cluster effect). The decrease in sampling costs often outweighs the loss of sampling efficiency (i.e. increases in standard error). Although most statistical software packages can account for complex survey designs, these analyses should be performed by a statistician or other person knowledgeable about the methods the packages use. A summary of software for survey analysis is available online.1 However, analysis of data from surveys using respondent-driven sampling (RDS) typically requires specialized software such as RDSAT or RDS-A, or specific additional packages to software such as Stata.
C-1-1 Developing a data analysis plan Data analysis turns raw survey data into actionable information that can be disseminated through presentations, technical reports, peer-reviewed publications, policy briefs, press releases and other dissemination tools (see Chapter C-2). To ensure that the survey objectives are satisfied when the data are analysed, develop a comprehensive data analysis plan along with the survey protocol described in Chapter A-2. A data analysis plan includes the following components: • survey objectives and research questions to be answered; • timeline for data analysis; • roles and responsibilities of those involved with data analysis; • type of product (e.g. technical reports or manuscripts, or both); • description of any variable transformations (e.g. grouping continuous variables into categories or creating new variables from a combination of variables); • sources of data (e.g. other datasets to be used for trend or comparative analysis); • description of the data-cleaning process; • strategies for addressing missing data; • description of the weighting process; • statistical methods and software to be used; • data-collection tools; and • table shells (templates) and a description of how data will be presented and which cross-tabulations will be made (including specific denominator definitions).
1
http://www.hcp.med.harvard.edu/statistics/survey-soft/
186
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
C-1.2 Preparing a dataset for analysis Chapter A-13 explained processes for transforming raw survey data into a single data file. This section describes the next step; that is, preparing the dataset for specific analyses to address survey objectives outlined in the data analysis plan. Before beginning weighting and analysis, data should be reviewed to assess the need for any transformations and to address the issue of nonresponse. Creating new variables Sometimes, it is necessary to transform or recode variables to fulfill survey objectives. Creating new variables in preparation for data analysis may include collapsing multiple response categories, combining multiple variables into one new variable, or categorizing continuous variables. One common situation for data transformation is viral load laboratory results. These have a wide range of possible values, making the results difficult to analyse and interpret. By transforming raw HIV RNA laboratory data into a logarithmic (log10) format, the resulting variable range is narrower, more manageable, and easier to analyse, interpret and graph. A log scale is based on the powers of 10 and is calculated as log10(# copies/mL) or 10y = # copies/mL. For example, if an HIV RNA laboratory result is 10 000 copies/mL, then log10(10 000), which is equivalent to 104 = 10 000. Thus, the transformed variable would be log10 = 4.0, a more manageable value for analysis than 10 000 copies/mL. Before starting this process of transforming or recoding variables, investigators should consider the number of responses in each resulting stratum (category). A sufficient number of responses in each stratum is important for producing meaningful estimates, especially if the data will be weighted. Collapsing response categories New variables can be created by collapsing response categories for existing variables. For example, consider condom use where consistent condom use is defined as “Always uses condoms with a partner”. If the three response categories for original condom use with a partner variable are “Always,” “Sometimes” and “Never”, the responses for “Always” would represent consistent use of condoms, and “Sometimes” and “Never” would be combined to represent not consistently using condoms with a partner. The original variable has three response categories and the new variable has two. Other examples of this process are collapsing the number of age categories (e.g. five age groups down to three age groups) or values for marital status (e.g. never married, currently married, divorced and widowed groups into “Ever married” and “Never married”). Looking at the data may also be helpful to determine how to collapse categories. For example, if two adjacent categories are similar regarding the measure of primary interest, it
may be appropriate to collapse these two categories. If the variable is numeric, it may be helpful to collapse according to quantiles, or according to commonly used public-health categories. Combining two or more variables New variables can also be created by combining responses from two or more variables into a single variable. For example, a new variable “men who have sex with men (MSM) who sell sex and have never been tested for HIV” would be created by combining the variables for MSM who sell sex and HIV-testing in MSM. Verify that new variables are created correctly by comparing the frequencies of the original and new variables. Categorizing continuous variables Sometimes, recoding continuous variables (e.g. age or number of partners) as categorical variables is useful. However, meaningful results depend on carefully defined categories. Categories may be defined according to published literature, biologically or clinically meaningful values, or statistically meaningful points such as a median (i.e. data are divided at the 50th percentile) or quartiles (i.e. data are divided into four equal increments at the 25th, 50th and 75th percentiles). Although categorizing continuous variables may make analysis and interpretation of results easier, when cutoffs are applied to continuous data, there is a loss of information, and a reduction in the statistical power to detect effects (4) and the precision of estimates (e.g. means or odds). To illustrate these limitations, consider a survey of respondents ranging in age from 18 to 54 years. A graph of the original age distribution might reveal two peaks (i.e. a bimodal distribution), one at 24 years and one at 38 years. Important information about age distribution might be lost if the age variable is categorized into two groups defined by the median or into four groups by quartile. The resulting estimates by age group might be misleading because interpretation may suggest a constant effect throughout each age group and an artificial change in effect between categories. In contrast, continuous variables such as CD4 count and viral-load data might be more meaningful if categorized according to clinical guidelines that can be interpreted in terms of disease severity or treatment adherence. Verify that new variables are created correctly by comparing the distributions of the original continuous and new categorical variables. Addressing nonresponse Missing or nonresponse data may influence estimates and must be evaluated before analysis begins, particularly with variables known to be correlated with the outcomes of interest. This evaluation should examine the types, patterns and magnitude of nonresponse. Nonresponse is of particular concern when it is not
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
187
random (i.e. individuals with certain characteristics are more likely to fail to provide an answer to a question than are other individuals); estimates will be biased if this issue is not addressed in analysis. Several types of nonresponse are possible in surveys: • Unit nonresponse. Sampled units (e.g. persons or venue) did not participate in the survey. The differences between those who participated and those who chose not to participate may have a significant impact on estimates of HIV prevalence or risk. Post-stratification weighting for unit nonresponse can reduce, but not eliminate, the influence of this bias on the survey results. The process uses a set of external population estimates (e.g. census data) to calculate weights that adjust the sample data to align with key characteristics (e.g. age, sex) of the population. For example, compared with younger MSM, older MSM may be less likely to attend certain venues, or less likely to be recruited into an RDS survey, which may make the population appear younger, on average, than it actually is, or appear to have a lower HIV prevalence than it actually has. • Item nonresponse. A participant gave no response or an invalid response to at least one survey item. Examining participant characteristics of item nonresponses provides context for interpreting estimates. Nonrandom item nonresponse might indicate problems with poorly constructed questions, response options or skip patterns, or might indicate differential sensitivities to questions. Depending on the size of the survey sample and the objectives of the analysis, item nonresponse might not affect the analysis. For small sample sizes or items that are directly related to analysis objectives, analysts should examine nonresponse patterns by participant characteristics and survey administration (e.g. with electronic data-collection tools). For example, individuals who refuse to be tested for HIV may be more likely to already know that they are HIV positive, which can make the HIV prevalence of the population appear lower than the actual prevalence. Some statistical software packages have commands that allow for assessment of patterns in missing data (e.g. “mvpatterns” in Stata). This assessment provides analysts with information to help decide which strategies are most appropriate to address the types of missing data in the sample. Weighting (e.g. post-stratification weighting for nonresponse) and imputation – replacing a missing value with an estimated value – are two methods of compensating for missing data during analysis. A sensitivity analysis measures the effect of different methods of dealing with non-response on the final results of the BBS. Several software packages offer options for running sensitivity analyses (e.g. PROC MI in SAS v 9.4). In brief, a sensitivity analysis examines violations of the assumption that data are missing at random by running different models and comparing
the consistency of the results of those models. If the results of multiple models produced by the sensitivity analysis are consistent, then it is likely that the strategy to deal with non-response was reasonable and produced reliable BBS results. If the sensitivity analysis produced inconsistent results across the multiple models, it would be necessary to consider a different strategy to address non-response data. These procedures are complex and are beyond the scope of these guidelines; they should be performed by a statistician or other person knowledgeable about the methods. Analysing and interpreting results from nonprobability surveys Because the selection probability of participants in nonprobability surveys is unknown, weighting data from such surveys is impossible. Probability-based analyses to assess associations are similarly impossible, because the data are representative of only the survey participants and not of the larger population. Data from such surveys should therefore be interpreted with caution. Findings may be used to describe a subgroup of the population but should not be extrapolated to the entire population. Creating sampling weights for BBS data Weighting adjusts the data from survey participants so that the estimates are representative of the population from which the participants were drawn and thus reduces some of the different types of bias that were defined in Chapter A-9. The weights being calculated here are the selection weights only. As mentioned in the sampling section (Chapter A-9), sometimes oversampling is used to ensure that enough people are selected to obtain stratum-specific estimates large enough to make inferences about that subgroup. In such cases, a design weight should be used to compensate for oversampling (or undersampling). This design weight ensures proportionate representation of the oversampled subgroup in the overall population. The computation of a design weight follows the same process as that used for other sampling weights below. As described in Chapter A-9, in a probability sample, each individual has a known chance of being selected for participation in a survey. Sampling weights are computed as the inverse of the probability of being sampled:
w = 1/P where w = sampling weight P = probability of selection If all participants have been assigned a weight equal to the inverse of their selection probability, the sum of the
188
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
weights will be equal to, or very close to, the size of the sample population. For example, if clusters of unequal sizes are selected via SRS, then the weights will probably not be equal to, but should be close to, the size of the sample population. For most BBS, however, participants have unequal probabilities of being selected, resulting in biased samples. When a sampling design results in unequal selection probabilities for participants, weighting techniques are used to give more weight in the analysis to those who had less chance of being selected. This process improves generalizability or representativeness of the estimates. Nonresponse and post-stratification weights may still be needed. In general, there will be one final analysis weight for each participant. This final analysis weight is the product of the individual sampling weight and other adjustments as needed (e.g. a nonresponse adjustment factor and a post-stratification adjustment factor, as shown in the example below).
components. The final wselection is the product of the inverse of all of these components. wnonresponse = a nonresponse adjustment factor, which is the inverse of the probability (i.e. P = 1 – nonresponse) that a selected individual will participate in the survey. wpost-stratification = a factor that uses population distribution data (e.g. census data) to adjust the distribution of survey participants. Weighting for cluster-based sampling: conventional cluster sampling and time-location sampling Table C-1.1 shows examples of cluster-based sampling stages in sampling designs. See Sections A-9.5 and 9.6 for more information. An additional stage might be the probability of selecting geographical units, if they are used in the sampling frame; for example, selecting a number of towns from a list of all towns. This would become the first stage and be included in the probability of selection:
wfinal = wselection × wnonresponse × wpost-stratification where wselection = individual sampling weight (“base weight”), which is the inverse of the probability of being selected. Selection probability may be influenced by more than one factor, and thus might have multiple
PI= pgeographic unit (if applicable) * ptime-location cluster * pindividual venue attendee Step 1: Calculate the sampling probability There are four common scenarios for cluster-based sampling: conventional cluster sampling (CCS) and timelocation sampling (TLS). The following examples provide guidance on calculating sampling probability based on how clusters are selected.
Table C-1.1 Examples of cluster-based sampling stages in sampling designs Conventional cluster sampling (CCS) First stage (cluster) Second stage (individual) Venue Individuals associated with venue
Time-location sampling (TLS) Time-location cluster Venue attendees during specified period
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
189
Scenario 1: • First stage: selection of clusters by probability proportional to size (PPS) • Second stage: equal number of individuals selected from each cluster
Pi = (m * Mi/M) * (ni/Ni) where
Pi = probability that individual in cluster i was selected for the survey m = number of sample clusters selected Mi = expected measure of population size for cluster i M = total measure of size for the survey universe (M = ΣMi = sum of the population sizes of all clusters) ni = number of subpopulation members selected in cluster i Ni = Total number of subpopulation members observed in cluster i. An equal number of individuals is selected from each cluster at the second stage. Thus, Mi and Ni are equal and they cancel each other out. Since ni is the same for each cluster, all clusters will have an equal sampling probability of ni /M. As such, this design results in a self-weighted sample. Thus it is not necessary to apply sampling weights to obtain valid point estimates, but it is still necessary to account for the sampling design when calculating the standard errors, CIs and P-values. In addition, adjustment for other issues such as nonresponse is still needed. Should sampling frames be updated during the survey, confirm that Mi and Ni remain equal. If not, then consider another, more appropriate weighting approach.
Scenario 2: • First stage: selection of clusters with equal probability • Second stage: “take all” strategy. In cases where the number of eligible individuals in a venue is less than the target number of participants at each sampling event, all eligible individuals in the venue will be sampled.
Pi = (m/M) where Pi = probability that a subpopulation member in cluster i was chosen for the survey m = number of sample clusters selected M = total number of sample clusters in the sampling frame Because all subpopulation members present on the randomly chosen day are selected for the sample, the second-stage sampling probability is equal to 1.0 and is not shown here. This design also results in a self-weighted sample, which means it is not necessary to apply sampling weights to obtain valid point estimates, but it is still necessary to account for the survey design when calculating the standard errors, CIs and P-values. In addition, adjustment for other issues (e.g. nonresponse) is still needed.
Scenario 3: • First stage: selection of clusters with probability proportional to size (PPS) • Second stage: subsampling
Pi = (m * Mi/M) * (ni/Ni) where
Pi = probability that an individual in cluster i was selected for the survey m = number of sample clusters selected Mi = expected measure of population size for cluster i M = total measure of size for the survey universe (M = ΣMi = sum of the population sizes of all clusters) ni = number of subpopulation members selected in cluster i Ni = total number of subpopulation members in the cluster i This design results in a non-self-weighting sample; thus it is necessary to apply sampling weights during analysis.
190
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Scenario 4: • First stage: selection of clusters with equal probability • Second stage: subsampling
Pi = (m/M) * (ni/Ni) where
Pi = probability that an individual in cluster i was selected for the survey m = number of sample clusters selected Mi = expected measure of population size for cluster i M = total number of sample clusters in the sampling frame ni = number of subpopulation members selected in cluster i Ni = total number of subpopulation members in the cluster i This design results in a non-self-weighting sample; thus it is necessary to apply sampling weights during analysis.
Step 2: Create the sampling weights specific for the sampling design After the appropriate sampling probability for the survey design has been calculated (Step 1), it is converted to a sampling weight as follows: where
probabilities, weighted analysis for RDS differs from the analysis for other sampling methods in several important ways: • it requires specialized software2 to analyse the data. • some estimators (e.g. RDS I and RDS II) take the RDS recruitment matrix into account; thus sampling probabilities (and, therefore, the weights) are generated for each variable and each record analysed by the software. For example, the sampling weight for a young female when analysing the “age” variable differs from the sampling weight when analysing the “sex” variable. The Gile SS estimator produces one weight that is applied to all variables for a single record. New estimators, with differing strengths and weaknesses, are being developed. Hence, the literature should be continually reviewed to find the appropriate estimator for each dataset.
wi = 1/Pi
wi = sampling weight in the ith cluster Pi = probability of selection in the ith cluster. Weighting for respondent-driven sampling Respondent-driven sampling (RDS) data are weighted using information from the participants’ self-reported social network size (or degree) and recruitment patterns. Because this weighting uses information from the recruitment process to calculate the sampling
Calculating response rates and weighting data There are many ways to calculate response rates for CCS and TLS. The same method should be used for all sampling events in the survey. The method should be clearly explained in all reports or publications. Some methods are: • number who take eligibility screener ÷ number approached; • number who take eligibility screener ÷ (number approached × % eligible among those taking the eligibility screener); • number who take survey ÷ number approached; and • number who take survey ÷ (number approached × % eligible among those taking the eligibility screener). Weighting in TLS is a complicated process: 1. The weight (1 ÷ probability) for a person at a given event is: [(pre-event count + entry count) × (eligible proportion among those taking screener)] ÷ (number interviewed) 2. Then, to generalize to venues, multiply those weights by: The number of eligible venue–day–times (VDTs) the event’s venue has ÷ the number of events at which sampling occurred. A venue with many VDTs would receive more weight than a venue with few VDTs. 3. Modify each person’s weight proportionally by their venue attendance frequency.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
191
Figure C-1.1 Probability of selection
• Person A: 10 ways to be recruited • Person B: 2 ways to be recruited • If all else is equal, Person A is 5 times more likely to be sampled than Person B
Person B Person A Source: Personal communication from Lisa Johnston, based on work done in 2012 at University of California, San Francisco
• there is no pre-existing information about the population (e.g. the sampling frames in cluster sampling) used in the weighted data analysis. Complex analytical techniques are used on the sample data to generate the sampling probabilities after the survey recruitment is finished. Social-network size (degree) The probability of selection is based on each participant’s social-network size. This is measured by the number of people the participant knows who fulfill the eligibility criteria for the survey, and whom they have seen in a predefined period. Each participant’s data are weighted by the inverse of the network size. Those with a small social-network size have a higher weight and those with a large social-network size have a lower weight. The response to the question of network size acts as an estimate of multiplicity: it is a proxy for a measurement of the number of persons who might be able to recruit the participant. Individuals with larger network sizes are more likely to receive a coupon (see Figure C-1.1); hence, their data are given a lower weight. Individuals with a smaller network size are less likely to receive a coupon; hence, their data are given a higher weight. Recruitment patterns Participants might be more likely to recruit people similar to themselves, thus introducing additional differences in the chance of being recruited for the survey. The coupon system in RDS is designed for researchers to trace who recruited whom in an RDS survey. This information is used to create weights – referred to below as “recruitment weights”– that adjust for possible selection bias from recruitment preferences of individuals and different preferences among groups. Assess the sample
for convergence and bottlenecks during implementation and after data collection has ended. The website of the Hard-to-Reach Population Methods Research Group3 website or the WHO/UNAIDS guide to using RDS Analyst (5) can help users learn how to use the tools in this software that assess convergence in RDS Analyst. 1. Recruitment effectiveness. This component of the recruitment weight adjusts for the differences between groups in the probability that a person with a certain characteristic will recruit someone, regardless of the characteristic of the person they recruit. Groups with better recruitment effectiveness might end up being overrepresented in the sample (6), which happens frequently when the population is similar or has high homophily (i.e. a tendency for individuals to know and socialize with people who have similar characteristics). 2. Differential recruitment between groups. This component of the recruitment weight adjusts for the probability that people with a certain characteristic will be recruited by someone with a certain characteristic, which may or may not be the same as that of the recruiter. The weighting attempts to account for the probability that someone will be recruited, based on the characteristic of the recruiter. These recruitment patterns are usually the result of homophily. For example, young people may have closer social-network ties among themselves than with older people, and therefore may differentially recruit other young people. Homophily is a diagnostic statistic that describes the mixing patterns in networks and is calculated by RDS software. In RDSAT, homophily can be positive or negative, ranging from –1 to 1 (7). In extreme cases, separate RDS samples within homophilous groups can be used (8). Ideally, investigators should monitor homophily during data collection. If homophily is deemed
2 3
For example, RDS-Analyst (http://hpmrg.org/software/ ) or RDSAT (http://www.respondentdrivensampling.org/). http://www.hpmrg.org/
192
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
high, investigators can improve recruitment training to encourage participants to recruit randomly from their networks, or can add new seeds. The homophily statistic also helps explain why the point estimate has a large CI if the homophily is high: • homophily = 1: always recruits from own group; • homophily >0 to 1: preferential recruitment of group with similar characteristics; • homophily = 0: no preferential recruitment; • homophily –1 to <0: preferential recruitment of groups with different characteristics; and • homophily = –1: always recruits from other than own group. Many analysts consider a value of between –0.3 and 0.3 as “normal” homophily. If homophily is outside this range (i.e. lower than –0.3 or higher than 0.3), differences between the unweighted sample data and weighted estimates may be more prominent. In RDS Analyst, homophily hovers around 1, indicating no homophily. A value of 1.3 or greater is considered high homophily (personal communication with Mark Handcock, University of California at Los Angeles, 26 January 2016). Differential recruitment among groups is usually limited to socially salient characteristics, that is, characteristics that can be observed in others and upon which people form friendships. For example, high homophily (above 0.3) is commonly observed in estimates of preferred drugs among people who use drugs, because they form friendships and drug-acquisition networks around the specific drugs they use. Conversely, homophily close to zero is observed in characteristics that are either not visible or not important for forming relationships, such as month of birth. Weighting for recruitment patterns makes RDS analysis very different from analysis of data with other sampling strategies. Several estimators can be used to create RDS weights, with more being developed. Some estimators, such as Gile’s Successive Sampling Estimator, create a single weight for each participant. Some, such as RDS I and RDS II, create a weight for each variable and each participant. Therefore, because the recruitment weights are based on the characteristics of the recruitments among groups, each record in a dataset will have a different weight for each of the variables analysed. For example, in a survey report with 500 respondents that has 100 indicators, 100 weight variables will be generated automatically by the specialized software (one for each indicator in each record).
C-1.3 Analytical methods for BBS data Currently, BBS are used to observe associations between the outcomes and behaviours or characteristics in the survey population. BBS data may be analysed within a single survey round, or across three or more rounds in the form of trend analysis, if methods and sampling approaches are similar among those rounds. Ideally, RDS survey results should be compared with subsequent RDS survey results in the same population with the same tools (the same is true for TLS). To be comparable, surveys must have been performed in similar locations and populations. However, if the different survey techniques (RDS and TLS) have been used correctly on the same sampling population, it should still be possible to assess trends, because both will simulate simple random sampling (SRS). Whatever the technique used, trends must be handled and interpreted with caution. The analyses described here require statistical software packages that allow analysts to account for complex survey designs with components such as primary sampling units (PSUs), stratifying variables and weights to generate appropriate estimates. Most survey reports present results as weighted means for continuous variables (e.g. age) or weighted proportions for categorical variables (e.g. marital status), with visual analysis (graphics) of variable distributions or HIVprevalence trends over time. Manuscripts might answer specific research questions that require more complicated statistical analyses (e.g. regression) that are beyond the scope of these guidelines. In a survey report, the unweighted denominator, weighted mean or proportion, and CI (e.g. 95% CI) for each variable should be reported in summary tables (9). Table C-1.2 is an example of a summary table of sociodemographic and risk characteristics. The design effect, calculated from the survey data, can also be presented for the main variables. This enables readers to compare the effective sample size used to a simple random sample.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
193
Table C-1.2 Example of a summary results table of sociodemographic and risk characteristics Characteristic n (unweighted) Sample % (unweighted) Estimated population % (weighted) 95% CI Design effect
Age (years) 15-19 20-24 25-29 30-34 35+ Etc. CI, confidence interval; n, sample size
A note on denominators Denominators must be well defined and adjusted when certain sequences of questions are asked, based on previous responses. In surveys, some questions are designed to prompt follow-up questions. These questions are referred to as “filter” questions and are often dichotomous (i.e. they have only two response values such as “Yes” or “No”). If a participant responds “Yes” to a filter question, this prompts one or more follow-up questions relevant only to those with “Yes” responses, acting as a type of filter. Conversely, a “No” response skips these follow-up questions because they are not appropriate for the participant based on previous responses, and the interviewer moves on to the next survey question. The term “skip pattern” describes this pattern of questions and responses, and provides information on denominators for each question. The resulting filter creates subsets of the data based on specific criteria (responses). Once applied, only participants who meet those criteria should respond. For filter questions, the denominator should be adjusted to reflect the participants and the criteria, depending on the intent of the analysis. Generally, respondents with missing values are excluded from the denominator to compute percentages. For example, if one wants to estimate the proportion of female sex workers (FSW) who used a condom the last time they had anal sex, the analyst might want to keep in the denominator only those who reported ever having had anal sex. However, in certain situations, it can be more meaningful to include participants with missing values in the denominator. For example, a survey might use the filter question “Have you ever been to school?” Participants who replied “No” are then excluded from the next question about the highest school level completed. When computing the proportion of respondents who completed secondary school or higher, the analyst should include in the denominator those who never went to school. Biomarker data can also present denominator challenges. One such example is indeterminate HIV test results. Such results should be analysed as their own category (e.g. positive or reactive, negative or nonreactive, or indeterminate) and should be included in the denominator. Another analytical challenge is presented with viral load, a clinical measure of the number of viruses in a person’s blood. Viral load is usually reported as copies of HIV/mL of blood, and an aggregate viral load (across all tested participants) can be calculated for the sample. However, when the viral load is below the threshold needed for detection by the specific test used, the result is classified as “undetectable” (e.g. <50 copies/mL with some current tests). Undetectable viral-load results need to be included in data analysis of viral load; the use of an imputed value representing the mid-point between zero and the threshold of detection is recommended (e.g. using 25 if the threshold of detection is 50 copies/mL). Alternatively, and more conservatively, the threshold value itself may be used for imputation (50 copies/mL, in this example). Depending on the survey objectives, viral load can be reported as an arithmetic mean, a geometric mean (mean logarithm of the viral load), a median, or categorically with several epidemiologically or clinically meaningful categories. For more information on measuring and analysing viral load, see the 2017 Global AIDS Monitoring guidelines (10).
194
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
C-1.3.1 Univariate analysis Descriptive statistics summarize the sample, but do not test any hypotheses. They include measurements of centre such as means and medians, with corresponding measures of uncertainty (spread) around the centre. The standard deviation measures the precision of the mean, the variance measures how far a data point deviates from the mean, and the range represents the lowest value to highest value. The distribution of continuous survey variables, such as age or number of sex partners, can be reported as weighted means with 95% CI. A proportion indicates the size of a part compared to the whole, with values falling between 0 and 1. Percentages, which are proportions multiplied by 100, are one way of expressing a proportion; for example, HIV prevalence. In survey reports, the distribution of categorical values are presented as weighted proportions with a 95% CI. The numerator is the count of participants with a specific characteristic (e.g. a positive HIV test result). The denominator includes all participants in whom the characteristic was measured (e.g. number tested for HIV). The counts in the numerator and denominator are multiplied by the appropriate weight (or weights).
may be stratified by sex (male or female), age group (<24 years, 25–34 years or 35+ years) or residence (urban or rural). Stratified analysis may be used to examine a subpopulation of interest, or as a method of controlling for a particular characteristic that might be a confounder. A confounder is a variable that correlates with both the independent (explanatory) variable and the dependent (outcome) variable. When stratum-specific odds ratios are relatively similar, it is acceptable to pool results, but when the ratios differ, results must be presented separately. Multivariable analysis Often, it is necessary to simultaneously control (adjust) for multiple confounding variables. If stratified analysis were used for this purpose, some strata would contain very small numbers, thus reducing the power to detect any associations. Multivariable analysis, or statistical modelling, is a method used to control for multiple confounders simultaneously. Multivariable regression can be performed after exporting the RDS data to another software program; some programs (e.g. RDS Analyst and Stata), which can use RDS analytical tools, can directly perform the regression. Multiple logistic regression, a type of multivariable analysis used with dichotomous outcomes, is one of the most common statistical operations performed with BBS data to generate adjusted odds ratios (measures of association between outcome and multiple explanatory variables) with 95% CIs. A consultation of statistical experts convened at the Centers for Disease Control and Prevention, Atlanta, Georgia, in 2012 recommended applying sampling weights that are created for RDS analysis during multivariable regression when analysing RDS data. Standard procedures for analysing sample survey data can be used to perform regression analysis of RDS data. The seed or shared recruiter should be specified as the primary sampling unit (i.e. cluster variable). Ideally, regression analysis should be performed using exponential-family random network models. These are computationally intense and not for the average user. The ERNM models can provide new approaches to network regression and multilevel modelling. Models include joint modelling of Y and X, exponential random graph models; Gibbs measure and conditional modelling of Y given X (personal communication from Berchenko Y, Frost S, Gile K, Handcock M, Heckathorn D, McFarland W, Salganik M, Szwarcwald C, Thompson S, White R on 22 August 2012). Other analysis methods More advanced investigations of repeated cross-sectional survey data include age–period–cohort analyses, which are based on the theory that cohort membership influences behaviour to a similar degree as other demographic factors. Other cohort or time-dependent statistical techniques that have been used to analyse repeated cross-sectional surveys include pooled logistic
C-1.3.2 Bivariate analysis Bivariate analysis assesses the relationship between two variables. Typically, one of the variables is considered a “dependent” (outcome) variable (e.g. HIV status); the other is considered an “independent” (explanatory) variable (i.e. a characteristic, behaviour or exposure that explains the change in the dependent variable). Bivariate analyses of simple “yes/no” explanatory and outcome variables can be reported as weighted proportions, standard errors and 95% CIs. The Chi-square test measures the relationship between two categorical variables (e.g. a 2 × 2 contingency table). Statistical tests are accompanied by a P-value, which is the probability of an event given the null hypothesis is true. Chi-square tests the odds or risk of an event, given one or more independent variables; results are reported as crude (unadjusted) odds ratios or relative risk (measures of association between an exposure and an outcome), with associated 95% CIs. Logistic regression is another method of testing relationships between a dichotomous dependent and one or more independent variables while controlling (or adjusting) for potential confounding variables. Logistic regression also yields adjusted odds ratios with 95% CIs. Linear regression may be used with continuous outcomes. Stratified analysis Stratification is the classification of a survey population into subgroups or categories (“strata”) on the basis of selected characteristics. For example, a survey population
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
195
regression, Cox proportional hazards regression, KaplanMeier curves, multilevel/hierarchical modeling, and generalized estimating equations to measure estimates over time. These analyses are beyond the scope of these guidelines; seeking assistance from a statistician is recommended when investigating these complex analytical approaches.
C-1.5 Comparing surveys Examining differences between two surveys Investigators commonly want to determine whether there are statistically significant differences between estimates (either percentages or means) from two surveys. If the surveys have independently selected samples with similar populations, and if the methods and instruments are alike, results from the two surveys can be used to formally test for differences in estimates, under certain assumptions. This comparison is often performed using a Z-statistic directly calculated from the point estimates and the associated CIs. To illustrate, consider the case in which we have conducted two RDS surveys at two different times, using the same methods and data collection instruments for the same populations. Data analysis of the first survey provides an estimate of HIV prevalence of p1 with a (1 − a)% CI of (l1, u1) after accounting for the weights and sampling design. Similarly, data analysis of the second survey gives an estimate of HIV prevalence of p2 with a CI of (l2, u2). If the CIs are roughly symmetric about the point estimates, then the standard errors (se) and variances (var) for the estimates are approximately:
C-1.4 Combining multiple surveys and subsetting data for analysis Combining multiple surveys In some situations, investigators need or want to combine data from multiple survey locations or periods to produce a single set of estimates. Combining data from multiple surveys should be performed with caution and only done when measures in each survey are comparable. For example, three cities in a country might have implemented surveys simultaneously and investigators want to combine survey data from each city for a national estimate. If combined without accounting for the unique characteristics (e.g. variance estimates) of the three surveys, the resulting estimates will be inaccurate. Thus, for each BBS sampling design, weighting needs to be adjusted to account for how the data will be combined (e.g. distinct geographical locations or periods have different variance estimations). Interpretation must consider that the final estimates are a combination of distinct samples with their own characteristics. Local data combined to produce national estimates need to be identified as such, rather than presented as nationally representative estimates. Subpopulation estimates There are situations that require estimates and inference for subpopulations. When performing a subpopulation analysis, the entire survey dataset should be retained to keep the integrity of the sampling-design information in order to correctly calculate the standard errors of the estimates. For example, if analysts want to report results for each step of the HIV-care cascade, they might select the subpopulation of survey participants with laboratory-confirmed HIV and then perform analyses on only those participants, rather than on all survey participants. Special commands for these subpopulation analyses are available in survey software packages and must be used on the entire survey dataset to maintain the integrity of weighting adjustments made on the entire survey dataset. Creating new datasets with only the subpopulation is not recommended because it will alter the weighting and may produce incorrect estimates, CIs and P-values.
where Z1−a/2 is the 1 − a/2 quantile for the standard normal distribution (e.g. for 95% CI, a = .05 gives Z.975= 1.96) The Z statistic for the difference in the survey estimates p2−p1 is calculated as:
The p-value for the Z-statistic is calculated using the quantile function of the standard normal distribution. If the (1 − a)% CIs of the point estimates of the first and the second surveys do not overlap, then the difference between the estimates is necessarily statistically significant, because the (1 − a)% CI about the difference between the two point estimates will not contain zero. However, even if the CIs overlap, the Z-statistic, as calculated above, may still be statistically significant.
196
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Table C-1.3 Example of a table for an analysis of differences between two surveys HIV Status Positive Negative
Survey A (2012) n=53 39 14
Survey B (2015) n=55 30 25
p1= 39/53 = 0.736
p2=30/55 = 0.545
For example, let HIV status results in Survey A (2012) be 39 positive and 14 negative, and in Survey B (2015) be 30 positive and 25 negative (Table C-1.3). To test for a significant difference in HIV-status estimates between 2012 and 2015, the Z-statistic is equal to 2.0593, which gives a p-value of 0.0394. Small p-values (e.g. p <0.05) indicate statistically significant differences between the surveys for the characteristics (point estimates). If we are estimating means rather than proportions, the procedure is similar, except that it uses the estimated mean from the two surveys, m1 and m2, and their CIs instead. Note that this procedure is only approximate and assumes that the CIs are symmetric about the point estimates and that the two surveys are independent (e.g. done at different times). If these assumptions cannot reasonably be made, then a statistician should be consulted about more appropriate methods for formally testing for significant differences.
Trend analysis Trend analysis measures changes in weighted BBS estimates among three or more points (surveys) in time. Comparisons should be made only when measures and methods are the same in all surveys; best practice in BBS calls for repeated surveys using the same instruments in identical populations. Most trend analyses should be conducted by producing weighted estimates for each variable for each round of a survey, and then determining whether there is a difference between rounds (Table C-1.4). Visual trend analysis of comparable data (e.g. weighted mean age of participants or HIVprevalence estimates) over time can be presented with 95% CIs (Figure C-1.2). Testing for statistical differences in comparable data over time can be performed using regression methods, with the guidance of a statistician. Combining results from surveys with different sampling strategies (e.g. RDS and TLS) or population definitions (e.g. SW and transactional sex) should be avoided.
Table C-1.4 Example of a table shell for an analysis of differences between two surveys Characteristic Survey A (2012) N = n HIV status Positive # positive % positive of all tested, p1 (95% CI) % negative of all tested (95% CI) # positive % positive of all tested, p2 (95% CI) % negative of all tested (95% CI) % (95% CI)
Survey B (2015) N = n % (95% CI)
Difference (p-value)
Negative
# negative
# negative
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
197
Figure C-1.4 Example of visual trend analysis of HIV prevalence with 95% confidence intervals from four rounds of BBS in the same population at Location X, 2005–2013 30
HIV Prevalence (%) with 95% Confidence Intervals
20
10
0
2005
2008
2011
2013
Year of BBS round
Table C-1.5 provides a comparative overview of the analytical approaches for BBS discussed above. Table C-1.5 Overview of common analytical approaches for BBS Approach Univariate analysis Bivariate analysis
Description Frequency of a variable value or indicator relative to one denominator Frequency of a variable value or indicator relative to two denominators (e.g. condom use at last sex among HIV+ MSM and HIV– MSM)
Purpose To describe the distribution of a characteristic, behaviour or outcome in a population To compare the distribution of a characteristic, behaviour or outcome in two groups; or to measure the association between two variables
Example 12.5% of 800 MSM surveyed were HIV+; the mean age of sex workers surveyed was 26 years Frequency of HIV infection was similar (P >0.05) among 200 women (18.3%) and 350 men (17.9%) in the sample; HIV+ MSM were 2.3 times more likely to report inconsistent condom use than were HIV– MSM (OR 2.3; 95% CI:1.9, 2.6) Among MSM aged <25 years, the odds of being HIV+ were 2.1 times higher in those reporting inconsistent condom use than in those reporting consistent condom use; whereas among MSM aged >25 years, the odds of being HIV+ were 4.2 times higher in those reporting inconsistent condom use After adjusting for age, marital status and urban/rural residence, HIV+ MSM were 5 times more likely to have an STI than were HIV–MSM The HIV prevalence among MSM in City X increased significantly over BBS rounds 1, 2 and 3 The HIV prevalence among MSM in City X increased significantly over BBS rounds 1, 2 and 3
Stratified analysis
Bivariate analysis for two or more strata or categories of a third variable (e.g. a subpopulation such as age group, marital status, or urban or rural residence); allows assessment of the measure of association in different strata.
To compare distribution of a characteristic, behaviour or outcome in two groups by multiple strata of a third variable (potential confounder)
Multivariate analysis
With dichotomous outcomes, use logistic regression to analyse potential association with multiple variables and indicators Comparison of a characteristic, behaviour or outcome variable or indicator over time or between two surveys Changes in same characteristic, behaviour or outcome variable or indicator over time
To assess independent associations between outcome (dependent) and exposure (independent) variables To assess whether significant differences exist in outcomes or characteristics over two or more BBS rounds To assess whether significant differences exist in outcomes or characteristics over three or more BBS rounds
Comparing surveys
Trend analysis
OR: odds ratio; STI, sexually transmitted infection.
198
1. Data analysis
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
C-1.6 Suggested resources University of California, Los Angeles (UCLA), UCLA stats: http://www.ats.ucla.edu/stat/ RDSAT 7.1 user manual: www.respondentdrivensampling.org Harvard summary of survey analysis packages: http://www.hcp.med.harvard.edu/statistics/survey-soft/
C-1.7 References 1 Global Health Sciences. Toolbox for conducting integrated HIV bio-behavioral surveillance (IBBS) in key populations. San Francisco: University of California; 2016 (http://globalhealthsciences.ucsf.edu/pphg/gsi/epidemiologicsurveillance/ibbs-toolbox, accessed 30 August 2016). CDC. National HIV behavioral surveillance system: men who have sex with men – Round 4: operations manual. Atlanta, GA: Centers for Disease Control and Prevention (CDC); 2014 (http://www.cdc.gov/hiv/pdf/statistics/ systems/nhbs/nhbs-msm4-operations-manual--version-date-05-16-2014.pdf, accessed 4 August 2016). Gardner MJ, Altman DG. Statistics with confidence. London, BMJ. 1994. Aiken LS, West SG. Multiple regression: testing and interpreting interactions. Sage Publications. 1991. WHO/UNAIDS. Introduction to HIV/AIDS and sexually transmitted infection surveillance: module 4 supplement. A guide to using RDS Analyst and NetDraw. World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS);(http://applications.emro.who.int/dsaf/EMRPUB_2014_EN_1686.pdf, accessed 5 November 2016). Erickson BH. Some problems of inference from chain data. Sociol Methodol. 1979;10:276–302. McPherson M, Smith-Lovin L, Cook JM. Birds of a feather: homophily in social networks. Annu Rev Sociol. 2001;27(1):415–444. Wejnert C, Pham H, Krishna N, Le B, DiNenno E. Estimating design effect and calculating sample size for respondent-driven sampling studies of injection drug users in the United States. AIDS Behav. 2012;16(4):797–806 (http://www.ncbi.nlm.nih.gov/pubmed/22350828, accessed 8 August 2016). White RG, Hakim AJ, Salganik MJ, Spiller MW, Johnston LG, Kerr L et al. Strengthening the reporting of observational studies in epidemiology for respondent-driven sampling studies: "STROBE-RDS" statement. J Clin Epidemiol. 2015;68(12):1463–1471.
2
3 4 5
6 7 8
9
10 UNAIDS/UNICEF/WHO Global AIDS Monitoring 2017: Indicators for monitoring the 2016 United Nations Political Declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2016. (http://www.unaids.org/sites/default/files/ media_asset/2017-Global-AIDS-Monitoring_en.pdf, accessed on 31 August 2017) 11 Heeringa SG, Brady TW, Berglund PA. Applied survey data analysis. Boca Raton, CRC Press. 2010. 12 Korn EL, Graubard BI. Analysis of health surveys. New York, Wiley. 1999.
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
1. Data analysis
199
Section
C
2. Use and dissemination of survey findings
Biobehavioural surveys (BBS) are only useful if their findings are used for action. BBS findings have many uses, ranging from programme planning and advocacy to guiding the development of future surveys. This chapter describes how to disseminate and use BBS data, including potential uses for BBS data, identifying a target audience, tailoring a message and determining a dissemination strategy. Investigators should decide early (during the planning phase) how the survey data will be used. Unless data use is planned before data collection begins, some information that might be of interest to decision-makers may not get collected. Example 2: a survey may find that sex workers (SW) who received peer-to-peer training are less likely to be victims of client-initiated violence than SW who did not receive the training. This finding could support expansion and increased funding of community-based training programmes. Various other studies have examined how surveys inform public-health action (1-3). Data from surveys with sufficiently large sample sizes (and sufficient numbers of respondents who are HIVpositive) may be of use in estimating the uptake of services along the continuum of HIV services, also known as the HIV prevention cascade and the “HIV cascade of care” (2, 4). The cascade displays the proportions (or percentages) of the participants who have been exposed to or have taken up certain HIV-related services, including having been offered HIV testing, enrolling in care, being in treatment, and having a suppressed viral load (Figure C-2.1). Surveys with probability samples may even enable estimation of such uptakes on a population level. Further, because population-based surveys sample individuals who do and do not use services, these data can be used to characterize those who are outside the continuum of care. UNAIDS has a goal of ensuring that, by 2020, 90% of people who are HIV-positive know their status, 90% of all people diagnosed with HIV are receiving sustained antiretroviral therapy, and 90% of all people on treatment have suppressed viral load. High-quality data on these indicators for key populations will help target efforts to ensure the UNAIDS goal is met. Advocacy Survey findings can support advocacy for increased funding, programmatic expansion and legal protection, and highlight issues related to stigma and violence. Community-based organizations are often particularly effective users of survey findings. Helping organizations to understand survey findings enables them to take ownership of the data, and to use them in workshops or funding applications.
C-2.1 Potential uses for BBS data Planning HIV programmes BBS using probability-based samples provide populationlevel estimates of target population demographics, risk behaviours, burden of disease, and service uptake to inform HIV programmes. BBS data may be used to estimate the proportion of the target population that has been exposed to a particular intervention or service. Survey data can also play a role in estimating programme access, coverage and acceptability (for suggested data measures, see the questionnaire module on services uptake and indicators in Sections II and III of the Supplemental Materials). Proven interventions that have been underused may be expanded. Measures of interest that appear unchanged despite efforts may need a new approach, perhaps one that pays more attention to the cultural or social context that determines why people behave in a particular way. Additionally, BBS data can be used to identify areas where subpopulations need more support or to identify successes in working with a population. Examples of the use of BBS data Example 1: a survey among people who inject drugs (PWID) may estimate the proportion who were tested for HIV in the past 12 months and, among those who tested HIV positive, the proportion enrolled into care. The survey data may inform investigators on factors associated with PWID who are HIV-positive enrolling into care (e.g. through outreach testing services or the availability of care services tailored to PWID). These findings may indicate that expanding such services could increase the proportion of these people enrolling into care. 200 2. Use and dissemination of survey findings
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Figure C-2.1 Schematic of a continuum of HIV care and treatment services populated through survey data
Not reached, not offered testing Undiagnosed Not linked to care Universe of HIV-infected population Reached out to, offered or sought testing Did not initiate ART Aware of HIVpositive status Not retained in ART Linked to care Initiated ART Retained, unsuppressed Retained in ART viral Suppressed viral load
ART, antiretroviral therapy
Presented appropriately to donors and international agencies, survey findings often play an important role in advocating for increased resources for activities that are not being adequately covered in government spending plans. The government health leadership or other government entities may implement policy changes based on new survey findings. Inform future surveys Previously completed surveys offer a wealth of practical experience that is invaluable for planning the next survey. For example, planners can learn about the sampling efficiency of the chosen design, whether the level of compensation was appropriate, and the usefulness of the collected data and laboratory measures.
Target audiences for BBS vary widely but should always start with the people from whom the data were collected: the survey’s target population. Other audiences include programme planners and policy-makers, service providers, donors, nongovernmental organizations (NGOs) and community-based organizations, as well as the wider scientific community. Target population of the survey Representatives of the surveyed population should be among the first to be informed about the survey’s findings in order to facilitate community-based organization work and advocacy, and to enable members to make informed choices about community-level efforts for HIV prevention. In addition, the target group might help the researchers interpret results and help plan the dissemination to other stakeholders. They may also have insight on how other stakeholders will use the results. When disseminating the survey’s findings, the language used should be appropriate and without epidemiological jargon. Service providers Surveys can inform service providers about their overall reach and their reach concerning particular subgroups, and how they might improve access, acceptability and quality of services. Surveys may also inform service providers about the characteristics of populations whom they are failing to reach. Policy-makers Policy-makers have many pressing and competing priorities, and investigators who inform them should adjust the focus of the presentations to the needs of and level of detail required for the specific policy-makers.
C-2.2 Identifying a target audience Target audiences vary in their scientific (or technical) literacy and interest in survey findings, and in their use of those findings. Therefore, the content of the message should be tailored to the intended audience. Some questions to consider before tailoring a message are: • who is the target audience? • what does the target audience already know about the subject? • what does the target audience need to know? • which findings are the most interesting or important for this audience? • what are the specific messages for the target audience? • what questions will the target audience have? • how will this audience use the information from the survey? • what is the best way to present and discuss the information with this audience? C. Data analysis and use B. Survey implementation & quality assurance
A. Survey preparation
2. Use and dissemination of survey findings
201
Sometimes policy-makers will consider issues such as funding priorities, the enforcement of laws on the ground and treatment eligibility based on the behaviour of atrisk groups. Policy-makers also need information on the absolute burden of disease and the need for services (e.g. population size of target population and estimated number requiring services). Scientific community Survey findings published in peer-reviewed literature have the widest reach among the scientific community. Publications reviewing survey findings on a regional or global level often include only surveys published in journals indexed in databases (e.g. PubMed and Medline). Hence, any survey of acceptable quality should be published to ensure its lasting place in the scientific literature. Technical reports targeting a scientifically trained audience should also focus on methodological issues such as sampling design and statistical tests used. Media and the public Journalists from newspapers, radio and, occasionally, television can be important for communicating a survey’s main findings. Investigators of surveys among vulnerable populations facing little or no stigma (e.g. students and transport workers) may choose to “go public” without putting the surveyed population at risk; however, stigmatized key populations should be protected from adverse media attention. For surveys of stigmatized populations, investigators should consult first with community-based organizations about the risks and benefits of briefing the media. When briefing, they should use simple language and focus on the survey’s findings, their interpretation and recommendations, and should avoid any messages that could further stigmatize the survey’s target population.
The following are examples of target audiences and targeted information: • a ministry of health will need to know which subpopulations are at greatest risk for acquiring HIV, and the population proportions undiagnosed and not enrolled into care and treatment, in order to prioritize the concept of “treatment as prevention”. • a community-based organization for SW will want to know about violence, stigma and access to services. • donors will be interested in how uptake of services and treatment coverage for a particular key population compare to those in the general population in the same locale. As a specific example, findings from an MSM survey that might be shared with policy-makers include the proportion of MSM who are HIV positive and who are on treatment, and the proportion of MSM who are bisexual and having sex with both men and women. These data might be compared to treatment uptake in the general HIV-positive population; the data may show that MSM are less likely to be on treatment and that these MSM have female sex partners who are therefore at increased risk. Together, these data can make a compelling case to policy-makers to see the HIV epidemic among MSM not in isolation but as part of a larger social and sexual network that facilitates HIV transmission among both MSM and the general population. Officials can draw conclusions from the data presented and suggest specific action to improve policies for HIV prevention and care.
C-2.4 Determining a dissemination strategy After the target audience and message have been determined, the next step is to decide how to deliver the message. As discussed below, dissemination methods include: • technical reports • indicator reporting • press releases • issues briefs and fact sheets • presentation slide sets • manuscripts • information cards for distribution to members of the population Technical reports A technical report is the most common information product of a BBS, but is rarely indexed in searchable databases. The report serves as a reference in subsequent publications and ensures that a record of the survey is publicly available. A technical report contains a complete description of the survey: the B. Survey implementation & quality assurance
C-2.3 Tailoring a message The careful selection of relevant data and its presentation in the appropriate way are fundamental to the audience’s acceptance and use of the findings. Different audiences are at different “stages of change” according to some health-behaviour models that encourage public-health workers to identify the attitudes and knowledge of the audience, and to tailor the message appropriately. If the audience is unknowledgeable about the target population, but is receptive to the idea of assisting the population, then more data on the situation and how to help them would be appropriate. If the audience is suspicious or hostile to the population, then care is needed in deciding what information to provide. In the latter case, it may be appropriate to focus on a message that humanizes the key population, or the benefit to the general population of helping this key population.
202
2. Use and dissemination of survey findings
A. Survey preparation
C. Data analysis and use
formative assessment findings; recruitment methods and survey domains; data-analysis approach; and findings, including a presentation as data tables and graphs. The report should also provide contact information for the authors and the owners of the final dataset should additional analysis be necessary. Technical reports usually present relatively simple descriptive statistics, including numerators, denominators and the prevalence and distribution of specific characteristics. They can provide bivariate analyses of the primary outcomes. A report should focus on indicators that track risk, prevention efforts, or both, and include recommendations for policy and service delivery. The more user friendly the presentation of the data, the more likely it will be used. The technical report should be released at a workshop to ensure stakeholders are aware of the data. See Appendix I-34 for a suggested table of contents for a hypothetical survey report. Indicator reporting Survey results provide the data for standardized reporting to national governments, United Nations (UN) agencies and donors. Checking during survey preparation whether the data instruments will provide the appropriate data for reporting is good practice. The standardized questionnaires provided in these guidelines include questions that meet reporting requirements for UNAIDS, the United States President’s Emergency Plan for AIDS Relief (PEPFAR), and the Global Fund to Fight AIDS, Tuberculosis and Malaria (the Global Fund). Section 3 of this document provides a list of common indicators. Press releases and interviews The mainstream media should be informed about a survey only if investigators can be assured that the safety of the population surveyed is not compromised. Ideally, investigators should discuss with representatives of the target population the benefits and risks of informing the public about the survey’s findings. If informing the public is in the interest of the target population, consider a press conference or a workshop aimed at journalists, who appreciate a prepared short summary with takeaway messages that facilitate their work and improve the accuracy of their reporting. Such press releases should contain minimal information on the methods used in the survey and focus on the findings, what they mean, and what to do about them. For interviews, predetermined communication objectives and prepared statements should be used. Whatever the question, conveying the main points in interviews is vital. See Appendix I-35 for a sample press release. Issues briefs and fact sheets The minister of health may not have time to read a 50page report about HIV-related behaviour that includes information on sampling methodology and statistical tests, but may well be interested in learning whether the country’s overall strategic plan for AIDS control is
working in a particular high-risk group. The issues brief may highlight gaps in service provision, or successes in the uptake of particular programmes such as condom use or HIV testing. Such information should be packaged with information from other sources into a one-page or two-page brief that makes a compelling case for more HIV-prevention activities in a particular at-risk population. Slide presentations Slide presentations are helpful for presenting information in person to a group of people. Manuscripts Manuscripts published in peer-reviewed journals target the scientific community and create a lasting record of the survey. Most journals are indexed; therefore, readers can easily find the manuscript through scientific search engines. Publishing important findings in a peer-reviewed journal contributes to the wider knowledge base about the target population and the state of the HIV epidemic. Presenting to challenging audiences One type of audience deserves special attention: any group that has demonstrated the power to obstruct effective HIV-prevention efforts. Information prepared for these groups should take into account the concerns that form the basis of their opposition to prevention activities in the target population. For example, if religious leaders or politicians are opposed to HIV services for SW (e.g. condom distribution or training in negotiating safe sex), bombarding them with data showing that HIV prevalence is high or that poverty is often the root cause for taking up sex work is unhelpful. These data must be sensitively and strategically presented; for example, by showing that HIV services in this population do not promote sex work, and that prevention services benefit the families of SW and the wider population by lowering the risk of HIV transmission between SW and clients and their families. See Appendix I-36 for more information on data presentation for clarity and impact.
C-2.5 Additional considerations for BBS data interpretation Cause and effect Inferences (interpretations) drawn from BBS data should be made in consultation with technical staff (e.g. statisticians and epidemiologists) to ensure that data are interpreted correctly. Demonstrating that behaviours or HIV prevalence change after programme implementation is one of the most effective ways to increase support for prevention and other activities. However, BBS are cross-sectional surveys. Data from a series of BBS over
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
2. Use and dissemination of survey findings
203
time cannot be used to infer cause and effect in the effectiveness of interventions, only to present ecological2 observations.1 One example is a change in HIV prevalence; such a change over two or more BBS rounds may or may not indicate a change in HIV incidence. HIV prevalence is a function of HIV incidence and mortality. Increased uptake of antiretroviral therapy (ART) should result in decreased mortality. Thus, an increase in people who are HIVpositive and are receiving ART may result in an increased prevalence despite a decline in incidence. In addition, numbers in some high-risk groups may fluctuate markedly; for example, in SW numbers through a rapid turnover of women or men in the industry or migration into or out of a city. Such changes can lead to fluctuations in the observed HIV prevalence. Changes in the prevalence of some sexually transmitted infections (STIs) (e.g. syphilis, gonorrhoea and chlamydia) may reflect changes in both risk behaviours (unprotected sex) and the uptake of treatment services. Correlating condom use with HIV prevalence is problematic, because condom-use behaviour is probed over a short period whereas HIV infection is a marker of chronic disease. Moreover, people aware of their HIV-positive status may be more likely to use condoms than those who are HIV-negative. Presenting the estimated size of the key population Most, or all, methods of estimating population size are based on assumptions that are hard to meet and difficult to assess. Therefore, the validity of these estimates is often unknown. When presenting size estimates, it is best to always present the accompanying confidence limits or uncertainty bounds, to present the results
of any sensitivity analysis, and to discuss the possible limitations and biases, comparing them to previous or regional estimates. The estimates of population size should be rounded, because exact numbers imply a level of accuracy that cannot be achieved. Timeliness Survey data and findings lose utility over time; therefore, BBS findings should be presented and published as soon as possible. Guidelines for the reporting of data from surveys using respondent-driven sampling form a useful starting point for planning reports and publications (5). Delays in disseminating survey findings can lead to delays in programme improvement, expansion or re-direction. Early dissemination may entail releasing results for key indicators through, for example, a presentation within 1–2 months of data collection, while report-writing is occurring. However, investigators should ensure that the preliminary findings will remain valid and will not change subsequently. Information may be discounted by an audience if it appears out of date or contradicts earlier reports. A full report can be distributed after data analysis and write-up is complete.
C-2.6 Relevant websites Research utilization toolkit: https://www.k4health.org/ toolkits/research-utilization Eight strategies for research to practice: http://www. fhi360.org/resource/eight-strategies-research-practice
C-2.7 References 1 Emmanuel F, Adrien A, Athar U, Imran M, Reza T, Blanchard J. Using surveillance data for action: lessons learnt from the second generation HIV/AIDS surveillance project in Pakistan. East Mediterr Health J. 2011;17(8):712–718 (http://www.ncbi.nlm.nih.gov/pubmed/21977576, accessed 4 August 2016). Hladik W, Benech I, Bateganya M, Hakim AJ. The utility of population-based surveys to describe the continuum of HIV services for key and general populations. Int J STD AIDS. 2016;27(1):5–12. Sweeney P, Gardner LI, Buchacz K, Garland PM, Mugavero MJ, Bosshart JT et al. Shifting the paradigm: using HIV surveillance data as a foundation for improving HIV care and preventing HIV infection. Milbank Q. 2013;91(3):558– 603 (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3790525, accessed 4 August 2016). Kilmarx PH, Mutasa-Apollo T. Patching a leaky pipe: the cascade of HIV care. Curr Opin HIV AIDS. 2013;8(1):59–64. White RG, Hakim AJ, Salganik MJ, Spiller MW, Johnston LG, Kerr L et al. Strengthening the reporting of observational studies in epidemiology for respondent-driven sampling studies: “STROBE-RDS” statement. J Clin Epidemiol. 2015;68(12):1463–1471.
2 3
4 5
1
Ecological studies are those in which the unit of observation is the population rather than the individual.
204
2. Use and dissemination of survey findings
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
C
3. Evaluating survey implementation
This chapter discusses survey evaluation, which is conducted after a survey has been implemented. Evaluation examines broad, overarching areas such as how well the survey was implemented, adherence to the protocol and standard operating procedures (SOPs), and the quality of survey and laboratory data.
Evaluation seeks to identify lessons that, if learned, can improve the next survey. In contrast, monitoring is the ongoing assessment of survey implementation, including sampling, recruitment and consent of survey participants, administration of data-collection instruments, data quality, specimen collection and processing, and referrals. Thus monitoring is conducted during a survey and results in immediate corrective action. Data from monitoring activities can also be used for evaluation. There are many different types of evaluation. To provide details on specific evaluation methodologies is beyond the scope of these guidelines. Table C-3.1 lists general evaluation topics that are useful for biobehavioural
surveys (BBS). Monitoring should have determined whether the concerns and questions listed occurred during a survey, whereas evaluation should determine how these items were or were not detected, their cause, and how they were addressed.
C-3.1 Data collection for evaluation Evaluation can be conducted internally (by the investigators) or externally by a third party. Data for evaluation can be generated qualitatively (e.g. obtaining feedback from field staff and stakeholders, or reviewing participant feedback) or quantitatively (e.g. examining data quality or reviewing test results).
Table C-3.1 Evaluation topics and related questions to consider for biobehavioural surveys Topic Sampling Recruitment and enrolment
Specific concerns and questions Sampling conducted according to protocol and SOPs Duplicate enrollees People posing as members of target population Fake or invalid coupons, coupon uptake (RDS) Inappropriate compensation, selling of coupons (RDS) Any respondents enrolled who were ineligible but received coupons (RDS) Missing interviews Partially completed interviews Identifying and rephrasing questions that: • were difficult to understand • led to refusals • led to extreme values
Data collection
Laboratory
Proportion of test results returned Proportion of test results delayed Missing and poor-quality specimens Poor-quality testing Insufficient supply of test kits Testing errors
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
3. Evaluating survey implementation
205
Topic Human subjects
Specific concerns and questions Breaches in confidentiality Informed consent procedures followed Time spent for survey participation Aborted interviews or survey procedures Actual vs budgeted data-collection costs Inclusion of external quality assessment to validate results (see Chapter A-6)
Participant burden and acceptance Data-collection costs Biomarker measurement
SOPs, standard operating procedures; RDS, respondent-driven sampling
Obtaining feedback from field staff and stakeholders Field staff are a valuable source of information because of their direct interaction with participants and their close involvement in survey activities. Together with the investigators, field staff can share what worked well, what challenges arose, and how they overcame those challenges. In addition to guiding planning for future surveys, this information can provide context for interpreting results. Stakeholders (e.g. ministry of health, donors and community-based organizations) should be asked whether the survey findings met their needs, and what additional data they want collected in future surveys. Reviewing participant feedback Participant satisfaction is important for a successful survey. Unhappy participants may discourage others from joining; thus, participant feedback should be monitored during survey implementation, reviewed again after implementation and synthesized with other data from other sources. Feedback may be gathered after participants complete the survey, or through interviews, a suggestion box or email. If such interviews are systematic, the procedure should be described in the survey protocol. Data-quality evaluation Data quality refers to the completeness, validity, consistency and accuracy of data. Data-quality assurance (e.g. checking for missing values, extreme values and logic errors) should form part of ongoing monitoring activities. Evaluation of data quality looks at overall data-quality issues. For example, if an evaluation finds considerable missing data, the reasons for these omissions must be determined (e.g. data-entry error, incorrect skip patterns, data manipulation or analysis error, or corrupt files), and how to avoid similar future occurrences. Poor data quality is a serious issue that may result in missing data or incorrect estimates of the
variables of interest. For guidance on assessing data quality, see Chapter B-2. Review of specimen test results At the end of a survey, investigators should determine the proportion of test results returned on time or delayed, the number of missing specimens, and the number of specimens in storage, if applicable. The evaluation should also examine the quality and quantity of specimens, the quality of test procedures, and the reporting of correct and incorrect test results. In addition, the evaluation should assess when and how problems were identified during the testing process, what corrective action was taken, when the action was taken, and the steps required to prevent similar problems in future surveys.
C-3.2 Using evaluation findings An evaluation report should be developed after the evaluation is complete. Findings from the evaluation should be shared with field-based and non-field-based survey staff. Survey staff must use the evaluation findings for action. Creating a table similar to Table C-3.1 may be useful, with an additional column on the right entitled “Recommendations for future surveys” that addresses each concern or problem identified during the evaluation.
206
3. Evaluating survey implementation
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
Section
C
4. Data sharing for public use
This chapter discusses the rationale for and means of providing third-party researchers access to individual-level data from biobehavioural surveys (BBS). Most BBS are publicly funded and data are collected from subpopulations of the public. Thus, the public has a right to be informed of the survey findings and, with certain safeguards, to have access to and use of the data. Sharing anonymized individual-level survey data free of charge is good public-health practice because it maximizes a survey’s utility. Data sharing allows other researchers to use the data in ways the original investigators may not have considered.
C-4.1 Data Sharing Sometimes a survey’s funding organization requires that survey data are made available to third parties. Examples of data sharing are rare for previous BBS, but best practices can be found in other survey areas. For example, data from the Demographic and Health Survey are routinely made available for interactive analysis or downloading from a website, because these data are from the US-based National Health and Nutrition Examination Survey (NHANES).1 2 Considerations when providing access to data Breaches of confidentiality and misuse of survey data may harm the target population and must be prevented. The ultimate responsibility for securing data and preventing misuse lies with the survey’s principal investigators. Individual-level survey data should be shared only if the risk of harm to the surveyed population can be minimized. This may require the vetting of potential data recipients by community-based organizations, and grouping data into larger categories to limit the potential for identification of participants. For example, participant age can be grouped into categories of 5–10 years. Survey protocols should indicate if survey data will be made available to other researchers. Regulations concerning human subjects may require the inclusion of similar information in the consent form. Investigators should determine whether the investigators’ or funders’ institutional regulations, or national policies, mandate, recommend or caution against data sharing. Knowledge
of national data-confidentiality regulations and standards should guide the decision to make data accessible to others (1). Data-use agreements should describe the requirements for access to and use of data. Preparing datasets for public use Investigators should clean raw data to make them error-free before sharing, or provide the raw data, cleaning code, and resulting clean data file, in order to increase the transparency of data cleaning and management. Secondary variables may be added to facilitate analysis. When making survey data accessible, reference documents should be included for context, including the data dictionary, the data instruments, the survey protocol, guidance for appropriate attribution, and investigator contact information. A data dictionary explains the response values and shows the variable labels; the data instruments show the full question-andresponse phrasing, as well as skip patterns; the survey protocol provides information about sampling design, eligibility criteria and other important design elements. Public-use datasets must be stripped of personally identifiable information; for example, dates of birth (day and month), photographic images, email addresses, phone numbers, street addresses, and social security or other identification numbers (IDs). Geographical information such as place of residence can be merged into larger area groupings or removed altogether. Reporting standards to guide third-party researchers may also be outlined in order to protect participant identity; for example, not presenting values for less than five participants in a table cell.
1 2
http://dhsprogram.com/Data/ http://www.cdc.gov/nchs/nhanes.htm
C. Data analysis and use
B. Survey implementation & quality assurance
A. Survey preparation
4. Data sharing for public use
207
Timing of data sharing Survey data should be made available as soon as possible. Some agencies may have a policy of releasing data a specific number of months after survey completion. Websites dedicated to data sharing Websites dedicated to hosting and sharing research and survey data enable third-party researchers to access data. Examples of such websites include the Princeton University Office of Population Research (OPR) Data Archive,3 Harvard University’s Dataverse Network,4 and the Inter-university Consortium for Political and Social Research at the University of Michigan.5 These websites include information about policies for accessing and using datasets, including the format in which the data should be made available, and which metadata and information should accompany them. Third-party researchers can then access the data by abiding by the data-sharing website’s rules and regulations. In most cases, a user must sign a data-use agreement in order to access the data; for some datasets, the user implicitly consents to the data-use agreement through the act of downloading the datasets. See Appendix I-37 for a sample data-use agreement.
C-4.2 References 1 Bernstein AB, Sweeney MH. Public health surveillance data: legal, policy, ethical, regulatory, and practical issues. MMWR Surveill Summ. 2012;61:30–34.
http://opr.princeton.edu/archive/ http://thedata.harvard.edu/dvn/ 5 http://www.icpsr.umich.edu/icpsrweb/ICPSR/ 3 4
208
4. Data sharing for public use
A. Survey preparation
B. Survey implementation & quality assurance
C. Data analysis and use
ISBN: 978-92-4-151301-2
Biobehavioural Survey Guidelines For Populations At Risk For HIV
Supplemental Materials
© World Health Organization 2017 Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). The mark “CDC” is owned by the US Dept. of Health and Human Services and is used with permission. Use of this logo is not an endorsement by HHS or CDC of any particular product, service, or enterprise. Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization. Suggested citation WHO, CDC, UNAIDS, FHI 360. Biobehavioral survey guidelines for Populations at Risk for HIV. Geneva: World Health Organization; 2017. Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data CIP data are available at http://apps.who.int/iris. Sales, rights and licensing To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see http://www.who.int/about/licensing. Third-party materials If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. Printed in Switzerland ISBN: 978-92-4-151301-2
Table of Contents Section I - Appendices 1
Appendix I-1 2 Protocol Checklist 2 Guide For General Protocol Checklist 4 Appendix I-2 Sample Budget Calculator and Gantt Chart (Survey Timeline) Appendix I-3 Sample Consent Form 10 10 30 30
Appendix I-4 33 Survey Design Checklist 33 Appendix I-5 39 Brief Demographic Survey for Key Informants and Focus Groups 39 Appendix I-6 41 Formative Interview Guide for Gatekeepers 41 Appendix I-7 46 Formative Assessment Questionnaire Guide for Target Population 46 Appendix I-8 59 Formative Mapping 59 Appendix I-9 Safety and Security Guide Appendix I-10 Specimens Used for HIV Testing Appendix I-11 Material Transfer Agreement Appendix I-12 Interpretation of Hepatitis Serologic Test Results Appendix I-13 Specimen Tracking Form Appendix I-14 Specimen Transportation Driver Log Appendix I-15 Comparison of HIV Testing Technologies 66 66 63 63 69 69 72 72 74 74 76 76 78 78
Appendix I-16 Specimen Temperature Monitoring Logs Appendix I-17 Sample HIV Rapid Testing Results Log Appendix I-18 Equal Probability Sampling (EPS) and Probability Proportional to Size (PPS) Cluster Sampling Guide Appendix I-19 Survey Implementation & Quality Assurance – Data Collection - TLS Appendix I-20 Venue Observation Forms Appendix I-21 Unique Participant Code (UPC) Appendix I-22 Sample Size Requirements to Detect Differences Appendix I-23 Calculator for the Number of Unique Objects for Estimating Population Size Appendix I-24 Qualitative Interview Guide Appendix I-25 Qualitative Interview for Target Population (FSW) Consent Form Appendix I-26 Formative Assessment Consent Form for Key Informants Appendix I-27 Data Management SOP for RDS Survey Using QDS Appendix I-28 Sample Data Dictionary Appendix I-29 Staff Data Agreement Appendix I-30 Specimen Handling (Collection, Processing, and Storage) SOP Appendix I-31 Recruitment Training Script
80 80 84 84 86 86 91 91 94 94 99 99 100 100 101 101 102 102 109 109 110 110 112 112 114 114 116 116 118 118 120 120
Appendix I-32 RDS Participant Checklist Appendix I-33 Data Error Audit Log Appendix I-34 Sample Table of Contents for Survey Report Appendix I-35 Sample Press Release Appendix I-36 Data Presentation Guide Appendix I-37 Data Use (Confidentiality) Agreement
122 122 125 125 126 126 128 128 130 130 132 132
Section II - Questionnaire Modules 1. Questionnaire Parameters 2. Eligibility 3. RDS Recruiter-Recruit Relationship 4. RDS Personal Network Size 5. Size Estimation 6. ACASI Tutorial 7. Demographics 8. Transgender Characteristics 9. Sex Work Characteristics 10. Clients Of Sex Workers Characteristics 11. Incarceration Characteristics 12. Transport Worker Characteristics 13. Reproductive Health 14. Sexual History
133 136 140 151 154 156 159 162 171 176 185 188 196 200 208
15. Recall Sexual Behavior 16. Partner Concurrency 17. Last Sex Act 18. Condom Use/Accessibility 19. Lubricant Use 20. Microbicides 21. Alcohol Use 22. Non-Injection Drug Use 23. Injection Drug Use 24. Drug Overdose 25. Sexually Transmitted Infections 26. Hepatitis B and C 27. Services Uptake 28. PEP and PrEP 29. Shame, Stigma, Harassment & Discrimination 30. Physical Violence 31. Sexual Violence: Coercion and Rape 32. Internalized Homophobia 33. Depression 34. Suicide Ideation and Attempts 35. Social Support 36. Social Cohesion, Social Participation and Collective Agency 37. Game of Contacts
218 239 243 249 256 261 264 270 277 295 301 306 309 341 346 369 383 403 406 410 415 431 437
38. HIV Knowledge and Perceptions 39. Questionnaire Feedback/Interview Status 40. RDS, Peer Recruitment
439 442 444 451 454 454 455 456 457 460 460 461 463 464 465 466 467 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493
Section III: Indicators Introduction What are indicators? How are data collected for indicators? How are indicators used? Indicator selection for these guidelines Indicators Sexual debut Condom use at last sex Consistent condom use Buying sex Selling sex Unprotected receptive anal intercourse Alcohol use Injecting drug use Received clean needles or syringes Sharing injecting equipment Sharing drug-preparation equipment Use of sterile injecting equipment Received medication for drug dependency Drug overdose Drug overdose prevention training Targeted information, education and communications Received condoms Never tested for HIV HIV test in last 12 months Tuberculosis screening Received tuberculosis treatment Currently in care CD4 count Currently on antiretroviral therapy Prevention of mother-to-child transmission Hepatitis B virus testing Hepatitis C virus testing Discussed pre-exposure prophylaxis Taken pre-exposure prophylaxis Current use of modern family planning methods Antenatal clinic access and HIV testing
Seeking care for sexually transmitted infection symptoms Sexually transmitted infection screening Received sexually transmitted infection treatment Received lubricants HIV knowledge and perceptions Incarceration Health-care stigma Discrimination and social exclusion Violence Condom use at last sex (sex workers only) INTRODUCTION OF BIOMARKER INDICATORS Awareness of HIV-positive status Eligible but not on HIV treatment HIV treatment coverage Viral load suppression Unsuppressed viral load Population viral load HIV/sexually transmitted infection coinfection HIV/hepatitis coinfection HIV prevalence Sexually transmitted infection prevalence
494 495 496 497 498 499 500 501 502 504
505 506 507 508 509 510 511 512 513 514 515
Section IV: Glossary
Supplemental Materials
Section
I II
APPENDICES
Section
QUESTIONNAIRE MODULES
Section
III
INDICATORS
Section
IV
GLOSSARY
Section
I
APPENDIX I-1
PROTOCOL CHECKLIST This checklist provides essential components of standard protocols.
1. PROJECT OVERVIEW Title (including protocol date and version) Table of contents Protocol summary Investigators/collaborators and roles/funding sources Abbreviations and acronyms
2. INTRODUCTION Literature review/current state of knowledge about topic Justification for survey Intended / potential use of survey findings Survey design / locations / timeline Objectives (including any size estimation activities) Hypotheses or questions General approach
3. ETHICAL CONSIDERATIONS Risks (physical, social, psychological) to participants Methods to reduce/minimize risks to participants Anticipated benefits Risk/benefit ratio Vulnerable populations (pregnant women, prisoners, children) Implementation/documentation of informed consent Justification for waiver/alteration of documentation of informed consent Implementation/documentation of assent (children) Implementation/documentation of parent/guardian permission (children)
2
APPENDIX I-1
Protection of privacy of individual Confidentiality of data Assurance/certificate of confidentiality (individuals and institutions) Consent process (participation, specimen collection, storage, potential future testing) Reimbursements and compensation
4. PROCEDURES / METHODS Design Survey design and how it addresses hypotheses and meets objectives Special considerations for sampling designs Formative assessment protocol Stakeholder participation Survey Population Description of survey population, catchment area Special considerations for population being surveyed Participant inclusion criteria Participant exclusion criteria Justification of exclusion of any sub-segment of population Sampling methods (including sampling design, sampling units/units of analysis, sample size calculation, statistical power) Recruitment Enrollment Notification/return of results Variables Survey variables and indicators (summarize) Survey data collection instruments, including eligibility screening, questionnaires, laboratory instruments, etc. Survey data and biological specimen collection, management and analysis Data collection procedures Biological specimen collection procedures (including national algorithms, tests, testing procedures, other relevant specimen-related processes, return of results, treatment provision, specimen transportation and storage) Method of interview administration: paper, ACASI, face-to-face, other Data analysis plan, including statistical methodology and planned tables and figures Data / information management and analysis software Data entry, editing, and management (including handling data collection forms, different versions of data and data storage and disposition) Survey and laboratory data quality control/assurance procedures Measurement / estimation and adjustment methods Data monitoring procedures Limitations of survey (including potential bias in data collection, measurement, analysis)
SECTION I: APPENDICES
3
5. NOTIFICATION, REPORTING AND OTHER DISSEMINATION OF SURVEY RESULTS Notifying participants and stakeholders of survey findings Preparing and publishing anticipated reports, presentations, or other materials to be published Disseminating results to public
6. REFERENCES References Additional resources
7. APPENDIX MATERIALS Survey instruments, including enrollment forms, questionnaires, laboratory instruments, etc. Consent forms Memoranda of understanding, data use agreements, confidentiality agreements Other relevant documents including proposed tables and figures
GUIDE FOR GENERAL PROTOCOL CHECKLIST PROJECT OVERVIEW Summarize the main idea under investigation. The title should be able to stand alone as an explanation of the survey. Include the protocol date and version.
Title
Table of Contents
List each chapter/section title and page number for reference.
Protocol summary
Give a concise overview of the project. Describe the purpose of the survey, including problem to be investigated and hypothesis (hypotheses) to be tested, the population, and the methods that will be used. Avoid the use of acronyms. Include the expected benefit of the survey.
Investigators/ collaborators/ funding sources
Include the names and degrees of all investigators and their roles in the project. Note any conflict of interest for each investigator and acknowledge all funding sources.
Abbreviations and acronyms
Define any abbreviations and acronyms referred to in the text.
4
APPENDIX I-1
INTRODUCTION Literature review/current state of knowledge about project topic: Discuss relevant information about the subject of the project based on a review of the literature. In the Reference section, attach a bibliography of the sources used. Justification for survey: Explain the public health and scientific importance of the survey. In the context of previous studies, describe the contribution this survey will make. Intended/potential use of survey findings: Define the primary target audiences and discuss the expected applicability of survey findings. Survey design/locations: Describe the survey design and the locations where the survey will be conducted. Objectives: Clearly and concisely list the objectives that the project will address. Hypotheses or questions: List the clear and focused question(s) that the survey will answer. State the type of hypothesis that will be explored or tested. General approach: Describe whether the approach used will be descriptive, exploratory (hypothesis-generating), confirmatory (hypothesis testing), or developmental (focused on corrective action).
ETHICAL CONSIDERATIONS Description of risks (physical, social, psychological) to the individual or group. Include methods to minimize risks: Define the nature, magnitude, probability, and duration of potential harms that a person may receive by participating in this research. Describe steps that have been taken to minimize risks, including the use of sound research design and by using procedures already being performed on the participant or other routine procedures that will be provided to the participant. Define the procedures that will be followed to identify, report and correct any social harm that may result from participating in the survey. Training: Describe the process for human subjects research ethics training for key staff and personnel. Description of anticipated benefits to the research participant: Discuss benefits to research participants resulting from the research. Describe the steps that have been, or will be, taken to maximize benefits. Description of the potential risks to anticipated benefit ratio: Justify that the potential risk are reasonable in relation to anticipated benefits and the importance of the knowledge that may reasonably be expected to result from the research. Justification for involving vulnerable participant populations: If survey participants include a special or vulnerable population, such as children, prisoners or mentally incompetent, provide justification for their use in terms of the purpose of the research. Procedures for implementing and documenting informed consent: Describe procedures for informing participants and methods to obtain and document consent.
SECTION I: APPENDICES
5
Justification for waiver or alteration of informed consent: If informed consent will not be obtained or will be altered, describe the justification for waiver. The justification must address the four criteria for waiving or altering consent: 1) the research involves no more than minimal risk to the participants, 2) the waiver or alternation will not adversely affect the rights and welfare of the participants, 3) the research could not practicably be carried out without the waiver or alteration, and 4) whenever appropriate, the participants will be provided with additional pertinent information after participation. Justification for waiver of documentation of informed consent: If written informed consent will not be obtained, provide justification for obtaining consent through other means. The justification must address one of the two criteria for waiving documentation: 1) that the only record linking the subject and the research would be the consent document and the principal risk would be potential harm resulting from a breach of confidentiality or 2) that the research presents no more than minimal risk of harm to participants and involves no procedures for which written consent is normally required outside of the research context. 3) If the first criterion is used, describe the procedures to ensure that participants’ wishes regarding documentation linking them to the research will be ascertained and honored. Description of procedures for implementing and documenting the assent process of children: Describe procedures for informing children and methods to document assent. Description of procedures for implementing and documenting parents or guardians permission: Describe procedures for informing participants and methods to document parental permission. Provisions for protecting privacy/confidentiality: Explain provisions for protecting survey participants from being identified either directly or indirectly. If for any reason data identifying subjects will be published or released to persons outside of the project, explain why this is necessary. Statement about need or lack of need for assurance or certificate of confidentiality: This refers to formal assurances and certificates of confidentiality. Consent Process: Describe procedures for informing participants about survey and methods and for obtaining consent. Description and justification of reimbursements or compensation that will be used: Self explanatory.
PROCEDURES / METHODS Design How survey design or surveillance system addresses hypotheses and meets objectives: Explain the appropriateness of the survey design to the project and to the questions and objectives previously outlined. Distinguish between procedures that are experimental and those that involve routine care. Identify specific design attributes that characterize the survey design (e.g. cross-sectional survey, case/control, cohort, focus group, chart review, etc.) or surveillance system (e.g. description of the system as active or passive, defining reported cases as individual versus aggregate and as laboratory confirmed or not). Include a formative assessment protocol. Provide a calendar with estimated dates for implementing and completing key activities. Audience and stakeholder participation: Define the primary audiences for the project. Assess the major stakeholders and describe ways they can (and cannot) participate in the survey. Explain the process by which those affected by the survey can express their views, clarify their needs, and contribute to the project.
6
APPENDIX I-1
Survey Population Description and source of survey population and catchment area: Demographically and in terms of the specific public health conditions to be studied, define the population from which the participants, sample, or surveillance subjects will be drawn and to what population inferences will be made. Include any special considerations for the population being surveyed. Participant inclusion criteria: Describe conditions or characteristics applicable to the identification and selection of participants in the survey and the conditions necessary for eligible persons to be included. Participant exclusion criteria: Describe characteristics that would disqualify otherwise eligible participants from the project. Justification of exclusion of any sub-segment of the population: If a sub-population as defined by gender, race/ethnicity, or age is excluded, provide reasons. In accordance with CDC’s policy for inclusion of women and minorities in research, state how these populations are included in the sponsoring CIO’s overall program of research if excluded in this particular survey. Sampling, including sample size and statistical power: Describe the sample (e.g. the sample will be one of convenience, a population-based representation, or systematically chosen for some other purpose). State the sampling units and units of analysis. Estimate required sample sizes to answer questions and test statistical hypotheses (based on available information from pilot studies or previous reports). Include statistical power estimates. Explain the conditions under which sampling estimates would be revised. If group-level or aggregate information will be collected (e.g. from focus groups), explain how the groups will be comprised, or what procedures will be followed to create appropriate groups. Recruitment: Describe the manner in which potential participants will be identified and approached. Enrollment: Describe the manner in which potential participants will be contacted, screened, and registered in the survey. Describe procedures for tracking the number of persons who withdraw from the survey. Explain the procedures for assigning participants to different groups. Include a discussion of how departures from the intended enrollment procedures will be handled and documented. Notifying participants of their individual results: Describe the process used to notify survey participants of their results, including those of immediate importance. Include precipitating circumstances and whether or not counselors will be used. Variables Survey variables: List and briefly describe the categories, topics, or domains of information to be explored and variables to be collected. Address consistency of definition of variables for data collected from multiple sources. Explain how the variables will be used and the process by which variables will be defined. Survey data collection instruments, including questionnaires, laboratory instruments, tests: Describe strategies to elicit information, including specific techniques and survey and laboratory instruments, and explain how they will be used. Describe the attributes of those strategies/ instruments as demonstrated in other studies, including appropriateness, validity, and reliability within the particular survey populations, sensitivity and specificity of instruments, how well they yield reproducible results, and whether any controversial methods are being used. Include a discussion of how changes to the survey instruments will be handled and documented.
SURVEY DATA AND BIOLOGICAL SPECIMEN COLLECTION, MANAGEMENT AND ANALYSIS Data collection: Describe data collection procedures, processes and documentation. For data emanating from a surveillance system, this would include frequency of reports. Biological specimen collection: Describe biological specimen collection procedures (including national algorithms, tests, testing procedures, other relevant specimen-related processes, return of results, treatment provision, specimen transportation and storage).
SECTION I: APPENDICES
7
Method of interview administration: Describe how the interview will be administered, i.e. paper, ACASI, face-to-face, or other method. Data analysis plan, including statistical methodology and planned tables and figures: Describe the sampling methods, information collection procedures, methods to maximize response rates, test procedures and relevant statistical quantities (e.g. variance, confidence intervals, and power based on data from the survey) in sufficient detail that the methods are reproducible. This includes calculation of relevant quantitative measures for tests and instruments, such as sensitivity and specificity. In outbreak investigations, it is common to employ an iterative process in the analysis (consisting of developing and testing hypotheses and planning and evaluating interventions) to identify the source of the outbreak and control it. For projects establishing or utilizing data from a surveillance system, this could include how and how often the surveillance system will be evaluated. Describe what tables and figures are planned to present survey results. Data/information management and analysis software: Provide the names of data entry, management and analysis software packages, and computer programming languages to be used for the project. Data entry, editing and management, including handling of data collection forms, different versions of data, and data storage and disposition: Describe the overall procedures for management of the data collected. Include in the description the process for entering and editing data. Describe how survey materials, including questionnaires, statistical analyses, unique reagents, annotated notebooks, computer programs and other computerized information, whether used for publication or not, will be maintained to allow ready, future access for analysis and review. Document operating procedures for managing and accessing different versions of data sets. State who the data belong to and any rights to and limitations to access for any primary and secondary data analyses and publications. Document procedures regarding confidentiality of the data, including how confidentiality will be preserved during transmission, use and storage of the data and the names of persons or positions responsible for technical and administrative stewardship responsibilities. Document what the final disposition of records, data, computer files, and specimens will be, including location for any relevant information to be stored. Records must be stored in compliance with CIO or Agency guidelines. Survey and laboratory data quality control/assurance: Describe the steps that will ensure no unintended consequences that could affect the quality of the data. Those steps might include methods to capture all reported data exactly as received, assuring logical consistency among all parts of a record, ensuring that manipulation or transformation of the data (e.g from audio tape to transcribed text) produces no unintended changes, and verifying that statistical and arithmetic calculations are performed as proposed in the data analysis plan. For outbreak investigations, this would include verifying diagnosis and confirming the outbreak. Describe procedures for ongoing data quality monitoring to assure that information of appropriate depth, breadth, and specificity is collected and remains consistent within and among staff over time, and acceptable levels of such attributes as validity, reliability, reproducibility, sensitivity and specificity are achieved. Measurement/estimation and adjustment for cross reactivity: Describe how cross reactivity will be measured, its potential effects on test results, and how it will be accounted/adjusted for in the analysis. Data Monitoring: Describe the ways that progress will be tracked and the survey will be evaluated prior to assessing final results. Limitations of survey and bias in data collection, measurement and analysis: Explain factors that might reduce the applicability of survey results. Discuss potential weak points or criticisms of the survey, including alternative methods. Describe the kinds of bias that may occur in collecting the data or in the measurement or analysis phases, and the steps that will be taken to avoid, minimize, and compensate for the bias. Include factors in the survey population or in survey personnel that could bias results, as well as the steps that will be taken to assure valid self-reporting or recording of observations. Include any randomization and blinding procedures that will be used to eliminate/minimize bias by investigators, other survey staff, or participants (e.g. in selection of participants, allocation to treatment groups, providing/receiving treatment).
8
APPENDIX I-1
DISSEMINATION, NOTIFICATION, AND REPORTING OF RESULTS Notifying participants of survey findings: Explain whether the participants will be offered the option of receiving overall survey findings and the form they will take. Disseminating results to public: Define effective communication channels and best formats for presenting information that will be used to disseminate project results to specific target audiences.
REFERENCES List bibliographic references and any additional resources used to create and delimit all aspects of the survey.
APPENDIX MATERIALS Survey instruments/data collection/other forms: Include any forms or documents used to collect data or from which data are abstracted. Proposed tables and figures: Provide table shells and examples of figures for presentation of data and survey results. Consent forms: Provide the consent forms used in the survey. Memoranda of understanding/data use agreements: Provide any MOUs or data use agreements pertaining to the study. Other relevant documents: Include any other relevant supplementary materials.
SECTION I: APPENDICES
9
Section
I
APPENDIX I-2
SAMPLE BUDGET CALCULATOR AND GANTT CHART (SURVEY TIMELINE) This Appendix includes a budget calculator template along with a sample budget for RDS and a sample Gantt Chart (depicting the survey timeline). This file is a reference as you prepare a budget and project timeline for a BBS. Please note that this calculator is for preliminary budgeting purposes; the line items should be adjusted accordingly for specific budgetary needs of the survey and implementing institution. The categories covered are core components of a budget and are not exhaustive. For example, the budget includes some point of care (POC) tests and some laboratory tests -- all of these will not necessarily be used in every survey. This template should guide planning for surveys that may be completely POC. This calculator was not designed to be used as a bookkeeping tool or to track expenses; it is a tool for estimating costs in the budgeting process. Discuss with your agency how to track the exact costs. For actual pricing, contact each manufacturer for a specific quote. It is important to review the RDS and TLS sections of the Guidelines to determine the kind and the number of staff needed and related costs. When resources are limited or a survey is on a small scale, one person may perform multiple tasks. This budget template will be available as an Excel file for download from the WHO website. Although not included in the Budget Calculator, the following two categories may need to be considered when preparing a budget: Consultant costs This category is appropriate when hiring an individual to give advice or services (e.g. training, expert consultant, etc.) for a fee but not as an employee of the organization. A subcontract may be required if a part of the survey effort will be performed by one or more other organizations.
Cost sharing
Description of the Budget Calculator Colour Scheme Description Items that may be needed depending on survey type, such as RDS or TLS Major budget categories Inputs Output Subtotal of each category Grand total of budget
10
APPENDIX I-2
Title A. Personnel salaries/fringe benefits
Notes and Instruction Category A combines personnel salaries and fringe benefits. When needed, they can be separated into two entries. Fringe benefits are any nonwage payment or benefit granted to employees by employers. Examples include pension plans, profit-sharing programs, vacation pay, and company-paid life, health, and unemployment insurance.
Number needed Annual salary No. of months
Enter the number of staff needed. Enter annual salary. Enter the length of the survey in months. Different staff may work for different lengths. For example, a PI works the entire survey, but a field staff works only while the survey is conducted in the field. Enter percentage of time to show the effort contributed during the survey period. For example, a PI may have several projects and other responsibilities (such as teaching as a professor) at the same time. If the PI will contribute part of her/his time to the current survey, say 4 hours a week (40 hours), her/his percentage of time effort is 4/40=10%. Salary = annual salary*No. of Month/12*Time (%). It will be auto-filled. A fringe benefit rate is the cost of an employee’s benefits divided by the wages paid for the hours working on the job. Some employees may not receive such benefit from their employer. Enter fringe rate for eligible staff. This can be 0 for non-eligible staff.
Time (%)
Salary total Fringe rate
Total salary and fringe benefits Field staff
This is the total of salary+fringe rate*salary. This will be auto-filled. The staff listed here are some essential personnel. Staff may vary depending on the type and scale of the survey. The list can be expanded to add a staff for one specific function, or reduced if one staff can perform multiple duties in a smaller scaled survey setting. For RDS, there are additional needs, such as testing counselors, quantitative interviewer, qualitative interviewer, transcribers and translators in qualitative interview. The list can be expanded based on survey need, such as the object distributor during size estimation, cleaner, and security staff. List staff for data management and data analysis. Some staff, such as a GIS developer, can be added depending on survey need. Or, in a small budgeted survey, data management and data analysis can be performed by one person. For a survey that involves biomarker testing only. The list can be expanded based on bio-survey need. Training expenses, including the cost of the conference room, handouts, and other expenses, apply to any type of training or meeting. Enter number of participants for handout cost. Enter number days of training for conference room. For simplicity, the calculator lists only one training example for fieldwork preparation/size estimation. Other trainings/meetings that may need to be included in the calculator: Formative assessment training Survey implementation training Data analysis, interpretation, and writing workshop Community forum Local/national dissemination meeting
Additional field staff for RDS
Data management/analysis staff
Laboratory staff B. Trainings and meetings
SECTION I: APPENDICES
11
Title Training expenses
Notes and Instruction Common expenses are bags for transporting materials, badges/IDs, pens and paper, large flip-chart paper and markers, communication (telephone and internet), refreshments for participants. For formative assessment training, also consider expenses for city/area map, copies of coupons, etc.
Number of participants No. of units needed Number of days Cost per unit C. Fieldwork preparation, formative assessment ,and size estimation implementation Formative assessment preparation Formative assessment supplies Transportation for key informants and focus group participants Phone vouchers for formative assessment staff Participants compensation (incentive) Field work size estimation implementation Unique event Unique objects for size estimation D. Survey implementation
Enter the number of participants. Enter the quantity of handouts needed, including extra copies (e.g. 5 extra for a group of 20 participants). Enter number of days for conference room rental. Enter unit price for handouts, or rent per day. Use this section to budget expenses for field preparation, formative assessment, and size estimation implementation.
Enter costs for maps, coupons, camera, audio recordings, batteries, etc. Enter number of participants and key informants, and unit price (cost per person).
Enter number of assessment staff and unit price.
Enter number of participants and unit price (amount paid per person). See above for Formative Assessment. Additional staff for distributing objects may be budgeted here also if they are not included in the Section A (Personnel). Enter number of days, and unit cost.
Enter unit price for objects and number of participants. Items listed in section D are items regularly used for survey implementation. This list may not include everything. Keep a detailed checklist for all supplies needed. Subcategories may be added for separate costs for RDS or TLS. Complete this section if any biomarker is used. The estimation of laboratory expenses is very complicated. Working with a laboratory professional to develop the budget is strongly recommended. Input assumed prevalence to generate numbers in “OUTPUT” below (number of specimens) and calculate cost of consumables. Enter number of participants.
E. Laboratory expenses
INPUT box
Assumed HIV prevalence
Enter assumed prevalence HIV prevalence based on survey population. This is used for budgeting HIV biomarker consumables that only test blood samples from HIVpositive participants, such as CD4 and VL.
12
APPENDIX I-2
Title Assumed syphilis prevalence
Notes and Instruction Enter assumed prevalence rate for syphilis among survey population to budget for syphilis biomarker consumables. Enter assumed prevalence rate for STI to be surveyed, such as HCV, to budget for STI biomarker consumables. Enter number of trainees in lab training, which may determine certain equipment needed, such as a timer. The number of labs determines the number of certain equipment needed.
Assumed STI prevalence
Number of trainees
Number of labs performing testing OUTPUT (Number of specimens) Expected number of HIV positive Expected number of syphilis positive Expected number of other STI positive Equipment
Calculated as: Assumed HIV prevalence*Number of participants
Calculated as: Assumed syphilis prevalence*Number of participants
Calculated as: Assumed STI prevalence*Number of participants
Equipment estimate assumes all required equipment needs to be purchased and therefore reflects a maximum cost. This calculator does not take into account equipment that may already be in place in individual laboratories. The equipment items may be expanded depending on the survey need, such as common equipment of plate washer, reagent reservoirs, etc. This section may be expanded according to survey needs. For example, for on-site testing and handling, mobile phones, tables, and chairs may be included.
Consumables for biomarkers
The consumables for laboratory assays listed can be revised according to bio-survey need, such as HIV genotyping, HIV DR, TB, chlamydia, HSV, HBV, etc. For example, testing kit for HCV among PWID. The list includes common items, It should be adjusted according to the survey.
Consumables for blood collection
Consumables for other specimen collection Other consumables
See above. Other items may be considered here are: distilled or de-ionized water, Band-Aids, cotton balls, latex-free gloves, latex gloves (powdered and powder free), paper towels and absorbent covers for lab benches, marking pens, pads for lab table, hand soap, facemasks, zip lock bags, protective eyewear, bins and liners, paper towels, spray bottles, etc.
Manufacture & unit description Enter brand/name of manufacture, catalog number, and description such as “10/pk, size large” for lab coat. Unit dimension No. of units needed Enter number of 1 unit here, such as “10” for 10-piece pack. Enter the number of units needed. The number of units for biomarkers is determined by estimating the number of positive participants in the OUTPUT box, and Unit dimension, then multipllying by 110%. This includes an extra 10% buffer. Enter price of unit.
Cost per unit
SECTION I: APPENDICES
13
Title For TLS Mobile van and clinic rentals for HIV testing Additional notes for point of care
Notes and Instruction Please refer to TLS chapter for other costs/items, although only mobile van is listed here. If the survey uses TLS, mobile van and clinic rentals for HIV testing may be needed. Please add any additional costs, such as gasoline and parking, if they are needed. The items listed in this section are not comprehensive list of health services/care/ treatment. Modify the budget as needed to account for specific point of care activities including STI treatment.
F. Travel
Cost for travel may be budgeted on an anticipated actual cost basis. The cost for consultant should be included in Consultant cost only. Enter other expenses in this category. Revise accordingly to include items such as software, toner, postage, delivery fees, etc. Refer to the RDS implementation chapter to add items. Including cleaning and bathroom: toilet cleaning, toilet paper, toilet brush, floor cleaner, mop, bucket, absorbent pad, washing liquid, dish drainer, sponge, and hand soap.
G. Other expenses
RDS survey sites Site maintenance
Description of the Gantt Chart of the Survey Timeline Gantt charts are used for project management. They are helpful for defining and scheduling tasks and deadlines as well as tracking progress.
14
APPENDIX I-2
BUDGET CALCULATOR TEMPLATE NOTE: Please refer to Instructions for use of this calculator
A. Personnel salaries/fringe benefits
Number needed
Annual salary
Number of months
Time (%)
Salary total
Fringe rate (% of salary total)
Total salary and fringe benefits
Principal investigator Project coordinator Field staff Site supervisor Enumerator Coupon manager Interviewer Counselor Community outreach worker Driver Additional field staff Data management/analysis staff Data manager Data analyst Laboratory staff Laboratory supervisor/coordinator Laboratory technician Phlebotomist
SECTION I: APPENDICES
15 SECTION I: APPENDICES
15
A. Personnel salaries/fringe benefits
Number needed
Annual salary
Number of months
Time (%)
Salary total
Fringe rate (% of salary total)
Total salary and fringe benefits $0 $0
Laboratory assistant Driver Subtotal of personnel salaries and fringe benefits
$0 $0
B. Trainings and meetings
Number of participants
Number of units needed
Number of days
Cost per unit
Total
Fieldwork preparation/size estimation training Conference room Handouts Refreshments Transportation for meeting attendees Subtotal: Trainings and meetings
C. Fieldwork preparation, formative assessment, and size estimation implementation Site selection visits (RDS) Car Hire + driver Fuel Compensation
Number of participants
Units needed
Number of days
Cost per unit
Total
16
APPENDIX I-2
16
C. Fieldwork preparation, formative assessment, and size estimation implementation Communication (Internet/phone calls, Mobile phones, Smart phones, External hard drive, USB GB, internet hub) Formative assessment Transportation for key informants and focus group participants, and site selection visits Unique objects for size estimation Formative assessment supplies Phone vouchers for formative assessment staff Field work size estimation implementation Unique event Unique objects for size estimation Phone vouchers for size estimation staff Subtotal: Fieldwork preparation
Number of participants
Units needed
Number of days
Cost per unit
Total
D. Survey implementation
Number of participants
Units needed
Number of days
Cost per unit
Total
General supplies and miscellaneous Tablets Fingerprint scanner Software, include antivirus Laptops $0 $0 $0 $0
SECTION I: APPENDICES
17 SECTION I: APPENDICES
17
D. Survey implementation
Number of participants
Units needed
Number of days
Cost per unit
Total
Scanner, printer and toner Participants compensation (incentive) Transport reimbursements for participants Health referral brochure Field team supplies (backpacks, clipboards, etc.) Transportation of DBS cards to central office Shipping and freight cost for lab equipment and consumables Other operational expenses (e.g. utilities, equipment based on security) For RDS Printing of survey coupons Secondary incentive RDS For TLS Clicker for enumerator Subtotal: Survey implementation
E. Lab expenses INPUT Number of participants Assumed HIV prevalence Assumed syphilis prevalence OUTPUT (Number of specimens) Expected number HIV positive Expected number syphilis positive Expected number other STI positive
18
APPENDIX I-2
18
INPUT Assumed other STI prevalence Number of trainees Number of labs performing testing
OUTPUT (Number of specimens)
Equipment
Manufacturer & unit description
Unit dimension
Number of units needed
Cost per unit
Total
Incubators Standard mini-centrifuge Countdown timers Pipettes (disposable transfer pipettes, precision pipettes, etc.) UPS Units Power cord extension leads Portable generator Cooler boxes Consumables for biomarker testing DBS cards Humidity monitoring cards Desiccants Consumables for CD4 testing Consumables for VL testing Consumables for syphilis testing Consumable for HBsAg testing Consumable for MTB testing $0 $0
SECTION I: APPENDICES
19 SECTION I: APPENDICES
19
Equipment
Manufacturer & unit description
Unit dimension
Number of units needed
Cost per unit
Total
Consumables for other STIs testing Consumables for blood collection Needles Tourniquets Alcohol Vacutainer tubes Vacutainer sleeves Vacutainer holder rack Centrifuge tube Consumables for other specimen collection Vaginal swabs Rectal swabs Oral fluid Sputum cup Urine collection jars Other consumables Biohazard bags Biohazard bag holder Disinfectant/bleach Sharps containers
20
APPENDIX I-2
20
Equipment
Manufacturer & unit description
Unit dimension
Number of units needed
Cost per unit
Total
Lab coats Labels (water- and freeze-proof labels) Gloves Cotton balls Bandages or adhesive tapes Other For TLS Mobile van rental for HIV testing Subtotal: Lab Days needed Cost per unit
F. Travel
Number of trips
Per diem & lodging
Mileage
Rate per mile
Total
Local transportation (site visit) Staff travel Subtotal: Travel
G. Other expenses
Units needed
Cost per unit
Total
IRB fees Cell phones
$0 $0
SECTION I: APPENDICES
21 SECTION I: APPENDICES
21
G. Other expenses Air time Internet Report printing and dissemination Other communication expenses (external hard drive, USB & internet hub) Stationery Paper per box (white and colored) Pens Highlighters Sticky notes White board White board markers and permanent markers Notebooks Files and tabs for cabinet Filing cabinet 4 drawer Folders Stapler Staples Tape Staple remover Paper clips
Units needed
Cost per unit
Total $0
22
APPENDIX I-2
22
G. Other expenses Equipment (TV, video players) RDS survey sites Site facility rental Utilities Site maintenance Furniture Filing cabinet locking Book shelf Refrigerator Subtotal: Other expenses TOTAL BUDGET
Units needed
Cost per unit
Total
SECTION I: APPENDICES
23 SECTION I: APPENDICES
23
SAMPLE BUDGET FOR RDS* Amounts are in US$ based on conservative cost estimates in a moderately expensive city. RDS at one site, sample size=400
A. Personnel salaries/fringe benefits
Number needed
Annual salary
Number of Months
Time (%)
Salary total
Fringe rate (% of salary total) 20% 15%
Total salary and fringe benefits $9,000 $57,500
Principal investigator Project coordinator Field staff Site supervisor Interviewer Counselor Community outreach worker Coupon manager Security Additional field staff (cleaner, receptionist) Data management/analysis staff Data manager/data analyst Laboratory staff Laboratory coordinator Laboratory technician
1 1
$75,000.00 $50,000.00
12 12
10% 100%
$7,500 $50,000
1 2 3 1 1 1 2
$15,000.00 $8,000.00 $8,000.00 $6,000.00 $6,000.00 $6,000.00 $6,000.00
3 3 6 3 6 3 6
100% 100% 100% 100% 100% 100% 100%
$3,750 $4,000 $12,000 $1,500 $3,000 $1,500 $6,000
0% 0% 0% 0% 0% 0% 0%
$3,750 $4,000 $12,000 $1,500 $3,000 $1,500 $6,000
1
$12,000.00
12
100%
$12,000
15%
$13,800
1 1
$20,000.00 $8,000.00
12 3
10% 100%
$2,000 $2,000
15% 0%
$2,300 $2,000
24
APPENDIX I-2
24
A. Personnel salaries/fringe benefits
Number needed
Annual salary
Number of Months
Time (%)
Salary total
Fringe rate (% of salary total)
Total salary and fringe benefits $2,750 $1,750 $120,850
Phlebotomist Driver Subtotal of Personnel Salaries and fringe benefits
1 1
$10,000.00 $7,000.00
3 3
100% 100%
$2,500 $1,750
10% 0%
B. Trainings and meetings
Number of participants
Number of units needed
Number of days
Cost per unit
Total
Fieldwork preparation/size estimation training (6 participants) Conference room Handouts Training expenses Formative assessment training (8 participants, 5 days) Conference room Handouts Training expenses RDS implementation training (10 participants, 10 days) Conference room Handouts Training expenses Subtotal: Trainings and meetings 14 140 10 $100.00 $10.00 $25.00 $1,000 $140 $3,500 $7,260 1 12 60 5 $100.00 $10.00 $25.00 $500 $120 $1,500 1 10 10 2 $100.00 $5.00 $25.00 $200 $50 $250
SECTION I: APPENDICES
25 SECTION I: APPENDICES
25
C. Fieldwork preparation, formative assessment, and size estimation implementation Formative assessment Unique event Unique objects for size estimation Formative assessment supplies Transportation for key informants and focus group participants Phone vouchers for formative assessment staff Field work size estimation implementation Unique event Transport for key informants and focus group participants Phone vouchers for formative assessment staff Subtotal: Fieldwork preparation
Number of participants
Units needed
Cost per unit
Total
1 500 1 60 18
$2,000 $2 $200 $4 $20
$2,000 $1,000 $200 $240 $360
1 60 18
$2,000 $4 $20
$2,000 $240 $360 $3,800
D. Survey implementation General supplies and miscellaneous Notebook Laptops Printing of survey coupons Printer and toner Reproduction/copies (6 months) Participant reimbursements
Number of participants
Units needed
Cost per unit
Total
6 1 400 1200 1 6 400 440
$500 $1,200 $2 $1,000 $100 $10
$3,000 $1,200 $2,400 $1,000 $600 $4,400
26
APPENDIX I-2
26
D. Survey implementation Participant secondary compensation x 3 Transport reimbursements for participants Health referral brochures Field team supplies (backpacks, clipboards, etc.) Transport of DBS cards to central office Other operational expenses (e.g. utilities, equipment based on security) Subtotal: Survey implementation
Number of participants 400 400 400
Units needed 1320 440 440 10 12 1
Cost per unit $4 $4 1 40 50 1000
Total $5,280 $1,760 $440 $400 $600 $1,000 $22,080
D. Lab Expenses INPUT Number of participants Assumed HIV prevalence Number of trainees Number of labs performing testing 400 25.0% 4 1 OUTPUT (Number of specimens) Expected number HIV positive 100
Equipment
Manufacture & unit description Fisher Scientific Fisher Fisher Scientific
Unit dimension 1 1 1
Number of units needed 1 1 4 2
Cost per unit
Total
Incubators Standard mini-centrifuge Countdown timers Cooler box Consumables for biomarkers
4800 425 50 50
$4,800 $425 $200 $100
SECTION I: APPENDICES
27 SECTION I: APPENDICES
27
Equipment
Manufacture & unit description Whatman, Pack of 100
Unit dimension 100
Number of units needed 5
Cost per unit
Total
DBS cards Other consumables All other lab supplies Subtotal: Lab
800
$4,000
1
5,000
$5,000 $14,525
F. Travel Local transportation (site visit) Subtotal: Travel
Number of trips 1
Per diem & lodging
Mileage
Rate per mile
Total $2,000.00 $2,000.00
G. Other expenses IRB fees Cell phones Air time (10 staff, 6 months) Internet RDS survey site Site facility rental Utilities Site maintenance Filing cabinet Bookshelf Refrigerator Subtotal: Other expenses TOTAL BUDGET
Unit needed 1 2 10 12
Cost per unit $200.00 $100.00 $120.00 $300.00
Total $200.00 $200.00 $1,200.00 $3,600.00
6 6 6 2 1 1
$1,500.00 $100.00 $50.00 $100.00 $100.00 $600.00
$9,000.00 $600.00 $300.00 $200.00 $100.00 $600.00 $16,000 $186,515
*Global Strategic Information. Toolbox for conducting integrated HIV bio-behavioral surveillance (IBBS) in key populations. University of California, San Francisco. 2014. pp. 131–133.
28
APPENDIX I-2
28
SAMPLE GANTT CHART - SURVEY TIMELINE Proposed Activity Month 2017 May Mar Nov Aug Dec Sep Apr Oct Jun Jan Jul 2018 Mar Apr Feb
Formative Protocol Approval Proposed Activity Formative Training Formative Implementation Formative Assessment analysis Pre-test tools Protocol Development SOPs adapted Submit protocols and receive approval from all IRBs, <Country> Ethics Committee Recruit study team Development of data collection systems Study Implementation Training Study Implementation Data cleaning/lab QA/data merging Data analysis and preliminary report Present preliminary findings to stakeholders Final Report National Dissemination Workshop
X
X X X
X X
X X X X X X X
X X X X
X
X X
X X X X
X X
X X
X X
X X
X X
X X
X X
X X
SECTION I: APPENDICES
29
Section
I
APPENDIX I-3
SAMPLE CONSENT FORM This is an example of an informed consent form to be shared with and signed by all participants before they begin survey procedures
INFORMED CONSENT FORM FOR <Survey Name> <TARGET POPULATION> INTRODUCTION AND OVERVIEW Hello, my name is <name of survey staff>. I am working with the <name of project> with <institutions affiliated with the project>. We are doing a survey with <target population> to better understand their HIV related health needs. This information will help improve HIV and other health services. This document is a consent form. It says what we do in this survey. That way you can make an informed decision about joining the survey. You are free to ask questions at any time. If you do not understand any words, please ask.
YOUR ROLE IN THIS SURVEY We ask you to take part in this survey, because <insert key eligibility criteria>. About <insert number> <insert target population> will take part in this survey. Interview: 1) If you join, you will do an interview on a computer. It is easy. Our staff will show you how to do this. The interview takes about 45–60 minutes to complete. 2) The interview asks questions about you and your sexual and drug use behavior. It also asks about your use of HIV services. You can refuse to answer any questions. Testing: 1) We will test your blood for HIV <and other diseases>. If you agree we will draw <XX> cc blood from a vein [and XX, indicate other specimens as appropriate]. 2) If you test positive for HIV, we will count your CD4. CD4 are the cells in your body that fight HIV. The more CD4 cells you have, the better for your health. If you have HIV, we will also count how much virus you have. This is called viral load. The fewer viruses you have in your body, the better for your health. We will do all these tests from the same blood specimen. <Insert other specimens and testing as appropriate.> 3) We will return the test results to you <by when>. If you test positive for HIV, we will refer you to a clinic for treatment. If you test positive for <STIs, e.g. syphilis>, we will give you treatment and refer you to services for follow-up.
30
APPENDIX I-3
After the Survey We will look at all the answers of all the people in the survey. We may share the survey records with other researchers. We make sure that no one will know who was in the survey. We will keep the leftover blood [and other specimens, as applicable] for other tests later. We do not know yet which tests we will do. These tests may be about other diseases, or about your body’s reaction to diseases. Because we do not know your name, we cannot return to you the results of tests done later.
POSSIBLE RISKS AND BENEFITS Risks Some questions may make you feel uncomfortable. You can refuse to answer any question. The blood draw may cause some pain and bruises. Benefits Taking part in the survey is free. For your time <and transport> we will give you <money value>. You can help make HIV services better for <target population> by joining this survey. We will also give you condoms, lubricants, and information on HIV and STIs. You will also receive your HIV and STI test results. For RDS surveys only: For completion of the survey you will get <primary compensation amount>. During your second visit, you may also get <secondary compensation amount> each for up to <maximum number of coupons> people whom you recruit for the survey.
CONFIDENTIALITY We do not record your name or anything else that shows who you are. We keep all your answers and test results private. We do not collect your name. Instead you will be given a survey participant number.
PARTICIPATION You are free to join or not join. There is no problem if you do not join. You can leave the survey at any time. You can refuse to answer any question. If you do not join, you can get an HIV test and other services at <name/location of service providers>.
YOUR RIGHTS <Name of relevant Research Ethic Committees (RECs)> gave us permission to do this survey. You may have a question about your rights in this survey. If you think you were hurt by taking this survey you may contact: <Name of relevant Research Ethic Committees (RECs)> gave us permission to do this survey. You may have a question about your rights in this survey. If you think you were hurt by taking this survey you may contact: <Name and contact information of RECs> If you have a questions about this survey, you may contact the survey’s investigator: <Name and contact information of PI/co-PIs>
PARTICIPANT AGREEMENT I understand what it means to join the survey. I understand my rights and risks. I had time to ask questions. I understand that I can join the survey at my free will. I understand that I can leave the survey at any time.
SECTION I: APPENDICES
31
Have all your questions been answered? Do you agree to do an interview? Do you agree to the blood draw and tests? Do you agree that we may keep the leftover blood for future testing?
YES YES YES YES
NO NO NO NO
(Circle answer) (Circle answer) (Circle answer) (Circle answer)
_____________________________________________ ______________________ Signature of survey staff Date
32
APPENDIX I-3
Section
APPENDIX I-4
I
SURVEY DESIGN CHECKLIST Consider using this checklist to inform survey design. Items on the checklist are organized by topic area. PREVIOUS SURVEYS Research on this target population and geographic area has been done before. If so: Method(s) used HIV testing performed Primary outcomes/results HIV prevalence How data were used What the previous research hasn’t answered or what findings are missing Any previous experiences with surveys that might inhibit participation
TARGET POPULATION CHARACTERISTICS Primary risk behavior Whether the target population is large enough to obtain the needed sample size Languages spoken Literacy rate Any laws that impact target population’s behaviors/sexual practices Consequences to being caught as someone who engages in risk behavior or gender/sexual identity (or other stigma that is related to target population) Perception of HIV risk among the target population Whether target population members have mobile phone? Whether they use internet Drugs commonly used and what they are called (non-injection and injection drug use questionnaire modules) Any other (non-HIV) conditions of interest for this population is diagnostic testing and/or treatment available for those conditions?
SECTION I: APPENDICES
33
SAMPLING AND PROCEDURAL CONSIDERATIONS Whether a complete list of individuals in this group (e.g. prisoners, schoolchildren) currently exists (conventional cluster sampling) Whether it is possible to create a complete list of individuals in the cluster once the study team arrives there (conventional cluster sampling) Whether the target population gathers in sufficient numbers at venues where they can be recruited (TLS) Whether the target population is socially networked or they are isolated from one another (RDS) Any changes to the environment that would influence the number of people in the target population, or the frequency or type of risk; any since last survey, if there was one
SERVICES AVAILABLE IN EACH CATCHMENT AREA Condoms available Lubricants available Clean needles available Clean syringes available HIV service providers the population can safely access and the services they provide
TARGET POPULATION PREFERENCES FOR IMPLEMENTATION Method Method/mode that should be used to administer interview: ACASI or face-to-face Whether interviewer should be member of target population. If this is not possible, who else would they be comfortable talking to? Whether recruiter should be member of target population. If this is not possible, who else would they be comfortable talking to? (TLS) Compensation amount and type Standard mode of transportation Suggestions to engage subjects in survey Assessment of exposure to services and for referrals to care, etc. for participants (from participants’ point of view) Participants’ confidentiality concerns Reasons people might not want to take an HIV test as part of the survey and ways to encourage them
LOCAL LEGAL AND POLICY Age of consent If including non-adult participants: Definition of emancipation Assent requirement Any special requirements from IRB regarding target population Whether the participant is required to sign the consent form
34
APPENDIX I-4
Use of a consent script rather than a consent form (i.e. talking points) Whether it is legal to provide an HIV test without requiring name and results Any policies in place that make it difficult for the target population to access (HIV) services HIV treatment policy with respect to foreigners: Can they get access to treatment? Do they risk deportation? How compensation will be determined
TARGET POPULATION SPECIFIC INFORMATION SEX WORKERS Whether sex work is legal Whether FSW are registered; if yes, are they tested for HIV on a regular basis? If yes, what happens to those who test HIV+? Brothel based Street based Internet based Phone/SMS based Pimps: If they exist, what is generally the relationship between pimp and SW? Between pimp and client? Whether SW are networked? If yes: How do they meet other sex workers? How do they recognize other sex workers? What activities do they engage in together? Proportion of sex workers considered “indirect” (i.e. meet clients while doing other activities such as working at a restaurant or hotel) MEN WHO HAVE SEX WITH MEN Relevant sub-types and the labels used to describe them (e.g. top, bottom, kothi, ponthi, yossi, etc.) Whether they are networked. If yes: How do they meet other MSM? Where do they meet other MSM? How do they recognize other MSM? What activities do they engage in together? TRANSGENDER WOMEN Relevant sub-types and the labels used to describe them (e.g. hijira, eunuch, heshes, etc.) How TG access hormone therapy, silicone/filler injections, plastic surgery Whether they are networked. If yes: How do they meet other TG? How do they recognize other TG? What activities do they engage in together?
SECTION I: APPENDICES
35
PEOPLE WHO INJECT DRUGS Where they get injection equipment Most commonly used drugs Typical drug prices (to detect non-PWID who are trying to pose as PWID for compensation) Are drug treatment services available? Whether methadone-assisted therapy is available How they get drugs Whether they are networked. If yes: How do they meet other drug users? How do they recognize other drug users? What activities do they engage in together?
CLIENTS OF SEX WORKERS Whether they are networked. If yes: How do they meet other clients of sex workers? How do they recognize other clients of sex workers? What activities do they engage in together?
IMPLEMENTATION INFORMATION RDS: Preliminary list of possible survey sites TLS: Preliminary list of possible venues TLS: Systematic method for finding venues, including internet, newspaper searches, key informants approach, etc.
LOCAL TERMINOLOGY Vaginal sex Anal sex Oral sex Drugs used (and names for them) Other: _____________
SELECTION/CHOICE OF SAMPLING METHODOLOGY CONVENTIONAL CLUSTER SAMPLING Primary stakeholders Estimated size of each cluster Whether a list of target population members is available at each site
36
APPENDIX I-4
TIME LOCATION SAMPLING Approach criteria Proportion of eligible respondents Whether to allow non-random events Method for executing simple random sample of venues Method for executing simple random sample of day-time periods Minimum number of eligible respondents in a venue during venue-day-time period Maximum number of participants per venue-day-time Number of venue-day-time periods per month Length of sampling event Plan for interviewing & testing outside venue Whether to allow other/next day appointments List of potential venues Venue types Other venues in the area that may be suitable (FG & KII)
RESPONDENT-DRIVEN SAMPLING Whether multiple interview sites needed Electronic or paper coupon tracking (electronic recommended) How potential participants know each other Whether population forms one large network Potential sub-networks Plan for recruiter training Whether sample size is over 4% of total target population size Number of seeds proposed Plan for finding seeds Plan for seed diversification Specific coupon tracking method Proposed ‘standard’ number of coupons Maximum number of coupons per participant Coupon description (style, material) Information needed to be included on coupon (address, phone number, potential incentive amount, directions) Network size questions Recruiter training System for identifying and paying recruiter for each recruit (unique ID) System for tracking recruiter payments Whether recruiters can be paid more than once (i.e. come in twice for two different recruits) Info regarding both compensation and how the type and amount was determined Whether people are willing to recruit one another
SECTION I: APPENDICES
37
Hours and days of week Location of study site Interview appointment, walk-in, or both
POPULATION SIZE ESTIMATION Lists that can be used as service or event-based multipliers What kind of unique object is recommended for capture/recapture or unique-object multiplier method Number of unique objects needed for distribution Whether other methods may be feasible for population size estimation
38
APPENDIX I-4
Section
APPENDIX I-5
I
BRIEF DEMOGRAPHIC SURVEY FOR KEY INFORMANTS AND FOCUS GROUPS This is a brief demographic survey to be administered to key informants and focus group participants (for target population only).
Survey ID: __________ Circle one: interview focus group
Date of interview or focus group: _____________________ DD/MM/YYYY Location of interview or focus group: __________________ Administer verbal consent first. If no consent is given, stop here and do not ask any further questions. This form may be self-administered. Question 6 should be asked only in surveys of sex workers. Question 7 should be asked only in surveys of PWID. Delete these questions for all other groups. Interviewer read: Before we begin the (interview/focus group), we would like to ask you a few questions. This information will help us describe some basic characteristics of our participants. We will not take your name. All your answers will be kept private. 1. What is your age? _____ 2. How long have you lived in <city>? ________ years ______ months 3. What is your highest level of school you completed? Please mark the line next to your answer. a. b. c. d. e. None _____ Primary school _____ Secondary school _____ University _____ Other _____
SECTION I: APPENDICES
39
4. What is your main source of income? Please mark the line next to your answer. a. b. c. d. e. f. g. No income _____ Student only _____ Local option A _____ Local option B _____ Local option C _____ Local option D _____ Other _____
5. What is your current marital status? Please mark the line next to your answer. a. b. c. d. Single, never married _____ Married ______ Separated/divorced _____ Widowed _____
6. For FSW only: How long have you been selling sex? _______ years (Write 0 if less than 1 year) 7. For PWID only: How long have you been injecting drugs? _______ years (Write 0 if less than 1 year)
40
APPENDIX I-5
Section
APPENDIX I-6
I
FORMATIVE INTERVIEW GUIDE FOR GATEKEEPERS Qualitative interviews should be conducted as guided conversations. This appendix provides a questionnaire guide for semi-structured individual interviews with gatekeepers during the formative assessment. The bulk of the questionnaire guide should be used in places where no BBS has previously been conducted with the target population. If a BBS has already been conducted, investigators may focus on specific questions whose answers may change over time. Before starting the interview, the interviewer should describe the purpose and procedures of the data collection activity, and then explain why the participant has been selected for the interview. Interviewers should stress that participants are the experts and interviewers are there to learn. This questionnaire guide includes topics relevant to all stakeholders and some relevant only to clinicians or HIV service providers.
Date
|____||____| / |____|| __| / |____||____||____||____| (DD/MM/YYYY)
Interviewer name
Venue
Start time
|____||____| : |____||____| (HOUR/MIN)
End time
|____||____| : |____||____| (HOUR/MIN)
Participant’s association with the target population
How was this participant referred to be interviewed?
Name of electronic audio file
SECTION I: APPENDICES
41
INTRODUCTION (To be read by interviewer to participant) START RECORDING HERE Before I start the interview, please turn off your cell phone and other mobile devices. I will ask you questions about yourself and your friends. Please do not use your name or anyone else’s. We want to improve health services for <target population> and need your help to do this. We want to learn about the HIV risks and health needs of <target population>. To do this we are planning a survey with <target population> in <city name>. What we learn from this future survey will help improve services for <target population> in <country name>. We will ask you some questions today about <target population>, about how they get together and spend time with one another and about what different groups of <target population> there are. We will also ask about what we can do to get <target population> here in <city name> to join the survey. We want to know what you think people will accept, find difficult to do, or refuse. We will also ask questions about how to improve services for <target population> and about your experience using these services. During the interview, we ask that you not use real names or anything that would identify others. Please feel comfortable to share your view and experiences. This will help us plan the survey and inform health services. It is okay if you do not want to answer certain questions. Also, when I say “friends”, “colleagues”, “peers” or “people like you”, I mean people you know who are <target population>. Do you have any questions before we start? (Take time to address all questions and concerns)
1. BACKGROUND QUESTIONS 1.1. Can you describe your relationship with the <target population> you meet? a. How have you come to know <target population>? In what situations do you meet them? b. How often do you interact with <target population>? 1.2. When you come into contact with <target population>, how do you know they are <target population>? a. If you guess, what makes you assume they are <target population>? b. What do you think about <target population>? 1.3. Regarding the <target population> you know, where do you see them or meet with them? a. Where do they get together? b. In what kind of places? c. Bars, clubs, brothels, parks, ceremonies, etc.? 1.4. How do people generally view your association with <target population>? 1.5. Without telling us names, do you know any <target population> who are influential among their peers? These would be people who know a lot of other <target population> and are well liked in this community. Would they be willing to talk to us? If yes, then sensitively ask the interviewee to refer the person to the survey team by giving them the contact card with the information of the Field Coordinator. 1.6. Can you describe the different kinds of <target population> in <city>? a. What names do they go by? 1.7. Have you noticed any changes or trends over the past year with regard to <target population>? (e.g. new populations/groups, new hangouts, new or changing risk behaviors)
42
APPENDIX I-6
2. FOR CLINICIANS/SERVICE PROVIDERS The following questions relate to your organization and their relation with <target population>. 2.1. What is your role at your clinic/organization? 2.2. What services does your organization provide? 2.3. What have been some of the challenges for you or your organization in providing services to or interacting with <target population>? 2.4. What have been some of the successes? What has worked well? 2.5. What are the main service needs for <target population>? 2.6. When you come into contact with <target population> for work, how do you know they are <target population>? a. If you guess, what makes you assume they are <target population>? b. What do you think about <target population>? 2.7. Please tell me how you interact with <target population> as part of the work in your clinic/organization. a. Can you discuss their problems openly with them? b. How do you get them to talk openly with you? Please give an example. 2.8. What health problems do <target population> have? a. Are there any health problems that <target population> have more than people who are not <target population>? b. How do they explain that they have such problems (directly or indirectly)? Please give an example. 2.9. What problems do <target population> have when they seek health care? a. Probe to see if there is stigmatization, perception of lack of confidentiality, etc. 2.10. Among your colleagues, what proportion is comfortable working with <target population>? a. Most, some, few? Probe to ascertain why certain healthcare practitioners might not be comfortable treating <target population>. b. How do health care providers ensure confidentiality? c. Can you give an example where confidentiality may have been breached?
3. Acceptability of Survey/ MSM Participation Say to participant: As you know, we are planning a future survey of <target population>. In the future survey we will interview <target population> about HIV and what they do to prevent it. We will ask questions about how they meet people and about things they do that might put them at risk for HIV. We also want to give free HIV and STI testing and counselling. People who choose to be tested for HIV and STIs will learn their test results. If they test positive for HIV, we will refer them for care and treatment. The future survey will take at least two hours for participation. 3.1. What are your initial thoughts about this survey? a. How would <target population> feel about participating in this survey? b. What would discourage them from participating in this survey? c. What would make them feel more comfortable about participating? 3.2. Only for clinicians: We plan to offer free testing and treatment for HIV and sexually transmitted infections as part of the study. The tests involve collecting some blood. What challenges have you experienced collecting blood and vaginal/rectal swabs from people in general and specifically from <target population> ?
SECTION I: APPENDICES
43
Probe for fear of learning they have HIV, fear of lack of confidentiality, infecting others, use in “witchcraft”, etc. a. What would make participants feel more comfortable providing specimens? b. Would people feel more comfortable with a male or female nurse? Any other things? 3.3. We will offer HIV and STI testing and referral. What other services would <target population> benefit from in your view? robe to see what services are currently lacking or how existing services could be improved (medical, psychoP social, etc.) 3.4. We want this survey to be helpful to <target population>. We also want to make it safe for people to join. What can we do to keep people safe? 3.5. Should we tell the police or other law enforcement agencies about the survey so they don’t bother people trying to join? Or is it better not to inform them? What about others (religious, military, local level government, etc.)? 3.6. How do you think more hidden <target population> will feel about joining the survey? a. What can we do to make it easier for such <target population> to join the survey? b. Are there any groups or individuals we can mobilize to make it easier for more hidden <target population> to participate? 3.7. What type of people do you know who are influential among <target population>? Peer leaders? Gatekeepers? This person need not be <target population>. This would be someone who knows a lot of other <target population> and is well liked by peers. Would they be willing to talk to us? That is the end of our interview. Thank you so much for sharing your thoughts with me. Do you have any questions, or is there anything that you would like to add before we end? If you have further thoughts about any of the issues we discussed today, please call [NUMBER WHERE INTERVIEWER CAN BE CONTACTED] END RECORDING HERE
44
APPENDIX I-6
TO BE COMPLETED BY THE INTERVIEWER: Please note your impressions about the session, its main themes, and the comments and reactions of participants. INTERVIEW WAS: ____ ROUTINE ____ NOT ROUTINE IF NOT ROUTINE, WHY:
ANY ADVERSE REACTIONS IN THE INTERVIEW: ____ YES ____ NO IF YES, SPECIFY:
OTHER OBSERVATIONS/COMMENTS:
SECTION I: APPENDICES
45
Section
I
APPENDIX I-7
FORMATIVE ASSESSMENT QUESTIONNAIRE GUIDE FOR TARGET POPULATION Qualitative interviews should be conducted as guided conversations. This appendix provides a questionnaire guide for semi-structured individual interviews with gatekeepers during the formative assessment. The bulk of the questionnaire guide should be used in places where no BBS has previously been conducted with the target population. If a BBS has already been conducted, investigators may focus on specific questions whose answers may change over time. Before starting the interview or focus group, the interviewer should describe the purpose and the procedures of the data collection activity, and then explain why the participant has been selected for the interview. Interviewers should stress that participants are the experts and interviewers are there to learn. This questionnaire guide covers the topics below. The topics are organized to go from those applicable to all methods to those applicable to only some methods. The guide starts with non-sensitive topics to help make participants comfortable and covers more sensitive topics later. 1. Background and rapport building (for all populations) 2. General information/trends (for all populations) 3. Information on population behaviors 4. Stigma, violence, and service provision 5. Survey participation 6. Biological testing 7. Interview administration 8. Survey logistics 9. RDS: compensation 10. RDS: location, days, times, etc. 11. RDS: coupon design and recruitment 12. Terminology Sections 9–11 should only be asked if RDS is likely to be the sampling method for the BBS. Questions specific to a target population (i.e. SW, MSM/TG, PWID) have been highlighted. Mapping is a key component of formative assessment. Appendix I-8 provides a guide for a mapping activity. Mapping is especially important if a cluster-based sampling method such as TLS will be used.
46
APPENDIX I-7
Date
|____||____| / |____|| __| / |____||____||____||____| (DD/MM/YYYY)
Interviewer name
Note Taker name
Venue
Start time
|____||____| : |____||____| (HOUR/MIN)
End time
|____||____| : |____||____| (HOUR/MIN)
How was this participant referred to be interviewed?
Name of electronic file
Introduction (To be read by interviewer to participant(s)) START RECORDING HERE For all: Before I start the interview, please mute your mobile phone. I will ask you questions about yourself and your friends. Please do not use your name or anyone else’s. For focus groups: Other rules are that we should respect what others say and not speak while others are speaking. After the focus group, do not share with others what people in the group say today. For all: We want to improve health services for <target population> and need your advice. We want to learn about the HIV risks and health needs of <target population>. To do this we are planning a survey with <target population> in <city>. What we learn from this future survey may help us improve services for <target population> in <country>. We ask for your advice on how to best prepare for the survey. We will ask questions about <target population>. About how they socialize with one another and about the different types of <target population> there are. We want to know what you think people will accept or find difficult in the future survey. We will also ask questions about how to improve services for <target population> and about your experience using these services. Today’s interview is completely separate from the future survey. It will not change your chances of participating in the future survey. During the interview, we ask that you not use real names or anything that would identify you or others. Please be comfortable to share your opinions and experiences. It is okay if you do not want to answer certain questions. Also, when I say “friends”, “colleagues”, “peers” or “people like you”, I mean people you know who are <target population>. Do you have any questions before we start? Take time to address all questions and concerns)
SECTION I: APPENDICES
47
1. Background and Rapport Building Please tell me about yourself. 1.1. 1.2. 1.3. 1.4. How old are you? Where are you from? What is your tribe/ethnicity? How did you come to live here in <insert city>?
2. General Information/Trends Let’s start by talking a bit about <target population> in <city>. 2.1 . 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8. 2.9. Some people use other words or terms to describe <target population> in <city>. What term would you like to use during our interview/focus group? Please tell me about <target population> in <city>. How large do you think the population of <insert risk behavior/population name> is in <insert city>? a. How many <target population> do you think there are? b. For SW: Are there more formal or informal sex workers? Are there more brothel-based or street-based sex workers? Where do most sex workers find clients (i.e. street, brothel, restaurant, bar, club, internet)? c. For MSM only: Out of every hundred men in <city>, how many men do you think are MSM, or have sex with other men? When you see a stranger, how do you know if he/she is also a <target population>? a. How do you recognize or identify other <target population>? b. If we wanted to find <target population> in a public place (bar, street, hotel, club), how would we be able to identify them? What are the different sub-groups or types of <target population>? How would you describe them? (i.e. age, socio-economic status, ethnicity/language, education) a. For SW: venue-based (i.e. brothel, restaurant, club, bar), street based, formal, informal, client type (working class, business class, trade, foreign/local) b. For MSM/TG: openness/identity (i.e. bi, gay, out, straight/closeted, married); sexual positions, public presentation (masculine/feminine, leather, sugar daddy), pre/post-operation, those who sell sex c. For PWID: sex, drug of choice, where they purchase drugs, where or with whom they use drugs, where they use drugs, those who sell sex. How much contact is there between the different sub-groups? How often and when do you see or meet <target population> who are different from you (i.e. different sub-group, age, ethnicity, places they sell sex, meet other MSM, buy/use drugs, etc.)? Gently prompt for any issues or tensions between different groups of <target population> in the city. Have you noticed any changes or trends among <target population> over the past year? For example: a. Populations/groups b. Hangouts c. Risk behaviors How do you think we can reach different sub-groups of <target population>?
48
APPENDIX I-7
For sex workers: 2.10. 2.11. 2.12. Could you please describe how the SW you know interact with each other? How often and how do SW see each other? (e.g. work, social activities, social or community organizations) a. Prompt for frequency of contact (i.e. daily, weekly, monthly, never), who these other sex workers are in terms of age, where they work, if they are from this town or travel in and out, etc. b. Probe further to see if this is reflective of all sex workers in the city. Do SW who work in different brothels have contact with each other? What about SW who work at different bars or streets? Do sex workers who work at brothels have contact with sex workers who don’t work at brothels? Where and how often do they see or meet each other?
For men who have sex with men or transgender women: Could you please describe how the MSM/TG you know interact with each other? 2.13. 2.14. 2.15. How often and when do MSM/TG see each other (e.g. work, social activities, social or community organizations)? Which bars, restaurants or similar places do you and your friends go to? Are these places mainly for MSM/TG or are they “mixed”? How old are most of your peers? Do MSM of different ages get together often or do they largely keep separate? What other factors might MSM social groups form around (e.g. where they live or work, drug use, whether they go to big parties, frequency of sexual activity, public presentation (gay/straight, male/female), interests [sports, parties, fashion, arts, MSM rights], income)?
For people who inject drugs: Could you please describe how the PWID you know interact with each other? 2.16. 2.17. 2.18. How often and when do PWID see each other? (e.g. buying drugs, using drugs, sleeping/living, social, looking for work, social or community organizations) Where do PWID hang out when they don’t use drugs? Where do they go to find drugs? Where do they use drugs? How old are most of your peers? Do PWID mix with PWID of different ages? Or do young and old PWID largely keep separate? What other factors might bring PWID?
For all populations: 2.19. 2.20. 2.21. 2.22. How often and in what situations do you meet with <target population> who are from other parts of the city? Do <target population> move in and out of <city> a lot? How many <target population> move to <city> from other regions? If so, why do they move here? How often? Besides <country>, from which countries are some of the <target population> you know? How many <target population> do you know? Please take your time to think about this. Don’t tell me their names; I just want to know about how many people you know. a. How many live in <city>? b. How many have you seen in the past month? c. How many of these people have you seen in the past two weeks?
SECTION I: APPENDICES
49
3. Information on the Population’s Behaviors Information in this section can be used to design the survey as well as develop response categories for the questionnaire. For sex workers: 3.1. 3.2. 3.3. 3.4. 3.5. Where do people in <city> sell sex? a. What hours of the day/night do they usually work in each of these places? b. What days of the week do they usually work in each of these places? c. How many sex workers usually work in in each of these places on a given day, evening, or night? Do most SW sell sex in just one place or many places? Where do SW normally find clients? [Probe] → Lodge/brothel, hotel, street, market, bar, club, restaurant, private home. Where do SW normally have sex with clients? [Probe] → At the place they meet or somewhere else? Where? Does anyone monitor the money SW make or the clients they bring in? [Probe] → Brothel owner, pimp, “mama”
For men who have sex with men, or transgender women: 3.6. 3.7. 3.8. 3.9. How do you meet other MSM/TG? How and where do you meet sex partners? What internet sites to people use to meet? What kinds of lubricants do MSM/TG use?
For people who inject drugs: 3.10. 3.11. 3.12. 3.13. Where do you normally find drugs? Where do you normally find injecting equipment? What kinds of drugs do people you know use? What kinds of drug mixtures do people use?
4. Stigma, Violence, and Service Provision 4.1. 4.2. 4.3. 4.4. 50
What is it like to be <target population> in <country><target population>? What about in <city>? What kind of problems do <target population> face in <country> because they are <target population>? What about in <city>? If an <target population> faces verbal abuse, stigma, or violence, who can they turn to? Who can help them? What <target population> support organizations exist? a. What is the name of the organization (formal or informal)? b. Is it only for <target population>? c. How does it give support?
APPENDIX I-7
4.5. 4.6. 4.7. 4.8. 4.9. 4.10.
How many of these groups do you belong to? How common is violence against <target population>? What healthcare services exist for <target population> a. What is the name of the organization/clinic (formal or informal)? b. Is it only for <target population>? c. What services can people get there? What kinds of HIV or STI prevention services are available for <target population> in <city>? a. Which of these HIV/STI prevention services did you get? b. What did you like or not like about the HIV/STI prevention services you got? What do the <target population> you know think about the HIV/STI services available here? Can you tell me about their experiences using these services? What specific advice do you have to improve HIV/STI services for <target population>?
5. SURVEY PARTICIPATION As you know, we are looking for advice for our future survey. In the future survey we will interview people about HIV and what they do to prevent it. We will ask questions about how they meet people and about things they do that put them at risk for HIV. We also want to give free HIV and STI testing and counselling. We will give back the test results to the people tested in the survey. If they test positive for HIV, we will tell them where they can get care and treatment. People will spend up to two hours in the future survey. 5.1. 5.2. 5.3. 5.4. What are your initial thoughts about this survey? a. How would you feel about joining this survey? b. What would keep you from joining this survey? c. What would make you feel more comfortable about joining? Do you think other <target population> would be willing to join in that future survey? Why, or why not? a. What would make your friends feel more comfortable about joining? b. How would you encourage a friend to join the survey? Especially one who is reluctant? Would you or your peers be willing to talk openly with an interviewer about personal sexual behavior if it is anonymous and in a private setting? What about your drug use? What do we need to know to make the survey a success? a. How can we let people know about the survey? b. What can we do to get a lot of people to join the survey? c. Who are the key people or groups we should talk to, to gain support for the survey?
6. BIOLOGICAL TESTING 6.1. 6.2. How do you feel about taking a free HIV test as part of this survey? a. How do you feel about giving blood from a vein in your arm or finger prick? b. Would free HIV testing make people more or less likely to join the survey? Why? c. How can we make HIV testing more acceptable? We are also thinking of offering testing for other STDs. This would involve collecting specimens from the <vagina>, <rectum>, <urine>. We can show people how to collect the specimens themselves or a trained health care worker can do it. Would you be willing to give a vaginal swab, rectal swab, or urine to test for other STDs? a. How would your peers feel about this? What may be the problems with this? b. Would it be better to have the healthcare worker collect the swab or for people to collect it SECTION I: APPENDICES 51
6.3. 6.4. 6.5.
themselves? (Explain how a swab is collected.) c. What might make STD testing more acceptable? Have you ever been checked for STDs before? Did the healthcare worker do a physical examination? a. How would you feel if a healthcare worker looked at your genitals to see if you have a STD as part of this survey? b. Would getting STD checkup be seen as a benefit? If you were told you had a STD, would you go to a clinic for treatment? a. Why or why not? b. Would getting STD treatment as part of the survey encourage participation? Which kind of person would you prefer to take swabs and test you for HIV? a. Would men and women both be okay? b. Are there people you would not be comfortable with?
7. Interview Administration Now I have some questions about how to collect personal information in the next survey. 7.1. 7.2. 7.3. 7.4. Would you feel comfortable being honest with an interviewer? Would you be more comfortable being honest to a computer? How would you feel about using a touchscreen computer or tablet to answer interview questions? Would you like the computer to read the questions to you? (It may be beneficial to show people what a tablet looks like.) How would you feel about a trained researcher asking you the interview questions? Describe the type of person with whom you would feel ok to answer personal questions in an interview. a. Would men and women both be okay as interviewers? b. Would you prefer other <target population> or people who are not <target population>? c. Are there people you would not be okay with?
Languages 7.5. 7.6. What languages do you speak? What is your main language? What do most of your peers speak? How many of your friends would speak neither of these languages? What languages can you read? Can most of your peers read? Are there <target population> who cannot read? What can you tell me about these people? Are they of a certain sub-group of <target population>? a. Is there a difference between the <target population> who can read and those who cannot read? (i.e. SW who are literate may be more likely to be brothel-based, MSM who are literate may self-identify differently from those who are not, PWID who are literate may use different drugs than those who are not). b. How can we make the study known amongst <target population> who cannot read?
8. Survey Logistics Now I would like to ask you a few questions about where and when we do the survey. For sex workers: 8.1. 8.2. For the future HIV survey, would you like us to come to your place of work to do the survey? For TLS: Would you be okay to leave to do the survey? Would you want to be invited to a specific private location like a clinic, private office/building, or other space where you can do the survey when you want?
52
APPENDIX I-7
For men who have sex with men, or transgender women: 8.3. 8.4. For the future HIV survey, would you like us to come to the places you socialize to do the survey? For TLS: Would you be okay to leave to do the survey? Would you want to be invited to a specific private location like a clinic, private office/building, or other space where you can do the survey when you want?
For people who inject drugs: 8.5. 8.6. For the future HIV survey, would you like us to come to the places you buy or use drugs to do the survey? For TLS: Would you be okay to leave to do the survey? Would you want to be invited to a specific private location, like a clinic, private office/building, or other space where you can do the survey when you want?
Contact with authorities 8.7. 8.8. We want this survey to be helpful to <insert target population>. We also want to make sure it is safe for people to be in the survey. What can we do to keep people safe? Should we tell the police or other law enforcement agencies about the survey so they don’t bother people trying to participate? What about others (religious, military, local-level government, etc.)?
9. RDS: Compensation We will want to give something to participants for their time and transport. We want to make sure we do not give too much to avoid people pretending that they are a <target population>. We also do not want to give something so big that people will join the survey just to get it. 9.1. 9.2. 9.3. 9.4. How do you think we can best compensate people for their time and transport? a. Would a gift (ie, phone credit/airtime, food voucher) be acceptable? b. Would money be acceptable? c. How much would it cost to get to the survey site? d. How much should we give for time spent doing the survey? Would health screenings make people more likely to join? We will ask participants to return to the survey office for a second visit about two weeks after the first. This second visit will take about 30 minutes. During the visit, participants will get their test results and will be given compensation for each of their recruits who participated. Would you and your peers be able to return for a second visit after two weeks? Why/why not? We would like to compensate participants who recruit peers. What do you think we should give participants for recruiting peers?
10. RDS: Location, Days, Times, Etc. 10.1. 10.2. What kind of place would you like to see as the main survey site location? What might be some convenient and safe locations? Where would you and your peers feel comfortable coming to participate in a survey? Rented apartment or office? Clinic? NGO? Other location?
SECTION I: APPENDICES
53
10.3. 10.4. 10.5. 10.6. 10.7. 10.8. 10.9.
How easy is it for people to get there? How do you think you and your peers would travel to the survey site? Probe for public transit, shared taxi, private taxi, private motorcycle, etc. About how much would you expect transport to the interview site to cost for most people? Probe to see if this cost is for public transit, shared taxi, private taxi, private motorcycle etc. What areas of town are unsafe and should be avoided for the survey? What types of <target population> would not feel comfortable going to certain areas? a. For SW: e.g, brothel-based, street-based, higher class, foreigners? b. For MSM: e.g. non-gay identified, men from other areas? c. For PWID: e.g., professionals, homeless? Are there any locations where you would not want to be surveyed? What would make a survey site most comfortable for other <target population>? What times of the day are best for <target population> to go to an interview site to take part in a survey? a. Morning (8 AM–12 PM)? b. Afternoon (12 PM–5 PM)? c. Evening (5 PM–10 PM)?
10.10. What days of the week do you think <risk behavior/population name> are best for people to join the survey? What about the weekend? 10.11. Would appointments or open-walk in times work better? 10.12. We may also ask other groups to participate in a similar study at the same time, such <risk behavior/ population name of other population to be studied>. How would you feel about being at the same survey site as they?
11. RDS: Coupon Design and Recruitment Let me describe one method we want to use to find people to join the survey. We would give participants three coupons to give out to peers who are also <target population>. For each friend who shows up to the study location with the coupon and participates, the person who referred them will get a small amount of money. Your friend would also be interviewed, get free HIV testing and STI screening and treatment, be told about HIV and how to prevent it, and get coupons to give out to his friends so they can also participate. Now we would like to know: 11.1. 11.2. 11.3. 11.4. How would you feel about giving a coupon to your peers and asking them to do the survey? a. Do you think these people would agree to join in the study? b. Do you think these people would be willing to refer others to participate? Do you know of any especially influential <target population> who may be good at referring people? This would be someone who knows a lot of other sex workers and is well liked by peers. a. Would they be willing to talk to us? b. Can you tell us a little bit about them? [Ask participants to ask identified peer leaders to contact survey coordinator. Give them referral card with contact information for survey coordinator.] What color should the coupon be? What information should on the coupon? Survey hours, contact phone number, survey location, survey name? a. Can you think of any information that really needs to be there to make it easy for participants to come see us? b. Can you think of anything that if on the coupon would make you not want to join? c. What languages should be on the coupon?
54
APPENDIX I-7
11.5
d. How could we adapt the coupon for those who can’t read? e. Do you think a map on the coupon is helpful for finding the survey site? We may want to collect fingerprints from two fingers [Show the respondent the fingers you are referring to] to make sure the same person does not try to join the survey several times. The computer will turn your fingerprint into a secret code. No one can use the code to identify you. How would you feel about giving your fingerprint for such a survey? What concerns might your peers have about this?
12. TERMINOLOGY Now I’m going to ask you to give me the local term of some terms related to <target population>. SEX WORKERS Terms Local terms/Slang Language 1 (specify) 1. Sex workers that are paid only in money 2. Sex workers that are paid only in goods and favors (e.g. food, lodging) 3. Sex workers that work on the streets 4. Sex workers who work in brothels/lodges 5. Young SW 6. SW that are new to selling sex 7. SW from other parts of the country 8. Foreign SW 9. SW who sell to “rich” clients or who charge a lot of money 10. People who have sex for money or goods and may act as long- term girlfriends/boyfriends. 11. Oral sex 12. Anal sex 13. Vaginal sex 14. Sex without a condom 15. Pimp or other person who controls SW or his/her money, may identify clients for SW 16. Other local terms that may come up when discussing sex workers or sex work Language 2 (specify)
SECTION I: APPENDICES
55
MEN WHO HAVE SEX WITH MEN/TRANSGENDER WOMEN: Terms Local terms Language 1 (specify) 1. Gay-identified man 2. Non-gay-identified man 3. Insertive or top partner 4. Receptive or bottom partner 5. Male SW 6. Club or venue for gay men 7. Oral sex 8. Anal sex 9. Older gay man 10. Younger gay man 11. Sex without a condom 12. Versatile, someone who is sometimes the insertive and sometimes the receptive partner 13. Out person who is effeminate 14. Gay-friendly person 15. Men who think of themselves as women, and who may dress as women, and have sex with other men (e.g. transgender) 16. Men who have cut/altered their genitals and have sex with other men (e.g. transgendered) 17. A trans person who sells sex on the street 18. Very effeminate MSM, especially young ones 19. Married men who have sex with other men (e.g. down low) 20. Men who have sex with men and women 21. Other local terms that may come up when discussing male sex workers or male sex work Language 2 (specify)
56
APPENDIX I-7
PEOPLE WHO INJECT DRUGS Terms Local terms Language 1 (specify) 1. Terms in the injecting process 2. Someone who has sex in order to get drugs 3. Someone who is wealthy and injects drugs 4. Overdosing 5. Methadone assisted therapy 6. Name of drugs 7. Other local terms/slang that may come up when discussing injection drug use Language 2 (specify)
That is the end of our interview. Thank you so much for sharing your thoughts with me. Do you have any questions, or is there anything that you would like to add before we end? If you have further thoughts about any of the issues we discussed today, please call [insert number where interviewer can be contacted]. END RECORDING HERE
SECTION I: APPENDICES
57
TO BE COMPLETED BY THE INTERVIEWER: Please note your impressions about the session, its main themes, and the comments and reactions of participants. INTERVIEW WAS: ____ ROUTINE ____ NOT ROUTINE IF NOT ROUTINE, WHY:
ANY ADVERSE REACTIONS IN THE INTERVIEW: ____ YES ____ NO IF YES, SPECIFY:
OTHER OBSERVATIONS/COMMENTS:
58
APPENDIX I-7
Section
APPENDIX I-8
I
FORMATIVE MAPPING Mapping can be conducted as part of a formative assessment interview or a focus group. It can also be an exercise on its own. If on its own, participants involved will need to be consented before they can participate. This guide can help investigators conduct the mapping exercise. Validation of Known Locations We are interested in learning more about the experiences of <target population>, but we are not sure we know where all <target population> are in <city>. I am going to show you a map. Please tell me where on the map groups of <target population> can be found. This can include places such as streets, markets, brothels, bars/clubs, hotels, barracks, or any other publicly accessible place. For each location, please confirm if <target population> can be found there, then tell me: a. What kind of <target population> are there (age, nationality, religion, <sub-group>, etc.)? b. About how many <target population> are at this location on a typical day, evening, or night? Please give me your best estimate or a range. c. Could we talk freely to <target population> there? d. Must we obtain permission from someone (owner, manager, madame, etc.) first? e. Could we conduct interviews there? f. Could we test people for HIV there? g. For SW: What types of clients go there?
Identification of New Locations There might also be places on this map that we have not identified, where [target population] can be found. Do you know of any other such places that are not on this map? For each new site identified, ask the following questions: a. b. c. d. e. f. g. What kind of [target population] are there (age, nationality, religion, <sub-group>, etc.)? About how many [target population] are at this location? Please give me your best estimate or a range. Could we talk freely to [target population] there? Must we get permission from someone (owner, manager, madame, etc.) first? Could we do interviews there? Could we test people for HIV there? For SW: What types of clients go there?
SECTION I: APPENDICES
59
Section
I
APPENDIX I-9
SAFETY AND SECURITY GUIDE This appendix provides guidance on safety issues that may be encountered while implementing a BBS. PHYSICAL SAFETY Survey staff must be alert to their own and survey participants’ safety. Advanced planning can help prevent problems. Generally, the survey site supervisor is responsible for maximizing the safety, security and privacy of survey participants and staff, and that clients feel comfortable at the site. It is vital that all survey investigators and site staff prioritize privacy and safety of all participants and survey staff. Emergency contact information for each member of the team should be easily accessible to the site supervisor.
SITE SAFETY An RDS survey will have at least one site where data collection takes place. For safety reasons, sites should ideally have more than one exit which can be opened from the inside without a key. For TLS, security guards should be considered, especially when data collection takes place at night. Receptionists should be stationed at any entrance during site hours, and anyone in the reception area should not be able to see or hear individuals while they are participating in survey procedures. For all surveys, develop a code word or phrase to call for assistance from a co-worker. For example, if a staff member needs help with an uncooperative participant, but does not want to alarm the individual in question or other people at the site, she/he would ask one of her/his co-workers to bring them the “blue folder” to indicate that they needed assistance. Also, if an unwanted visitor comes into the survey site, staff could use a code word to alert other staff that a breach of confidentiality or possible danger to clients may occur. This should trigger the supervisor to come to assist and enact all other emergency security actions. In general, people who are not survey staff or participants should not be allowed into the survey site. Sometimes participants will want to bring friends or their young children to the survey site. Investigators should predetermine if this constitutes an acceptable risk and institute their decision as a policy. However, under no circumstances should friends or family be allowed in the interview or testing area.
How to handle visitors or callers in different scenarios The best way to deal with a problem is to prevent a sensitive situation from escalating. Be polite and friendly to any visitor or caller to the site. If a visitor or client becomes angry: • • • • Remain as calm as possible and assure them that you are trying to help them Listen to and acknowledge concerns Lower your voice, tone, and tempo Avoid becoming defensive
60
APPENDIX I-9
Questions about the survey Any individual who comes to or calls the site asking about sensitive matters regarding the survey should be referred to the supervisor. The supervisor should follow the following guidelines: • Obtain the following information from the visitor(s) before disclosing ANY information o Who do they represent (legal, law enforcement, concerned or curious citizen, relative, participant, or others) o What is the purpose of the visit and how you can help them o What information they are seeking Provide only basic information about the survey. o Indicate that the survey provides HIV testing, counseling and referrals for ALL people (This is the single, overriding communication objective regarding questions about specific risk groups.) o Do not talk about specific key populations participating in the survey or their risk behaviors o Avoid specifying what types of questions are asked or data are collected o If pressed, refer the individual to the principal investigator or Ministry of Health representative • • If the individual makes implied or direct threats, state that the site does not collect information or knowingly serve any particular group including the group under question (e.g. MSM or FSW), but provides non-discriminatory services to everyone.
•
Prevent or limit access to data or client information o Stress and show IRB approvals. State that survey staff are obliged to abide by the data access and confidentiality procedures outlined in the protocol and may not breach them without informing/seeking permission from these bodies. o If the country’s laws allow: Request a warrant, state that data are not stored on location, and that participants have a right to privacy under the law o State that data are stored at a central location and not onsite. Refer the individual to the principal investigator or Ministry of Health representative for more information o If a law enforcement officer provides a warrant OR makes direct threats of physical harm and refuses to provide a warrant: • Inform them that protocol dictates that you are required to contact the funding agency and the local IRB before releasing information • Provide the individual with what he or she requests, making an attempt to de-identify all information first and immediately contact individuals on the chain of command afterwards • Instruct staff to emergency security actions if necessary
Proper procedures to report incidents Not every staff member can be expected to know how to respond to requests for information or access to the site. Therefore, there should be standard operating procedures (SOP) to direct staff how to refer questions or requests to a higher authority. The chain is: 1. Supervisor/Field staff coordinator: If asked for any information on the survey, all staff should refer questions to the site supervisor. The site supervisor can answer basic questions about the survey activities, but should not release data, participate in interviews with the media, or allow visitors to interact with participants. 2. Principal investigators: The principal investigators are the only individuals who should respond to requests for data, conduct interviews with the media, or allow visitors to interact with participants. In cases of serious events, or for legal and security recommendations, the PI should contact the appropriate government official (e.g. ethics committees and Ministry of Health representative). Discuss the SOP as a team and adapt it accordingly. Ensure that all staff members know the SOP. All staff should know who to inform in case of an emergency and how they can be contacted.
SECTION I: APPENDICES
61
Emergency security actions In the event of a possible security breach or safety concern, the following positions have specific responsibilities. Supervisor/Field staff coordinator: 1. 2. 3. 4. 5. Speak to the individuals who attempt to enter the facility. Minimize risk to participants and staff. Secure all personally identifiable data (names, addresses, phone numbers). Contact offsite management staff in the chain of command. Evacuate the facility if necessary. In case of fire or other emergency, the site supervisor should be the last person to leave the facility
Interviewers/HIV counselor/nurse: 1. Ensure that participants are safe and protected from access by intruders. Escort them to safe exits if necessary. 2. Secure personally identifiable information (names, addresses, phone numbers), or deliver sensitive data to data entry/data management staff. Data entry/data management staff: 1. Shut down all computers and other data capture devices with password encryption. 2. Lock up all paper forms with sensitive information and portable data devices (audio recorders, PDAs, laptops, etc.) in locking cabinets or safes. 3. Lock rooms where possible.
Security with participants sexual harassment If a participant makes inappropriate sexual advances or sexually harasses staff, staff should feel comfortable to address the participant as needed. If a participant is behaving inappropriately, first remind him/her of the survey objectives and procedures and indicate that sexual offers are not appropriate. If the participant continues, tell it is acceptable to terminate survey participation. Immediately terminate participation if a staff member feels uncomfortable or unsafe. Make note of it and report it to the site supervisor.
Drunk, high, or drowsy participants A participant may not be able to complete the interview, give accurate answers, or may appear to be very drowsy during survey participation due to sleep deprivation, alcohol, or drugs. Such a person may not be able to provide informed consent and therefore is ineligible for survey participation. Explain to the participant that the computer has not selected him/her to participate in the survey and thank the individual for his/her time. Document the incident accordingly. It is acceptable to terminate participation partway through if a participant stops being coherent or becomes mentally unable to participate. Thank the participant for their time and document what happened.
Friends or family of participants Individuals accompanying participants or purporting to be friends/family of participants must be asked to remain in the reception area or other location where they cannot observe or disrupt data collection activities.
62
APPENDIX I-9
Section
APPENDIX I-10
I
SPECIMENS USED FOR HIV TESTING The following information, excerpted from the UNAIDS/WHO Guidelines for Using HIV Testing Technologies in Surveillance: Selection, Evaluation, and Implementation – 2009 update,1 provides an overview of specimens used for HIV testing; advantages and disadvantages of specimen types; specimen collection, processing, storage, and testing; and labeling and recording specimens. This appendix can be used to supplement information in Chapter 6: Biomarker Considerations.
Specimens Used for HIV Testing Many types of specimens can be used for biological surveillance of HIV: plasma, serum, whole blood, DBS and oral fluid. The choice of specimen collected depends on the logistics, populations and sites selected, and the HIV testing strategy and algorithm. Specimens must be collected, tested and stored in an appropriate manner in order to obtain accurate and reliable results. For serosurveillance activities, specimens are usually collected and stored prior to HIV testing at a regional or national laboratory. Serum, plasma and DBS can be stored and tested at a later date; specifications for storage will depend on the type of specimen collected. Specimens not tested on site at the local level will need to be transported to a regional or national laboratory for testing. The methods by which specimens are transported will depend on the country’s infrastructure. Few countries may have courier systems linking health care facilities and laboratories. More frequently, the field surveillance staff members themselves transport the specimens from the local to the national laboratory. However, other options such as public transportation can be explored.
Selecting Specimens Advantages and disadvantages of serum and plasma HIV testing of serum and plasma, which can be collected by venipuncture (see below), have the following advantages and disadvantages: Advantages • • • • Have higher concentrations of HIV antibodies than oral fluids Have the potential for additional routine testing (e.g. syphilis, hepatitis B, hepatitis C) from a single specimen Have the potential for special studies (e.g. testing for recent infections, HIV typing [HIV-1 vs. HIV-2], HIV subtyping, antiretroviral [ARV] resistance) Are easy to collect and test with a trained phlebotomist and a laboratory
1
http://www.unaids.org/en/media/unaids/contentassets/dataimport/publications/irc-pub02/jc602-hivsurvguidel_en.pdf
SECTION I: APPENDICES
63
Disadvantages • • • • • Require invasive collection technique Require skilled staff to collect and process Compared with oral fluid, require more equipment (e.g. needles, tubes or lancets) and biohazard waste facilities May be challenging to collect in non-clinical settings as venipuncture is required Pose a greater risk to health-care workers and staff through inadvertent exposure, both because of higher HIV concentrations and the use of sharp collection devices
Advantages and disadvantages of whole blood HIV testing of whole blood, which can be collected by venipuncture or finger-stick (see below), has the following advantages and disadvantages: Advantages • • • • Has higher concentrations of HIV antibodies than oral fluids Has the potential for additional routine testing (e.g. syphilis, hepatitis B, hepatitis C) from a single specimen Is easy to collect and test with trained phlebotomist, if venous whole blood Is easy to collect in non-clinical settings if finger-stick
Disadvantages • • • • • Requires invasive collection technique Requires skilled technician for collection Compared with oral fluid, requires more equipment (e.g. needles, tubes or lancets) and biohazard waste facilities May be challenging to collect in non-clinical settings if venipuncture is required Poses a greater risk to health-care workers and technicians through inadvertent exposure, both because of higher HIV concentrations and the use of sharp collection devices
Advantages and disadvantages of dried blood spots Dried blood spots can be prepared for HIV testing by collecting venous or finger-stick whole blood and dropping an amount onto a filter paper. DBS have the following advantages and disadvantages: Advantages • • • • • Are easy to collect in a clinical or non-clinical setting, depending on whether venipuncture is available Do not require a centrifuge or other equipment for processing the blood specimen Once dried, can be stored at room temperature for a short time Can be transported easily to the central laboratory for further testing Facilitate testing for prevalence, incidence and special studies such as resistance testing
Disadvantages • • • • • Require specific filter paper (see below) for preparation Potential for less accuracy (false-positives and/or false-negatives) if the test is not optimized for DBS Modified procedure required for DBS elution and HIV testing Make the testing process more lengthy as an elution step needs to be performed Limited number of tests validated for use with DBS specimens
Advantages and disadvantages of oral fluid Other specimens besides blood and blood products can be used for HIV testing. For linked testing, where informed consent must be obtained, oral fluid may be used.
64
APPENDIX I-10
Advantages • • • • Does not require a trained laboratory technician for specimen collection and processing, can be collected by a trained health worker Does not require contact with possibly contaminated laboratory materials, e.g. used needles or lancets that need biohazard waste facilities for sharps disposal Can be collected in a variety of field settings, including non-clinical settings May be more acceptable to hard-to-reach populations than specimen collection requiring venipuncture or fingerstick. Therefore, a greater percentage of the target population may agree to be tested. Disadvantages • • • • • May require special collection devices Currently available testing technologies used for oral fluid specimens are limited but additional new tests are being validated Cannot be used to perform additional testing for special studies (e.g. testing for recent infections, HIV subtyping, ARV resistance) Same specimen cannot be used to confirm initial reactivity with a second test; therefore, a second specimen must be taken, i.e. whole blood, serum, plasma for further testing (this is specific to the OraQuick HIV rapid test as the oral fluid collection device and test are one and the same) Should not be used for confidential linked testing (i.e. with the return of results to the individual)
Recommendation Blood (serum, plasma, DBS) is the preferred specimen for testing because it has a higher concentration of HIV antibodies than oral fluid. It also allows for additional testing, including for syphilis, hepatitis B and hepatitis C, and for special studies of HIV type and subtype, and ARV resistance.
Collecting, processing and storing blood, serum and plasma specimens Blood needed for an HIV test can be collected either by venipuncture (whole blood, serum, plasma) or by finger-stick (whole blood). To ensure proper specimen labeling, label the blood collection device (either vacutainer, cryovial, or DBS card) before collecting the specimen. Put the participant at ease by explaining the blood collection process and which test(s) will be performed. Processing blood collected by venipuncture To collect blood by venipuncture, follow local clinical or laboratory procedures and adhere to safety procedures. The following steps are recommended for processing blood collected by venipuncture: Collect up to 10 ml of blood from the patient’s vein into a sterile 10 ml tube. For serum, blood is collected in a red-top tube (without anticoagulants). For plasma, blood is collected in a purple-top tube (with anticoagulants, e.g. EDTA). For safety reasons, the use of an evacuated blood collection system (e.g. Vacutainer® tube) is recommended. Note: Obtaining an additional tube of blood during routine blood collection solely for the purpose of unlinked anonymous testing is considered unethical and is not advised. 1. If the blood specimen will not or cannot be processed immediately (e.g. no centrifuge is available), collect the blood in a red-top tube and allow it to stand at room temperature for at least 20–30 minutes, and then remove the serum. Usually, plasma takes longer (~1 hour) to separate without centrifugation than serum collected in a red-topped tube. Process (see Step #3) and test within 24 hours to avoid hemolysis of the specimen.
SECTION I: APPENDICES
65
2. Centrifuge the specimen to separate the serum (without EDTA) or plasma (with EDTA). If blood is collected for serum, allow the blood to stand for at least 20–30 minutes so that a clot forms before the specimen is centrifuged. In general, the specimen should be centrifuged at 300–400 g or 1200–1500 rpm for 10 minutes. After the specimen is centrifuged or has had time to separate, use a clean plastic pipette (do not pour) to remove an aliquot of 0.5–2.0 ml off the top layer. Transfer this to another sterile labelled plastic tube or cryovial (1.5–2.0 ml with a screw cap) and tighten the cap. The specimen is ready for storage and testing. Storing serum and plasma collected by venipuncture To store serum and plasma, consider the following: • • • • Make sure the cap is tight on the labelled cryovial or plastic tube. Do not use glass tubes for storing specimens. Place the cryovials in a cardboard freezer box with a partitioned insert. If the specimens are to be transported to the testing laboratory, store the specimens at 4–8° C for up to a maximum of 1 week. For longer-term storage, the specimens should be frozen at –20° C or below. Pack the specimens upright in a cooler containing cold packs for transport to the testing facility. Limit the number of freeze/thaw cycles because it may impact the HIV test results and subsequent additional testing.
Collecting blood by finger-stick Blood collected by finger-stick can be used to perform rapid tests or make DBS on filter paper. Finger-stick collection may be preferred in rural and non-clinical settings, which often do not have trained phlebotomists or laboratory facilities with appropriate equipment (e.g. centrifuges). 1. To obtain a finger-stick specimen, massage the finger (preferably the middle or ring finger), which will cause blood to accumulate at the tip of the finger. 2. Cleanse the finger pad (not just the tip or side of the finger) with 70% isopropyl (rubbing) alcohol. Wipe away the alcohol with a sterile gauze pad. 3. Use a sterile lancet to firmly prick the finger pad. Wipe the first drop of blood off the finger with sterile gauze before collecting subsequent blood using a collection device to place on the rapid test device or on the filter paper for the DBS. If the original puncture is inadequate, the same site should not be reused; another site or finger should be used. Avoid milking or squeezing the puncture as this may cause hemolysis of the specimen and could invalidate the test result. The ear lobe may be pricked instead of the finger. Preparing and storing DBS specimens for an HIV test Blood from a finger- or ear-lobe stick can be used to make DBS. Although finger-stick is the most typical method, DBS can also be prepared by using blood collected in a tube with an anticoagulant. DBS have the advantage of being easy to transport, without the need for a cold chain. Apply blood directly from a finger or a pipette onto special filter-paper (Schleicher and Schuell Grade 903 filter paper). The paper may come with preprinted circles that will contain approximately 100 µl blood when completely filled. If the paper does not have preprinted circles, place blood on the paper so that it makes a circle with a 1.5 cm diameter. Allow the blood to soak through and fill the entire circle. Caution: If the blood does not saturate the filter-paper, do not use that paper. 1. Label the side of the filter paper with a specimen reference code after the filter paper is saturated with blood (circle is filled). 2. Suspend filter-paper strips containing the filled circles during the drying process to allow air to circulate around the paper. Stands for holding the strips are commercially available. However, strips may also be dried by placing them between two books (taping the edges of the strips to the books with sticky tape) on a table or a laboratory bench top so that the blood-containing part of the paper is not in contact with the surface of the table or laboratory bench top. Be sure not to get tape on the blood spots.
66
APPENDIX I-10
3. Let the blood spots air dry horizontally at room temperature for at least 4 hours or(and for at least 24 hours in humid climates). Do not heat or stack blood spots, do not allow them to touch other surfaces while they are drying, and keep them away from direct sunlight, dust, and insects while they are drying. 4. After blood spots have been adequately dried, wrap the strip in one sheet of glassine paper or plastic to prevent carryover of specimen from one sheet to another. 5. Place the wrapped strips in a gas-impermeable bag (zip-lock bag) with desiccant and humidity indicator cards. Approximately 20 strips may be placed in each bag. Bags may be kept at room temperature for up to 30 days and then stored at 4° C for up to 90 days. If the DBS in their plastic bags are to be stored for more than 90 days, they should be maintained at –20° C. Properly stored DBS have been shown to be stable for at least two years. The bags should be placed with appropriate documentation in a sturdy envelope for shipment. If additional testing, such as resistance testing, is anticipated, DBS must be stored at –20° C or below immediately after the DBS specimens are dry.
Collecting and storing oral fluid Collecting oral fluid For specimen collection, follow test instructions as part of a standard operating procedure. Oral fluid collection devices are available and may be used, if indicated. Some rapid test devices contain an oral fluid collection pad at one end which facilitates collection and testing. Oral fluid can only be used with certain EIAs and rapid tests designed for oral fluid specimens, such as the OraQuick brand. Additional rapid tests using oral fluid are currently under field evaluation. The following are the general steps for collecting a specimen: 1. Use a specially treated absorbent pad attached to a plastic stick (usually provided by the test kit manufacturer). 2. Collect the specimen. Collection procedures are specified by the manufacturers of collection devices and must be carefully followed. Then place the pad into a vial containing a preservative solution (usually provided by the test kit manufacturer). 3. If an oral fluid-specific rapid test (e.g. OraQuick) is performed, storage and transport are not necessary. Due to the complexity of the test, oral fluid specimens collected for EIAs should be sent to a laboratory performing EIAs for analysis. Storing oral fluid Oral fluid specimens should be stored at 4° C for a short period of time. They should be refrigerated during shipment. Specimens should be frozen (–20° C or below) if stored for an extended period of time. Once thawed, they can be refrozen once. Consult the test kit insert for more specific storage information.
Labelling and recording collected specimens Labelling specimens The plastic tube, cryovial or filter-paper containing the specimen must be labelled with a specimen code at the time of collection and processing. If labels are used, make sure the label is placed on the side of the tube, not on the cap. Preprinted cryolabels designed to adhere during freezer storage should be used when specimens are stored in cryovials. It is important that freezing does not affect the visibility of the printing on the label. Surveillance coordinators should provide the field staff responsible for specimen collection with a series of labels or permanent markers and the codes to be used. For unlinked anonymous testing, label the tube only with a new specimen code unlinked to personal identifying information.
SECTION I: APPENDICES
67
Recording specimens and test results A separate laboratory logbook or line-listing for surveillance activities should be maintained to record HIV test results by the corresponding code. The logbook should be accessible only to laboratory and surveillance staff; it should be secured in a locked drawer or cabinet when not in use to ensure confidentiality of the persons’ test results as well as their participation in surveillance activities. For unlinked anonymous testing, the logbook or line-listing should contain only the new specimen codes and corresponding HIV test results; no personal identifying information on the participants whose specimens are tested should be included. HIV test results can be matched by the new specimen code to the demographic information abstracted earlier on the surveillance form.
68
APPENDIX I-10
Section
APPENDIX I-11
I
MATERIAL TRANSFER AGREEMENT The purpose of this appendix is to provide a template of a material transfer agreement for off-site testing.
This Agreement is by and between the: • • <INSTITUTION A>: Address <INSTITUTION B/TESTING FACILITY>: Address
PURPOSE 1. This Agreement establishes the terms and conditions under which <INSTITUTION A> will provide <TYPE OF BIOLOGIC SPECIMENS> to <INSTITUTION B/TESTING FACILITY> in order to evaluate the specimens collected in the survey <”SURVEY NAME”> and how <INSTITUTION B/TESTING FACILITY> will store, maintain, and use the <BIOLOGICAL SPECIMENS> collected from <COUNTRY> citizens. 2. This Agreement covers the sharing of <TYPE OF BIOLOGIC SPECIMENS> from <INSTITUTION A> with <INSTITUTION B/TESTING FACILITY> for the specific purpose of <SURVEY OBJECTIVES> and the sharing of the findings by <INSTITUTION A> with its <COUNTRY> stakeholders: <ALL PARTIES>. 3. The investigators and collaborating institutions as outlined in the protocol <ALL PARTIES> retain shared ownership of the data and coauthor rights. 4. The terms of this Agreement can be changed only by a written modification of the agreement by the authorized signatories of the Parties (or their designated representatives) to this Agreement or by the parties adopting a new agreement in place of this Agreement. 5. Any other uses will be subject to prior approval by <ALL PARTIES> and the survey’s Principal Investigators (PIs). PERIOD OF AGREEMENT The material transfer agreement will begin immediately upon the signatures of all responsible parties and will be terminated upon the mutual agreement between <INSTITUTION(S)> and <INSTITUTION B/TESTING FACILITY>.
SECTION I: APPENDICES
69
BACKGROUND 1. The above named survey relies on an approved protocol. The goal of this survey is to <SURVEY OBJECTIVES>. <INSERT SHORT SUMMARY OF PURPOSE AND SIGNIFICANCE OF THE SURVEY/RESULTS>. 2. This is a/an <SURVEY DESIGN>. <INSERT SHORT DESCRIPTION OF STUDY PARTICIPANTS, LOCATION, ENROLLMENT INFORMATION AND CONSENT PROCESS, BLOOD DRAW PROCESS, AND DATA MEASURES>. 3. Overview of enrolled participants (EXAMPLE BELOW IS BY SEX, ART STATUS AND DURATION). Category Male Female Total
ART Naive
6 months
12 months
24 months
Total
4. No personal identifiers are used to label specimens. This survey is not expected to have clinical utility for the survey participants. There will be no (further) contact with human subjects. There is minimal or no risk involved with this transfer since serum specimens have previously been collected and contain no patient identifiers. Justification for testing outside <Country> Estimating the <SURVEY MEASURES> provides important benchmark data for future studies in <COUNTRY> that estimate <DISEASE PREVALENCE/INCIDENCE> using <SPECIFY TESTS>. It is therefore critical to <INSERT PROPOSED PROCESS/TESTING HERE>. PREPARATION AND TRANSPORT OF THE SPECIMENS FROM <CITY>, <COUNTRY> TO <CITY>, <COUNTRY> 1. All specimens will be shipped to <CITY>, <COUNTRY> 2. Each specimen will be <SPECIFY PREPARATION FOR SHIPPING PROCEDURES >. Any remaining <TYPE OF BIOLOGIC SPECIMEN> will be <SPECIFY HOW THIS WILL BE HANDLED AND STORED>. 3. The <SPECIFY SPECIMEN CONTAINERS> will be transferred from the central repository to <INSTITUTION B/TESTING FACILITY> laboratories in <CITY>, <COUNTRY> in <SPECIFY PACKAGING> with appropriate documentation for shipping. 4. <SPECIFY BIOLOGICAL SPECIMENs> will be shipped following <NAME OF TRANSPORTING ENTITY> and United Nations packaging recommendations, with valid <REGULATING BODY FOR BIOLOGICAL SPECIMENS AS APPROPRIATE> permits for the shipping and receiving of tissues from human origin. INTENDED USE AND CONSTRAINTS ON USE OF THE SERUM SPECIMENS <INSTITUTION B/TESTING FACILITY> in <CITY>, <COUNTRY> will perform the laboratory testing. 1. <SPECIFIC TESTING> using <SPECIFIC TESTS> such as <EXAMPLES>. 2. Additional testing may include <SPECIFIC INFORMATION>. GOVERNING LAW This agreement is subject to the governing Laws of <COUNTRY>. DISPUTE RESOLUTION 1. Amicable Settlement The parties shall use their best efforts to settle amicably any dispute arising from or in connection with this Agreement or the interpretation thereof. Each Party shall select its representatives to carry out such amicable discussions
70
APPENDIX I-11
2. Arbitration If the dispute has not been settled amicably within fourteen days or such longer period as may be agreed upon between the parties from when the discussions were commenced, any party may elect to commence arbitration. Such arbitration shall be referred to arbitration by a single arbitrator to be appointed by agreement between the Parties. Such arbitration shall be conducted in <CITY>, <COUNTRY>. To the extent permissible by <COUNTRY> Law, the determination of the arbitrator shall be final and binding upon the parties. On behalf of both parties the undersigned individuals hereby attest that he or she is authorized to enter into this Agreement and agrees to all the terms specified herein. Signed for and on behalf of: <INSTITUTION A>
_____________________________________________ ______________________ <Name of Institution A Director>, Director Date:
<INSTITUTION B/TESTING FACILITY> _____________________________________________ <Name of Institution B Director>, Director
______________________ Date:
SECTION I: APPENDICES
71
Section
I
APPENDIX I-12
INTERPRETATION OF HEPATITIS SEROLOGIC TEST RESULTS Hepatitis B serologic testing involves measurement of several hepatitis B virus (HBV)-specific antigens and antibodies. Different serologic biomarkers or combinations of biomarkers are used to identify different phases of HBV infection and to determine whether a person has acute or chronic HBV infection, is immune to HBV as a result of prior infection or vaccination, or is susceptible to infection. This chart can be used for classifying phases of HBV infection when analyzing BBS HBV test results.
Antigen/Antibody HBsAg Anti-HBc Anti-HBs HBsAg Anti-HBc Anti-HBs HBsAg Anti-HBc Anti-HBs
Result Negative Negative Negative
Interpretation
Susceptible
Negative Positive Positive Immune due to natural infection
Negative Negative Positive Immune due to hepatitis B vaccination
HBsAg Anti-HBc IgM anti-HBc Anti-HBs
Positive Positive Positive Negative Acutely infected
HBsAg Anti-HBc IgM anti-HBc Anti-HBs
Positive Positive Negative Negative Chronically infected
72
APPENDIX I-12
Antigen/Antibody
Result
Interpretation Interpretation unclear; four possibilities: 1. Resolved infection (most common) 2. False-positive anti-HBc, thus susceptible “Low level” chronic infection Resolving acute infection 3. “Low level” chronic infection 4. Resolving acute infection
HBsAg Anti-HBc Anti-HBs
Negative Positive Negative
Definitions: • Hepatitis B surface antigen (HBsAg): A protein on the surface of hepatitis B virus; it can be detected in high levels in serum during acute or chronic hepatitis B virus infection. The presence of HBsAg indicates that the person is infectious. The body normally produces antibodies to HBsAg as part of the normal immune response to infection. HBsAg is the antigen used to make hepatitis B vaccine. Hepatitis B surface antibody (anti-HBs): The presence of anti-HBs is generally interpreted as indicating recovery and immunity from hepatitis B virus infection. Anti- HBs also develops in a person who has been successfully vaccinated against hepatitis B. IgM antibody to hepatitis B core antigen (IgM anti-HBc): Positivity indicates recent infection with hepatitis B virus (≤6 mos). Its presence indicates acute infection. Total hepatitis B core antibody (anti-HBc): Appears at the onset of symptoms in acute hepatitis B and persists for life. The presence of anti-HBc indicates previous or ongoing infection with hepatitis B virus in an undefined time frame.
•
• •
Adapted from: (1) A Comprehensive Immunization Strategy to Eliminate Transmission of Hepatitis B Virus Infection in the United States: Recommendations of the Advisory Committee on Immunization Practices. Part I: Immunization of Infants, Children, and Adolescents. MMWR 2005;54(No. RR-16). (2) CDC Division of Viral Hepatitis fact sheet on Interpretation of Hepatitis B Serologic Test Results: http://www.cdc.gov/hepatitis/HBV/PDFs/SerologicChartv8.pdf
SECTION I: APPENDICES
73
Section
I
APPENDIX I-13
SPECIMEN TRACKING FORM This form can be used to track a group of specimens (a “batch”) that are moved from one site (such as a survey site) to another (such as a laboratory for testing or storage). This form is used along with the Specimen Transportation Log and should be adapted for local use. The form should be completed as follows: INSTRUCTIONS: Line 1: Fill in-a. The survey site or sending laboratory code number or name. b. The driver log batch number. Each batch of specimens sent with a driver should have a number that can be matched with the Specimen Transportation Log. c. Time of departure from the sending site. Circle a.m. or p.m. as appropriate. Line 2: Fill in-a. The receiving lab code number or name. b. Temperature in coolbox (or container in which specimens are carried) on arrival at the receiving lab. Table: a. Column 2—Place labels with pre-printed participant or specimen ID numbers in the cells as indicated. If labels were not used, write in the participant or specimen ID number. b. Column 3—Staff who placed specimen in the batch should enter initials. c. Column 4—Staff (from sending site) who verified that each specimen listed on the form is included in the batch should enter initials before transferring the batch to the driver. d. Column 5—Staff (from receiving laboratory) who verified that each specimen listed on the form was received should enter initials and indicate the date the specimens were received. e. Column 6—Receiving laboratory staff may enter comments about specimens when received (e.g. dried blood spot specimen moldy; plasma hemolyzed; tube cracked and leaking; serum thawed; etc.) f. Column 7— Receiving laboratory staff should note where the specimen is being stored so that it may be retrieved as necessary (e.g. refrigerator number 2, -800C freezer, etc.).
74
APPENDIX I-13
74
SPECIMEN TRACKING FORM Specimens sent from SURVEY SITE or LAB #: DRIVER LOG Batch No: Time of departure from SENDING SITE: _____________ a.m./p.m. Temp on arrival at RECEIVING LAB: C
RECEIVING LAB #:
o
Serial No.
Participant / Specimen ID (Place Label Here)
STAFF who prepared batch (initials)
STAFF who checked and transferred batch to DRIVER (initials)
RECEIVING LABORATORY STAFF who checked specimens in and date received (initials/date)
Comments about specimens received
Specimen storage location at RECEIVING LAB
1
2
3
4
SECTION I: APPENDICES
75 SECTION I: APPENDICES
75
Section
I
APPENDIX I-14
SPECIMEN TRANSPORTATION DRIVER LOG The purpose of this form is to track the transportation of a group of specimens (referred to as a “batch”) that are moved from one site (such as a survey site) to another (such as a laboratory for testing or storage). This form is intended to be used along with the Specimen Tracking Form and should be adapted for local use. INSTRUCTIONS: Line 1: Fill in-a. Driver name (each driver should maintain his or her own log) b. Supervisor name Line 2: Fill in-a. Survey site or lab code number or name from which driver received batch b. Date the transportation log was reviewed by a supervisor Table: a. Column 1—Enter the batch number that corresponds with the batch number on line 1 of the Specimen Tracking Form b. Column 2—Enter the number of specimens included in the batch c. Columns 3 and 4—Enter the date and time of departure from the sending survey site or laboratory d. Column 5—The sending site staff who transferred batch to the driver should enter their initials e. Column 6—Enter time that batch arrived at receiving lab. If date of arrival is different than date of departure, it should be entered here. If this is a common occurrence a column for date of arrival should be added to the form. f. Column 7—Enter receiving lab staff initials
76
APPENDIX I-14
76
g. Column 8—Driver should sign when the batch has been delivered h. Column 9—Any comments (by driver, sending site staff, or receiving lab staff) about the batch or delivery should be entered here. Comments about individual specimens should be entered on the Specimen Tracking Form.
Specimen Transportation Driver Log Driver name: _________________ Superviser name: ________________ Batch picked up from survey site or lab #: ________________ Date reviewed: __________________ Batch No. No. of specimens in batch Date Time of departure Sending site staff (DD/MM/YY) from survey site initials or lab Time of arrival at receiving lab Receiving lab staff initials Driver signature Additional comments
When log is full, please give to receiving Lab Manager who will keep logs in a binder for record keeping. Please use a new form to log additional shipments.
SECTION I: APPENDICES
77 SECTION I: APPENDICES
77
Section
I
APPENDIX I-15
COMPARISON OF HIV TESTING TECHNOLOGIES The table below compares various HIV testing technologies for survey investigators to consider when planning a survey. Investigators need to consider which methods are appropriate for their survey. For example, rapid tests may be a good choice if specimens will be tested one at a time for providing same-day test results to participants. If specimens will be batched for testing larger quantities at a time by skilled technicians, EIAs may be a better choice.
Table. Comparison of HIV testing technologies: Enzyme immunoassays (EIAs) and rapid tests1. HIV testing technology Specimens Advantages Limitations Cost per test (US$) 0.5–1
Complexity, from simple (1) to highly complex (4)
Enzyme • Serum immuno assays (EIAs) • Plasma
• Can be batched: good for ≥90 specimens at a time
• Dried blood spots • Can be automated • Oral fluid • May identify HIV positives earlier in infection; highly sensitive—window period reduced in fourth-generation EIAs • Standardized classification (i.e. dependent on OD value rather than qualitative (subjective) assessment of rapid test results) • Quality control may be easier as control specimens are tested on plate with test specimens
• Requires skilled, trained lab technicians to perform testing and calculate results • Can take ±2 hours • Requires special equipment • Kits require refrigeration
(4)
78
APPENDIX I-15
HIV testing technology
Specimens
Advantages
Limitations
Cost per test (US$) 1–3
Complexity, from simple (1) to highly complex (4)
Rapid test
• Serum • Plasma • Whole blood • Oral fluid
• Requires minimal equipment and reagents • Can be performed outside a laboratory (on-site testing) • Test results easy to interpret • Results in 30 min or less • Most kits can be stored at up to 30°C
• Not suitable for large numbers of specimens • Positive and negative control specimens often not included in the kit • May cost more per test than EIA (unless EIA plate used to test small numbers of specimens)
For tests based on: • Immunochromatography (1) • Dipstick and flowthrough devices (1–2) • Agglutination (2–3)
Adapted from: 1
WHO/UNAIDS Guidelines for Using HIV Testing Technologies in Surveillance: Selection, Evaluation, and Implementation—2009 update
SECTION I: APPENDICES
79
Section
I
APPENDIX I-16
SPECIMEN TEMPERATURE MONITORING LOGS Each of the tables below is an example of a monthly temperature log for freezers, refrigerators, and incubators. Each table can be copied, pasted, adapted as necessary, and printed on a separate sheet of paper. Usually each piece of equipment being monitored has a log either attached to it or very close by (such as on the wall next to the equipment). Persons responsible for monitoring the temperature should record the temperature (on the line marked Temp.) at least once a day. If a protocol requires twice-daily monitoring, an extra line may be added, so that the log includes a line for a.m. and p.m. (see last table below for example). The person who checks and records the temperature should also record their initials in the bottom row (marked Initials). If temperatures are found to be outside of the acceptable range or the thermostat is not working properly, corrective action, date, and supervisor’s initials should be recorded for each incident. *Templates adapted from the Global Laboratory Initiative Example -80°C freezer log Temperature monitoring form Month/year: Equipment: Critical temperature range: Day Temperature Initials January 2017 -80°C Freezer [INSERT ROOM NUMBER AND FREEZER NAME OR NUMBER] ≥ -65°C (acceptable variation) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
80
APPENDIX I-16
80
Example -20°C freezer log Temperature monitoring form Month/year: Equipment: Critical temperature range: Day Temperature Initials January 2017 -20°C Freezer [INSERT ROOM NUMBER AND FREEZER NAME OR NUMBER] –18 °C ± 2 °C (acceptable variation) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
Example refrigerator log Temperature monitoring form Month/year: Equipment: Normal temperature range: Day Temperature Initials January 2017 Refrigerator [INSERT ROOM NUMBER AND REFRIGERATOR NAME OR NUMBER] +6 °C ± 2 °C (acceptable variation) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
SECTION I: APPENDICES
81 SECTION I: APPENDICES
81
Example incubator log Temperature monitoring form Month/year: Equipment: Normal temperature range: Day Temperature Initials January 2017 Incubator [INSERT ROOM NUMBER AND INCUBATOR NAME OR NUMBER] 36 °C ± 1 °C (acceptable variation) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
Example generic log Temperature monitoring form Month/year: Equipment: Normal temperature range: Day Temperature Initials January 2017 [INSERT EQUIPMENT TYPE, ROOM NUMBER, AND EQUIPMENT NAME OR NUMBER] [INSERT EQUIPMENT NORMAL TEMPERATURE RANGE AND ACCEPTABLE VARIATION] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
82
APPENDIX I-16
Example generic log Temperature monitoring form Month/year: Equipment: Normal temperature range: Day Temperature Initials January 2017 [INSERT EQUIPMENT TYPE, ROOM NUMBER, AND EQUIPMENT NAME OR NUMBER] [INSERT EQUIPMENT NORMAL TEMPERATURE RANGE AND ACCEPTABLE VARIATION] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Example generic log Temperature monitoring form Month/year: Equipment: Normal temperature range: Day a.m. p.m. Temperature Initials January 2017 [INSERT EQUIPMENT TYPE, ROOM NUMBER, AND EQUIPMENT NAME OR NUMBER] [INSERT EQUIPMENT NORMAL TEMPERATURE RANGE AND ACCEPTABLE VARIATION] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Supervisor’s initials: _________________________________________
Corrective action: ________________________________________________________________________________________________________________________________________________ Date (DD/MM/YY): ________________________________________ Supervisor’s initials: _________________________________________
SECTION I: APPENDICES
83 SECTION I: APPENDICES
83
Section
I
APPENDIX I-17
SAMPLE HIV RAPID TESTING RESULTS LOG This is an example of a test results form for HIV rapid testing using a two-test serial algorithm. This form should be adapted for the HIV testing algorithm being used or other biomarker testing.
Instructions: 1. Survey name (e.g. Country X BBS among PWID) and survey site (e.g. RDS site name or number) should be filled in. This information may be filled in ahead of time so that copies may be made and distributed to the survey sites. 2. The first three rows in the table show examples of how to complete the form. It may be helpful to color-code or highlight reactive results to indicate that a second test is required. 3. In example 1, a second test is required because test 1 was reactive. If test 2 is reactive, the final result is interpreted as HIV-positive and recorded as shown. Survey staff also should record whether the result was provided to the participant and where they were referred for services and any other relevant notes (e.g. specimen quality, participant left without receiving results, etc.) 4. In example 2, the second test was non-reactive. Depending on the national guidelines, a third (tie-breaker test) or a repeat of the first two tests may be required. Some guidelines specify the result as indeterminate and refer the participant for re-testing after a specified period of time. 5. In example 3, a non-reactive result for test 1 is interpreted as HIV-negative, and in this two-test algorithm a second test would not be done.
84
APPENDIX I-17
HIV Rapid Testing Results Log SURVEY NAME: _________________________ Survey ID Lab ID number (if different from Survey ID in first column) Testing staff ID number or initials Date of test (DD/ MM/YY)
SURVEY SITE: _________________________ Specimen type (serum, plasma, whole blood or fingerprick whole blood) Test 1 name Test kit lot no. Result Test 2 name Test kit lot no. Result Final result Result provided to participant Referral to Notes
Example 1
123
XYZ
15/08/17
Whole blood
Determine
999999
Reactive
UniGold
888888
Reactive
Positive
Y
Care facility X
N/A
Example 2
456
XYZ
15/08/17
Whole blood
Determine
999999
Reactive
UniGold
888888
NonReactive
Indeterminate
Y
Testing site X per national guidelines
N/A
Example 3
789
XYZ
15/08/17
Whole blood
Determine
999999
Non- reactive
N/A
N/A
N/A
Negative
Y
N/A
Specimen moderately hemolyzed
SECTION I: APPENDICES
85 SECTION I: APPENDICES
85
Section
I
APPENDIX I-18
EQUAL PROBABILITY SAMPLING (EPS) AND PROBABILITY PROPORTIONAL TO SIZE (PPS) CLUSTER SAMPLING GUIDE This appendix describes steps for sampling clusters using Equal Probability Sampling (EPS) and Probability Proportional to Size (PPS) sampling. It is meant to be used in conjunction with the cluster sampling section with CCS.
The number of clusters to be selected must first be determined before selecting clusters by PPS or EPS. The figure on the next page provides a decision tree for selecting between EPS and PPS.
Terms used in this Appendix EMoS Estimated Measure of Size: The estimated number of individuals who meet the approach criteria in a specific venue at a specific time. The word “Estimated” is used to designate the number of people observed in a venue during venue observation, i.e. before sampling has started.
AMoS
Actual Measure of Size: The actual numberof individuals who meet the approach criteria during a sampling event.
CMoS
Cumulative Measure of Size: The cumulative sum of the estimated number of individuals who meet the approach criteria in a specific venue at a specific time.
RS
Random Start: A randomly selected number between 1 and the Sampling Interval.
SI
Sampling Interval: The selection of every Nth person as a participant in a study. The SI allows for random selection throughout the population.
86
APPENDIX I-18
Decision tree for choosing between EPS and PPS EPS No
No
Sampling frame with venue, time, and EMoS
Information about EMoS available?
Yes
Large variation in EMoS
Yes
PPS
Once clusters have been selected using either EPS or PPS, save the selected clusters with their unique cluster IDs on a separate worksheet for planning fieldwork. Maintain the original sampling frame for reference and in case a second round of cluster selection is required. Document the sampling information in accordance with the steps above; this information is required to calculate sampling probabilities and non-response.
Selecting Clusters with Equal Probability Sampling (EPS) M = Number of clusters in the sampling frame EMoS= Estimated measure of size a = Number of clusters to select SI = Sampling interval RS = Random start 1. Finalize the sampling frame and prepare a sequential list of clusters with unique identification codes for venues and time periods. 2. Calculate the number of clusters to select: Number of clusters to select = a = Target sample size / Average cluster size EMoS For example: a = 450 / 10 = 45 3. Calculate the sampling interval (SI) by dividing the total number of clusters in the sampling frame (M) by the number of clusters to be selected (a). Sampling interval =M/a For example: SI = 150 / 45 = 3.3 (If the SI is not a whole number, create a column that rounds the SI to the nearest whole number.) 4. Select the random start (RS): this is a randomly selected number between 1 and the SI. (One method for selecting the random start is to write numbers from 1 to the SI on small pieces of paper, put them in a container, and select one piece of paper from the container. 5. Start at the first cluster in the sampling frame and count to the cluster in the RS position. For example, if the selected RS is 3.3, select the third cluster in the sampling frame. 6. Select the subsequent clusters by adding the SI to the RS. In other words: Cluster 1 = RS = 2 Cluster 2 = RS + SI = 5.3 = 5 Cluster 3 = RS + 2*SI = 8.6 = 9 Cluster 4 = RS + 3*SI = 11.9 = 12 (and so on)
SECTION I: APPENDICES
87
Example: Sampling frame for EPS
# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 (etc)
Location Venue 1 Venue 1 Venue 1 Venue 2 Venue 2 Venue 3 Venue 3 Venue 3 Venue 3 Venue 4 Venue 5 Venue 5 Venue 5 Venue 6 Venue 7 Venue 8 Venue 8 Venue 8
Time Time 1 Time 2 Time 3 Time 1 Time 2 Time 1 Time 2 Time 3 Time 4 Time 1 Time 1 Time 2 Time 3 Time 1 Time 1 Time 1 Time 2 Time 3 Example: Equal Probability Sampling of Time Location Clusters
Target Sample Size M a SI RS Total Number of Clusters Number of Clusters to Select Sampling Interval Random Start
450 150 45 3.3 2
Average Estimated Measure of Size = 10
Equation for Sampled Cluster Random Start (RS) RS + 1* SI RS + 2* SI RS + 3* SI RS + 4* SI RS + 5* SI RS + 6* SI RS + 7* SI Y RS + 8* SI RS + 9* SI RS + 10* SI
Result 2 5.3 8.6 11.9 15.2 18.5 21.8 25.1 28.4 31.7 35
Selected Cluster 2 5 9 12 15 19 22 25 28 32 35
Example: Sampling frame for EPS with selected clusters
Order # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Location Venue 1 Venue 1 Venue 1 Venue 2 Venue 2 Venue 3 Venue 3 Venue 3 Venue 3 Venue 4 Venue 5 Venue 5 Venue 5 Venue 6 Venue 7 Venue 8 Venue 8 Venue 8
Time Time 1 Time2 Time 3 Time 1 Time2 Time 1 Time2 Time 3 Time 4 Time 1 Time 1 Time2 Time 3 Time 1 Time 1 Time 1 Time2 Time 3
Selected? Y
Y
Y
Y
88
APPENDIX I-18
Selecting Clusters with Probability Proportional to Size (PPS) a = Number of clusters to select SI = Sampling interval M = Number of clusters in the sampling frame RS = Random start EMoS= Estimated measure of size CMoS = Cumulative measure of size To select the clusters using PPS, follow these steps: 1. Finalize the sampling frame and prepare a sequential list of clusters with unique identification codes for venues and time periods and EMoS. 2. Starting at the top of the list, calculate the cumulative MoS (CMoS), and enter these figures in a column next to the MoS for each cluster. 3. Calculate the number of clusters to select: Number of clusters to select = Required sample size / Average EMoS 4. Calculate the SI: SI= MoS / Number of clusters to select 5. Select an RS between 1 and the SI. 6. Select subsequent clusters by adding the sampling interval to the previously calculated interval. For example: Cluster 1 = RS Cluster 2 = RS + SI Cluster 3 = RS +2SI (etc.) 7. Repeat step 6 until the list has been exhausted.
SECTION I: APPENDICES
89
Example: Sampling frame for PPS
Order # Location 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 (etc.) Sum Venue 1 Venue 1 Venue 1 Venue 2 Venue 2 Venue 3 Venue 3 Venue 3 Venue 3 Venue 4 Venue 5 Venue 5 Venue 5 Venue 6 Venue 7 Venue 8 Venue 8 Venue 8
Time Time 1 Time2 Time 3 Time 1 Time2 Time 1 Time2 Time 3 Time 4 Time 5 Time 1 Time2 Time 3 Time 1 Time 1 Time 1 Time2 Time 3
EMoS CMoS 8 15 10 5 15 4 18 26 15 10 15 5 30 12 18 14 11 40 8 23 33 38 53 57 75 101 116 126 141 146 176 188 206 220 231 271 2 250 M AMoS a SI RS CMoS
Example: Probability proportional to size of time-location clusters
Target Samples Size Total Number of Clusters Average Measure of Size Number of Clusters to Select Sampling Interval Random Start Cumulative Measure of Size
450 150 15 30 22.5 11 2250
Equation for Sampled Cluster Random Start RS + 1* SI RS + 2* SI RS + 3* SI RS + 4* SI RS + 5* SI RS + 6* SI RS + 7* SI RS + 8* SI RS + 9* SI RS + 10* SI
Result
Selected Cluster (rounded) 11 34 56 79 101 124 146 169 191 214 236
11 33.5 56 78.5 101 123.5 146 168.5 191 213.5 236
Example: Sampling frame for PPS
Order # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 (etc.)
Location Venue 1 Venue 1 Venue 1 Venue 2 Venue 2 Venue 3 Venue 3 Venue 3 Venue 3 Venue 4 Venue 5 Venue 5 Venue 5 Venue 6 Venue 7 Venue 8 Venue 8 Venue 8 Sum
Time Time 1 Time2 Time 3 Time 1 Time2 Time 1 Time2 Time 3 Time 4 Time 5 Time 1 Time2 Time 3 Time 1 Time 1 Time 1 Time2 Time 3
EMoS 8 15 10 5 15 4 18 26 15 10 15 5 30 12 18 14 11 40
CMoS 8 23 33 38 53 57 75 101 116 126 141 146 176 188 206 220 231 271 2250
CMoS Range 1 9 24 34 39 54 58 76 102 117 127 142 147 177 189 207 221 232 8 23 33 38 53 57 75 101 116 126 141 146 176 188 206 220 231 271
Selected? Y Y Y Y* Y Y Y* Y Y Y*
* Clusters with an Estimated Measure of Size that is larger than the sampling interval are selected with certainty.
90
APPENDIX I-18
Section
APPENDIX I-19
I
TIME-LOCATION SAMPLING INTERCEPT FORM INTERCEPT FORM The Intercept Form is an example of the information that recruiters should collect during an intercept (or recruitment attempt) in a venue. It is designed to allow field staff to quickly record the result of any one intercept/recruitment attempt using tick marks. Adapt this form to include all of the information needed to calculate response rates (i.e. the percentage of people intercepted who agreed to be screened for eligibility) for the specific data needs of the survey. It should be printed in large enough font. Each recruiter must have his/her own Intercept Form. Local Adaptation Adapt this form to fit local procedures and protocol. The information collected here provides non-response and eligibility information needed to validate survey results. The eligibility information is used in data analysis. Be sure to update the instructions once the form is adapted. See the “SURVEY PREP TLS” section for more information and an example of how this form would be used.
Instructions for the Intercept Form Step 1: Fill in the information about the recruitment event in the form’s header. Step 2.1: Recruiter approaches venue attendee as directed by counter and attempts to engage him/her in conversation. • If the person ignores the recruiter, mark a tick next to “No” on line A.
Step 2.2: If the person is willing to talk, briefly describe the survey to the prospective participant and determine whether he/she previously participated in the survey. • If the person has previously participated, mark a tick next to “Yes” on line B.
Step 3: Invite the person to participate in eligibility screening for the survey. • • If the person is not willing to be screened for eligibility, mark a tick next to “No” on line C. If the person is willing to be screened for eligibility, mark a tick next to “Yes” on line D. The number of marks on line D should equal the number of eligibility screening records in the data.
Step 4: Add tick marks to calculate number of people approached. Write total on line E.
SECTION I: APPENDICES
91
INTERCEPT FORM Event Data Date: Event #: Venue ID: Recruiter ID:
Line
Intercept data Accepted Intercept?
Sum
A
No: Previous Participant?
B
Yes: Recruited to be Screened for Eligibility?
C D E
No: Yes: Total:
Instructions for the Summary Intercept Information Form Once the recruitment event is over, collect all the intercept forms from the recruiters. To calculate recruitment rates for each recruiter and for the event as a whole: Step 1: Add the number of ticks on each line of each intercept form. Step 2: Enter the event information on the Summary Intercept Information Form for each event. Step 3: Copy the sum of ticks on each line from the Intercept Form to the Summary Intercept Information form for each recruiter. Step 4: Calculate the sum of lines A, B, C, and D to determine the total number of approaches for each recruiter. Step 5: Calculate the event total for lines A, B, C, and D by summing all recruiters’ totals for each line. Step 6: Calculate the recruitment percentages for each recruiter and for the event.
92
APPENDIX I-19
SUMMARY INTERCEPT INFORMATION FORM Calculate recruitment rates for each recruiter and for the event by using the table below (or a similar one adapted for the survey). Review response rates with recruiters and provide tips for improvement. Summary Intercept Information Form for each event Date Event # Venue ID Recruiter ID: # A B C D E # Did not accept intercept # Previous participant # Refused to be screened # Screened Total # of approaches (A+B+C+D) % # % # %
Total for event # %
% values
(A/E)% (B/E)% [C/(E–B)]% [D/(E–B)]%
SECTION I: APPENDICES
93
Section
I
APPENDIX I-20
VENUE OBSERVATION FORMS Multiple forms (Forms A–I) are included in this appendix to help organize the venue information necessary for inclusion in the final sampling frame. In addition to venue details needed for the sampling frame, there are a number of forms that may be further tailored to describe logistical issues and population characteristics that the survey team may find useful to plan and set up sampling events. Please note that the information collected by each form is needed for each individual venue being considered for inclusion in the sampling frame. Form A. Venue Sampling Frame Details THIS FORM SHOULD CONTAIN ALL DETAILS NEEDED FOR EACH VENUE TO CONSTRUCT THE SAMPLING FRAME. IT IS A SUMMARY FORM THAT COMPILES INFORMATION COLLECTED ON EACH INDIVIDUAL VENUE IN THE FORMS LISTED BELOW.
Venue name Describe details needed for stratification of sampling frame:
Venue type Venue code
Contact Person Final time-location combinations to enter into Sampling Frame
Contact Phone #
Complete the final day/time combinations below after observing the venue (and filling out the rest of the forms) and reaching consensus among informants. Add more rows if needed. Day of the week 1 2 3 4 5 6 Time period (24-hour clock) Estimated Measure of Size (EMoS) Special event?
From: From: From: From: From: From:
: : : : : :
To To To To To To
: : : : : :
94
APPENDIX I-20
Form B. Venue Geographical and Layout Details Use this form to record venue location information for GIS mapping and to describe the geography and physical layout of the venue. Venue Location GPS coordinates: Name / Code Name/code of province Name/code of district Name/code of town Name of local area Address or description of area Indoor or outdoor venue? Indoor Outdoor Both / / / /
Draw the layout of the venue below. Please include the following items: • Landmarks, street names, intersections, and other identifying information. • Boundaries of the site (indicating the general area encompassing the venue/site). • Use arrows to indicate the direction of entry to the venue/site. • Any other information that will help the field team when they arrive at the sampling event. For example, if the site is a room in a building, be specific about which floor and side of the hall the room is on and which room it is by showing all rooms on that floor in the drawing and highlighting the specific one(s) to be used for sampling. • If the site is an area along a street, show the side(s) of the street where the target population members are found.
Form C. Venue Observations for Consideration in Sampling Frame Use this form to record all the days and times the venue was observed to determine eligibility for inclusion in the sampling frame. Add more lines as needed to the bottom of the form. Venue Observations Date 1. 2. 3. 4. 5. 6. 7. Day of week Time period observed Staff initials Should more observations take place at this day/time? Y Y Y Y Y Y Y N N N N N N N
SECTION I: APPENDICES
95
Form D. Venue Challenges Use this form to keep track of any barriers that may be encountered by survey staff during recruitment at the venue. Also indicate which items are sufficient reasons for removing the venue from the sampling frame. Venue Challenges Include notes below for each item encountered Remove from frame? If yes, check
Safety concerns?
Potential interview/HIV testing location outside venue needed?
Barriers to recruiting?
Barriers to interviewing?
Barriers to HIV testing?
Parking for staff or project vehicle?
Form E. Venue Type Determination Use this form to document how key informants characterize the venue for final venue-type determination The information in this section should be completed once key informants provide enough information to allow the team to characterize what type of venue this represents. For example, after talking to key informants at the site, Venue A is characterized as a brothel. Adapt the list as needed. Venue Type Final venue type characterization: Tally votes of informants Brothel Home Massage parlor Hair salon Bar Restaurant Disco City block Street Public Park Bathhouse Railway platform Beach Other, specify:
96
APPENDIX I-20
Form F. Estimated Measure of Size of Target Population for the Venue Use this form to summarize various EMoS estimates of interest to plan productive sampling events. Number Of Target Population Members Encountered (Estimated Measure Of Size) EMoS Maximum number of target pop members present at the site Minimum number of target pop members present at the site Operational days (days of the week when the target population members are likely to be at the site) Peak days and times (days of the week when the maximum number of target pop members are likely to be found) and estimated measure of size on those days Lean days and times (days of the week when the minimum number of target pop members are likely to be found) and estimated measure of size on those days N/A Notes
Form G. Listing of Potential Time Periods to Include in Sampling Frame Use the form below to record the range of days and time periods that may be included in the final sampling frame. In addition to days and time periods, it is important to document the Estimated Measure of Size (EMoS) of the target population encountered during the observation (adjusted to your full sampling event duration). The lower part of the form is to document any special sampling events that may be included in the frame. Time Information Which days of the week and time periods do you find <target population> at this site? (Select all that apply) Please identify time periods that are the same length as the length of the sampling events. Time 1 (24-hour clock) Mon Tues Wed Thur Fri Sat Sun From: From: From: From: From: From: From: : : : : : : : To To To To To To To : : : : : : : EMoS Incl. in sampling frame? From: From: From: From: From: From: From: Time 2 (24-hour clock) : : : : : : : To To To To To To To : : : : : : : EMoS Incl. in sampling frame?
Are there any special or one-time events at this venue that should be on the sampling frame? Date Time period EMoS Description of event
SECTION I: APPENDICES
97
Form H. Characteristics of Target Population at the Venue Document any characteristics of the target population that may be encountered by the survey team at the venue. This will help the team consider special needs for language capacity, matching of team members to population characteristics (e.g. age group, language group, etc.). Population Characteristics Notes Potential language barriers
Observed behaviors
Observations about demographic characteristics
Seasonal variations
Mobility
Sub-groups of interest at venue
Form I. Venue Details for Data and Specimen Collection Please provide specific details to describe how the team should set up data collection and specimen collection areas associated with the sampling venue. Location For Data And Biological Specimen Collection Notes Name and address of the space (if different)
Nearest landmark
Contact person
Number of rooms available
Is a toilet available for collecting urine samples?
Is running water available?
98
APPENDIX I-20
Section
APPENDIX I-21 UNIQUE PARTICIPANT CODE (UPC)
I
A unique participant code (UPC) in an RDS survey is used on the second visit to verify that the person presenting for the second visit is the same person who participated in the first visit. The UPC is created using a series of questions that can only be answered by the participant. This form should be filled out by a member of the survey staff. After the UPC is created, record the UPC in the coupon tracking system. To recreate the code and confirm participant identity at the second visit, simply ask the participant the questions below. Instructions: Ask each question to each participant and fill in the boxes with one or two letters or digits, as appropriate.
1
What are the first two letters of your family name? L1 What are the first two letters of the village, town, or city you were born? If unknown, enter “ZZ”. L3 L4 L2
2
3
What are the first two letters of your mother’s first name? If unknown, enter “ZZ”. L5 L6
What month were you born?
4
Use two digits for month of birth. For example, if the participant was born in March, use “03”. If unknown, enter “99”.
D1
D2
5
What are the first two letters of the village, town, or city of your mother’s birth? If unknown, enter “ZZ.” L7 L8
Instructions: Transfer the letters and numbers from the boxes directly to the boxes below. Match the appropriate codes from each question (L1 to L1, L2 to L2, etc.) to create the unique identifier code. UPC: L1 L2 L3 L4 L5 L6 D1 D2 L7 L8
SECTION I: APPENDICES
99
Section
I
APPENDIX I-22
SAMPLE SIZE REQUIREMENTS TO DETECT DIFFERENCES This table displays various scenarios showing the necessary sample sizes (for each survey round) to detect a given change in a desired proportion (e.g. HIV testing in the last 12 months) between two survey rounds. The required sample sizes shown consider a 95% confidence level, 80% power, and design effects ranging from 2.0 to 4.0. Survey sample sizes need to be further inflated for assumed missing data (not shown here). A sample size calculator will be available as an Excel file for download from the WHO website.
Round 1 (P1) 0.10 0.10 0.20 0.20 0.30 0.30 0.40 0.40 0.50 0.50 0.60 0.60 0.70 0.70 0.80 0.80
Round 2 (P2) 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95
P1 and P2 0.10 0.15 0.10 0.15 0.10 0.15 0.10 0.15 0.10 0.15 0.10 0.15 0.10 0.15 0.10 0.15
2.0 398 200 587 276 712 325 775 346 775 339 712 304 587 241 398 151
3.0 597 300 881 414 1068 488 1163 519 1163 509 1068 456 881 362 597 227
4.0 796 400 1174 552 1424 650 1550 692 1550 678 1424 608 1174 482 796 302
Adapted from WHO. Regional Office for the Eastern Mediterranean. Introduction to HIV/AIDS and sexually transmitted infection surveillance: Module 4: Introduction to respondent driven sampling. p. WHO-EM/STD/134/E
100
APPENDIX I-22
Section
APPENDIX I-23 CALCULATOR FOR THE NUMBER OF UNIQUE OBJECTS FOR ESTIMATING POPULATION SIZE
I
This appendix can be used to determine the number of unique objects needed for distribution to estimate the size of a population with a certain level of precision. This unique object calculator will be available as an Excel file for download from the WHO website.
Step 1: Estimate the number of unique objects to distribute Estimated population size Desired precision (½ width of CI as % of estimate) 33.0% 33.0% 25.0% 10.0% 25% 10.00%
Survey sample size
Survey design effect
Precision on N (½ width of CI)
Number of objects to distribute
Expected number of objects recaptured 64 69 138 544 110 359
Verification: population size estimate 13000 2500 2500 2500 2500 2500
13000 2500 2500 2500 2500 2500
1000 1000 1000 1000 500 500
2.00 2.50 3.00 3.50 2.5 3.5
±4290 ±825 ±625 ±250 625 250
826 172 344 1359 551 1793
Step 2: Determine the precision with which population size can be estimated given the number of objects Number of objects distributed 826 172 344 1359
Survey sample size 1000 1000 1000 1000
Number of objects found in survey 27 56 91 290
Survey design effect 2.00 2.50 3.00 3.50
Population size estimate 30588 3079 3785 4686
Std error
Lower bound
Upper bound
Relative bounds
8144.5 589.6 625.8 420.3
14625 1923 2559 3862
46552 4235 5012 5510
±52.2% ±37.5% ±32.4% ±17.6%
SECTION I: APPENDICES
101
Section
I
APPENDIX I-24
QUALITATIVE INTERVIEW GUIDE Qualitative interviews should be conducted as guided conversations. This appendix provides a guide for semi-structured interviews of target population members. The order and wording of questions should be changed based on the local context. Questions may also be added or deleted to meet local needs and circumstances. The purpose of a domain and guidance as to when it should be included are presented in italics. They should not be read to participants. Domain selection should consider the local context as well as participant responses in the quantitative interview. It is recommended that only a few key domains be covered in each interview. For each interview domain we list possible probes. These domains and probes are not meant to be exhaustive, and it may not be necessary to probe them in their entirety. The first probe in each domain helps introduce the domain and should be asked when the domain is used. Before starting the interview, the interviewer should describe the survey goals and process to the participant. Then explain why the participant has been selected for the interview. Interviewers should emphasize that participants are the experts and interviewers are there to learn from participants. This questionnaire guide covers the following domains: • • • • • • • • • Introduction and starting the conversation Upbringing and self-perception Alcohol and drug use Violence Relationships and sexual attraction and behaviors. Transgender women HIV and <target population> HIV protective behaviors and access to HIV and STI services Ending the interview
102
APPENDIX I-24
Date
|____||____| / |____|| __| / |____||____||____||____| (DD/MM/YYYY)
Interviewer name
Note Taker name
Venue
Start time
|____||____| : |____||____| (HOUR/MIN)
End time
|____||____| : |____||____| (HOUR/MIN)
How was this participant referred to be interviewed?
Name of electronic (audio) file
INTRODUCTION (To be read by interviewer to participant) I want to thank you again for participating in this survey. Today we will talk about some of the same topics you were asked about during the main interview. What you tell us will help us understand and put in context the responses of all participants. Before I start the interview, please turn off your cell phone and other mobile devices. I will ask you questions about yourself. Please do not use your name or anyone else’s. Please feel comfortable to share your opinions and experiences. Also, when I say “friends”, “colleagues”, “peers” or “people like you”, I mean people you know who are <target population>. Do you have any questions before we start? (Interviewer: Take time to address all questions and concerns)
1. Domain: Introduction and starting the conversation Purpose: To make participants comfortable and build rapport by having the participants tell the interviewer some basic information about themselves. The purpose is not to learn everything at once but to start a conversation, a story. Guidance: Recommended to ask of all participants. 1.1 1.2 1.3 1.4 1.5 Let’s start by learning a little about you. Can you tell me something about yourself? Where were you born? How long have you lived in this town? Where have you spent most of your adult life? What things do you and your friends do here to relax? How do you earn money? Tell me about someone special in your life.
SECTION I: APPENDICES
103
2. Domain: Upbringing and self-perception Purpose: To elicit what it was like growing up in their home and community and how they were viewed as a child. To understand some of the significant moments in their childhood that may impact their current lives. Guidance: Useful to include when seeking to learn more about self-perception and the roots of shame, selfstigmatization, or risky behaviors. Possible probes: 2.1 2.2 2.3 2.4 2.5 Can you tell me about growing up? How would you describe your childhood? Tell me about the things that you enjoyed? What didn’t you like as a child? What things were important to you as a child? Who were the most important people in your life and why? What challenges did you face growing up? What was your family life like growing up? What was it like growing up in your community? Can you tell me about the best moment as a child? What about the worst moment of your childhood?
3. Domain: Alcohol and drug use Purpose: To focus on personal experiences with alcohol and drugs, with a particular focus on injecting drug use and access to clean syringes/needles. Guidance: This domain should only be asked of people who indicated in the quantitative interview that they use alcohol and drugs. Most useful in settings where high alcohol consumption is common. Should be asked of PWID. Probes: 3.1 3.2 Can you tell me about your use of alcohol and drugs? How would you describe your drinking/drug use? How old were you when you started using drugs and drinking alcohol? Who introduced you to alcohol and/or drugs? o Why did you start drinking alcohol/using drugs? How do you feel when drinking/using drugs? o How has your use of alcohol/drugs changed over the years? o Are there certain times/moments of your life that you drink/use more often? Have you ever tried to stop drinking/using? What was it like trying to stop? How accessible is drug treatment? o Have you ever gone on treatment? Can you tell me about your experience with it? o What might encourage you to seek treatment? Can you tell me about how people access drugs here? o Where can you go to get clean/sterile syringes/needles? o In what situations do you share syringes/needles? o What types of issues do people face trying to access clean/sterile syringes/needles?
3.3 3.4
3.5
4. Domain: Violence Purpose: To focus on experiences of violence, particularly intimate partner violence and other types of violence that are common in the community and the person’s individual experience of violence, particularly if related to gender, sexuality or involvement in the selling and exchanging of sex.
104
APPENDIX I-24
Guidance: This domain is recommended in settings where violence is common. It may be asked of people who have and have not experienced violence. Probes: 4.1 4.2 Some people experience different types of violence – physical, emotional, sexual – can you tell me about violence in your community? How do people view violence in your community? o How frequent are these acts of violence? o Who usually carries out this kind of violence? o Are some forms of violence more common than others? Which people experience more violence than others and why? What services are there for people who have been raped/sexually assaulted/ physically abused? How are people who have experienced these types of violence treated in the community? How are people who are violent treated in the community? Are there some types of violence that are not really seen by people as violence? What is it like for people to report violence to the authorities and what happens if they do? What other ways is violence resolved in your community? What do people in your community do to reduce the risk of violence? What has been your personal experience with violence? (perpetrator or victim or both) o If you have been violent to another person, can you tell me about what provoked you to be violent and what you did? o If you have been the victim, can you tell me what happened and about the type of support you received?
4.3 4.4 4.5 4.6 4.7 4.8
5. Domain: Relationships and sexual attraction and behaviors Purpose: To elicit information on relationships, including but not exclusive to sexual relationships. Guidance: This domain includes questions that target all populations as well as others that target only SW, MSM, or TG. Probes: 5.1 5.2 5.3 I would like to talk to you about the different types of relationships. Can you tell me about the most important person in your life and why? Can you tell me about your first memory of being sexually attracted to someone? Who was it? How old were you? Have you always been attracted to the same type of people? o How has it changed over the years? o Who are you most attracted to now (men, women, both) and why? How do you describe your sexual attraction? Can you tell me about the last time you had sex? o Who was it with (main partner/casual partner/client/sex worker)? o Where did you have sex? Did you use a condom? o How did you talk about using a condom? What do you think about the number of people you have had sex with in the past 6 months?
5.4 5.5
5.6
SECTION I: APPENDICES
105
5.7 5.8
How come you have different partners? Are there ever times when you receive money or goods in exchange for sex? Can you tell me more about that? o How often, to whom, where, types of sex, etc.? o How do you find people who will give you something for sex? What challenges do you face exchanging sex? o How do the authorities treat people who exchange sex? Are there ever times when you give money or goods in exchange for sex? Can you tell me more about it? Include how often, to whom, where, types of sex, etc.
5.9
6. Domain: TG Purpose: To elicit information about the lives of transgendered persons. Guidance: Should be asked of all TG due to the limited data on this population. Probes: 6.1 6.2 6.3 6.4 6.5 What do you say when people ask your gender? How do they respond? How did you decide whether to live as a man or a woman? What is it like being a TG in <city>? What are some challenges you face as a TG? How do you meet other TG? How do you choose your sex partners? Have you had a gender enhancement/transition procedure or used hormones? What kind of person helps you with this?
7. Domain: HIV and <target population> Purpose: To elicit information about the social context of HIV, how people talk about HIV, and acceptance of people living with HIV. Guidance: Useful when policy makers and service providers have little experience with the target population in the survey city. Probes: 7.1 7.2 7.3 7.4 7.5 Can you tell me about HIV in your community and among <target population>? What do people say about HIV in your community? Prompt for government, churches/mosques, leaders, friends, family, etc. How common is it for you to hear, see, or read something about HIV in your community? What types of messages are usually shared? How relevant are the messages to you and other <target population>? How do people think HIV is transmitted? What do you think? How are people with HIV treated in your community? Are all people with HIV treated the same? How long ago did you first hear about HIV?
106
APPENDIX I-24
7.5 7.6
Do you know people who have HIV? How open are people in your community about having HIV?
8. Domain: HIV protective behaviors and access to HIV and STI services Purpose: To focus on the role of HIV in individuals’ lives and learn about the availability, accessibility, and quality of HIV and STI services. Guidance: Recommended for use in all surveys. Probes: 8.1 8.2 8.3 8.4 8.5 What role does HIV play in your life? How much do you think about HIV / STIs when you have sex with others? Why/Why not? Can you tell me about how you take care of yourself in terms of HIV and other STIs? o How do your protect yourself from becoming infected with HIV/STIs? Probe: only have sex with certain kinds of people, always use a condom, always insertive partner, take PreP, always use clean syringes or needles How often do you use condoms during sex? o How do you decide when to use a condom during sex? o Tell me about the last time you did not use a condom. What was the reason for that? Have you ever heard of preexposure prophylaxis or PrEP to prevent HIV? o What are your thoughts about using PrEP? o What has been your experience using PrEP? How do people respond when they learn you use PrEP? o What do you think of your peers who use PrEP? And those who don’t? How do you decide when to use a clean needle or syringe?
8.6 What things make it easier or harder for you to protect yourself or others? o If HIV-positive, can you tell me if and how you tell your sexual partner(s) about your HIV status? What things make it harder/easier for you to tell someone you have HIV? 8.7 8.8 8.9 8.10 Can you tell me about HIV and STI services in your community? o What types of HIV services are available in your community? o What other sexual and reproductive health services are there? Who uses them? o What do people do when they get an STI? Which of the HIV and STI services in your community have you accessed/not accessed and why? o What was your experience like accessing these services? How do you feel telling the health care worker that you <risk behavior>? How do health care workers treat you when you tell them that you <risk behavior>? o How often do you use these services? o What do you like about them? What do you not like about them? How could these services be improved to meet your needs and that of your friends? What things are important for you when deciding where to get services from? What challenges do people face accessing condoms? What about water-based lubricants? What about clean needles and syringes?
8.11 How easy is it for someone to have a free and confidential HIV test in your community? Have you ever been tested for HIV before being in this survey? o If no, how come? What might encourage you to get tested in the future? o If yes, can you tell me more about that? (counseling, consent, and results) How did you decide to get tested?
SECTION I: APPENDICES
107
8.12 8.13 8.14 8.15
How often do you get tested? How come? If HIV-positive, how has HIV impacted your life? o What happened when you learned your test result? What did you do? o What types of support have been made available to you? Who have you told about your status? How did/why did you make this/these decisions? o How important is HIV in your daily life in terms of how you feel emotionally, your relationships with family and friends, peers and your health? o How do people treat you when they learn you are living with HIV? What types of issues have you faced living with HIV? If HIV-positive, how long after learning your HIV status did you seek care or treatment? o How come you went so quickly? o How come you delayed in going? o Are you on treatment now? How come you are not on treatment? How long have you been on treatment? If on treatment, what is your experience of accessing HIV treatment services? o How has it affected your life? o What challenges do you face being on ART? o In what ways, if any, has living with HIV affected your sexual behavior?
9. Ending the interview Purpose: To ask the person to reflect on the future rather than only the past. It can help participants leave the interview feeling more in control of their future. Guidance: This domain is especially useful for emotional interviews. Possible prompts: 9.1 9.2 9.3 Can you tell me about what you see as your future? Can you describe what you would like to see happen in your future? What changes would you like to see for your population? What things need to change in your community and in your life to ensure you are happy and healthy in the future?
108
APPENDIX I-24
Section
APPENDIX I-25
I
QUALITATIVE INTERVIEW FOR TARGET POPULATION (FSW) KEY INFORMANT CONSENT FORM This appendix provides a sample consent form for a qualitative interview among FSW. OVERVIEW: Hello, my name is <interviewer name>. I am working with <institution name>. Thanks for taking the time to speak with us. We invite you to join a detailed interview. This interview will help us understand how we can do our work better. You are free to join. You can stop taking part at any time. PROCEDURES: We would like you to take part in a survey. If you agree, we will do an interview with you. You and approximately <number of participants> others are being asked to participate in this survey, because we think you have ideas and opinions that can help us plan a larger survey. Please feel comfortable to be honest. One of our staff will write down your answers and record the interview with a recording device. CONFIDENTIALITY: What you tell us will be kept confidential. We will not ask for your name, address, or any other personal information that can be used to identify you. We will code your interview with a number to assist in tracking our notes. You can ask the interviewer to turn off the recorder temporarily if you are not comfortable having what you say recorded. However, if during the interview you say you are planning to harm yourself or others, or that you are a victim of human trafficking, we are required to report that information to the <name of local authority, for example, Ministry of Health>. RISKS AND BENEFITS OF THE INTERVIEW: We will ask about activities that are sensitive or stigmatized. Some interview questions are of a personal nature. You do not have to answer a question if you do not want to. There is no direct benefit to you for taking part in this activity. You may find the interview interesting and learn more about yourself. What we learn from the surveys may help us to make plans for how to improve health services for sex workers in <city>. This may help sex workers protect themselves and their partners from HIV and infections spread by sexual contact. COMPENSATION: You will receive <compensation amount> for taking part in this survey to compensate you for your time. If you travel to the interview you will receive <compensation amount> for your transportation. You will also get condoms, lubricants, and HIV prevention information. YOUR RIGHTS: If you have questions about the study or believe that you have been harmed by being in the study, you may call <principal investigator> at Tel. xxx-xxxxxxx. If you have questions about your rights as a research subject, you may contact <name> of the Research Ethics Committee, Tel. xxx-xxxxxxx. I confirm that this information has been reviewed with the survey participant, who has had a chance to ask questions and agrees to take part in this survey.
_______________________________ _______________________ Signature of Data Collector Date
SECTION I: APPENDICES
109
Section
I
APPENDIX I-26
FORMATIVE ASSESSMENT CONSENT FORM FOR NON-TARGET POPULATION KEY INFORMANTS This appendix provides a sample consent form for a qualitative interview among non-target population key informants.
DESCRIPTION OF THE INTERVIEW: Hello, my name is <interviewer name>. I am working with <institution name>. We are conducting a survey that will interview <target population> in <city>. The interviews may help us plan a future survey of HIV risk and health behaviors. PROCEDURES: We would like you to take part in a survey. You and approximately <number of participants> others are being asked to participate in this survey, because we think you have ideas and opinions that can help us plan a larger survey. If you agree to be interviewed, we will ask you about: • • • • • • Characteristics of <target population> in <city>; The best way to find <target population> who may want to take part in the survey; Ways that we can increase the number of <target population> willing to take part in the future survey; What kind of services are available <target population> to prevent or treat HIV; Where we should have the future survey; What you think about the future survey.
We will take notes on paper. We will use an audio recorder if you agree. We will keep these notes in a locked place. No one outside the study will have access to the notes. RISKS AND BENEFITS OF THE INTERVIEW: We will ask about activities that are sensitive or stigmatized. Some of the questions might make you uncomfortable. You do not have to answer a question if you do not want to. There is no direct benefit to you for taking part in this activity. What you share may help us to plan a better survey for <target population>. What we learn from the surveys may help us to make plans for how to improve health services for <target population> in <city>. This may help <target population> protect themselves and their partners from HIV and other infections. COMPENSATION: You will receive <compensation amount> for taking part in this survey to compensate you for your time. If you travel to the interview you will receive <amount of compensation> for your transportation. You will also get condoms, lubricants, and HIV prevention information. CONFIDENTIALITY: What you tell us will be kept confidential. We will not collect any information that can be used to identify you. We will code your interview with a number to assist in tracking our notes. You can ask the interviewer to turn off the recorder temporarily if you are not comfortable having what you say recorded. However, if during the
110
APPENDIX I-26
interview you indicate that an individual is plan to harm yourself or others, or that someone is a victim of human trafficking, we are required to report that information to the <Name of local authority, for example, Ministry of Health>. YOUR RIGHTS: It is your choice if you want to be in the survey or not. You are free to choose not to be in the survey. There is no penalty for this. You may choose not to answer any question. You are free to stop at any time. You do not have to give us a reason for stopping. You do not need to talk about anything you do not want to. CONTACTS: We will give you a copy of this consent form. If you have questions about your rights for taking part in this study, you can contact <principal investigator and institution> at Tel. xxx-xxxx. I confirm that this information has been reviewed with the survey participant who has had a chance to ask questions and agrees to take part in this survey.
_______________________________ Signature of Data Collector
_______________________ Date
SECTION I: APPENDICES
111
Section
I
APPENDIX I-27
DATA MANAGEMENT SOP FOR RDS SURVEY USING QDS <SURVEY NAME> SOP #: DATA MANAGEMENT FOR RDS SURVEY USING QDS: These SOPs are written for: SOPs referred to: SOP date: Data managers X MM/DD/YYYY
MAIN PROCEDURES IN INTERVIEW ROOM • • • Maintain the laptops and power supply. Ensure that power outages cannot interrupt the interview. If the laptops run out of power, organize computerassisted personal interviews (CAPI) with appropriate staff. If Coupon Managers (CM) are available, they could also help to conduct CAPI. Perform daily backups and data archiving.
MAIN PROCEDURES OUTSIDE INTERVIEW ROOM • • • Maintain, and repair as necessary, the IT infrastructure – servers, laptops not used for audio-computer-assisted self-interviews (ACASI), inverters, uninterruptible power supply (UPS), printer, barcode readers, and fingerprint scanners. Alert survey coordinator to IT supplies that need to be procured. Complete the relevant checklist parts.
FIRST-VISIT PROCEDURES • • • • Have a coupon manager accompany participants to the interview room. If all ACASI laptops are busy, have the participant wait near your desk or in the waiting area. If some ACASI laptops are vacant, see if anyone in the waiting area is ready for the interview. Welcome the participant and validate his/her survey ID. Lead the participant to a vacant laptop. Log in, choose the questionnaire, choose the appropriate language, and demonstrate how to do the interview by using the tutorial questions. If you are satisfied that the participant will be able to handle the ACASI, have him/her start the interview. • In the provided sheet, log in the participant’s ID and start time to see how long the interview takes. If the participant is taking a long time but is not finished, see if he/she is having any trouble. • If the participant feels uncomfortable with the interview questions: o Say that the interview questions are the same for everybody and do not assume everyone has the same (sex/ drug use) behaviors. o Reassure the participant that this interview is completely confidential and anonymous and ask him/her to answer all questions truthfully.
112
APPENDIX I-27
• •
If the participant quits the interview, ask what the problem is. If the participant is unable to use the computer, organize a CAPI instead. A data manager should administer the CAPI within the same visit. • If the participant refuses to continue, close the interview and record the reason in the checklist. • When the interview is complete: o Ask the participant to wait with you for a few minutes while you fill out the forms. o Verify that the interview was completed. o Escort the participant to an interviewer. If all interviewers are busy, ask the participant to wait in the waiting room.
SECOND-VISIT PROCEDURES • • • A participant’s second visit includes a short peer recruitment interview. It is done in the same mode as the first visit (ACASI, CAPI, HAPI). A reception staff brings a participant to the interview room for a peer recruitment interview or non-response interview. Register the participant on the second-visit log. Check the room regularly for anyone waiting to do the peer recruitment interview non- response interview.
DATA BACKUP AND SAFETY • Protect the laptops from viruses: o Do not use private flash drives on the survey laptops. o Do not install software that is not used for the survey. o Do not turn off the installed antivirus software. o Do not share the administrator password with others. Perform data backup at the end of every day: o Back up all participant data collected, including QDS and Epi Info. Transfer data into the data warehouse manager at the end of every day on the respective drives. Merge data into one main file for the day’s data to be stored on the network drive (to be named with date suffix for the day’s data). Back up these data on a CD-RW, DVD-RW, or flash drives. Erase the backed up data from the previous day and use the same CD/DVD, or flash drive again. Place the CD/DVD in the locked CD box and note the date of data backup on a piece of paper placed inside the CD box. Store the CD/DVD at the agreed-upon place.
• • • •
DATA SECURITY As most BBS data have no personal identifiers and participants are usually linked by their unique participant ID number: • Password protect all data and allow access only by authorized users. • Keep in a locked and secure file cabinet all entered and un-entered questionnaires, logs of data activities, data printouts, data encryption, etc.
DATA LINKING • Link all data using the study ID number to make up the final data analysis dataset. This includes data collected by the ACASI, CAPI, and laboratory results. This should be done upon completion of data collection and before data cleaning.
DATA QUALITY • Check data systematically for completeness and consistency every week if not more often. This gives interviewers an opportunity to correct or fill in missing information when the participant comes back for the second visit.
SECTION I: APPENDICES
113
Section
I
APPENDIX I-28
SAMPLE DATA DICTIONARY The following is an example of a data dictionary using MODULE 7: DEMOGRAPHICS. A data dictionary should be created for each dataset. At a minimum, a data dictionary should include: • • Variable name Description of the variable o Usually the interview question from the survey • Acceptable values o Include all valid responses for the variable o Field is not applicable for text responses • Variable type o Text/string ▪ Can be open (i.e. name of participant) ▪ Can be choice (i.e. categorical responses such as marital status) o Numeric o Date/time o Boolean- only two possible responses (i.e. YES/NO, TRUE/FALSE) • Checks or restrictions o Includes ranges, including minimum and maximum values o Logic checks o Skip patterns
Variable Name INT_DATE INT_CODE SUBJECT_ID
Description Date of interview Interviewer code Participant unique ID Do you consider yourself as male, female, transgender, or other?
Acceptable Values N/A N/A N/A Male = 1 Female = 2 Transgender = 3 Other = 4 Male = 1 Female = 2 Other = 3 N/A
Variable Type Date Text (Open) Text (Open)
Checks or Restrictions
DESEXNOW
DESEXBTH
What was your sex at birth? In what year were you born?
Text (Choice)
DEAGEY
Numeric
114
APPENDIX I-28
Variable Name
Description How old were you at your last birthday?
Acceptable Values
Variable Type
Checks or Restrictions Current Year – Deagy = Deagenum
DEAGENUM DEEDEVER
N/A
Numeric
Have you ever attended Yes = 1 school? No = 2 What is the highest level/grade/year of school you completed? Primary = 1 Secondary = 2 Higher = 3 Single, Never Married =1
Yes/No
DEEDHIGH
Text (Choice)
Asked If Deedever =1
DEMARSTA
What is your current marital status?
Text (Choice) Married = 2 Separated/Divorced = 3 Widowed = 4 Yes = 1 No = 2
DELIVESX
Are you currently living with a sexual partner?
Yes/No
SECTION I: APPENDICES
115
Section
I
APPENDIX I-29
STAFF AGREEMENTS A confidentiality agreement is an understanding between two or more parties that outlines rules and regulations for how the parties involved will manage. The agreement should be signed by each member of the survey team. The rules of conduct guide general staff behaviors and should also be signed by all staff. CONFIDENTIALITY AGREEMENT I, _______________________ in my role as a team member on the <survey name>, representing <organization name>, understand and agree to comply with each of the following requirements: 1. I will treat all information collected for this survey as confidential before, during, and after the survey period. I will not use such information for any purposes other than for the work assigned to me during this study. 2. I will NOT tell anyone outside the study about any of the participants, including what I know about them and their HIV test results. 3. I will NOT remove participant information, whether in paper, audio, or electronic format, from authorized storage area unless I have explicit permission of the Study Coordinator. 4. I WILL refer all data-related questions asked of me that are not within my permission to disclose to the necessary survey team leader, study coordinator, supervisor, and/or study investigators. 5. I WILL maintain all related survey data/material in a secured location at all times. I will also make sure that persons not involved in this study do not have access to survey material. 6. I WILL report the loss of any survey data/material or corruption of any computer files containing survey data immediately to my immediate supervisor, who is responsible for reporting this information to the Principal Investigator. 7. If I use a computer to enter or store collected information, I WILL keep that information in password- protected electronic files only in a computer that has current virus protection software. 8. I will NOT misuse any information security privileges that I may have from working on this survey. 9. I WILL comply fully with any other data confidentiality procedures that I am instructed to follow for this survey. I understand that failure to comply with these rules and regulations could result in disciplinary action, including termination. 10. I WILL protect all electronic study data with passwords. 11. I WILL protect all electronic study data on portable media (e.g. flash drives, CDs) using passwords and encryption. 12. I will NOT give access or password(s) to study data to any person other than study personnel. 13. I will NOT produce copies or back-up of study data sets except as required for the maintenance of the systems. I WILL ensure that the back-up datasets are also stored according to the confidentiality guidelines mentioned above. 14. If I cause a breach or become aware of a breach in confidentiality, I will take immediate steps to secure the sensitive information and inform the survey coordinator/PI. 15. I WILL help facilitate any investigations into breaches of client confidentiality to the best of my ability. _____________________________________________ ______________________ Name/signature of staff person Date: _____________________________________________ ______________________ Name/signature of survey coordinator Date:
116
APPENDIX I-29
RULES OF CONDUCT As a team member on the <survey name>, I will: 1. 2. 3. 4. 5. Treat all participants with respect and in a professional manner. Keep accurate data information and never forge data. Keep truthful records of compensation returned by any participant who declines to be paid. NOT interview a participant (or potential participant) if I know him/her personally. Request other staff to conduct the interview with a participant if I have any issues or foresee potential problems with him/her. 6. NOT have sex with participants during the survey period, or collect contact information from participants for personal use during or after the survey period. 7. NOT drink and/or take drugs while performing survey duties. 8. NOT trade illicit drugs or engage in any commercial activities with participants. My signature below indicates I have reviewed, understand, and accept the above requirements. _____________________________________________ ______________________ Name/signature of staff person Date: _____________________________________________ ______________________ Name/signature of survey coordinator Date:
SECTION I: APPENDICES
117
Section
I
APPENDIX I-30
SPECIMEN HANDLING (COLLECTION, PROCESSING, AND STORAGE) SOP This appendix provides a sample SOP for specimen handling.
<SURVEY NAME> SOP #: DATA MANAGEMENT FOR RDS SURVEY USING QDS These SOPs are written for: Counselor These SOPs may be modified by: Investigator(s) SOP date: Phlebotomist, Nurse Counselor Project Manager MM/DD/YYYY
VENIPUNCTURE Steps for processing blood collected by venipuncture: 1. Collect 10 ml of blood from the patient’s vein into a sterile 10 ml tube. 2. If the blood specimen will not or cannot be processed immediately (e.g. no centrifuge is available), collect the blood in a red-top tube and allow it to stand at room temperature for at least 20–30 minutes, and then remove the serum. Usually, plasma takes longer (~1 hour) to separate without centrifugation than serum collected in a red-top tube. 3. Centrifuge the specimen to separate the serum (without EDTA) or plasma (with EDTA). In general, centrifuge the specimen at 300–400 g or 1200–1500 rpm for 10 minutes. a. If blood is collected for serum, allow blood to stand for 20–30 minutes so that a clot forms before specimen is centrifuged. 4. After specimen is centrifuged or has had time to separate: a. Use clean plastic pipette (do not pour) to remove an aliquot of 1.0–2.0 ml off top layer. b. Transfer aliquot to another sterile labelled tube (plastic, not glass) or cryovial (1.5–2.0 ml) with a screw cap and tighten cap. c. Specimen is now ready for storage and testing Steps for storing serum and plasma: 1. 2. 3. 4. 5. 6. Make sure cap is closed tightly on labelled cryovial or plastic tube. Place cryovial in cardboard freezer box with partitioned insert. For specimens to be transported to testing laboratory, store specimens at 4–8°C for up to 1 week. For longer-term storage, freeze specimens at -20°C or below. Pack specimens upright in cooler containing cold packs for transport to testing facility. Limit number of freeze/thaw cycles.
118
APPENDIX I-30
FINGER-STICK Blood collected by finger-stick can be used to perform rapid tests and for storage and transport using as dried blood spot (DBS) on filter-paper. Steps for collecting blood by finger-stick: 1. 2. 3. 4. 5. 6. Massage the finger (preferably the middle or ring finger), causing blood to accumulate at the tip of the finger. Cleanse finger pad (not just the tip or side of the finger) with 70% isopropyl (rubbing) alcohol. Wipe away alcohol with sterile gauze pad. Use finger prick collection kit or sterile lancet to firmly prick finger. Wipe first drop of blood off finger with sterile gauze. Collecting subsequent blood using a collection device (pipette) to place on rapid test device or on filter paper for DBS. 7. If original puncture is inadequate, repeat finger stick using another finger. 8. Avoid milking or squeezing puncture, which may cause hemolysis of specimen and could invalidate test result. 9. The earlobe may be pricked instead of the finger.
STEPS FOR PREPARING AND STORING A DRIED BLOOD SPOT FOR AN HIV TEST 1. Apply blood directly from finger or pipette onto special filter paper (Schleicher and Schuell Grade 903 filter paper). Paper may come with preprinted circles that will contain approximately 100 µl blood when completely filled. If paper does not have preprinted circles, place blood on paper so that it makes a circle 1.5 cm diameter. 2. Allow blood to soak through and fill the entire circle. 3. Label side of the filter paper with specimen code after filter paper is saturated with blood (circle is filled). Suspend filter-paper strips containing filled circles during drying process to allow air to circulate around paper. Stands for holding strips are commercially available. However, strips may be dried by placing them between two books (taping the edges of strips to the books with sticky tape) on a table or lab bench top so that the blood-containing part of the paper is not in contact with surface of the table or lab bench top. Be sure not to get tape on the blood spot. 4. Let blood spots air dry at room temperature for at least 4 hours (and for at least 24 hours in humid climates). 5. Do not heat or stack blood spots, and do not allow them to touch other surfaces while they are drying. 6. After blood spots have been adequately dried, place filter papers between sheets of glassine paper or plastic to prevent carryover of specimen from one sheet to another. 7. Place wrapped strips in a gas-impermeable bag (zip-lock bag) with desiccant and humidity indicator cards. Approximately 20 strips may be placed in each bag. Bags may be kept at room temperature for up to 30 days. 8. Place bags in a designated envelope for shipment. 9. If DBS in their stored plastic bags will be stored for more than 90 days, they should be maintained at -20°C. Properly stored DBS have been shown to be stable for at least two years. 10. If additional testing, such as resistance testing, is anticipated, DBS must be stored at or below -20°C immediately after the DBS specimens are dry.
SECTION I: APPENDICES
119
Section
I
APPENDIX I-31
RECRUITMENT TRAINING SCRIPT The following is an example of a script to train participants in how to recruit peers for a survey using respondent-driven sampling (RDS). The use of a script can minimize recruitment selection bias by standardizing the way different staff members train participants in peer recruitment. This example is adapted from the NHBS-HET2 Operations Manual. Investigators should fill in the information within the “<>” (i.e. <number of coupons>) to tailor this script to their survey.
SCRIPT: Who to give the coupons to We will give you <number of coupons> coupons to give to friends, relatives, or people you are close to so that they can be in the survey, too. The coupons will be used to recruit them to be in the survey. You should give the coupons to people you know who are <target population>. You should only give the coupons to people who live in <city>. Because people can be in the survey only once, do not give the coupons to anyone who has already participated. Most importantly, you should NOT give the coupons to strangers. Coupons To be in the survey, everyone has to have a coupon. Be sure to tell the people you give a coupon to that they need to have the coupon with them when they come in or call to make an appointment. The first thing we will do is check to see if their coupon is valid. Your coupons cannot be replaced if they are lost or stolen or if the person you recruited is not eligible. Each coupon has an expiration date, and after that date it cannot be used anymore. Process When you give the coupons out, be sure to tell the people you recruit about the survey. For example, tell them that the survey is confidential and that the information will be used to provide better services for <target population>. Mention that there is an interview and a test for HIV and that the whole survey process should take about 1 to 2 hours. People you give coupons to who complete the survey will be given <primary compensation amount>, like you will today. We will not admit anyone into the survey that is drunk or high on drugs. Reward You will get paid <secondary compensation amount> for each person you recruit who is eligible and completes the survey. But, it is not guaranteed that you will get the <secondary compensation amount> just for recruiting someone. • You will not be paid for someone who is not eligible. • You will not be paid for recruiting someone who has already participated in the survey. • You will not be paid for someone who does not complete the entire survey. Not everyone in this survey gets the opportunity to recruit others, and not everyone gets the same number of coupons. Our computer decides who gets to recruit other people for the survey and how many people they get to recruit. The survey is time-limited, so eventually there will be no more coupons given out.
120
APPENDIX I-31
Recruiter Information In order for us to be sure that we give the reward to the right person, we are going to ask you a few questions and enter the information into the computer to create a participant code that is unique to you. When you come in to get paid, we will ask you those same questions again to create the code and check it in the computer. The coupons we give you are linked to you so we will know which ones to pay you for and when. We will schedule a second visit for you after your coupons have expired. The people you recruited should have participated by this date. When you come in for your second visit you will be reimbursed for each recruit who had participated. We will only pay you, so do not send someone else in to collect the money for you.
Closing Do you have any questions? Okay, remember, give the coupons to people you know. Thanks for helping us!
SECTION I: APPENDICES
121
Section
I
APPENDIX I-32
RDS PARTICIPANT CHECKLIST This checklist allows survey staff to track the steps completed by each participant in a respondentdriven sampling (RDS) survey. There is one checklist for the participant’s first visit and another for the second visit. These checklists provide documentation that the participant completed the required steps to receive compensation for completing any survey-related procedures (first visit) and referring peers to participate in the RDS survey (second visit). Additional rows may be added, as appropriate, for additional coupons, other specimen collection, treatment provided, and referrals to qualitative interviews. RDS Participant Checklist: 1st Visit This form should be filled out only by authorized personnel. Each step should be initialed by the appropriate survey staff member.
Coupon ID/Survey ID: Date: Site location: Unique participant code (UPC):
Activity
Completed
Initials*
Comments
1
Arrival time
_____ : _____ am/pm
CM/REC
2
Presented valid coupon
□ Yes □ No → FILL INELIGIBILITY FORM □ □ Yes No → FILL INELIGIBILITY FORM
CM/REC INT
3
Eligible to participate
4
Provided informed consent
Interview: □ Y □ N → FILL REFUSAL FORM
□Y□N Specimen storage: □ Y □ N Specimen/testing:
INT
5
Answered network size questions
□ □
Yes No
INT
122
APPENDIX I-32
Activity
Completed
Initials*
Comments
6
Completed the questionnaire
□ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □
Yes No Yes No Yes No Yes No Yes No Yes No Yes → NUMBER GIVEN: _______ No Yes No Yes → AMT RECEIVED: _______ No Yes → DATE: ______________ No
INT
7
Received (HIV) pre-test counseling
CN
8
Provided biological specimen(s) Received (HIV) post-test counseling and all rapid test results Referred to appropriate services Received prevention commodities and supplies (such as informational brochures, condoms, and lubricants) Issued RDS coupons
CN/PB
9
CN
10
CN
11
CN
12
CM
13
Received recruitment training
CM
14
Received primary compensation
CM/SS
15
Date/time scheduled for second visit
CM/REC
16
Time of departure
_____ : _____ am/pm
CM/REC
*CM = coupon manager; REC = receptionist; INT = interviewer; CN = counselor; PB = phlebotomist; SS = site supervisor
SECTION I: APPENDICES
123
RDS Participant Checklist: 2nd Visit This form should be filled out only by authorized survey personnel. Each step should be initialed by the appropriate survey staff member.
Coupon ID/Survey ID: Date: Site location: Unique participant code (UPC):
Activity 1 2 Arrival time Validated UPC Completed peer recruitment questionnaire Received post-test counseling and results for non-rapid tests Verified coupons returned
Completed _____ : _____ am/pm
Initials* CM/REC CM/REC
Comments
□ □ □ □ □ □ □ □ □ □
Yes No → LEAVE SITE Yes No Yes No Yes No Yes → AMT RECEIVED: No
3
INT
4
CN
5
CM
6
Provided secondary compensation
CM
7
Time departed site
_____ : _____ am/pm
CM
*CM = coupon manager; REC = receptionist; INT = interviewer; CN = counselor
124
APPENDIX I-32
Section
APPENDIX I-33 DATA ERROR AUDIT LOG
I
Use this form in the field or during the data checking phase to document errors, whether they can be corrected or not. Correct errors in a systematic and controlled way based on a pre-determined data cleaning plan. Make sure that the errors are corrected according to the data cleaning plan. Document all modifications made to original values.
Survey Site: ___________________________________________ Data Source: __________________________________________ Survey ID (range of IDs) Variable name (s) Description error Resolution description Original value New value Date of final resolution Initials of data manager
SECTION I: APPENDICES
125
Section
I
APPENDIX I-34
SAMPLE TABLE OF CONTENTS FOR SURVEY REPORT This sample table of contents can guide development of the survey report. I. Foreword, Acknowledgement, Preface II. Executive summary III. Table of contents IV. List of tables V. List of figures VI. List of abbreviations and terms VII. List/map of survey locations 1. Background a. Context of HIV in <country> among <population> during <time period> b. Survey rationale, goals, objectives, whether this is first or subsequent survey round 2. Methods a. Formative assessment b. Staff training c. Community mobilization d. Target population and study sites e. Sample size and sampling design f. Data collection g. Laboratory procedures h. Supervision and monitoring i. Data entry, cleaning, and management j. Data analysis k. Human subjects considerations l. Funding and collaborating institutions 3. Findings a. Survey response b. Sample (unweighted) and population (weighted) characteristics c. Uptake of services, exposure to interventions d. HIV-related risk behaviors e. HIV infection i. Univariate analysis ii. Bivariate analysis iii. Multivariate analysis iv. Trend analysis f. Other outcomes i. Univariate analysis ii. Bivariate analysis iii. Multivariate analysis g. Size estimation
126
APPENDIX I-34
4. Discussion a. Summary of key findings b. Interpretation of findings c. Limitations 5. Recommendations a. For MOH and other policy makers b. For service providers c. For donors d. For surveyed population e. For future surveys 6. References 7. Appendices a. Consent language b. Data instruments c. Forms d. Data quality assessment results e. Testing quality (quality control testing) f. Access to individual survey data g. List of survey staff h. List of investigators
SECTION I: APPENDICES
127
Section
I
APPENDIX I-35
SAMPLE PRESS RELEASE This is sample text for a press release. A press release announces a media event and informs journalists of the main survey findings, implications, and recommendations. Investigators should carefully consider whether “going public” through a press release will be beneficial to the surveyed population (e.g., services, funding) without causing harm (e.g., stigma, police raids). MINISTRY OF HEALTH <Name of AIDS Control Program>
FOR IMMEDIATE RELEASE PRESS RELEASE <Date> HEADER (Describe one or two key points to highlight survey’s key findings) Survey of men who have sex with men finds high level of HIV infection but low uptake of treatment, frequent mental health challenges
SUB-HEAD: Complete survey results to be disseminated on <weekday, date>
SAMPLE TEXT: The Ministry of Health, in collaboration with <organization name> and <organization name>, will release the results of the <name of survey> Survey on HIV and related risk behaviors among <key population surveyed> during a workshop and press briefing to be held at the <hotel or conference location name> in <city>on <day of week, date>, at <time>. “The Ministry of Health is pleased to disseminate these findings and thanks the <donor names> for its support of this research,” said Dr. <name>, <position>, MoH. “The results of this important survey will be used not only to measure progress toward achieving health and other program and policy goals, but also to assess trends in HIV prevalence and risk behaviors/other biomarker results among <target population> and help understand the impact of interventions geared towards this population.” The Survey examined the burden of HIV disease, uptake of HIV-related services, and a broad range of HIV-related risk factors among <target population> in <city>. In addition to sexual health and HIV infection, the survey measured <other HIV-related biomarkers>, mental health, alcohol use, drug abuse, uptake of HIV services, as well as exposure to <other issues, such as homophobia and violence>. A total of <number> respondents participated and completed <computer-based> anonymous interviews.
128
APPENDIX I-35
Among the key findings, the survey revealed that <number> in <number> <target population> are infected with HIV, and only <X%> of these are aware of their HIV infection. Of those HIV-infected, only <number> in <number> <X%> are on lifesaving anti-retroviral treatment. And of those on treatment, X% are virally suppressed. Almost <number> in <number> have a low CD4 count (below 500 cells), indicating that they are in urgent need of treatment. Levels of <depression> and <anxiety> are high. Non-injecting drug use is common (<number> in <number> consumed <drug names> in the last six months). Condom use varies by partner and is low within <kind of partnership> partnerships even if the partner’s serostatus is not known. The survey was carried out by <name of organization> between <month> and <month> <year> as part of the overall key population based HIV surveillance system, a collaborative project among <organization name>, the <organization name>, and <organization name>. <Organization name> provided technical assistance in the implementation of the survey.
BACKGROUND The < survey name > is an HIV and health-related risk behaviors surveillance project in <city> that focuses on generating strategic HIV-related information with an aim of informing <country> ’s health policy makers, service delivery providers, and the public. Since the establishment of <Survey activity> in <year>, <institution name> has focused on supporting the Government of <country> and the Ministry of Health in its efforts to scale up HIV prevention, care, treatment, and support services by providing both critical funding and key technical assistance at the national, district, community, and civil society levels. The full survey report is downloadable at:
(QR code below linking to website)
http://websiteaddress.xxx
SECTION I: APPENDICES
129
Section
I
APPENDIX I-36
DATA PRESENTATION GUIDE The physical presentation of survey findings can increase their utility and impact. This appendix provides suggestions to improve data presentation. SLIDE PRESENTATIONS • • • • • • • • Provide a printout of your slides before presentation. Choose a basic template/background color for your slides. Use a text color that is distinguishable from the background. As a rule, the amount of text displayed in a slide should be minimal. Use “telegram style”; there should be fewer words on the slide than in your speech. Try to limit the text to no more than 5 lines of text per slide using bullets and no more than 25 words per slide. If necessary, split the slide’s text into two or more slides. Make the font size large enough so that people in the back of the room (or using printed handouts) can easily read it. Do not use animation, or use it sparingly. Do not use a busy (distracting) background. When presenting the slides, make your spoken words accompany the slides’ text, so that people hear and read the same content at the same time. Try to avoid speaking about things that are not shown on the slide, and do not include text on the slide without addressing it. Present data (numbers) in tables or as a graph (chart). It is better to have more graphs or tables with less information on each slide than to try to put all the data into one graph/table.
FIGURES Figures (graphs and charts) are an ideal way to present survey findings to an audience using slide presentations, and may also be used in technical reports. Survey findings presented through simple and clear graphs (or charts) are often much more persuasive to the majority of people than tables or text. Bear in mind that many people may want to reproduce graphs to use in their own presentations, so they should be self-contained and self- explanatory. A good graph contains: • • • • A title that clearly specifies what is being shown, including the population in the numerator and denominator, the location and the time (e.g. year) of data collection. A data source (survey name). Axis labels in charts, explaining what each axis is displaying. Be sure to include units. Value labels for each data point, if not too many. This allows people to read values and either recreate the graph or refer to them in a text or speech. Where there are many data points on a graph and data labels are likely to cause confusion, the base data for the graph may accompany it in a small table. Truncate any decimals that imply unwarranted precision. Legends should be used in charts with more than one series. Significance levels and/or confidence intervals. In technical reports, the p-value itself may be reported on the data series for the graph. For more general presentations, relevant data series can simply be labeled “statistically significant.” Alternatively, indicate statistical significance through the use of one or more asterisks.
• •
130
APPENDIX I-36
DATA TABLES Tables are an ideal way to present data in technical reports and manuscripts. In contrast to figures, tables are well suited to display many and detailed data points, using exact decimals, confidence limits, and p values. When presenting data in tables, many of the same standards for graphs hold true. Each table should stand on its own; it should have a full title (person, place, time) and clear row and column headings. Special care should be taken to clearly show who is included or excluded in the denominator. The table should include the number of observations that make up proportional or relative risk values. When presenting percent values, be clear whether these are row, column, or cell percent values. Indicate missing data points.
SUPPORTING EVIDENCE No single survey can provide definitive evidence or answers, and all survey data are estimates. People are more likely to be persuaded that observations are real if different data from various data sources all point in the same direction. A graph showing a significant rise in self-reported HIV testing will be more powerful if it is presented in conjunction with other data showing a rise in the uptake of care or treatment. These different data may not necessarily come from the same source, and may include other surveys, routine health system reporting, or marketing data.
SECTION I: APPENDICES
131
Section
I
APPENDIX I-37
DATA USE (CONFIDENTIALITY) AGREEMENT The purpose of this sample text is to provide a template for an agreement to share data between investigators and other researchers. Investigators may want to share all or part of the collected data. Investigators may ask to be acknowledged and/or informed of any publication as a result of a data sharing agreement. Adapt the text below as needed. DATA USE AND CONFIDENTIALITY AGREEMENT Parties to this agreement: <Institution Name>, represented by: <Person name> (“Data provider” or “DP”) <Institution Name>, represented by: <Person name> (“Data recipient” or “DR”) A. B. C. This agreement refers to: <Survey Name, Year, Location> This agreement serves for the purpose of: <Data analysis to evaluate correlates of HIV infection among survey group> This agreement expires on DD/MM/YYYY.
Through this agreement, the Data Recipient: 1. Agrees to safeguard the received information against any use inconsistent with the Data Use Agreement and against any further disclosure or dissemination. The received data will be stored on password-protected computers and/or in locked cabinets. 2. Agrees to maintain a list of those persons who are provided with access to the data within their own organization. 3. Agrees not to use the provided data beyond the purpose stated in B. 4. Agrees to make no use of the identity of any person discovered, inadvertently or otherwise, and advise the principal investigator of any such discovery (<email address>). 5. Agrees not to link these data files with individually identifiable data from any other data set. 6. Pledges not to publish the received individual data on the World Wide Web, and not to distribute them to any other organization or individual. 7. Agrees to refer any request for these data by third parties to the DP. 8. Agrees to inform the DP in writing within X days of any request by authorities to have access to these data. 9. Agrees not to use these data in any way that would put the surveyed population at risk or in harm’s way, or stigmatize them. 10. Agrees to inform the DP in writing of any breach of this agreement, including but not limited to any possible or confirmed unauthorized transfer of these data to third parties. 11. Agrees to acknowledge and inform the DR when publishing or disseminating results and findings stemming from this data analysis. Signature: _______________________________________ Date: ___________________________________________ Type DR representative name Type DR organizational name
132
APPENDIX I-37
Supplemental Materials
Section
I
APPENDICES
Section
II Section
QUESTIONNAIRE MODULES
III
INDICATORS
Section
IV
GLOSSARY
Section II - Questionaire Modules 1. Questionnaire Parameters 2. Eligibility 3. RDS Recruiter-Recruit Relationship 4. RDS Personal Network Size 5. Size Estimation 6. ACASI Tutorial 7. Demographics 8. Transgender Characteristics 9. Sex Work Characteristics 10. Clients Of Sex Workers Characteristics 11. Incarceration Characteristics 12. Transport Worker Characteristics 13. Reproductive Health 14. Sexual History 15. Recall Sexual Behavior 16. Partner Concurrency 17. Last Sex Act 18. Condom Use/Accessibility 19. Lubricant Use 20. Microbicides 21. Alcohol Use 22. Non-Injection Drug Use 23. Injection Drug Use 24. Drug Overdose 25. Sexually Transmitted Infections 136 140 151 154 156 159 162 171 176 185 188 196 200 208 218 239 243 249 256 261 264 270 277 295 301
134
SECTION 2: QUESTIONAIRE MODULES
26. Hepatitis B and C 27. Services Uptake 28. PEP and PrEP 29. Shame, Stigma, Harassment & Discrimination 30. Physical Violence 31. Sexual Violence: Coercion and Rape 32. Internalized Homophobia 33. Depression 34. Suicide Ideation and Attempts 35. Social Support 36. Social Cohesion, Social Participation and Collective Agency 37. Game of Contacts 38. HIV Knowledge and Perceptions 39. Questionnaire Feedback/Interview Status 40. RDS, Peer Recruitment
306 309 341 346 369 383 403 406 410 415 431 437 439 442 444
SECTION 2: QUESTIONAIRE MODULES
135
Section
Il
1. QUESTIONNAIRE PARAMETERS
This module collects administrative information about the survey activity and participant to facilitate data management. It can be used at the start of all electronic data collection instruments including screening, size estimation, interview, exit interview, and RDS second visit interview. Unlike other modules, this module is not asked of participants. The data are either computer generated or entered by survey staff.
Investigators may choose to program the questionnaire to automatically fill in information that will be the same for every participant, including questionnaire version, survey-design, and city. NA QP Questionaire Parameters Response Options Core or Secondary* C
Data Generator
Comments
1.
QPVER
Questionnaire version
<TEXT>
Computer-generated
2.
QPDE
Device name
<TEXT>
S
Staff
3.
QPIDATE
Date
<DATE>
C
Computer-generated
4.
QPITIME
Interview start time
<TIME>
C
Computer-generated
136
1. QUESTIONNAIRE PARAMETERS
NA 5.
QP QPSURVD
Questionaire Parameters What survey design is being used?
Response Options NON-PROBABILITY RDS TLS CCS OTHER 1 2 3 4 5
Core or Secondary* C
Data Generator Staff
Comments This is a filter question.
6.
KEYPOP
What population is this person sampled for?
SEX WORKERS (SW) MEN WHO HAVE SEX WITH MEN (MSM) TRANSGENDER WOMEN (TG) PEOPLE WHO INJECT DRUGS (PWID) INCARCERATED POPULATION TRANSPORT WORKERS CLIENTS OF SEX WORKERS OTHER
1 2 3 4 5 6 7 8
C
Staff
7.
QPMODE
What mode of interview administration is being used?
ACASI CASI FACE-TO-FACE CAPI FACE-TO-FACE HAPI PAPER OTHER
1 2 3 4 5 6
C
Staff
SECTION 2: QUESTIONNAIRE MODULES
1 3 1 737
NA 8.
QP QPLANG
Questionaire Parameters What interview language is being used?
Response Options LANGUAGE A LANGUAGE B LANGUAGE C OTHER LANGUAGE 1 2 3 6
Core or Secondary* C
Data Generator Staff
Comments
9.
QPCITY
In which city is this survey being implemented?
CITY A CITY B CITY C OTHER CITY
1 2 3 6
C
Staff
Required only if survey is being conducted in multiple cities.
10.
QPSITE
(RDS) At which site is this survey being implemented?
SITE A SITE B SITE C
1 2 3
C
Staff
ONLY IF QPSURVD = 2 Required only if an RDS survey has multiple sites or offices in the same city.
11.
QPVENUE
TLS/CCS Venue/Cluster At which venue is this survey being implemented?
VENUE A VENUE B VENUE C
1 2 3
C
Staff
ONLY IF QPSURVD = 3 or 4 Required for TLS or CCS.
12.
QPICODE
Please enter interviewer code.
[ | | | ]
C
Staff
Required if an interviewer is used. SKIP IF ELIGIBILITY SCREENING
13.
QPID
Survey ID
[ | | | ]
C
Staff
138
1. QUESTIONNAIRE PARAMETERS
NA 14.
QP QPIDCK
Questionaire Parameters Re-enter Survey ID
Response Options [ | | | ]
Core or Secondary* C
Data Generator Staff
Comments SKIP IF ELIGIBILITY SCREENING Should be entered twice.
*C=Core, S= secondary
SECTION 2: QUESTIONNAIRE MODULES
1 3 1 939
Section
Il
2. ELIGIBILITY
All survey participants must meet the eligibility criteria. Responses to the eligibility screening questions are used to determine who is eligible to participate in the survey. This is the only module that does not allow the participant to refuse to answer a question. This module also includes questions to help survey staff confirm that the prospective participant is in fact part of the target population. For example, the eligibility questions for persons who inject drugs include asking how to prepare or inject drugs. All eligibility questions should be asked before a prospective participant is informed whether he/she is eligible to participate in the survey. The interviewer can then determine eligibility based on the responses to all questions. If using electronic data collection instruments, the interviewer can inform the prospective participant that the computer determines who is or is not eligible to participate and that there are no ‘correct’ answers. For RDS surveys, ensure that prospective participants have a valid coupon before screening them for eligibility. This module should be conducted face-to-face by the interviewer. N/A Variable ID: EL N/A EL1MSG
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M T P O
[Instructions to interviewer]: Before asking population-specific questions to determine eligibility, all prospective participants should be asked the first few questions on age, sex and geographic locations. [Interviewer should read]: Thank you for being here. We will now ask you some questions to determine if you meet the requirements to take the full survey. Please enter participant’s RDS coupon number. ONLY IF QPSURVD = 2 Required for RDS. Should be entered twice to avoid data entry errors.
N/A
EL2MSG
1.
ELCIN
140
2. ELIGIBILITY
N/A Variable ID: EL 2. ELCINCK
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M T P O
Please re-enter participant’s RDS coupon number.
ONLY IF QPSURVD = 2 Required for RDS.
3.
ELAGEL
How old were you at your last birthday?
AGE IN COMPLETED YEARS: [ | ] DON’T KNOW 97
Age must be above pre-determined minimum value.
C
C
C
C
C
4.
ELLOC
Do you live or work in [insert name of city]?
YES NO
1 2
BACKGROUND/ RISK Only for RDS/VBS/TLS
S
S
S
S
S
5.
ELLONG
How long have you lived or worked in [insert # YEARS: [ | ] name of city]? DON’T KNOW 97 CODE 00 IF <1 YEAR
IF KEYPOP=1 → BACKGROUND/ RISK ELSWSX Only for RDS/VBS/TLS IF KEYPOP=2 → ELMSMM IF KEYPOP=3 → ELTGNOW IF KEYPOP=4 → ELINJ
S
S
S
S
S
PWID 6. ELINJ
PWID Eligibility Have you ever in your life shot up or injected YES any drugs other than those prescribed for NO you? By shooting up, I mean anytime you might have used drugs with a needle, either DON’T KNOW by mainlining, skin popping, or muscling. 1 2 7 ‘2’, ‘7’ → SKIP TO EL5MSG RISK/ VULNERABILITY N/A N/A N/A C N/A
SECTION 2: QUESTIONNAIRE MODULES
1 4 141
N/A Variable ID: EL 7. ELINJMN
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
How long ago did you last inject?
LESS THAN 1 MONTH AGO 1 OR MORE MONTHS AGO BUT LESS THAN 12 MONTHS AGO 1 OR MORE YEARS AGO DON’T KNOW
1 2 3 7
‘1’,→SKIP TO ELINJMND ’2’ →SKIP TO ELINJMNM ’3’→SKIP TO ELINJMNY ‘7’ → SKIP TO EL5MSG
RISK
S
8.
ELINJMND
How many DAYS ago did you last inject?
# DAYS [ | ] MAX: 31 DON’T KNOW 97
RISK
S
S
S
C
S
9.
ELINJMNM
How many MONTHS ago did you last inject?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
S
S
S
C
S
10.
ELINJMNY
How many YEARS ago did you last inject?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97
RISK
S
S
S
C
S
N/A
EL3MSG
[Instructions for interviewer]: The following questions are to validate injection behavior.
142
2. ELIGIBILITY
N/A Variable ID: EL 11. ELINJDR
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P S
O N/A
The last time you injected, which drug did you Inject?
HEROIN COCAINE MORPHINE/OPIUM SPEEDBALL (HEROIN & COCAINE MIX) OTHER DON’T KNOW
1 2 3 4 6 7
BACKGROUND/ RISK/ VULNERABILITY Use locally defined responses.
N/A
12.
ELINJB
Where on your body do you usually inject? [Interviewer: Have participant show ALL injection areas on body. Check for physical signs of injection and indicate accordingly]
FRESH TRACK MARKS NEEDLE-SIZE SCABS ABSCESSES OLD TRACK MARKS OR SCABS NO PHYSICAL SIGNS/INJECTS IN COVERED AREA
1 2 3 4 5
’1’,’2’,’3’,’4’ →SKIP TO ELINJBT
BACKGROUND/ RISK/ VULNERABILITY
N/A
N/A
N/A
S
N/A
N/A
EL4MSG
[Instructions for interviewer]: If the participant does not have any visible injection marks, then ask the following questions:
13.
ELINJDES
Can you please tell me, step by step, how you prepare your drugs? Description could include: Mix drugs with water or lemon juice/vinegar Use cooker/heat drugs Use filter
DESCRIPTION OK DESCRIPTION NOT OK
1 2
BACKGROUND/ RISK/ VULNERABILITY
N/A
N/A
N/A
S
N/A
SECTION 2: QUESTIONNAIRE MODULES
1 4 1 343
N/A Variable ID: EL 14. ELINJSTP
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P S
O N/A
Can you please describe to me, step-bystep, how you inject your drugs? Description could include: Tie off and find vein (IVDU) Clean injection site Register (IVDU)
DESCRIPTION OK DESCRIPTION NOT OK
1 2
N/A
15.
ELINJSYR
What type of syringe do you usually inject with? Description could include: Syringe size (in cc’s or units) Needle size (gauge, length) Cap (color, number)
DESCRIPTION OK DESCRIPTION NO OK
1 2
RISK/ PREVENTION
N/A
N/A
N/A
S
N/A
16.
ELINJBT
The last time when you injected [insert response from ELINJDR], did you buy the drug or did someone give you the drug?
BOUGHT MYSELF SOMEONE GAVE TO ME DON’T KNOW
1 2 7
‘2’, OR ’7’→ SKIP TO EL5MSG
RISK/ PREVENTION
N/A
N/A
N/A
S
N/A
17.
ELINJMNY
How much did you pay to buy the amount of drugs you got?
[insert local currency] DON’T REMEMBER 9997
RISK/ PREVENTION
N/A
N/A
N/A
S
N/A
N/A
EL5MSG
[Interviewer should read]: We have finished the first series of questions. Now the computer will determine whether you’ve been selected to participate in the survey.
N/A
ELINJNO
[Interviewer should read]: The computer has determined that you are not eligible to participate in the study. Thank you for your time and interest.
SKIP TO END OF SURVEY
Eligibility is determined based on survey criteria.
N/A
N/A
N/A
C
N/A
N/A
ELINJYES
[Interviewer should read]: The computer has determined that you are eligible to participate in the study.
Skip to QPID
Eligibility is determined based on survey criteria.
N/A
N/A
N/A
C
N/A
144
2. ELIGIBILITY
N/A Variable ID: EL MSM 18. ELMSMW
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M T P O
MSM Eligibility Have you ever had sex? By sex, we mean either vaginal sex or anal sex? With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus YES NO DON’T KNOW 1 2 7 ‘2’, ’7’ → SKIP TO EL6MSG N/A S N/A N/A N/A
19.
ELMSMM
Have you ever had anal sex with a man? With anal sex we mean a penis enters a person’s anus.
YES NO DON’T KNOW
1 2 7
N/A
C
N/A
N/A
20.
ELMSMSXT
How long ago did you last have anal sex with LESS THAN 1 MONTH AGO a man? 1 OR MORE MONTHS AGO BUT LESS THAN 12 MONTHS AGO 1 OR MORE YEARS AGO DON’T KNOW
1 2 3 7
‘1’→SKIP TO ELMSMSXTD ‘2’ → SKIP TO ELMSMSXTM ‘3’ → SKIP TO ELMSMSXTY ‘7’ → SKIP TO EL6MSG
RISK
S
C
S
S
S
21.
ELMSMSXTD
How many DAYS ago did you last have anal sex with a man?
# DAYS [ | ] MAX: 31 DON’T KNOW 97
RISK
S
C
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 4 1 545
N/A Variable ID: EL 22.
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M C
T S
P S
O S
ELMSMSXTM How many MONTHS ago did you last have anal sex with a man?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
S
23.
ELMSMSXTY
How many YEARS ago did you last have anal sex with a man?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97
SKIP TO END OF SURVEY
RISK
S
C
S
S
S
N/A
EL6MSG
[Interviewer should read]: We’ve finished the first series of questions. Now the computer will determine whether you’ve been selected to participate in the survey. [Interviewer should read]: The computer has determined that you are not eligible to participate in the study.Thank you for your time and interest. [Interviewer should read]: The computer has determined that you are eligible to participate in the study. SKIP TO END OF SURVEY Skip to QPID Eligibility is determined based on survey criteria Eligibility is determined based on survey criteria N/A C N/A N/A N/A
N/A
ELMSMNO
N/A
ELMSMYES
N/A
C
N/A
N/A
N/A
TG 24. ELTGNOW
TG Eligibility Do you consider yourself as male, female, or other? MALE FEMALE OTHER 1 2 3 N/A N/A C N/A N/A
146
2. ELIGIBILITY
N/A Variable ID: EL 25. ELSXBORN
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M N/A
T C
P N/A
O N/A
What was your sex at birth?
MALE FEMALE OTHER
1 2 3
If responses to ELTGNOW and ELSXBORN differ the person meets the TG-specific eligibility criteria. ‘2’ → SKIP TO EL7MSG Given relatively small population of TG individuals, consider not basing eligibility on having had sex in a given recall period.
N/A
26.
ELTGSX
Have you ever had sex? By sex, we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus.
YES NO
1 2
N/A
N/A
S
N/A
N/A
27.
ELTGSXM
Have you ever had anal sex with a man? With anal sex we mean a penis enters a person’s anus.
YES NO DON’T KNOW
1 2 7
‘2’, ‘7’ → SKIP TO EL7MSG
N/A
N/A
S
N/A
N/A
28.
ELTGSXT
How long ago did you last have anal sex with LESS THAN 1 MONTH AGO a man? 1 OR MORE MONTHS AGO BUT LESS THAN 12 MONTHS AGO 1 OR MORE YEARS AGO DON’T KNOW
1 2 3 7
‘1→SKIP TO ELTGSXTD ‘2’ → SKIP TO ELTGSXTM ‘3’ → SKIP TO ELTGSXTY ‘7’→ SKIP TO EL7MSG
RISK
S
C
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 4 1 747
N/A Variable ID: EL 29. ELTGSXTD
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M C
T S
P S
O S
How many DAYS ago did you last have anal sex with a man?
# DAYS [ | ] MAX: 31 DON’T KNOW 97
RISK
S
30.
ELTGSXTM
How many MONTHS ago did you last have anal sex with a man?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
S
C
S
S
S
31.
ELTGSXTY
How many YEARS ago did you last have anal sex with a man?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97
RISK
S
C
S
S
S
N/A
EL7MSG
Interviewer should read: We’ve finished the first series of questions. Now the computer will determine whether you’ve been selected to participate in the survey. [Interviewer should read]: The computer has determined that you are not eligible to participate in the study. Thank you for your time and interest. [Interviewer should read]: The computer has determined that y ou are eligible to participate in the study. SKIP TO END OF SURVEY Eligibility is determined based on survey criteria. Eligibility is determined based on survey criteria. N/A N/A C N/A N/A
N/A
ELTGNO
N/A
ELTGYES
N/A
N/A
C
N/A
N/A
148
2. ELIGIBILITY
N/A Variable ID: EL SW 32. ELSWSX
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M T P O
SW Eligibility Have you ever had sex with someone for money? YES NO DON’T KNOW 1 2 7 ‘2’, ‘7’ → SKIP TO EL8MSG C N/A N/A N/A N/A
33.
ELSWSXT
How long ago did you last have sex with someone for money?
LESS THAN 1 MONTH AGO 1 OR MORE MONTHS AGO BUT LESS THAN 12 MONTHS AGO 1 OR MORE YEARS AGO DON’T KNOW
1 2 3 7
‘1’→SKIP TO ELSWSXTD ‘2’ →SKIP TO ELSWSXTM ‘3’ →SKIP TO ELSWSXTY ‘7’→ SKIP TO EL8MSG
RISK
C
S
S
S
S
34.
ELSWSXTD
How many DAYS ago did you last have sex with someone for money?
# DAYS [ | ] MAX: 31 DON’T KNOW 97
RISK
C
S
S
S
S
35.
ELSWSXTM
How many MONTHS ago did you last have sex with someone for money?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
C
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 4 1 949
N/A Variable ID: EL 36. ELSWSXTY
Eligibility
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
How many YEARS ago did you last have sex with someone for money?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97
RISK
C
N/A
EL8MSG
Interviewer should read: We’ve finished the first series of questions. Now the computer will determine whether you’ve been selected to participate in the survey. [Interviewer should read]: The computer has determined that you are not eligible to participate in the study. Thank you for your time and interest. [Interviewer should read]: The computer has determined that you are eligible to participate in the study. FOR ALL POPULATIONS SKIP TO END OF SURVEY Eligibility is determined based on survey criteria. Eligibility is determined based on survey criteria. C N/A N/A N/A N/A
N/A
ELSWNO
N/A
ELSWYES
C
N/A
N/A
N/A
N/A
37.
QPID
Survey ID
C
C
C
C
C
38.
QPIDCK
Re-enter Survey ID
Enter twice
C
C
C
C
C
39.
ELSWCONT
Are you ready to proceed with the rest of the questions?
YES NO
1 2
Only skip to end of survey if participant refuses to continue. Otherwise, allow participant a short break.
C
C
C
C
C
*C=Core, S= secondary
150
2. ELIGIBILITY
Section
II
3. RDS RECRUITER-RECRUIT RELATIONSHIP
One assumption of respondent driven sampling (RDS) about recruitment is that the population being recruited must know one another as members of the target population. Without satisfying this assumption, investigators cannot successfully use RDS to produce a sample of the target population. This assumption can be verified by determining if there is an existing relationship between the recruiter and the recruit. The first few questions verify the relationship between the recruiter and the recruit as well as the strength of that relationship. The responses for relationships can be modified according to the target population and the context of the survey. The last few questions in this module determine whether the participant was coerced to participate in the survey. These questions should be used for all RDS surveys. This module should be conducted face-to-face by interviewer.
N/A
Variable ID: RC RC1MSG
RDS Recruiter-Recruit Relationship
Response Options
Skips
Comments
Core or Secondary*
N/A
[Interviewer should read]: We’re going to start by asking you about the person who gave you the coupon and about others that [insert eligibility behavior] you know in [insert eligibility locality]. Please remember that your answers will be kept private. Which of the following describes how you know the person who gave you this coupon? CHECK ALL THAT APPLY. A RELATIVE OR FAMILY MEMBER A SEX PARTNER A PERSON YOU USE DRUGS WITH OR BUY DRUGS FROM A FRIEND/AN ACQUAINTANCE A STRANGER OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z ‘A’, ‘C’, ‘D’, ‘E’, ‘X’, ONLY FOR RDS. ‘Y’, ‘Z’→SKIP TO RCLENTH C
1.
RCRELA
SECTION 2: QUESTIONNAIRE MODULES
1 5 151
N/A
Variable ID: RC RCSXTYP
RDS Recruiter-Recruit Relationship
Response Options
Skips
Comments
Core or Secondary* S
2.
What kind of a sex partner is this person? A main sex partner is someone you are committed to, for example could be your spouse, live-in sex partner, or boyfriend. There is no payment or exchange of goods or services for sex with these partners. A casual sex partner is someone you have sex with but don’t feel committed to. There is no payment or exchange of goods and services for sex with these partners.
MAIN CASUAL COMMERCIAL - I PAID HIM/HER COMMERCIAL - HE/SHE PAID ME DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
ONLY FOR RDS.
3.
RCLENTH
Do you know this person for some days, or some months, or for years?
DAYS MONTHS YEARS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
ONLY FOR RDS.
S
4.
RCSEE
How many times did you see this person in the last 30 days?
# TIMES [ | | ] DON’T KNOW 997 REFUSE TO ANSWER 998
ONLY FOR RDS.
S
152
3. RDS RECRUITER-RECRUIT RELATIONSHIP
N/A
Variable ID: RC RCEXCH
RDS Recruiter-Recruit Relationship
Response Options
Skips
Comments
Core or Secondary* C
5.
Was your coupon exchanged for something?
NO, NOTHING MONEY OR GOODS FAVORS OR SERVICES SEXUAL FAVORS OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 6 7 8
ONLY FOR RDS.
6.
RCREAS
What is the main reason you joined this study?
INTERESTED IN THE STUDY WANT TO KNOW MY HIV STATUS WANT TO KNOW IF I HAVE AN STD WANT THE INCENTIVE WAS FORCED TO COME OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
ONLY FOR RDS.
S
*C=CORE, S=SECONDARY for all KP groups.
References: Heckathorn DD. Extensions of respondent driven sampling: analysing continuous variables and controlling for differential recruitment. Sociological Methodology. 2007: 37(1): 151-207
SECTION 2: QUESTIONNAIRE MODULES
1 5 1 353
Section
Il N/A
4. RDS PERSONAL NETWORK SIZE
The weighting of estimates derived from respondent-driven sampling primarily depends on self-reported network sizes. Therefore, because these questions are crucial to obtain adjusted estimates in a respondent-driven sampling survey, this module should not be self-administered. Participants are asked about their networks broadly at first and with each question network sizes should decrease based on survey-specific eligibility characteristics. These questions should be used for all populations using RDS for recruitment.
Variable ID: DG DG1MSG
RDS Personal Network Size
Response Options
Skips
Comments
Core or Secondary*
N/A
[Interviewer should read]: Now I’m going to ask you a few questions about how many people you know.
1.
DGBEHAV
How many people [insert number of males/females] # PEOPLE [ | | ] do you know who [insert eligibility behavior]? MAX: 99 DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’ → SKIP TO END OF CURRENT SECTION
ONLY FOR RDS.
C
2.
DGLOCAL
Of [insert number from DGBEHAV] people, how many live or work in [eligibility locality]?
# PEOPLE [ | | ] MAX: NUMBER FROM DGBEHAV DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’ → SKIP TO END OF CURRENT SECTION
ONLY FOR RDS. Only applicable if locality is an eligibility criterion.
C
154
4. RDS PERSONAL NETWORK SIZE
N/A
Variable ID: DG DGAGE
RDS Personal Network Size
Response Options
Skips
Comments
Core or Secondary* C
3.
Of [insert number from DGLOCAL] how many are aged [insert eligibility age range]?
# PEOPLE [ | | ] MAX: NUMBER FROM DGLOCAL DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’ → SKIP TO END OF CURRENT SECTION
ONLY FOR RDS.
4.
DG14SEEN
Of [insert number from DGAGE], how many have you seen in the past two weeks?
# PEOPLE [ | | ] MAX: NUMBER FROM DGAGE DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’ → SKIP TO END OF CURRENT SECTION
ONLY FOR RDS. Consider changing time based on coupon validity and local context.
C
5.
DGINVITE
Finally, of [insert number from DG14SEEN], who would you consider inviting to participate in this survey?
# PEOPLE [ | | ] MAX: NUMBER FROM DG14SEEN DON’T KNOW 997 REFUSE TO ANSWER 998
ONLY FOR RDS. Should be used to determine network size for RDS.
C
*:C=CORE, S= SECONDARY for all KP groups.
SECTION 2: QUESTIONNAIRE MODULES
1 5 1 555
Section
Il
5. SIZE ESTIMATION
An estimate of the size of the population helps in the design of an appropriate programmatic response, advocate for resources or policy changes, and facilitate program monitoring and evaluation. One version of the multiplier method compares two independent sources of data for populations to estimate the total number in the population. The first source is a count or listing from program data including only the population whose size is being estimated, and the second source is a representative survey of the populations whose size is being estimated. Another version of multiplier method involves the distribution of a unique object to members of the population. It relies on access to members of the key populations where a particular unique object is distributed broadly. During the survey the participants are asked if they had received services at the particular institution or if they had received the object that had been distributed prior to implementation of the survey. This module should be conducted face-to-face. Details of these methods are provided in the guidelines or other references. Information collected using these questions will help estimate the size of the population being surveyed. N/A Variable ID: SZ Size Estimation Response Options Skips Comments Core or Secondary*
N/A A QPSIZEST
Service multiplier method Size estimation method CHECK ALL THAT APPLY SERVICE MULTIPLIER UNIQE EVENT MULTIPLIER UNIQUE OBJECT MULTIPLIER 1 2 3 C
N/A
SZ1MSG
[Interviewer should read]: The next question is about a service or event that may or may not be familiar to you.
ONLY ASK IF QPSIZEST = ‘2’ OR ‘3’
156
5. SIZE ESTIMATION
N/A
Variable ID: SZ SZSMLT
Size Estimation
Response Options
Skips
Comments
Core or Secondary* C
1.
Did you go to or attend [insert facility/location/event] for [insert service] during [insert time period]?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
ONLY ASK IF QPSIZEST = ‘A’ OR ‘B’
N/A N/A SZ2MSG
Unique object multiplier method The next set questions are asked of survey participants who are being recruited into survey, and unique object has been distributed to members of the key (source) population from which the sample is to be drawn prior to the start of the recruitment (with the intention of using the information as a multiplier to calculate the size of the population). ONLY ASK IF QPSIZEST = ‘C’ IF QPSIZEST=’A’ OR ‘B’ SKIP TO END OF CURRENT SECTION ONLY IF QPSIZEST = ‘4’
2.
SZKEY
Before we started this survey we passed out some [insert distributed object] in the community. This was done by [insert outreach workers/healthcare worker/community member] and they may have told you to keep it and not to give anyone. Did you get a [insert distributed object]?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 3
IF ‘2’,’7’,’8’→SKIP TO END OF CURRENT SECTION.
ONLY IF QPSIZEST = ‘4’
C
3.
SZSHOW
Can you show it to me?
CANNOT SHOW SHOWS THE WRONG OBJECT
Interviewer should verify and key in response.
SHOWS THE RIGHT OBJECT
‘2’→SKIP TO END OF CURRENT SECTION ‘3’ → SKIP TO SZ4MSG
ONLY IF QPSIZEST = ‘4’
C
SECTION 2: QUESTIONNAIRE MODULES
1 5 1 757
N/A
Variable ID: SZ SZ3MSG
Size Estimation
Response Options
Skips
Comments
Core or Secondary*
N/A
Show the participant a number of unique objects including the ones distributed and then ask:
ONLY IF QPSIZEST = ‘4’
4.
SZSEEN
We have distributed some of these objects. Can you identify the one you may have received?
IDENTIFIED THE CORRECT OBJECT COULD NOT IDENTIFY CORRECTLY
1 2
‘2’ → SKIP TO END OF CURRENT SECTION
ONLY IF QPSIZEST = ‘4’
S
N/A
SZ4MSG
Probe participant for the following questions:
ONLY IF QPSIZEST = ‘4’
5.
SZNUM
How many [insert distributed object] did you receive?
#[ | ] MIN: 1 DON’T KNOW REFUSE TO ANSWER 7 8 1 2 3 7
‘0’ → SKIP TO END OF CURRENT SECTION.
ONLY IF QPSIZEST = ‘4’
C
6.
SZREC
When did you receive the [insert distributed object]?
BEFORE DISTRIBUTION DURING DISTRIBUTION AFTER DISTRIBUTION DON’T KNOW
ONLY IF QPSIZEST = ‘4’
S
7.
SZBUY
Did you buy or pay for the object?
YES NO DON’T KNOW
1 2 7
ONLY IF QPSIZEST = ‘4’
S
*C=CORE, S=SECONDARY for all KP groups.
158
5. SIZE ESTIMATION
Section
II
6. ACASI TUTORIAL
(Audio) Computer-Assisted Self-administered Interviews (ACASI) provide more privacy than face-to face-interviews. In doing so, ACASI provides participants more comfort when answering questions about sensitive issues or behaviors and reduces social desirability bias. Furthermore, the ACASI interview mode minimizes data entry by interview staff. The purpose of this tutorial is to teach the participant how to use ACASI on a computer, tablet, or smart phone. Most or all questions in this tutorial are sample questions that do not collect hard data for analysis.
This tutorial should precede the actual interview. It is recommended that a survey staff member is present during the tutorial in order to answer any questions and to determine whether the respondent is capable of conducting an ACASI. After demonstrating that the respondent is capable of conducting the ACASI, the survey staff should leave the respondent to him/herself but remain within earshot to answer any further questions. If the tutorial suggests that the respondent may not be competent enough to conduct an ACASI, staff may opt to administer the questionnaire as a computer-assisted personal interview (CAPI, i.e. face-to-face, electronic). N/A Variable ID: TU TUT1MSGA
ACASI Tutorial
Response Options
Skips
Comments
Core or Secondary* C
N/A
Welcome. First we will tell you how to use this computer for the interview. You can listen to and read the questions on the screen and then give your best answer. Use the “NEXT” button to go to the next question. Use the “BACK” button to return to the previous question if you need to.
Survey staff to observe and assist respondent.
N/A
TUTWELC
[Interviewer should read]: Thank you for your answers so far. We would like you to finish the rest of the survey on your own if possible. The following section will teach you the basics of using this device. I will remain here to help you and answer any questions you may have. Please give me the tablet after a few question when you are instructed to do so.
C
SECTION 2: QUESTIONNAIRE MODULES
1 5 1 959
N/A
Variable ID: TU TUT1MSG
ACASI Tutorial
Response Options
Skips
Comments
Core or Secondary* C
N/A
Welcome to the device tutorial, where we will tell you how to use this device for the interview. You can read the questions on the screen (or listen if this is available) and then give your best answer. Use the “right arrow” button to go to the next question. Use the “left arrow” button to return to the previous question if you need to. You may also swipe left or right to switch between questions.
Survey staff to observe and assist respondent.
1.
TUTYES
Now we will explain to you the different types of questions and answers. Some questions you answer by clicking either “Yes” or “No”. Try it and click on “Yes”.
YES NO
1 2
IF ‘2’, PRESENT ERROR MESSAGE AND ASK TO REENTER CORRECT VALUE. IF ANSWER NOT ‘THURSDAY’, PRESENT ERROR MESSAGE AND ASK TO REENTER CORRECT VALUE.
C
2.
TUTBEST
For some questions you choose the best or correct answer. For example, what is the day after WEDNESDAY?
MONDAY TUESDAY WEDNESDAY THURSDAY FRIDAY SATURDAY SUNDAY
1 2 3 4 5 6 7 Y N 2 2 2 2 2
C
3.
TUTMLT
For some questions you can choose more than one answer. For example, here is a list of things people do. Select the first three activities and then go on to the next question.
C
LISTEN TO MUSIC REPAIR THINGS READ BOOKS WATCH TV GO ON A HOLIDAY
1 1 1 1 1
160
6. ACASI TUTORIAL
N/A
Variable ID: TU TUTNUM
ACASI Tutorial
Response Options
Skips
Comments
Core or Secondary* C
4.
For some questions you answer by entering a number, for example when we ask how old one is or how many children one has. Let’s try that: Type the number “18”.
<NUM>
IF ANSWER IS NOT ‘18’, ASK TO REENTER CORRECT VALUE.
N/A
TUT2MSG
Please give the tablet to the staff.
C
5.
TUTPASS
[Instruction to staff]: The respondent passes the tutorial if he/ she correctly answers each question. Did the respondent pass the tutorial?
YES NO
1 2
‘1’ → PROCEED TO TUT3MSG ‘2’ → SKIP TO TUT4MSG
C
N/A
TUT3MSG
[Instruction to staff]: Since the respondent passed the tutorial, please give the tablet back to him or her so that he or she may continue answering the questions individually.
C
N/A
TUT4MSG
[Instruction to staff]: Since the respondent did not pass the tutorial, please continue the interview by asking the questions aloud and recording the participant’s responses.
C
N/A
TUT5MSG
Thank you. Remember that you can refuse to answer questions if you want to. For some questions if you do not know the exact answer, try to estimate it as best as you can. This completes the exercise. Please continue with the survey.
C
*C=CORE, S=SECONDARY for all KP groups.
SECTION 2: QUESTIONNAIRE MODULES
1 6 161
Section
Il N/A
7. DEMOGRAPHICS
This module is intended to capture key demographic information about participants. The questions presented here broaden understanding of social and economic determinants found to impact risk and vulnerability of key populations.
Variable ID: DE DESEXNOW
Demographics
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P C
O C
1.
Do you consider yourself as male, female transgender, or other?
MALE FEMALE TRANSGENDER OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
BACKGROUND
C
2.
DESEXBTH
What was your sex at birth?
MALE FEMALE OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
BACKGROUND
C
C
C
C
C
162
7. DEMOGRAPHICS
N/A
Variable ID: DE DEIDENT
Demographics
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P S
O S
3.
Would you describe yourself as:
GAY/ HOMOSEXUAL/ LESBIAN BISEXUAL STRAIGHT/ HETEROSEXUAL OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 6 7 8
BACKGROUND ADAPT TERMS TO LOCAL CONTEXT.
S
4.
DEATTRA
What sex would you say you are most sexually attracted to? ONLY OR MOSTLY MALE EQUALLY MALE AND FEMALE ONLY OR MOSTLY FEMALE NONE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
BACKGROUND
S
C
C
S
S
5.
DEAGENUM
How old were you at your last birthday?
AGE IN COMPLETED YEARS: [ | ] MIN: ELIGIBILITY CRITERIA DON’T KNOW 97 REFUSE TO ANSWER 98
Answer provided must match answer for “ELAGEL”. Provide assistance if necessary.
BACKGROUND
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
1 6 1 363
N/A
Variable ID: DE DEBORN
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
6.
In what country/nation were you born?
COUNTRY A COUNTRY B COUNTRY C OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 6 7 8 1 2 3 7 8 1 2 7 8 1 2 ‘2’,’7’ OR ‘8’→SKIP TO DEMARSTA
BACKGROUND Use locally defined responses. More options can be included.
S
7.
DEREADWR
Can you read and write?
CANNOT READ OR WRITE CAN READ ONLY CAN READ AND WRITE DON’T KNOW REFUSE TO ANSWER
BACKGROUND
S
S
S
S
S
8.
DEEDEVER
Have you ever attended school?
YES NO DON’T KNOW REFUSE TO ANSWER
BACKGROUND Use locally defined responses. More options can be included.
C
C
C
C
C
9.
DEEDHIGH
What is the highest level/grade/year of school you attended?
PRIMARY SECONDARY HIGHER (UNIVERSITY OR OTHER POST-SECONDARY SCHOOL) DON’T KNOW REFUSE TO ANSWER
BACKGROUND Use responses based on local context.
C
C
C
C
C
3 7 8
164
7. DEMOGRAPHICS
N/A
Variable ID: DE DEEMPL
Demographics
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P C
O C
10.
What best describes your current employment status?
EMPLOYED FULL-TIME EMPLOYED PART-TIME FULL-TIME STUDENT RETIRED UNEMPLOYED OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
BACKGROUND
C
11.
DEMARSTA
What is your current marital status?
SINGLE, NEVER MARRIED MARRIED SEPARATED/ DIVORCED WIDOWED DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
BACKGROUND
C
C
C
C
C
12.
DELIVESX
Are you currently living with a sexual partner?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
BACKGROUND
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
1 6 1 565
N/A
Variable ID: DE DEPOLYGM
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
13.
Do you have more than one wife or live with other women as if married?
YES NO DON’T KNOW REFUSES TO ANSWER
1 2 7 8
MEN ONLY. ‘2’, BACKGROUND ‘7’ OR ‘8’→ SKIP Appropriate in setting TO DEREG. where polygamy is practiced.
S
14.
DEWIVESM
How many wives or live-in partners do you have?
# WIVES [ | ] MIN: 2 DON’T KNOW 97 REFUSE TO ANSWER 98
MEN ONLY.
BACKGROUND Appropriate in setting where polygamy is practiced.
S
S
S
S
S
15.
DEPOLYGW
Does your husband have more than one wife or live with other women as if married?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
WOMEN ONLY. ‘2’, ‘7’ OR ‘8’→ SKIP TO DEREG
BACKGROUND Appropriate in setting where polygamy is practiced.
S
S
S
S
S
16.
DEWIVESW
How many wives or live-in partners does your husband have, including you?
# WIVES [ | ] MIN: 2 DON’T KNOW 97 REFUSE TO ANSWER 98
WOMEN ONLY.
BACKGROUND Appropriate in setting where polygamy is practiced.
S
S
S
S
S
166
7. DEMOGRAPHICS
N/A
Variable ID: DE DEREG
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
17.
Do you have a regular place to sleep at night?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
BACKGROUND Adapt term ‘regular place’ to local context.
S
18.
DEHOMEL
Are you currently homeless?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 3 7 8
BACKGROUND
S
S
S
S
S
19.
DECURLIV
In which [select community/town/neighborhood/village] AREA A do you currently live? AREA B AREA C DON’T KNOW REFUSE TO ANSWER
BACKGROUND & SAMPLING Recommended for RDS/TLS Adapt to local context. More options can be included. BACKGROUND
S
S
S
S
S
20.
DERESIDE
How long have you lived here in [insert name of community/town/neighborhood/village from DECURLIV]?
# YEARS: [ | ] MIN: 0 MAX: 96
S
S
S
S
S
TYPE ‘0’ IF LESS THAN 1 YEAR.
DON’T KNOW 97 REFUSE TO ANSWER 98
SECTION 2: QUESTIONNAIRE MODULES
1 6 1 767
N/A
Variable ID: DE DEAWAYT
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
21.
In the last 6 months, how many times have you been away from home for one or more nights? By “home”, we mean the place you usually live. By “away from home”, we mean spending one or more nights away from the place you usually live. TYPE ‘0’ IF NONE.
# TIMES [ | | ] MIN: 0 MAX: 180 DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’, ‘997’ OR ‘998’→ SKIP TO DEEMPSRC
BACKGROUND Can be used to determine mobility.
S
22.
DEAWAYM
In the last 6 months, have you been away from home for more than one month at a time?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
BACKGROUND Can be used to determine mobility.
S
S
S
S
S
23.
DEEMPSRC
What is your main source of income?
NO INCOME OPTION A OPTION B OPTION C OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 6 7 8
BACKGROUND Use locally defined responses.
S
S
S
S
S
168
7. DEMOGRAPHICS
N/A
Variable ID: DE DEINCOME
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
24.
What is your monthly income?
AMOUNT _______ DON’T KNOW 9997 REFUSE TO ANSWER 9998
BACKGROUND Use local currency. MIN and MAX values of numeric field should be set as per local currency value.
S
25.
DECHILIV
How many children do you have currently living with you # CHILDREN [ | ] that you are responsible for (including those you may MIN: 0 have adopted or care for)? TYPE ‘0’ IF NONE. MAX: 96 DON’T KNOW 97 REFUSE TO ANSWER 98
BACKGROUND
S
S
S
S
S
26.
DELANG
What is your primary spoken language?
LANGUAGE A LANGUAGE B LANGUAGE C OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 6 7 8
BACKGROUND Use locally defined responses.
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 6 1 969
N/A
Variable ID: DE DEETHNIC
Demographics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
27.
To which [ethnic group/race/tribe] do you belong?
ETHNICITY/RACE A ETHNICITY/RACE B ETHNICITY/RACE C MIXED ETHNICITY/RACE OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 6 7 8 1 2 3 6 7 8
BACKGROUND Use locally defined responses.
S
28.
DERELIG
What religion are you?
RELIGION A RELIGION B RELIGION C OTHER DON’T KNOW REFUSE TO ANSWER
BACKGROUND Use locally defined responses.
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
170
7. DEMOGRAPHICS
Section
II N/A
8. TRANSGENDER CHARACTERISTICS
This module should only be asked of people identifying as transgender individuals in the Demographics module. Some surveys do not exclusively recruit people identifying as transgender. In surveys where transgender individuals are included in surveys of men who have sex with men, questions from this module should be included.
Variable ID: TG TGLIVE
Transgender Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T N/A
P N/A
O N/A
1.
In the last 6 months, have you lived as a woman? By living as a woman, I mean dressing and presenting yourself as a woman.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO END OF CURRENT SECTION
BACKGROUND
N/A
N/A
TG1MSG
Thank you. Now we are going to ask you some questions about your use of hormones and any gender enhancement/transition procedures you may have had.
SECTION 2: QUESTIONNAIRE MODULES
1 7 171
N/A
Variable ID: TG TGHORM
Transgender Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T C
P N/A
O N/A
2.
Which type of hormones have you used?
NONE INJECTION
A B C D E X Y Z 1 2 3 4 7 8
CHECK ALL THAT APPLY.
GEL PILLS PATCH OTHER DON’T KNOW REFUSE TO ANSWER
IF ‘A, ‘Y’ OR ‘Z’ SELECTED, NO OTHER CATEGORY CAN BE CHOSEN AS WELL. ‘A’, ‘X’, ‘Y’, ‘Z’ → SKIP TO TGENHANC ‘C’, ‘D’, ‘E’ → TGPROF
BACKGROUND & RISK
N/A
3.
TGINJ
In the last 6 months, how often did you get hormone injections?
NEVER AT LEAST ONCE A DAY ONCE A WEEK ONCE A MONTH DON’T KNOW REFUSE TO ANSWER
ONLY ASK IF TGHORM=’B’
RISK
N/A
S
C
N/A
N/A
4.
TGSHARE
In the last 6 months, with how many people # PEOPLE [ | ] did you use a needle after someone else had MIN: 0 injected hormones with it? MAX: 96 CODE ‘00’ IF NONE DON’T KNOW 97 REFUSE TO ANSWER 98
ONLY ASK IF TGHORM=’B’
RISK
N/A
S
C
N/A
N/A
172
8. TRANSGENDER CHARACTERISTICS
N/A
Variable ID: TG TGPROF
Transgender Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T C
P N/A
O N/A
5.
Does a health care provider know about or supervise your use of hormones?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
ONLY ASK IF TGHORM=’B’, ‘C’, ‘D’, OR ‘E’
COVERAGE
N/A
6.
TGENHANC Other than hormones, have you had a gender enhancement/or transition procedure?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,‘8’’→ SKIP TO BACKGROUND TGAWARE
N/A
S
C
N/A
N/A
7.
TGPROC
What type of procedures did you have? CHECK ALL THAT APPLY.
COLLAGEN INDUSTRIAL SILICONE GEL SEX OR GENDER REASSIGNMENT SURGERY IMPLANTS OTHER DON’T KNOW REFUSE TO ANSWER
A B C D X Y Z
BACKGROUND
N/A
S
C
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
1 7 1 373
N/A
Variable ID: TG TGBODY
Transgender Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P N/A
O N/A
8.
What parts of your body did you have procedures on? CHECK ALL THAT APPLY.
GENITALS NOSE BUTTOCKS THIGHS BREASTS FOREHEAD CHEEKBONES LIPS CHIN HIPS DON’T KNOW REFUSE TO ANSWER
A B C D E F G H I J Y Z 1 2 7 8 ONLY ASK IF DESEXBTH=’1’ AND DESEXNOW = ’2’ ‘2’,’7’,’8’→ TGAWARE
BACKGROUND
N/A
9.
TGVAGINA
Do you have a surgically constructed vagina? YES NO DON’T KNOW REFUSE TO ANSWER
BACKGROUND
N/A
S
C
N/A
N/A
174
8. TRANSGENDER CHARACTERISTICS
N/A
Variable ID: TG TGESTAB
Transgender Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P N/A
O N/A
10.
Where was the surgical procedure carried out?
PRIVATE HEALTH ESTABLISHMENT PUBLIC HEALTH ESTABLISHMENT PRIVATE HOME OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 6 7 8 1 2 7 8 1 2 7 8 ‘1’,’7’ OR ‘8’→ SKIP TO END OF CURRENT SECTION
RISK
N/A
11.
TGAWARE
Think about the last time you had sex with a male partner. At that time was this partner aware you were transgender?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
S
S
N/A
N/A
12.
TGTELL
Did you tell him that you were transgender?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
S
S
N/A
N/A
*: S= Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: Silva-Santisteban, A; Raymond, HF; Salazar, X; et al. Understanding the HIV/AIDS Epidemic in Transgender Women of Lima, Peru: Results from a Sero- Epidemiologic Study Using Respondent Driven Sampling. AIDS and Behavior. 2012, 16:872–881.
SECTION 2: QUESTIONNAIRE MODULES
1 7 1 575
Section
Il
9. SEX WORK CHARACTERISTICS
Engaging in sex work—the exchange of sex for money— places individuals at high risk of HIV. This module asks participants about their history of sex work, current engagement in sex work, and about vulnerability related to sex work. The module is essential for all surveys targeting sex workers. Investigators are encouraged to include questions from this module in other surveys to obtain information on sex workers who belong to other target groups as well. For survey populations that are not sex workers, this module should be placed after it has been determined that the participant has sold sex for money in the Recall Sexual Behavior module. It is appropriate to decrease the recall period for sex workers given the higher expected number of partners he/she may have had creating difficulties in recalling events over a longer time period. This recall period should be modified depending on the frequency of selling sex in the sex worker population. N/A Variable ID: SW SW1MSG
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A
Now we are going to ask you some questions about getting money to have sex.
1.
SWAGE
How old were you when you first had sex with someone for money?
YEARS: | | | MIN: 1 MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY
C
S
S
S
S
176
9. SEX WORK CHARACTERISTICS
N/A
Variable ID: SW
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
2.
SWREASON What is the main reason you started selling sex for money?
MONEY FOR SELF OR FAMILY WAS FORCED/PRESSURED/COERCED LIKE TO/PLEASURE/SELF-ESTEEM FRIENDS/FAMILY DOING IT MARITAL ISSUES OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK, VULNERABILITY & PREVENTION
C
3.
SWDURAT
For how many years have you been selling sex for money? TYPE ‘0’ IF LESS THAN 1 YEAR.
YEARS: | | | MAX: CURRENT AGE - XX DON’T KNOW 97 REFUSE TO ANSWER 98
RISK Value of “XX” depends on eligibility criteria.
C
S
S
S
S
4.
SWDURAT2
Just to be sure, could you please reenter the number of years you have been selling sex for money? TYPE ‘0’ IF LESS THAN 1 YEAR.
YEARS: | | | MAX: CURRENT AGE - XX DON’T KNOW 97 REFUSE TO ANSWER 98
RISK Value of “XX” depends on eligibility criteria.
C
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 7 1 777
N/A
Variable ID: SW SWLOC
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
5.
In the last 6 months, in what part of town or neighborhood have you exchanged sex? CHECK ALL THAT APPLY
NEIGHBORHOOD A NEIGHBORHOOD B NEIGHBORHOOD C OTHER DON’T KNOW REFUSE TO ANSWER
A B C X Y Z
PREVENTION/ DESIGNRDS Adapt question and responses to local context.
C
6.
SWVENUE
Where do you usually meet or find clients?
BROTHEL, GUESTHOUSE, MASSAGE PARLOR HOTEL, CLUB, BAR, RESTAURANT STREET, PARK, OTHER PUBLIC PLACES INTERNET, PHONE BROKER, ESCORT AGENCY OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK & PREVENTION Adapt responses to local context.
C
S
S
S
S
178
9. SEX WORK CHARACTERISTICS
N/A
Variable ID: SW SWPLACES
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
7.
Where do you usually have sex with clients?
BAR, PUB, CLUB LODGE, HOTEL BROTHEL, MASSAGE PARLOR SEX WORKER’S HOME CLIENT’S OR OTHER HOME OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
PREVENTION Adapt responses to local context.
S
8.
SWAGENT
Do you have someone who helps you meet clients?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
C
S
S
S
S
9.
SWDECIDES Who usually decides where to have sex?
YOU CLIENT BOTH PIMP OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 6 7 8
VULNERABILITY Adapt term “pimp” to local context.
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 7 1 979
N/A
Variable ID: SW
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
10.
SWINCOME Is selling sex your main source of income?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 A B C D X Y Z IF ‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
RISK & PREVENTION
C
11.
SWEMPL
Apart from sex work, what other work do you do to earn income? CHECK ALL THAT APPLY
NOTHING OPTION A OPTION B OPTION C OTHER DON’T KNOW REFUSE TO ANSWER
BACKGROUND Adapt responses to local context.
S
S
S
S
S
12.
SWEARNA
What is the approximate amount you earn per sex act?
VALUE: [ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK Use local currency.
C
S
S
S
S
13.
SWEARNW
In the last week, how much money did you earn from selling sex?
VALUE: [ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK Use local currency and adjust time period as needed. 1 2 7 8 VULNERABILITY
C
S
S
S
S
14.
SWNOPAY
In the last 6 months, did a client ever refuse to pay you?
YES NO DON’T KNOW REFUSE TO ANSWER
S
S
S
S
S
180
9. SEX WORK CHARACTERISTICS
N/A
Variable ID: SW SWPAYLES
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
15.
In the last 6 months, did a client ever pay you less than agreed?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 3 4 7 8
VULNERABILITY
S
16.
SWNOCON
In the last 6 months, how often were you able to negotiate condom use, when a client told you he did not want to use a condom?
NEVER RARELY SOMETIMES FREQUENTLY DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
S
17.
SWAVOID
In your experience, what do clients do to avoid using a condom?
THEY SIMPLY REFUSE OFFER MORE MONEY THREATEN ME BEAT ME IF I REFUSE OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 6 7 8
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 8 181
N/A
Variable ID: SW SWCUSE
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
18.
What do you do to make clients use condoms? CHECK ALL THAT APPLY.
NOTHING TALK TO HIM REFUSE TO HAVE SEX UNLESS HE USES A CONDOM ASK FOR MORE MONEY OTHER DON’T KNOW REFUSE TO ANSWER
A B C D X Y Z A B C D E X Y Z 1 2 3 7 8
IF ‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY
S
19.
SWINSI
Are there reasons why you do not insist on using a condom with a client? CHECK ALL THAT APPLY.
I ALWAYS USE A CONDOM I CAN GET MORE MONEY SEX DOESN’T TAKE AS LONG WITHOUT A CONDOM I DON’T INSIST WHEN I AM AFRAID ASK FOR MORE MONEY OTHER DON’T KNOW REFUSE TO ANSWER
IF ‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY
S
S
S
S
S
20.
SWCARRY
In the past six months, have you stopped carrying condoms with you because you feared the police would identify you as a sex worker?
YES NO I DO NOT USUALLY CARRY CONDOMS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
S
182
9. SEX WORK CHARACTERISTICS
N/A
Variable ID: SW SWABUSE
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
21.
In the last 6 months, has a client abused or threatened you?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8 A B C D Y Z 1 2 3 4 7 8 ‘A’, ‘Y’, OR ’Z’ → SKIP TO END OF CURRENT SECTION
VULNERABILITY
C
22.
SWFORCE
In the last 6 months, has a client forced you to have sex?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
C
S
S
S
S
23.
SWMOB
In the last 12 months, where else outside this town/city have you sold sex? CHECK ALL THAT APPLY.
NOWHERE ELSE OUTSIDE THE TOWN/CITY BUT INSIDE PROVINCE/DISTRICT OUTSIDE THE PROVINCE/DISTRICT BUT INSIDE COUNTRY OUTSIDE COUNTRY DON’T KNOW REFUSE TO ANSWER
PREVENTION
C
S
S
S
S
24.
SWOUT
In the last 12 months, how often did you travel outside of this town/city to sell sex?
DAILY WEEKLY MONTHLY A FEW TIMES A YEAR DON’T KNOW REFUSE TO ANSWER
PREVENTION
C
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
1 8 1 383
N/A
Variable ID: SW SWCLNR
Sex Work Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
25.
In the last week, how many different men paid you money for sex that were regular clients? A regular client is someone you have had sex with on more than one occasion.
[ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
‘0’, ‘9997’, OR RISK ‘9998’ à SKIP TO SWCLNO
C
26.
SWCLNRC
The last time you had sex with a regular client, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
S
S
S
S
27.
SWCLNO
In the last week, how many different men paid you money for sex that were one-time clients? A one-time client is someone you have had sex with only once.
[ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
‘0’, ‘9997’, OR ‘9998’ à SKIP TO END OF CURRENT SECTION.
RISK
C
S
S
S
S
28.
SWCLNOC
The last time you had sex with a one-time client, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
184
9. SEX WORK CHARACTERISTICS
Section
II
10. CLIENTS OF SEX WORKERS CHARACTERISTICS
Clients of sex workers are at high risk for HIV infection in most settings. A sex worker is someone who receives money for sex. While clients of sex workers often are not considered a key population, their HIV risk is defined by having sex with sex workers. As clients can be difficult to reach, surveys sometimes focus on occupationally-defined groups (e.g. transport workers or migrants) when formative research suggests that a high proportion of them have sex with sex workers. Other modules in these guidelines probe such occupational characteristics, whereas this module probes client-specific sex behaviors. Any population group that is surveyed as a proxy for clients of sex workers should be asked the questions from this module.
Variable ID: CL N/A CL1MSG
Clients Of Sex Workers Characteristics
Response Options
Skips
Comments S
Core or Secondary* M T P O
Thank you for your answers so far. Next, a few questions about paying anyone for sex with money.
Only for anyone who paid for sex
1.
CLAGE
How old were you when you first paid for sex? By sex, I mean vaginal and anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus (butt).
# YEARS [ | ] MIN: (CURRENT AGE) - (AGE AT FIRST SEX) MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY ‘Current age’ from DEMOGRAPHICS
S
S
S
S
C
SECTION 2: QUESTIONNAIRE MODULES
1 8 1 585
Variable ID: CL 2. CLMEET
Clients Of Sex Workers Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O C
Where is the main place you find sex workers, either male, female, or transgender? A sex worker is someone who receives money for sex.
BAR, PUB, CLUB LODGE, HOTEL BROTHEL, MASSAGE PARLOR STREET, PUBLIC OR OPEN SPACE BY PHONE, SMS, INTERNET OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK & PREVENTION
S
3.
CLSEX
Where do you have sex with sex workers most often?
BAR, PUB, CLUB LODGE, HOTEL BROTHEL, MASSAGE PARLOR SEX WORKER’S HOME MY OR OTHER HOME OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK & PREVENTION
S
S
S
S
S
4.
CLSREG
Do you usually go to the same sex worker for sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
S
S
S
S
S
186
10. CLIENTS OF SEX WORKERS CHARACTERISTICS
Variable ID: CL 5. CLSRAPE
Clients Of Sex Workers Characteristics
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
Have you ever used or threatened to use force to have sex with a sex worker?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK, VULNERABILITY
S
6.
CLHIVQ
The last time you paid for sex, did you ask the sex worker whether she or he has HIV?
YES NO, I DIDN’T ASK NO, I ALREADY KNEW STATUS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
RISK
S
S
S
S
S
7.
CLTEST
There are home HIV test kits available. Have you ever tested a sex worker for HIV before having sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
PREVENTION & VULNERABILITY In settings where home- based, selftesting is available
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, C=Client/Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
1 8 1 787
Section
Il
11. INCARCERATION CHARACTERISTICS
This module captures information specific to the incarcerated population. Here, “detained” describes temporary, often pre-trial custody or police holding cells and “imprisoned” refers to being convicted of a crime and confined in a prison as punishment. “Incarceration” describes either type of confinement, detention or imprisonment. Formative assessment is crucial to understanding important factors including but not limited to the structural (i.e. how many prisoners share facilities, cells) and organizational (i.e. how are prisoners segregated—by sentence, crime, sex) environment, whether conjugal visits are allowed, and access to health care. Other modules, such as those on violence, drug use, alcohol use, and prevention services should be reviewed and modified for prisoners (male or female). This module focuses on risk during incarceration, thus consideration should be given to whether a minimum amount of time incarcerated (e.g. 6 months or 1 year) should be used as part of eligibility criteria to participate in biobehavioral surveys. Here, a recall period of six months was used, but the recall period should be based on the local context and should not exceed the time spent in prison to capture information specific to risk during incarceration. There are two sections to this module. The first part addresses historical incarceration, whereas the second part contains questions about current incarceration. Reference periods, response categories and topics used in this module were based on those suggested by the United Nations Office on Drugs and Crime (UNODC) tool kit on HIV and prisons (2010). N/A Variable ID: IN N/A IN1MSG INHISTOR
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A N/A 1.
HISTORICAL INCARCERATION The next questions are about any incarceration before this survey. Have you ever been detained or imprisoned before this time? YES, DETAINED YES, IMPRISONED NO DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 ‘3’,’7’,‘8’ →SKIP BACKGROUND TO IN2MSG S S S S C
188
11. INCARCERATION CHARACTERISTICS
N/A
Variable ID: IN INTIMES
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O C
2.
Before this time, how many times have you been incarcerated?
ONCE TWO OR MORE TIMES DON’T KNOW REFUSE TO ANSWER
1 2 7 8 IF NOT 96→SKIP TO IN2MSG
BACKGROUND
S
3.
INLASTM
The last time you were incarcerated, for how long were you detained or imprisoned?
[locally defined time period here] # MONTHS | | | CODE ‘00’ IF LESS THAN ONE MONTH MORE THAN 12 MONTH 96 DON’T KNOW 97 REFUSE TO ANSWER 98
BACKGROUND Time period should be locally defined.
S
S
S
S
S
4.
INLASTY
If more than 12 months, how many years?
# YEARS | | | DON’T KNOW 97 REFUSE TO ANSWER 98
BACKGROUND
S
S
S
S
S
N/A N/A IN2MSG
CURRENT INCARCERATION The next questions are about this incarceration.
5.
INSTATUS
Are you currently…
SENTENCED/CONVICTED REMANDED/AWAITING TRIAL OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
BACKGROUND
S
S
S
S
C
SECTION 2: QUESTIONNAIRE MODULES
1 8 1 989
N/A
Variable ID: IN INREASON
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
6.
For what crime are you charged?
[insert locally defined responses] DON’T KNOW REFUSE TO ANSWER 7 8
RISK & VULNERABILITY
S
7.
INSENTNY
For how many years is your detention/remand or sentence this time?
# YEARS | | | CODE ‘00’ IF LESS THAN ONE YEAR DON’T KNOW 97 REFUSE TO ANSWER 98
‘00’→SKIP TO INSENTNM
BACKGROUND & RISK
S
S
S
S
S
8.
INTIMYR
For how many years have you been incarcerated this time?
# YEARS | | | CODE ‘00’ IF LESS THAN ONE YEAR DON’T KNOW 97 REFUSE TO ANSWER 98
BACKGROUND & RISK
S
S
S
S
S
9.
INSENTNM
For how many months is your detention/remand or sentence this time?
[insert locally defined time period] # MONTHS | | | CODE ‘00’ IF LESS THAN ONE MONTH DON’T KNOW 97 REFUSE TO ANSWER 98
ONLY ASK IF BACKGROUND & INSENTNY = 00 RISK Time period should be locally defined.
S
S
S
S
S
190
11. INCARCERATION CHARACTERISTICS
N/A
Variable ID: IN INTIMMN
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
10.
For how many months have you been incarcerated this time?
[insert locally defined time period] # MONTHS | | | CODE ‘00’ IF LESS THAN ONE MONTH MORE THAN 12 MONTHS 96 DON’T KNOW 97 REFUSE TO ANSWER 98
ONLY ASK IF BACKGROUND & INSENTNY = 00 RISK Time period should be locally defined.
S
N/A
IN3MSG
The next few questions are about sex since you have been incarcerated this time.
11.
INSHARSP
With how many other people do you share sleeping quarters (i.e. cell)?
# PERSONS | | | DON’T KNOW 97 REFUSE TO ANSWER 98
RISK
S
S
S
S
C
12.
INSXVAGN
Have you had vaginal sex or anal sex with women since being detained/incarcerated this time?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’,‘8’ →SKIP RISK TO INSXANAL
S
S
S
S
C
13.
INSXFEML
With how many female partners have you had vaginal or anal sex during the past six months since you were detained/incarcerated this time?
# SEXUAL PARTNERS | | | DON’T KNOW 997 REFUSE TO ANSWER 998
RISK
S
S
S
S
C
SECTION 2: QUESTIONNAIRE MODULES
1 9 191
N/A
Variable ID: IN INSXANAL
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O C
14.
Have you had anal sex during the past six months since you were incarcerated this time?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’, ‘8’ →SKIP TO INSXCOND
RISK
S
15.
INSXMALE
With how many male partners have you had anal sex during the past six months since you were incarcerated this time?
# SEXUAL PARTNERS | | | DON’T KNOW 997 REFUSE TO ANSWER 998
RISK
S
S
S
S
C
16.
INSXRAPP
While incarcerated during the past six months, have you been forced or coerced to have sex by another prisoner?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8 1 2 7 8
‘2’,’7’,‘8’ →SKIP RISK & TO INSXEXCH VULNERABILITY
S
S
S
S
C
17.
INSXRAPS
While incarcerated in the past six months, have you been forced or coerced to have sex by prison staff/ authority figure?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK & VULNERABILITY
S
S
S
S
C
18.
INSXCSTF
While incarcerated during the past six months, have YES you had sex with prison staff when it was not forced? NO DON’T KNOW REFUSE TO ANSWER
RISK & VULNERABILITY
S
S
S
S
C
192
11. INCARCERATION CHARACTERISTICS
N/A
Variable ID: IN INSXFOBJ
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O C
19.
While incarcerated in the past six months, have you had an object inserted into your anus/vagina by another prisoner or staff against your will?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK & VULNERABILITY
S
20.
INSXEXCH
While incarcerated during the past six months, have you had sex with someone for money, goods, access, protection, or services?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK & VULNERABILITY
S
S
S
S
C
21.
INSXCOND
Are you able to obtain condoms in this prison?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
PREVENTION
S
S
S
S
C
N/A
IN4MSG
Thank you for your answers. The next few questions are about drug use since you have been incarcerated. Have you used drugs to get high while in prison this time? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 ‘2’,’7’ OR ‘8’ →SKIP TO INBLADE RISK This can be used if specific drug use is not of interest; otherwise, refer to the questions below S S S S C
22.
INEVRUSE
SECTION 2: QUESTIONNAIRE MODULES
1 9 1 393
N/A
Variable ID: IN ININJECT
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P C
O C
23.
Have you injected drugs while in prison this time?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO INBLADE
RISK Use if not interested in specific drugs; if specifics are needed see INIDUEVR
C
24.
INIDULOC
Did you share your injection equipment with others during the past six months since you were incarcerated this time?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
N/A
IN5MSG
Thank you. This section is about experiences in prison.
25.
INBLADE
In the last six months since you were incarcerated this time, have you ever shared blades to shave or cut hair?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
RISK
S
S
S
S
C
26.
INTATTOO
In the last six months while incarcerated, have you ever been tattooed?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
C
194
11. INCARCERATION CHARACTERISTICS
N/A
Variable ID: IN INPIERCE
Incarceration
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O C
27.
In the last six months while incarcerated, have you ever made a piercing or put in an earring?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
S
*: S= Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary References: United Nations Office on Drugs and Crime (UNODC). HIV and Prisons: Situation and Needs Assessment Tool Kit, 2010. http://www.unodc.org/documents/hiv- aids/publications/HIV_in_ prisons_situation_and_needs_assessment_document.pdf
SECTION 2: QUESTIONNAIRE MODULES
1 9 1 595
Section
Il
12. TRANSPORT WORKER CHARACTERISTICS
This module is intended to capture characteristics of transport workers such as truckers and their assistants. Assistants are defined as any person employed either through informal or formal means to assist the driver of a vehicle in the transport, loading, off-loading, maintenance or other support provided during the transportation of goods from one site to another. This excludes persons who are receiving a ride to any point or destination that are not there exclusively to assist the driver in the transport of his cargo (i.e., vendors, friends, hitchhikers, etc.). The definition of an assistant must be clear when determining eligibility. Formative research should inform local questions regarding specific routes or truck rest stops. Key concepts here are related to transport including the length of time spent away from home, multiple concurrent partnerships, condom use, and alcohol/drug use before sex. When conducting biobehavioral surveys among truckers, it is recommended to also include modules about alcohol, drug use, condom use, multiple concurrent partnerships, and STIs. N/A Variable ID: TR TR1MSG
Transport Worker Characteristics
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A
Thank you for your answers so far. This next section is about the truck route including distance and time away from home and family. Are you a [insert local name for transport worker] or [insert local name for assistant]? TRANSPORT WORKER ASSISTANT DON’T KNOW REFUSE TO ANSWER 1 2 7 8 ‘7’,‘8’ → SKIP TO END OF CURRENT SECTION. BACKGROUND Use locally defined term for transport worker and assistant and on occupation. Should be excluded if already included in eligibility. PREVENTION N/A N/A N/A N/A C N/A N/A N/A N/A C
1.
TROCCU
2.
TRORIG
On your last trip, which city was your origin?
<NUM>
196
12. TRANSPORT WORKER CHARACTERISTICS
N/A
Variable ID: TR TRDEST
Transport Worker Characteristics
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O C
3.
On your last trip, which city was your destination?
<NUM>
PREVENTION
N/A
4.
TRDISTND
On average, how many days does it take for you to travel between these two cities? TYPE ‘0’ IF LESS THAN A DAY.
# DAYS [ | ] DON’T KNOW 97 REFUSE TO ANSWER 98 LESS THAN 1 MONTH 1 OR MORE MONTHS BUT LESS THAN 12 MONTHS 1 OR MORE YEARS DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 ‘7’ OR ‘8’→SKIP TO TR2MSG ‘1’→ GO TO TRLONGD ‘2’ → SKIP TO TRLONGM ‘3’→ SKIP TO TRLONGY
PREVENTION
N/A
N/A
N/A
N/A
C
5.
TRLONG
For how long have you been a [insert local name for transport worker]/[insert local name for assistant]?
RISK
S
S
S
S
C
6.
TRLONGD
For how many DAYS have you been a [insert local name for transport worker]/[insert local name for assistant]?
# DAYS [ | ] MAX: 31 DON’T KNOW 97
RISK
S
S
S
S
C
7.
TRLONGM
For how many MONTHS have you been a [insert local name for transport worker]/[insert local name for assistant]?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
S
S
S
S
C
8.
TRLONGY
For how many YEARS have you been a [insert local name for transport worker]/[insert local name for assistant]?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97
RISK
S
S
S
S
C
SECTION 2: QUESTIONNAIRE MODULES
1 9 1 797
N/A
Variable ID: TR TR2MSG
Transport Worker Characteristics
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A 9.
This section is about the time away from home. # NIGHTS [ | ] MIN: 0 MAX: 30 DON’T KNOW 97 REFUSE TO ANSWER 98 ’0’,’ 97’ OR ‘98’→ SKIP TO TRTRIPS RISK N/A N/A N/A N/A C
TRWAYHMD In the past 30 days, how many nights have you spent away from home on work travel?
10.
TRWYNT
In the past 30 days, when you travel, where do you take your night rest? CHECK ALL THAT APPLY.
IN VEHICLE AT A LODGE/REST HOUSE OTHER DON’T KNOW REFUSE TO ANSWER
A B X Y Z
PREVENTION & RISK
N/A
N/A
N/A
N/A
C
11.
TRTRIPS
In the past six months, how many trips did you do?
# TRIPS [ | ] DON’T KNOW 97 REFUSE TO ANSWER 98
RISK Time period can change depending on context.
N/A
N/A
N/A
N/A
C
N/A 12.
TR3MSG TRAWAYSX
This section is about sex you may have had while on a trip. In the past 6 months, have you had sex with anyone on any of your trips? By sex, I mean vaginal and anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus (butt). YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 ‘2’,’7’ OR ‘8’→SKIP TO END OF CURRENT SECTION RISK N/A N/A N/A N/A C
198
12. TRANSPORT WORKER CHARACTERISTICS
N/A
Variable ID: TR TRSXWITH
Transport Worker Characteristics
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O C
13.
In the past 6 months, with whom have you had sex on any of your trips?
LIVE-IN PARTNER/SPOUSE GIRLFRIEND/BOYFRIEND SEX WORKER OTHER DON’T KNOW REFUSE TO ANSWER
A B C X Y Z A B X Y Z
RISK Adapt responses to local context.
N/A
CHECK ALL THAT APPLY.
14.
TRSXWR
In the past 6 months, where have you had sex during your trips?
INDOORS (HOTEL, HOME) OUTDOORS (BUS PARK, FOREST, STREET, TRUCK) OTHER LOCATION DON’T KNOW REFUSE TO ANSWER
PREVENTION
N/A
N/A
N/A
N/A
C
CHECK ALL THAT APPLY.
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: Atiola GO, Akpab OM, Komolafea IOO. HIV/AIDS and the long-distance truck drivers in south-west Nigeria: A cross-sectional survey on the knowledge, attitude, risk behaviour and beliefs of truckers. J Infect Pub Hlth 2010; 3:166-178. Agrawal KK, Agrawal L, Agrawal VK, Chaudhary V. Behavior surveillance survey amongst truck drivers regarding HIV/AIDS. J Behav Hlth 2012; 1(3):196-200. ASHA/FHI. Integrated Biological and Behavioral Surveillance Survey (IBBS) among Truckers in 22 Terai Highway Districts of Nepal, 2009. http://pdf.usaid.gov/pdf_docs/PNADU489.pdf
SECTION 2: QUESTIONNAIRE MODULES
1 9 1 999
Section
Il
13. REPRODUCTIVE HEALTH
The intent of this module is to capture information about family planning, pregnancies, spontaneous and induced abortions, and biological children. This module is intended only for females and is appropriate for sex workers (SW) and people who inject drugs (PWID).
N/A Variable ID: RH N/A RH1MSG
Reproductive Health
Response Options
Skips SKIP ENTIRE SECTION FOR MALES ‘2’,’7’,‘8’ → SKIP TO RHTRYPRG
Comments
Core or Secondary* S M T P O
Now some questions about family planning, pregnancies, and children. First, we would like to ask you about pregnancy. Have you ever been pregnant? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8
1.
RHEVRPRG
BACKGROUND
C
N/A
N/A
C
C
2.
RHPRENUM
How many times have you been pregnant?
[ | ] MIN: 1 DON’T KNOW 97 REFUSE TO ANSWER 98
BACKGROUND
S
N/A
N/A
S
S
200
13. REPRODUCTIVE HEALTH
N/A Variable ID: RH 3. RHPRGNOW
Reproductive Health When were you last pregnant?
Response Options I AM PREGNANT NOW WITHIN THE LAST 12 MONTHS BETWEEN 12 MONTHS AND 3 YEARS AGO LONGER THAN 3 YEARS AGO DON’T KNOW REFUSE TO ANSWER 1 2 3 4 7 8
Skips
Comments BACKGROUND
Core or Secondary* S S
M N/A
T N/A
P S
O S
4.
RHPRGMIS
Have you ever had a pregnancy that miscarried?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’,‘8’ → SKIP TO RHPRGABO
BACKGROUND
S
N/A
N/A
S
S
5.
RHPRGMIR
Were any of these miscarriages within the last 12 months?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
BACKGROUND
S
N/A
N/A
S
S
6.
RHPRGABO
Have you ever had a pregnancy that was aborted or terminated?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’,‘8’ → SKIP TO RHPRGSTL
BACKGROUND
S
N/A
N/A
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 0 2 101
N/A Variable ID: RH 7. RHPRGABR
Reproductive Health Were any of these abortions or terminations within the last 12 months? YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 7 8 1 2 7 8
Skips
Comments BACKGROUND
Core or Secondary* S S
M N/A
T N/A
P S
O S
DON’T KNOW REFUSE TO ANSWER 8. RHPRGSTL Have you ever had a pregnancy that ended in a stillbirth? YES NO DON’T KNOW REFUSE TO ANSWER 9. RHPRGSTR Were any of these stillbirths within the last 12 months? YES NO DON’T KNOW REFUSE TO ANSWER 10. RHEVBRTH Have you ever given birth? YES NO DON’T KNOW REFUSE TO ANSWER 11. RHNUMKID How many living sons and daughters do you have? # CHILDREN [ | ] CODE ‘00’ IF NONE DON’T KNOW 97 REFUSE TO ANSWER 98
‘2’,’7’,‘8’→SKIP TO BACKGROUND RHEVBRTH
S
N/A
N/A
S
S
BACKGROUND
S
N/A
N/A
S
S
‘2’,’7’,‘8’→ SKIP TO RHTRYPRG
BACKGROUND
S
N/A
N/A
S
S
BACKGROUND
S
N/A
N/A
S
S
202
13. REPRODUCTIVE HEALTH
N/A Variable ID: RH N/A RH2MSG
Reproductive Health
Response Options
Skips
Comments
Core or Secondary* S M T P O
The next few questions are about the last pregnancy that resulted in a birth. This does not include a current pregnancy. Think about the last time you were pregnant that resulted in a live birth. A live birth is when the baby shows signs of life, such as breathing, heart beating of the movement. Did you go to an antenatal care (ANC) clinic? YES NO DID NOT HAVE A PREGNANCY THAT RESULTED IN A LIVE BIRTH DON’T KNOW REFUSE TO ANSWER 1 2 ‘2’,’3’,’7’,‘8’→ SKIP COVERAGE TO RHHIVTST C N/A N/A C S
12.
RHANCUSE
3 7 8 1 2 7 8 COVERAGE C N/A N/A C S
13.
RHANCHIV
During the last time you were pregnant, were you offered an HIV test at any of your ANC visits?
YES NO DON’T KNOW REFUSE TO ANSWER
14.
RHHIVTST
During the last time you were pregnant, how many times NEVER did you test for HIV? ONE TIME TWO OR MORE TIMES DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
‘1’,’7’,‘8’ → SKIP TO RHANCTP
COVERAGE
C
N/A
N/A
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 0 2 303
N/A Variable ID: RH 15. RHHIVTRI
Reproductive Health At what months during your last pregnancy, did you test for HIV? CHECK ALL THAT APPLY.
Response Options 0-3 MTHS/FIRST TRIMESTER 4-6 MTHS/SECOND TRIMESTER 7-9 MTHS/THIRD TRIMESTER DON’T KNOW REFUSE TO ANSWER A B C Y Z 1 2 3 4 7 8
Skips
Comments COVERAGE & PREVENTION
Core or Secondary* S S
M N/A
T N/A
P S
O S
16.
RHHIVRES
During the last time you were pregnant, what was the result of the last HIV test?
POSITIVE NEGATIVE UNCLEAR / NEITHER POSITIVE OR NEGATIVE DID NOT RECEIVE RESULT DON’T KNOW REFUSE TO ANSWER
IF ’2’,‘3’,‘4’,‘7’,‘8’ → SKIP TO RHTPTEST
RESULT
C
N/A
N/A
C
S
17.
RHHIVARV
Some HIV-positive women who are pregnant take ARV medicine to reduce the risk of passing HIV on to their baby. Did you take ARVs before giving birth?
YES, TOOK ARVS BEFORE BIRTH YES, ALREADY ON ARVS NO DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 7 8 ’2’,’7’,‘8’ → SKIP TO RHTRYPRG
COVERAGE
C
N/A
N/A
C
S
18.
RHANCTP
During the last time you were pregnant, were you offered a syphilis test?
YES NO DON’T KNOW REFUSE TO ANSWER
COVERAGE
C
N/A
N/A
C
S
204
13. REPRODUCTIVE HEALTH
N/A Variable ID: RH 19. RHTPTEST
Reproductive Health The last time you were pregnant, were you tested for syphilis? YES NO
Response Options 1 2 7 8
Skips ‘2’,’7’,‘8’ → SKIP TO RHTRYPRG
Comments COVERAGE
Core or Secondary* S C
M N/A
T N/A
P C
O S
DON’T KNOW REFUSE TO ANSWER
20.
RHTPRSLT
The last time you were pregnant and had a syphilis test, what was the result of that syphilis test?
POSITIVE NEGATIVE UNCLEAR / NEITHER POSITIVE OR NEGATIVE DID NOT RECEIVE RESULT DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
’1’,’3’,’4’,’7’,’8’ → SKIP TO RHTRYPRG
RESULT
C
N/A
N/A
C
S
21.
RHTPTX
Did you get treatment for your syphilis infection during the last time you were pregnant?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
C
N/A
N/A
C
S
22.
RHSELL
Did you sell sex during the last time you were pregnant?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’,‘8’ →SKIP TO RHTRYPRG
C
N/A
N/A
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 0 2 505
N/A Variable ID: RH 23. RHRETRN
Reproductive Health How long after giving birth did you wait before selling sex again?
Response Options LESS THAN 1 YEAR 1 YEAR OR LONGER DON’T KNOW REFUSE TO ANSWER 1 2 7 8
Skips ‘1’,→SKIP TO RHRETRNM ’2’ →SKIP TO RHRETRNY ‘7’,’8’→SKIP TO RHTRYPRG
Comments
Core or Secondary* S C
M N/A
T N/A
P C
O S
24.
RHRETRNM
How many MONTHS after giving birth did you wait before selling sex again?
# MONTHS [ | ] MAX: 12 DON’T KNOW 97
RISK
C
N/A
N/A
C
S
25.
RHRETRNY
How many YEARS after giving birth did you wait before selling sex again? Type ‘0’ if less than one year.
# YEARS [ | ] DON’T KNOW 97
RISK
C
N/A
N/A
C
S
26.
RHTRYPRG
Are you currently trying to get pregnant?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
‘1’,’7’,‘8’ → SKIP TO END OF CURRENT SECTION
BACKGROUND
C
N/A
N/A
C
S
27.
RHCNNOW
Do you use any family planning method?
YES NO DON’T KNOW REFUSE TO ANSWER
‘2’,’7’,‘8’ → SKIP TO END OF CURRENT SECTION
PREVENTION
C
N/A
N/A
C
S
206
13. REPRODUCTIVE HEALTH
N/A Variable ID: RH 28. RHTYPNOW
Reproductive Health Which method do you mainly use? PILL
Response Options 1 2 3 4 6 7 8
Skips ‘1’,’2’,’3’,’4’,’7’, ’8’ → SKIP TO END OF CURRENT SECTION
Comments COVERAGE
Core or Secondary* S C
M N/A
T N/A
P C
O S
INJECTION NORPLANT IUD OTHER DON’T KNOW REFUSE TO ANSWER
29.
RHTYPOTH
You said “other”. Which of the following family planning methods do you mainly use?
I AM STERILIZED/PARTNER STERILIZED PATCH VAGINAL RING CONDOM OTHER REFUSE TO ANSWER
1 2 3 4 6 8
COVERAGE
C
N/A
N/A
C
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
2 0 2 707
Section
Il
14. SEXUAL HISTORY
The first part of this module should be used with male/TG participants only and is intended to capture information about first sex and lifetime history of sex. Because MSM and TG may be recruited into surveys targeting sex workers or people who inject drugs, participants should be asked about their sexual experiences with males, females, and TG. The second part of this module should be used with non-TG female participants only.
When asking about first sex, countries should utilize formative research and knowledge of local customs to determine whether young ages of first sex should be coded as sexual abuse in the violence module or as sex in this module. This should be decided before survey implementation. N/A N/A
Variable ID: LI LI1MSG
Sexual History
Response Options
Skips FOR MALES AND TG ONLY. IF FEMALE SKIP TO LI8MSG
Comments
Core or Secondary* S M T P O
The next few questions are about your lifetime sexual history. This includes vaginal and anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus (butt).
N/A
LI2MSG
First, we will ask you some questions about your sexual experiences with women.
1.
LIMFVAG
Have you ever had vaginal sex? By vaginal sex we mean your penis enter a woman’s vagina.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
C
C
C
C
208
14. SEXUAL HISTORY
N/A 2.
Variable ID: LI LIMFANAL
Sexual History Have you ever had anal sex with a woman? By anal sex we mean you put your penis into a woman’s anus (butt). YES NO
Response Options 1 2 7 8
Skips IF LIMFVAG=’2’,’ 7’ OR ‘8’ AND LIMFANAL=’2’, ’7’ OR ‘8’ → SKIP TO LI4MSG RISK
Comments
Core or Secondary* S C
M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER
3.
LIMFPART
In your lifetime with how many women have you had vaginal or anal sex?
# WOMEN [ | | | ] MIN: 1 DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK
S
S
S
S
S
N/A
LI3MSG
Now we will ask you some questions about the first time you had sex with a woman.
4.
LIMF1AGE
How old were you when you first had vaginal or anal sex with a woman?
# YEARS [ | ] MIN: 10 MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY
C
S
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 0 2 909
N/A 5.
Variable ID: LI LIMFAGPA
Sexual History Approximately how old was this woman at the time you first had sex? PLEASE GIVE YOUR BEST GUESS.
Response Options MORE THAN 10 YEARS YOUNGER THAN ME 5-10 YEARS YOUNGER THAN ME ABOUT THE SAME AGE 5-10 YEARS OLDER THAN ME MORE THAN 10 YEARS OLDER THAN ME DON’T KNOW REFUSE TO ANSWER
Skips
Comments VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
1 2 3 4 5 7 8 1 2 7 8 1 2 7 8
6.
LIMFGIVE
Did you pay or give her something to have sex?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
7.
LIMFGET
Did she pay you or give you something in exchange for sex?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK & VULNERABILITY
S
S
S
S
S
N/A
LI4MSG
Now we will ask you about your sexual experiences with TG women.
8.
LIMTANAL
Have you ever had anal sex with a TG woman?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
2’,’7’ OR ‘8’→ SKIP TO LI6MSG
RISK & VULNERABILITY
S
S
S
S
S
210
14. SEXUAL HISTORY
N/A 9.
Variable ID: LI LIMTPART
Sexual History In your lifetime with how many TG women have you had anal sex?
Response Options # TG WOMEN [ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
Skips
Comments RISK & VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
N/A
LI5MSG
Please think back to the first time you had sex with a TG woman.
10.
LIMT1AGE
How old were you when you first had anal sex with a TG female partner?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY
S
S
S
S
S
11.
LIMTAGPA
Approximately how old was your TG female sex partner at the time you first had sex? PLEASE GIVE YOUR BEST GUESS.
MORE THAN 10 YEARS YOUNGER THAN ME 5-10 YEARS YOUNGER THAN ME ABOUT THE SAME AGE 5-10 YEARS OLDER THAN ME MORE THAN 10 YEARS OLDER THAN ME DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 7 8
RISK & VULNERABILITY
S
S
S
S
S
12.
LIMTGIVE
Did you pay or give her something to have sex?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK & VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 1 2 111
N/A 13.
Variable ID: LI LIMTGET
Sexual History Did she pay you or give you something in exchange for sex? YES NO
Response Options 1 2 7 8
Skips
Comments RISK & VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
N/A
LI6MSG
Now we will ask you about your sexual experiences with men.
14.
LIMMANAL Have you ever had anal sex with a man?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→ SKIP TO END OF CURRENT SECTION
RISK
C
C
C
C
C
15.
LIMMPART In your lifetime with how many men have you had anal sex?
# MEN [ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK
S
S
S
S
S
N/A
LI7MSG
Please think back to the first time you had sex with a man.
16.
LIMM1AGE How old were you when you first had anal sex with a male partner?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY
C
C
C
C
C
212
14. SEXUAL HISTORY
N/A 17.
Variable ID: LI LIMMAGPA
Sexual History Approximately how old was your male sex partner at the time you first had sex? PLEASE GIVE YOUR BEST GUESS.
Response Options MORE THAN 10 YEARS YOUNGER THAN ME 5-10 YEARS YOUNGER THAN ME ABOUT THE SAME AGE 5-10 YEARS OLDER THAN ME MORE THAN 10 YEARS OLDER THAN ME DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 7 8
Skips
Comments VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
18.
LIMMGIVE
Did you pay or give him something to have sex?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
S
19.
LIMMGET
Did he pay you or give you something in exchange for sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 1 2 313
N/A 20.
Variable ID: LI LIMMID
Sexual History How would you describe the first man you had sex with?
Response Options BOYFRIEND/PARTNER FRIEND/ACQUAINTANCE/ COWORKER RELATIVE STRANGER ‘AUTHORITY FIGURE’ (GOVERNMENT OFFICIAL, RELIGIOUS LEADER, TEACHER, EMPLOYER, MILITARY, POLICE, PRISON GUARD) OTHER DON’T KNOW REFUSE TO ANSWER 1 2 3 4
Skips ‘1’,’2’,’3’,’4’,’6’,’7’,’8’ → SKIP TO END OF CURRENT SECTION
Comments VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
5 6 7 8 1 2 3 4 5 6 7 8 FOR FEMALES ONLY. IF MALE, SKIP TO END OF CURRENT SECTION. VULNERABILITY S S S S S
21.
LIMMAUT
What kind of “authority figure” was he?
PROFESSOR/TEACHER RELIGIOUS LEADER EMPLOYER MILITARY MAN/POLICE OFFICER PRISON GUARD OTHER DON’T KNOW REFUSE TO ANSWER
N/A
LI8MSG
The next few questions are about your lifetime sexual history. This includes vaginal and anal sex. Vaginal sex is when the penis is inserted into the vagina. Anal sex is when a penis is inserted into the anus.
214
14. SEXUAL HISTORY
N/A 22.
Variable ID: LI LIFMVAG
Sexual History Have you ever had vaginal sex? By vaginal sex I mean where a man inserts his penis into a woman’s vagina. YES NO
Response Options 1 2 7 8
Skips RISK
Comments
Core or Secondary* S C
M N/A
T N/A
P C
O C
DON’T KNOW REFUSE TO ANSWER
23.
LIFMANAL
Have you ever had anal sex? By anal sex I mean YES where a man inserts his penis into another NO person’s anus. DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
N/A
N/A
C
C
24.
LIFMPART
In your lifetime with how many different men have you had penetrative (vaginal or anal) sex?
# MEN [ | | | ] MIN: 1 DON’T KNOW 9997 REFUSE TO ANSWER 9998
IF LIFMVAG=’2’,’ 7’ OR ‘8’ AND LIFMANAL=’2’, ’7’ OR ‘8’ → SKIP TO END OF CURRENT SECTION
RISK
N/A
N/A
N/A
S
S
N/A
LI9MSG
Please think back to the first time you had sex.
25.
LIFM1AGE
How old were you when you first had vaginal or anal sex with a male partner?
# YEARS [ | ] MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
RISK & VULNERABILITY
C
N/A
N/A
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 1 2 515
N/A 26.
Variable ID: LI LIFMAGPA
Sexual History Approximately how old was your male sex partner at the time you first had sex? PLEASE GIVE YOUR BEST GUESS.
Response Options MORE THAN 10 YEARS YOUNGER THAN ME 5-10 YEARS YOUNGER THAN ME ABOUT THE SAME AGE 5-10 YEARS OLDER THAN ME MORE THAN 10 YEARS OLDER THAN ME DON’T KNOW REFUSE TO ANSWER
Skips
Comments VULNERABILITY
Core or Secondary* S S
M N/A
T N/A
P S
O S
1 2 3 4 5 7 8 1 2 7 8
27.
LIFMGET
Did your first male sex partner pay you or give you something in exchange for sex?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
N/A
N/A
S
S
28.
LIFMGIVE
Did you pay or give your first male sex partner something in exchange for sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
N/A
N/A
S
S
216
14. SEXUAL HISTORY
N/A 29.
Variable ID: LI LIFMID
Sexual History How would you describe the first man you had sex with?
Response Options BOYFRIEND/PARTNER FRIEND/ACQUAINTACE/ COWORKER RELATIVE STRANGER ‘AUTHORITY FIGURE’ (GOVERNMENT OFFICIAL, RELIGIOUS LEADER, TEACHER, EMPLOYER, MILITARY, POLICE, PRISON GUARD) OTHER DON’T KNOW REFUSE TO ANSWER 1 2 3 4
Skips ‘1’,’2’,’3’,’4’,’6’,’7’,’8’ → SKIP TO END OF CURRENT SECTION
Comments VULNERABILITY
Core or Secondary* S S
M N/A
T N/A
P S
O S
5 6 7 8
30.
LIFMAUT
What kind of “authority” was he?
PROFESSOR/TEACHER RELIGIOUS LEADER EMPLOYER MILITARY MAN/POLICE OFFICE PRISON GUARD OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
VULNERABILITY
S
N/A
N/A
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary.
SECTION 2: QUESTIONNAIRE MODULES
2 1 2 717
Section
Il
15. RECALL SEXUAL BEHAVIOR
This instrument probes sexual behavior over defined recall time periods, focusing on the type and number of partners, type of sexual intercourse, as well as condom use. The recall period needs to be chosen with careful consideration. Shorter recall time periods (e.g. 1 or 3 months) benefit from better participant recall at the expense of fewer events of interest compared with longer recall periods (e.g. 6 or 12 months). Investigators need to weigh the need for better recall over shorter period of times against the risk of some participants recalling fewer or, in extreme cases, no events (e.g. number of sex acts) where the recall period is short. When determining the recall period, keep in mind the corresponding time frame used for eligibility. Eligibility criteria using a longer time period (e.g. did the participant have anal sex in the past 12 months?) may lead to zero values in an interview using a shorter recall period (e.g. did the participant have anal sex in the past 3 months) in the key behavior of some participants.
Suggested sex partner definitions include main partners, casual partners, clients and sex workers. These partner types may be changed depending on local context but in all cases should be clearly delineated from each other while capturing all possible types to prevent missed counts and duplicate reporting of sex partners. This module includes three measures for each partner type: number of partners, type and number of sex acts, and condom use. The recall time periods vary by measure: The number of partners is probed for the last 6 months, the number of sex acts (“times”) for the last 4 weeks. This module includes a separate section on TG sexual partners. Depending upon the local context, you may opt to not include this section. If you opt to not include specific TG questions, MSM questions should be modified to include TG. This module starts with male partners, followed by questions about TG women partners, and female partners. For male participants this order may be reversed so that the interview starts with questions about female partners due to the sensitivity required in certain contexts for questions regarding same-sex behavior. N/A N/A
Variable ID: RC
Recall Sexual Behavior
Response Options
Skips
Comments
Core or Secondary* S M T D O
This next section is about your sexual behavior in the last 6 months. The questions are about different sex partners, how many times you had sex, and condom use. With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus (butt). We will ask you about different sex partner types: main sex partners, casual sex partners, people you pay money for sex, and people who pay you money for sex. Each person you have had sex with in the last 6 months should be counted in one of these categories.
218
15. RECALL SEXUAL HISTORY
N/A N/A 1.
Variable ID: RC
Recall Sexual Behavior MALE SEX PARTNERS
Response Options
Skips
Comments (SECTION)
Core or Secondary* S M T D O
RCMAPA
First we will ask you some questions about any male sex partners you may have had. In the last 6 months, did you have sex with any males?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8→SKIP RCFEPA
RISK
C
C
C
C
C
N/A 2. RCMAMNPA
MAIN MALE PARTNERS As mentioned earlier, we will ask you about different types # [ | ] of sex partners. First, we will ask you about any main male sex partners you may have had sex with in the last 6 MAX: 20 months. DON’T KNOW 97 A main sex partner is someone you are committed to, for REFUSE TO ANSWER 98 example your spouse, live-in sex partner, or boyfriend. In the last 6 months, with how many different main male partners did you have sex? Type ‘0’ if none. ‘0’,’97’ OR ‘98’→SKIP TO RCMACSPA
(SUBSECTION) RISK C C C C C
3.
RCMAMNSX
In the last 4 weeks, what is the total number of times you # [ | | | ] had sex with your main male partners? By “times”, we mean number of sex acts. For example, you could have sex DON’T KNOW 9997 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none. The last time you had sex with a main male partner, what type of sex did you have? VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 WOMEN ONLY.
RISK
S
S
S
S
S
4.
RCMAMTYP
RISK
C
N/A
N/A
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 1 2 919
N/A 5.
Variable ID: RC RCMAMNR
Recall Sexual Behavior The last time you had anal sex with your main male partner, did you have receptive or insertive anal sex? Receptive anal sex is when he puts his penis in your anus (butt) and insertive anal sex is when you put your penis in his anus (butt).
Response Options RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8
Skips MEN/TG WOMEN ONLY.
Comments RISK
Core or Secondary* S C
M C
T C
D C
O C
6.
RCMAMNRC The last time you had sex with a main male partner, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
C
C
C
C
7.
RCMAMNFQ In the last 6 months, how often did you use condoms with your main male partners?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
C
C
C
C
C
220
15. RECALL SEXUAL HISTORY
N/A N/A 8.
Variable ID: RC
Recall Sexual Behavior CASUAL MALE PARTNERS
Response Options
Skips
Comments (SUBSECTION)
Core or Secondary* S M T D O
RCMACSPA
These next questions are about any casual male partners you may have had sex with in the last 6 months. A casual male partner is a man you have sex with but don’t feel committed to.
#[ | | ] MAX: 100 DON’T KNOW 997
‘0’,’997’ OR ‘998’→ RISK SKIP TO RCMASWPA
C
C
C
C
C
There is no payment of money for sex with these partners. REFUSE TO ANSWER 998 In the last 6 months, with how many different casual male partners did you have sex? Type ‘0’ if none. 9. RCMACSSX In the last 4 weeks, what is the total number of times you # [ | | ] had sex with your casual male partners? By “times”, we mean number of sex acts. For example, you could have sex DON’T KNOW 997 5 times with the same partner. REFUSE TO ANSWER 998 Type ‘0’ if none. The last time you had sex with a casual male partner, what VAGINAL type of sex did you have? ANAL BOTH DON’T KNOW REFUSE TO ANSWER 11. RCMACSAT The last time you had anal sex with your casual male partner, did you have receptive or insertive anal sex? Receptive anal sex is when he puts his penis in your anus (butt) and insertive anal sex is when you put your penis in his anus (butt). RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 1 2 3 7 8 MEN/TG WOMEN ONLY. RISK C C C C C WOMEN ONLY. RISK S S S S S
10.
RCMACSA
RISK
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 2 2 121
N/A 12.
Variable ID: RC RCMACSAC
Recall Sexual Behavior The last time you had sex with a casual male partner, did you use a condom? YES NO
Response Options 1 2 7 8 1 2 3 4 5 7 8
Skips
Comments RISK
Core or Secondary* S C
M C
T C
D C
O C
DON’T KNOW REFUSE TO ANSWER 13. RCMACSFQ In the last 6 months, how often did you use condoms with casual male partners? ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
RISK
C
C
C
C
C
N/A 14. RCMASWPA
MALE SEX WORKER PARTNERS These next questions are about buying sex with money. In the last 6 months, how many different men did you give money in exchange for sex? Type ‘0’ if none. #[ | | ] MAX: 100 DON’T KNOW 997 REFUSE TO ANSWER 998 ‘0’,’997’ OR ‘998’ → SKIP TO RCMACLPA
(SUBSECTION) RISK S C C C C
15.
RCMASWSX
In the last 4 weeks, how many times did you have sex with male partners to whom you gave money? By “times”, we mean the number of sex acts. For example, you could have sex 5 times with the same partner. Type ‘0’ if none.
#[ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK
S
S
S
S
S
222
15. RECALL SEXUAL HISTORY
N/A 16.
Variable ID: RC RCMASWA
Recall Sexual Behavior The last time you had sex with any male you gave money for sex, what type of sex did you have?
Response Options VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 1 2 3 7 8 1 2 7 8 1 2 3 4 5 7 8
Skips WOMEN ONLY.
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
17.
RCMASWNR The last time you had anal sex with any male you gave money for sex, did you have receptive or insertive anal sex? Receptive anal sex is when he puts his penis in your anus (butt) and insertive anal sex is when you put your penis in his anus (butt).
RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER
MEN/TG WOMEN ONLY.
RISK
S
S
S
S
S
18.
RCMASWAC
The last time you had sex with any male you gave money for sex, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
19.
RCMASWFQ
In the last 6 months, how often did you use condoms with men you gave money for sex?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 2 2 323
N/A N/A 20.
Variable ID: RC
Recall Sexual Behavior MALE CLIENT PARTNERS
Response Options
Skips
Comments (SUBSECTION)
Core or Secondary* S M T D O
RCMACLPA
Now we will ask you about getting money or goods for sex. In the last 6 months, how many different men gave you money or goods for sex? Type ‘0’ if none.
#[ | | ] MAX: 800 DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’,’997’ OR ‘998’→ SKIP TO RCFEPA
RISK
N/A
C
C
C
C
21.
RCMACL
In the last 4 weeks, how many times did you have sex with # [ | | | ] any male partner who gave you money? By “times”, we mean number of sex acts. For example, you could have sex DON’T KNOW 9997 5 times with the same partner. REFUSE TO ANSWER 9998 The last time you had sex with any male who gave you money for sex, what type of sex did you have? VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 WOMEN ONLY.
RISK
N/A
S
S
S
S
22.
RCMACLA
RISK
N/A
C
C
C
C
23.
RCMACLAT
The last time you had anal sex with any male who gave you money for sex, did you have receptive or insertive anal sex? Receptive anal sex is when he puts his penis in your anus (butt) and insertive anal sex is when you put your penis in his anus (butt).
RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
MEN/TG WOMEN ONLY.
RISK
N/A
C
C
C
C
224
15. RECALL SEXUAL HISTORY
N/A 24.
Variable ID: RC RCMACLAC
Recall Sexual Behavior The last time you had sex with a man who gave you money, did you use a condom? YES NO
Response Options 1 2 7 8 1 2 3 4 5 7 8
Skips
Comments RISK
Core or Secondary* S N/A
M C
T C
D C
O C
DON’T KNOW REFUSE TO ANSWER 25. RCMACLFQ In the last 6 months, how often did you use condoms with men who gave you money for sex? ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
RISK
N/A
C
C
C
C
N/A 26. RCFEPA
FEMALE PARTNERS Now we will ask you about female sex partners. In the last 6 months, did you have sex with any females? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8
SKIP FOR FEMALE PARTICIPANTS ‘2’,’7’ OR ‘8’→SKIP TO RCTAPA
(SECTION) RISK C C C C C
SECTION 2: QUESTIONNAIRE MODULES
2 2 2 525
N/A N/A 27.
Variable ID: RC
Recall Sexual Behavior MAIN FEMALE PARTNERS
Response Options
Skips
Comments (SUBSECTION)
Core or Secondary* S M T D O
RCFEMNPA
As mentioned earlier, we will ask about different types of sex partners. First we will ask about any main female sex partners you may have had sex with in the last 6 months. A main sex partner could be your spouse, live-in sex partner, or girlfriend. There is no payment for sex with these partners. In the last 6 months, with how many different main female partners did you have sex? Type ‘0’ if none.
#[ | ] MAX: 20 DON’T KNOW 97 REFUSE TO ANSWER 98
‘0’,’97’ OR ‘98’→ SKIP TO RCFECSPA
RISK
C
C
C
C
C
28.
RCFEMNSX
In the last 4 weeks how many times did you have sex with # [ | | | ] these female main partners? By “times”, we mean number DON’T KNOW 9997 of sex acts. For example, you could have sex 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none. The last time you had sex with a female main partner, what type of sex did you have? VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8
RISK
S
S
S
S
S
29.
RCFEMNA
RISK
S
S
S
S
S
226
15. RECALL SEXUAL HISTORY
N/A 30.
Variable ID: RC RCFEMNAC
Recall Sexual Behavior The last time you had sex with a female main partner, did you use a condom? YES NO
Response Options 1 2 7 8
Skips
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
DON’T KNOW REFUSE TO ANSWER
31.
RCFEMNFQ
In the last 6 months, how often did you use condoms with main female partners?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
N/A 32. RCFECSPA
CASUAL FEMALE PARTNERS These next questions are about casual female partners. A female casual partner is woman you have sex with but don’t feel committed to. There is no payment for sex with such partners. In the last 6 months how many different female casual partners did you have sex with? Type ‘0’ if none. #[ | | ] MAX: 100 DON’T KNOW 997 REFUSE TO ANSWER 998 #[ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998 ‘0’,’997’ OR ‘998’→ SKIP TO RCFESWPA
(SUBSECTION) RISK C C C C C
33.
RCFECSSX
In the last 4 weeks how many times did you have sex with these casual female partners? By “times”, we mean number of sex acts. For example, you could have sex 5 times with the same partner. Type ‘0’ if none.
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 2 2 727
N/A 34.
Variable ID: RC RCFECSA
Recall Sexual Behavior The last time you had sex with any casual female partner, what type of sex did you have?
Response Options VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8
Skips
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
35.
RCFECSAC
The last time you had sex with a casual female partner, did YES you use a condom? NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
S
S
S
S
S
36.
RCFECSFQ
In the last 6 months, how often did you use condoms with casual female partners?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
228
15. RECALL SEXUAL HISTORY
N/A N/A 37.
Variable ID: RC
Recall Sexual Behavior FEMALE SEX WORKER PARTNERS
Response Options
Skips
Comments (SUBSECTION)
Core or Secondary* S M T D O
RCFESWPA
These next questions are about buying sex. In the last 6 months how many different women did you pay money for sex? Type ‘0’ if none.
#[ | | ] MAX: 100 DON’T KNOW 997 REFUSE TO ANSWER 998
‘0’,’997’ OR ‘998’→ SKIP TO RCFECLPA
RISK
C
C
C
C
C
38.
RCFESWSX
In the last 4 weeks how many times did you have sex with these females you gave money? By “times”, we mean number of sex acts. For example, you could have sex 5 times with the same partner. Type ‘0’ if none.
#[|||] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK
S
S
S
S
S
39.
RCFESWA
The last time you had sex with any female you gave money VAGINAL for sex, what type of sex did you have? ANAL BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
RISK
S
S
S
S
S
40.
RCFESWAC
The last time you had sex with a female you gave money for sex, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 2 2 929
N/A 41.
Variable ID: RC RCFESWFQ
Recall Sexual Behavior In the last 6 months, how often did you use condoms with females you gave money for sex?
Response Options ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments RISK
Core or Secondary* S C
M C
T C
D C
O C
N/A 42. RCFECLPA
FEMALE CLIENT PARTNERS Some people get money in exchange for sex. In the last 6 months how many different women gave you money for sex? Type ‘0’ if none. #[ | | ] MAX: 800 DON’T KNOW 997 REFUSE TO ANSWER 998 ‘0’,’997’ OR ‘998’ → SKIP TO SKIP TO RCTAPA
(SUBSECTION) RISK N/A C C C C
43.
RCFECL
In the last 4 weeks how many times did you have sex with # [ | | | ] these females who gave you money for sex? By “times”, we mean number of sex acts. For example, you could have DON’T KNOW 9997 sex 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none. The last time you had sex with any female who gave you money for sex, what type of sex did you have? VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8
RISK
N/A
S
S
S
S
44.
RCFECLA
RISK
N/A
S
S
S
S
230
15. RECALL SEXUAL HISTORY
N/A 45.
Variable ID: RC RCFECLAC
Recall Sexual Behavior The last time you had sex with a female who gave you money for sex, did you use a condom? YES NO
Response Options 1 2 7 8
Skips
Comments RISK
Core or Secondary* S N/A
M S
T S
D S
O S
DON’T KNOW REFUSE TO ANSWER
46.
RCFECLFQ
In the last 6 months, how often did you use condoms with females who gave you money for sex?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
N/A
S
S
S
S
N/A 47. RCTAPA
TG PARTNERS These next questions are about any TG women sex partners you may have had. In the last 6 months, did you have sex with any TG women? ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8 ‘2’,’7’ OR ‘8’→ SKIP TO END OF CURRENT SECTION
(SECTION) RISK S S S S S
SECTION 2: QUESTIONNAIRE MODULES
2 3 2 131
N/A N/A 48.
Variable ID: RC
Recall Sexual Behavior MAIN TG PARTNERS
Response Options
Skips
Comments (SUBSECTION)
Core or Secondary* S M T D O
RCTAMNPA
First we will ask you about any main TG women sex partners you may have had sex with in the last 6 months. A main sex partner is someone you are committed to, for example your spouse, live-in sex partner, girlfriend, or boyfriend. In the last 6 months, with how many different main TG women partners did you have sex? Type ‘0’ if none.
#[ | ] MAX: 20 DON’T KNOW 97 REFUSE TO ANSWER 98
‘0’,’97’ OR ‘98’→ SKIP TO RCTACSPA
RISK
S
S
S
S
S
49.
RCTAMNSX
In the last 4 weeks, what is the total number of times you # [ | | | ] had sex with your main TG women partners? By “times”, we mean number of sex acts. For example, you could have DON’T KNOW 9997 sex 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none. The last time you had sex with a main TG woman partner, what type of sex did you have? VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 1 2 3 7 8 MEN/TG WOMEN ONLY. WOMEN ONLY.
RISK
S
S
S
S
S
50.
RCTAMTYP
RISK
S
N/A
N/A
S
S
51.
RCTAMNR
The last time you had anal sex with your main TG woman partner, did you have receptive or insertive anal sex? Receptive anal sex is when she puts her penis in your anus (butt) and insertive anal sex is when you put your penis in her anus (butt).
RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
232
15. RECALL SEXUAL HISTORY
N/A 52.
Variable ID: RC RCTAMNRC
Recall Sexual Behavior The last time you had sex with a main TG woman partner, did you use a condom? YES NO
Response Options 1 2 7 8
Skips
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
DON’T KNOW REFUSE TO ANSWER
53.
RCTAMNFQ
In the last 6 months, how often did you use condoms with your TG women partners?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
N/A 54. RCTACSPA
CASUAL TG PARTNERS These next questions are about any casual TG women partners you may have had sex with in the last 6 months. A casual TG woman partner is a TG woman you have sex with, but don’t feel committed to. There is no payment or exchange of goods and services for sex with these partners you have sex? In the last 6 months, with how many different casual TG women partners did you have sex? Type ‘0’ if none. #[ | | ] MAX: 100 DON’T KNOW 997 REFUSE TO ANSWER 998 ‘0’,’997’ OR ‘998’→ SKIP TO RCTASWPA
(SUBSECTION) RISK S S S S S
SECTION 2: QUESTIONNAIRE MODULES
2 3 2 333
N/A 55.
Variable ID: RC RCTACSSX
Recall Sexual Behavior
Response Options
Skips
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
In the last 4 weeks, what is the total number of times you # [ | | | ] had sex with your casual TG women partners? By “times”, we mean number of sex acts. For example, you could have DON’T KNOW 9997 sex 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none.
56.
RCTACSAT
The last time you had sex with a casual TG woman partner, VAGINAL what type of sex did you have? ANAL BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
WOMEN ONLY.
RISK
S
S
S
S
S
57.
RCTACSAC
The last time you had anal sex with your casual TG woman partner, did you have receptive or insertive anal sex? Receptive anal sex is when she puts her penis in your anus (butt) and insertive anal sex is when you put your penis in her anus (butt).
RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 7 8
MEN/TG WOMEN ONLY.
RISK
S
S
S
S
S
58.
RCTACSAC
The last time you had sex with a casual TG woman partner, YES did you use a condom? NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
234
15. RECALL SEXUAL HISTORY
N/A 59.
Variable ID: RC RCTACSFQ
Recall Sexual Behavior In the last 6 months, how often did you use condoms with casual TG women partners?
Response Options ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
N/A 60. RCTASWPA
TG SEX WORKER PARTNERS Now we ask questions about buying sex with money. In the last 6 months, how many different TG women did you give money for sex? Type ‘0’ if none. #[ | | ] MAX: 100 DON’T KNOW 997 REFUSE TO ANSWER 998 ‘0’,’997’ OR ‘998’→ SKIP TO RCTACLPA
(SUBSECTION) RISK S S S S S
61.
RCTASWSX
In the last 4 weeks, how many times did you have sex with TG women partners to whom you gave money? By “times”, we mean the number of sex acts. For example, you could have sex 5 times with the same partner. Type ‘0’ if none
#[ | | | ] DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK
S
S
S
S
S
62.
RCTASWA
The last time you had sex with any TG woman you gave money for sex, what type of sex did you have?
VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
WOMEN ONLY.
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 3 2 535
N/A 63.
Variable ID: RC RCTASWNR
Recall Sexual Behavior The last time you had anal sex with a TG woman you gave money for sex, did you have receptive or insertive anal sex? Receptive anal sex is when she puts her penis in your anus (butt) and insertive anal sex is when you put your penis in her anus (butt).
Response Options RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8 1 2 7 8 1 2 3 4 5 7 8
Skips MEN/TG WOMEN ONLY.
Comments RISK
Core or Secondary* S S
M S
T S
D S
O S
64.
RCTASWAC
The last time you had sex with a TG woman you gave money for sex, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
65.
RCTASWFQ
In the last 6 months, how often did you use condoms with TG women you gave money for sex?
ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
N/A 66. RCTACLPA
TG CLIENT PARTNERS Some people get money for sex. In the last 6 months, how many different TG women paid you money for sex? Type ‘0’ if none. #[ | | ] MAX: 800 DON’T KNOW 997 REFUSE TO ANSWER 998 ‘0’,’997’ OR ‘998’ → SKIP TO END OF CURRENT SECTION
(SUBSECTION) RISK N/A S S S S
236
15. RECALL SEXUAL HISTORY
N/A 67.
Variable ID: RC RCTACL
Recall Sexual Behavior
Response Options
Skips
Comments RISK
Core or Secondary* S N/A
M S
T S
D S
O S
In the last 4 weeks, how many times did you have sex with # [ | | | ] the TG women partners who gave you money? By “times”, we mean number of sex acts. For example, you could have DON’T KNOW 9997 sex 5 times with the same partner. REFUSE TO ANSWER 9998 Type ‘0’ if none.
68.
RCTACLA
The last time you had sex with any TG woman who gave you money for sex, what type of sex did you have?
VAGINAL ANAL BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
WOMEN ONLY.
RISK
N/A
S
S
S
S
69.
RCTACLAC
The last time you had anal sex with any TG woman who gave you money for sex, did you have receptive or insertive anal sex?
RECEPTIVE INSERTIVE BOTH
1 2 3 7 8
MEN/TG WOMEN ONLY.
RISK
N/A
S
S
S
S
Receptive anal sex is when she puts her penis in your anus (butt) and insertive anal sex is when you put your penis in her anus (butt).
DON’T KNOW REFUSE TO ANSWER
70.
RCTACLAC
The last time you had sex with a TG woman who gave you money, did you use a condom?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK
N/A
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 3 2 737
N/A 71.
Variable ID: RC RCTACLFQ
Recall Sexual Behavior In the last 6 months, how often did you use condoms with TG women who gave you money for sex?
Response Options ALWAYS MOST OF THE TIME SOMETIMES RARELY NEVER DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments RISK
Core or Secondary* S N/A
M S
T S
D S
O S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
238
15. RECALL SEXUAL HISTORY
Section
II
16. PARTNER CONCURRENCY
Partner concurrency is suspected to be a key driver of HIV infection in generalized epidemics. Definitions of partner concurrency vary; the UNAIDS Reference Group on Estimates, Modeling, and Projections describes it as “Overlapping sexual partnerships where sexual intercourse with one partner occurs between two acts of intercourse with another partner.”Multiple approaches exist to measure partner concurrency. This module presents a modified version of the UNAIDSproposed measurement questions. It aims at describing the duration of sexual relationships for the last three sexual partners of the participant. This instrument can easily be expanded to capture additional relationships. It may be helpful to have a calendar available with key time points (i.e. holidays) to aid the participant in determining the duration of each relationship.
N/A 1.
Variable ID: CC CCNUM
PARTNER CONCURRENCY Think about all the sex partners you had in the last 6 months. With how many of these do you still have active sexual relationships? With ‘active relationships’ we mean you currently have sex with or expect to have sex again in the near future. With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus.
Response Options PARTNERS [ | | ] DON’T KNOW 997 REFUSE TO ANSWER 998
Skip ‘0’,’1’, ‘997’ OR ‘998’ → SKIP TO END OF CURRENT SECTION RISK
Comments
Core or Secondary* S N/A
M S
T S
P S
O S
This question is a stand- alone question, separate from the remaining questions which follow a UNAIDS recommended question set.
2.
CC1LASTM
MONTHS [ | ] Think about the person you had sex with most recently. How many months ago did you last have sex MAX: 12 with this person? Type 0 if less than one month. With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus. DON’T KNOW 97 REFUSE TO ANSWER 98
RISK
N/A
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 3 2 939
N/A 3.
Variable ID: CC CC1FIRST
PARTNER CONCURRENCY How long ago did you first have sex with this person? With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus.
Response Options WITHIN THE LAST 12 MONTHS 12 MONTHS OR MORE THAN 12 MONTHS AGO DON’T KNOW REFUSE TO ANSWER 1 2 7 8
Skip ‘2’,’7’OR ‘8’ → SKIP TO CC1Y RISK
Comments
Core or Secondary* S N/A
M S
T S
P S
O S
4.
CC1M
How many months ago?
MONTHS [ | ] MIN: RESPONSE FROM CC1LAST MAX: 12 DON’T KNOW 97 REFUSE TO ANSWER 98
SKIP TO CC2LASTM
RISK
N/A
S
S
S
S
5.
CC1Y
How many years ago?
YEARS [ |
]
RISK ‘Current age’ from module DEMOGRAPHICS. ‘Age at first sex’ from LIFETIME SEX HISTORY.
N/A
S
S
S
S
MIN: 1 MAX: CURENT AGE – AGE AT FIRST SEX DON’T KNOW 97 REFUSE TO ANSWER 98 6. CC2LASTM Think about the 2nd person you recently had sex with. How many months ago did you last have sex with this person? Type 0 if less than one month. With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus. MONTHS [ | ] MAX: 12 DON’T KNOW 97 REFUSE TO ANSWER 98
N/A RISK
S
S
S
S
240
16. PARTNER CONCURRENCY
N/A 7.
Variable ID: CC CC2FIRST
PARTNER CONCURRENCY How long ago did you first have sexual intercourse with this person? With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus.
Response Options WITHIN THE LAST 12 MONTHS MORE THAN A YEAR AGO DON’T KNOW REFUSE TO ANSWER 1 2 7 8
Skip ‘2’,’ 7’,’ 8’ → SKIP TO CC2Y RISK
Comments
Core or Secondary* S N/A
M S
T S
P S
O S
8.
CC2M
How many months ago?
MONTHS [
| ]
MIN: RESPONSE FROM CC2LAST MAX: 12 DON’T KNOW 97 REFUSE TO ANSWER 98
SKIP TO CC3LASTM
RISK
N/A
S
S
S
S
9.
CC2Y
How many years ago?
YEARS [ MIN: 1
|
]
RISK ‘Current age’ from module DEMOGRAPHICS. ‘Age at first sex’ from LIFETIME SEX HISTORY.
N/A
S
S
S
S
MAX: CURRENT AGE-AGE AT FIRST SEX DON’T KNOW 97 REFUSE TO ANSWER 98
10.
CC3LASTM
Think about the 3rd person you recently had sex with. MONTHS [ | How many months ago did you last have sex with this MIN: 1 person? Type 0 if less than one month. With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus. MAX: 12
]
IF CCNUM=’2’ → SKIP TO END OF CURRENT SCECTION
RISK
N/A
S
S
S
S
DON’T KNOW 97 REFUSE TO ANSWER 98
SECTION 2: QUESTIONNAIRE MODULES
2 4 2 141
N/A 11.
Variable ID: CC CC3FIRST
PARTNER CONCURRENCY How long ago did you first have sexual intercourse with this person? With sex we mean either vaginal sex or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus.
Response Options WITHIN THE LAST 12 MONTHS MORE THAN A YEAR AGO DON’T KNOW REFUSE TO ANSWER MONTHS [ | ] 1 2 7 8
Skip ‘2’,’ 7’ OR ‘8’ → SKIP TO CC3Y RISK
Comments
Core or Secondary* S N/A
M S
T S
P S
O S
12.
CC3M
How many months ago?
MIN: RESPONSE FROM CC3LAST MAX: 12 DON’T KNOW 97 REFUSE TO ANSWER 98 13. CC3Y How many years ago? YEARS [ | ] MIN: 1 MAX: CURRENT AGE-AGE AT FIRST SEX DON’T KNOW 97 REFUSE TO ANSWER 98 *: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SKIP TO END OF CURRENT SECTION
RISK
N/A
S
S
S
S
‘Current age’ from module DEMOGRAPHICS. ‘Age at first sex’ from LIFETIME SEX HISTORY.
N/A
S
S
S
S
References: 1. UNAIDS Reference Group on Estimates Modelling and Projections: Working Group on Measuring Concurrent Sexual Partnerships. Consultation on concurrent sexual partnerships: recommendations from a meeting of the UNAIDS Reference Group on Estimates, Modelling and Projections held in Nairobi, Kenya, 20–21 April 2009. http://www.epidem.org/ Publications/Concurrency%20meeting%20recommendations_Final.pdf. 2. UNAIDS Reference Group on Estimates Modelling and Projections: Working Group on Measuring Concurrent Sexual Partnerships. HIV: consensus indicators are needed for concurrency. Lancet 2010; 375:621–622.
242
16. PARTNER CONCURRENCY
Section
II
17. LAST SEX ACT
The last sex act is often the sex act a participant best remembers. This instrument is applicable to male, female and TG participants. Many or most of the probes below are part of or may be incorporated in other modules. Additional skip patterns apply, related to, for example, participant sex, and HIV status. Investigators may want the participant to recall the last sex act with a particular partner type; for example, for a survey among men who have sex with men, the ‘last sex act’ may want to probe the last time the participant had sex with a man even if the chronologically last sex act was with a woman. Similarly, in a survey of sex workers, the last sex act module may want to focus on the last sex act with a client. This module can be repeated multiple times if investigators want to probe e.g. the last three sex acts, or want to probe the last sex act with a range of different partner types, e.g. a main and casual partner. If this module will be used to probe the last sex act with a specific type of partner, the questions and the responses must be modified accordingly.
N/A 1.
VARIABLE ID: LS LSATIM
LAST SEX ACT Which answer best describes when you last had sex? By sex, I mean vaginal and anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus (butt). NEVER
Response Options 1 2 3 4 5 7 8
Skip
Comments S RISK ‘1’→ SKIP TO END OF SECTION ‘Never’ may be removed if having had sex is part of the eligibility criteria S
Core or Secondary* M S
T S
P S
O S
LAST 7 DAYS LAST 4 WEEKS LAST 3 MONTHS LONGER THAN 3 MONTHS AGO DON’T KNOW REFUSE TO ANSWER
SECTION 2: QUESTIONNAIRE MODULES
2 4 2 343
N/A 2.
VARIABLE ID: LS LSASEX
LAST SEX ACT The last time you had sex, was this with a man, a MAN woman, a transgender, or another person? WOMAN
Response Options 1 2 3 6 7 8
Skip
Comments S RISK C
Core or Secondary* M C
T C
P C
O C
TRANSGENDER OTHER DON’T KNOW REFUSE TO ANSWER 3. LSAAGE The last time you had sex, how old was this person? Please give your best guess. YEARS: [ |__] DON’T KNOW 97 REFUSE TO ANSWER 98 4. LSAPRT1 What kind of sex partner was that person? Remember, a main sex partner is someone you are committed to, for example your spouse, live-in sex partner, girlfriend or boyfriend. There is no payment or exchange of goods or services for sex with these partners. A casual partner is a man or woman you have sex with but don’t feel committed to. There is no payment or exchange of goods and services for sex with these partners. MAIN SEX PARTNER CASUAL SEX PARTNER SOMEONE I PAID MONEY OR GOODS FOR SEX SOMEONE WHO GAVE ME MONEY OR GOODS FOR SEX DON’T KNOW REFUSE TO ANSWER 5. LSACC1 Do you have a sexual relationship with this person, meaning you have had sex with this person on more than one occasion? YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
S
S
1 2 3 4 7 8 1 2 7 8
‘3’,’4’,’7’ OR ‘8’ → SKIP TO LSASXTY
RISK
N/A
S
S
S
S
‘2’→ SKIP TO LSASXTY
RISK
N/A
S
S
S
S
244
17. LAST SEX ACT
N/A 6.
VARIABLE ID: LS LSACC2
LAST SEX ACT As far as you know, during the time you were having a sexual relationship with this partner did this person have sex with other people
Response Options DEFINITELY DID NOT PROBABLY DID NOT PROBABLY DID DEFINITELY DID DON’T KNOW REFUSE TO ANSWER 1 2 3 4 7 8 1 2 7 8 1 2 3 7 8
Skip
Comments S RISK N/A
Core or Secondary* M S
T S
P S
O S
7.
LSACC3
During the time you were having a sexual relationship with this person did you have sex with other people?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
N/A
S
S
S
S
8.
LSASXTY
The last time you had sex with this person, what kind of sex did you have: vaginal, anal or both?
VAGINAL SEX ANAL SEX BOTH VAGINAL AND ANAL DON’T KNOW REFUSE TO ANSWER
SKIP IF DESEXNOW=’1’ AND LSASEX=’1’ OR DESEXNOW=’2’ AND LSASEX=’2’ ‘1’, ‘7’, OR ‘8’ → SKIP TO LSALBR MEN ONLY.
RISK
S
N/A
S
S
S
9.
LSAANAL
What kind of anal sex did you have at that time— receptive or insertive? Receptive means your partner put his penis in your anus. Insertive means you put your penis in your partner’s anus (butt).
RECEPTIVE INSERTIVE BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 4 2 545
N/A 10.
VARIABLE ID: LS LSACON
LAST SEX ACT Did you use a condom the last time you had sex with this person? YES NO
Response Options 1 2 7 8 1 2 3 4
Skip ‘1’, ‘7’, or ‘8’ → SKIP TO LSALBR
Comments S PREVENTION C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 11. LSAREAS The last time you had sex, what was the main reason you did not use a condom? ALWAYS USE A CONDOM I DIDN’T HAVE ONE/ I DIDN’T THINK OF IT I DON’T LIKE THEM MY PARTNER OBJECTED I DIDN’T FEEL AT RISK BECAUSE I’M IN A MONOGAMOUS RELATIONSHIP OTHER DON’T KNOW REFUSE TO ANSWER 12. LSALBR Did you or your partner use a lubricant the last time you had sex? YES NO DON’T KNOW REFUSE TO ANSWER
RISK Adapt responses based on local context
S
S
S
S
S
5 6 7 8 1 2 7 8 RISK, PREVENTION S S S S S
246
17. LAST SEX ACT
N/A 13.
VARIABLE ID: LS LSAARV
LAST SEX ACT
Response Options 1 2 3 7 8 1 2 3 7 8 1 2 3 7 8 1 2 3 7 8
Skip
Comments S PREVENTION Skip if PrEP not available in country S
Core or Secondary* M S
T S
P S
O S
Some people take ARVs or HIV medications YES – BEFORE before or after sex so that they don’t get HIV. Did you take ARVs (PrEP or HIV treatment) by mouth YES – AFTER before or after you had sex with this person? NO DON’T KNOW REFUSE TO ANSWER
14.
LSAMBC
Did you use a microbicide before you had sex the last time with this person?
YES NO NO, I AM HIV POSITIVE DON’T KNOW REFUSE TO ANSWER
PREVENTION Skip if microbicides not available in country.
S
S
S
S
S
15.
LSAALC
Did you or your partner drink alcohol before having sex?
YES I DID YES MY PARTNER DID WE BOTH DID DON’T KNOW REFUSE TO ANSWER
RISK, VULNERABILITY
S
S
S
S
S
16.
LSADRG
Did you or your partner take drugs to get high before you had sex?
I TOOK DRUGS MY PARTNER TOOK DRUGS WE BOTH TOOK DRUGS DON’T KNOW REFUSE TO ANSWER
RISK, VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 4 2 747
N/A 17.
VARIABLE ID: LS LSARAP1
LAST SEX ACT The last time you had sex, did the person force you? YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 3 4 7 8 1 2 7 8
Skip ‘1’→ SKIP TO LSAHIV1
Comments S RISK, VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER 18. LSARAP2 The last time you had sex, did you force that person to have sex with you? YES NO DON’T KNOW REFUSE TO ANSWER 19. LSAHIV1 With this sex partner, did you ever disclose each other’s HIV status? YES – ONLY I DISCLOSED YES – ONLY MY PARTNER DISCLOSED WE BOTH DISCLOSED NO DON’T KNOW REFUSE TO ANSWER 20. LSAHIV2 What is that person’s HIV status? HIV-POSITIVE HIV-NEGATIVE DON’T KNOW REFUSE TO ANSWER *: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
RISK, VULNERABILITY
S
S
S
S
S
IF ‘1’, ‘4’, ‘7’, OR PREVENTION ‘8’ → SKIP TO END OF CURRENT SECTION
S
S
S
S
S
RISK, PREVENTION
S
S
S
S
S
248
17. LAST SEX ACT
Section
II N/A N/A
18. CONDOM USE/ACCESSIBILITY
This module assesses access to condoms and their use. It should be asked of all groups. It is useful to provide images of male and female condoms to participants so they will better understand what the questions refer to. Questions concerning female condoms may be skipped if it is believed that a small proportion, such as less than 5% of people, use female condoms.
Variable ID: CO CO1MSG
Condom Use
Response Options
Skips
Comments S
Core or Secondary* M T P O
Now we are going to ask you some questions about condoms and your experience using them Where can a person get male condoms? CHECK ALL THAT APPLY. CLINIC/HOSPITAL KIOSK/SHOP PHARMACY LOCAL FREE DISPENSER FRIENDS/PEERS SEXUAL PARTNER NGO OTHER DON’T KNOW REFUSE TO ANSWER A B C D E F G X Y Z IF ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL. KA Adapt responses to local context. S S S S S
1.
COMLOC
SECTION 2: QUESTIONNAIRE MODULES
2 4 2 949
N/A 2.
Variable ID: CO COMGET
Condom Use If you wanted to, could you get yourself a male condom? YES NO
Response Options 1 2 7 8 A B C D E F G X Y Z 1 2 7 8
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER 3. COFLOC Where can a person get female condoms? CHECK ALL THAT APPLY. CLINIC/HOSPITAL KIOSK/SHOP PHARMACY LOCAL FREE DISPENSER FRIENDS/PEERS SEXUAL PARTNER NGO OTHER DON’T KNOW REFUSE TO ANSWER 4. COFGET If you wanted to, could you get yourself a female condom? YES NO DON’T KNOW REFUSE TO ANSWER
IF ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
KA Adapt responses to local context.
S
S
S
S
S
PREVENTION
S
S
S
S
S
250
18. CONDOM USE/ACCESSIBILITY
N/A 5.
Variable ID: CO COBRND
Condom Use What brands of condoms do you prefer to use? CHECK ALL THAT APPLY.
Response Options DO NOT USE CONDOMS BRAND A BRAND B OTHER DON’T KNOW REFUSE TO ANSWER A B C X Y Z
Skips IF ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
Comments S COVERAGE & PREVENTION Use brands from local context. More brands can be added. C
Core or Secondary* M C
T C
P C
O C
6.
COCARRY
Are you carrying any condoms with you right now?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
PREVENTION
S
S
S
S
S
7.
COASK
Could you ask your main sex partner to use a condom if you wanted? A main sex partner is someone you are committed to, for example your spouse, live-in sex partner, or boyfriend. There is no payment or exchange of goods or services for sex with these partners.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILI TY
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 5 2 151
N/A 8.
Variable ID: CO COVUSE
Condom Use Under what circumstances do you tend not to use condoms during vaginal sex? With vaginal sex we mean a penis enters a vagina. CHECK ALL THAT APPLY.
Response Options WHEN I’M DRUNK OR HIGH WHEN I AM AFRAID TO ASK MY PARTNER TO USE A CONDOM OR THEY REFUSE WHEN HAVING SEX WITH A REGULAR PARTNER WHEN HAVING SEX WITH A NON-REGULAR PARTNER WHEN THE PERSON DOES NOT EJACULATE INSIDE ME OTHER DON’T KNOW REFUSE TO ANSWER A
Skips ONLY ASK IF LIMFVAG=’1’ OR LIFMVAG=’1’ RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
B C D E X Y Z
Do not include option 5 for men.
252
18. CONDOM USE/ACCESSIBILITY
N/A 9.
Variable ID: CO COANO
Condom Use Under what circumstances do you tend not to use condoms during anal sex? With anal sex we mean a penis enters a person’s anus. CHECK ALL THAT APPLY.
Response Options WHEN I’M DRUNK OR HIGH WHEN I AM AFRAID TO ASK MY PARTNER TO USE A CONDOM OR THEY REFUSE WHEN HAVING SEX WITH A REGULAR PARTNER WHEN HAVING SEX WITH A NON-REGULAR PARTNER WHEN I AM THE INSERTIVE (TOP) PARTNER WHEN I AM THE RECEPTIVE (BOTTOM) PARTNER WHEN THE PERSON DOES NOT EJACULATE INSIDE ME I CANNOT GET ONE OTHER DON’T KNOW REFUSE TO ANSWER A
Skips ONLY ASK IF LIMFANAL=’1’ OR LIFMANAL=’1’S RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
B C D E F G H X Y Z
Do not include questions on insertive or receptive partners for females. Adapt responses based on population and local context.
10.
COLIKELY
Are you more likely to use a condom when a man inserts his penis into your anus (butt) or when you put your penis in his or equally likely
WHEN HIS PENIS IS IN ME WHEN MY PENIS IS IN HIM/ HER EQUALLY LIKELY DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
MALES ONLY. SKIP IF LIMMANAL =’2’, ‘7’, OR ‘8’
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 5 2 353
N/A 11.
Variable ID: CO COFREE
Condom Use Thank you. Now a question about getting condoms. In the last 12 months, have you been given condoms for free? For example, through an outreach YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 3 7 8
Skips ‘2’,’7’,’8’→ CONOGT
Comments S COVERAGE C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 12. COFREEU Do you use free condoms? YES NO DON’T KNOW REFUSE TO ANSWER 13. COPREF Do you prefer free condoms or branded condoms that you pay for? FREE BRANDED NO PREFERENCE DON’T KNOW REFUSE TO ANSWER
COVERAGE
S
S
S
S
S
KA
S
S
S
S
S
254
18. CONDOM USE/ACCESSIBILITY
N/A 14.
Variable ID: CO CONOGT
Condom Use In the last 12 months, what are some reasons you couldn’t get condoms when you needed them? CHECK ALL THAT APPLY.
Response Options CAN ALWAYS GET CONDOMS COSTS TOO MUCH NOT CONVENIENT CLINIC DOES NOT PROVIDE THEM EMBARRASSED TO GET CONDOMS DO NOT KNOW WHERE TO GET CONDOMS CONDOMS NOT AVAILABLE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E F G X Y Z
Skips
Comments S S
Core or Secondary* M S
T S
P S
O S
IF ‘A’, ‘Y’, OR COVERAGE ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
15.
COINFO
In the last 12 months, have you received information on condom use and safe sex? For example, through an outreach service, drop-in center or health clinic.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
C
C
C
C
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
2 5 2 555
Section
Il N/A N/A
19. LUBRICANT USE
This module assesses respondents’ access to lubricants and their use. This section is especially important for studies with MSM and TG. Questions concerning dry sex (when a woman dries her vagina before sex), and types and brands of lubricants should be adapted to the local context based on information collected during formative assessment.
Variable ID: LU LU1MSG
Lubricant Use
Response Options
Skips
Comments S
Core or Secondary* M T P O
Some people use lubricants during vaginal or anal sex. With vaginal sex we mean a penis enters a vagina. With anal sex we mean a penis enters a person’s anus. Lubricants make your penis or your partner’s penis more slippery and easier to insert into the vagina or anus. Lubricants also prevent the condom from breaking. Now we will ask you some questions about using lubricants in the last 6 months.
1.
LUDRY
Do you dry out your vagina before sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
WOMEN ONLY.
RISK ADAPT TERMINOLO GY OF ‘DRY SEX’ TO LOCAL CONTEXT
S
N/A
N/A
S
S
2.
LU6LUB
In the last 6 months, have you used a lubricant during anal or vaginal sex? Examples include water-based lube, oil-based lube and saliva.
YES NO DON’T KNOW REFUSE TO ANSWER
‘2’,’7’ OR ‘8’ → SKIP TO END OF CURRENT SECTION
RISK ADAPT TERMINOLOGY BASED ON LOCAL CONTEXT
C
C
C
C
C
256
19. LUBRICANT USE
N/A 3.
Variable ID: LU LUTYPE
Lubricant Use In the last 6 months, which lubricant did you use during vaginal or anal sex? CHECK ALL THAT APPLY
Response Options WATER-BASED LUBE, KY JELLY, VENDOME SALIVA VASELINE, POMADE OR OTHER PETROLEUM JELLY PRODUCT BODY LOTION, SHEA NUT BUTTER, OR BABY OIL COOKING OIL, MAYONNAISE, BUTTER OR MARGARINE OTHER DON’T KNOW REFUSE TO ANSWER
Skips IF ‘A’ NOT SELECTED → SKIP TO LUAVAIL
Comments S RISK & PREVENTION MODIFY RESPONSE OPTIONS BASED ON LOCAL CONTEXT C
Core or Secondary* M C
T C
P C
O C
A B
C
D
E X Y Z PREVENTION & COVERAGE S S S S S
4.
LUGET
In the last 6 months, where did you usually get water-based lubricants? CHECK ALL THAT APPLY
SHOP OR SUPERMARKET PHARMACY OR CLINIC NGO/ORGANIZATION FRIEND SEX PARTNER OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
SECTION 2: QUESTIONNAIRE MODULES
2 5 2 757
N/A 5.
Variable ID: LU LUAVAIL
Lubricant Use Are you able to get water-based lubricants when you need them? YES NO
Response Options 1 2 7 8
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
6.
LU3USEVA
In the last 6 months, how often did you use a lubricant during vaginal sex?
ALWAYS SOMETIMES NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
ONLY ASK IF LIMFVAG=’1’ OR LIFMVAG=’1’
RISK & PREVENTION
C
S
S
S
S
7.
LUUSEAN
In the last 6 months, how often did you use a lubricant during anal sex?
ALWAYS SOMETIMES NEVER DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
ONLY ASK IF LIMFANAL=’1’ OR LIFMANAL=’1’ ‘1’,’7’ OR ‘8’→SKIP TO LUCOST
RISK & PREVENTION
C
C
C
S
S
258
19. LUBRICANT USE
N/A 8.
Variable ID: LU LUNOTUSE
Lubricant Use In the last 6 months, what is the main reason you do not always use a water- based lubricant during anal sex? SELECT ONE ONLY.
Response Options CAN’T GET THEM EASILY/TOO EXPENSIVE DO NOT LIKE LUBRICANTS PARTNER DOESN’T LIKE THEM I’VE NEVER HEARD OF IT I’M ASHAMED/ EMBARRASSED TO BUY IT BECAUSE IT IS ASSOCIATED WITH HOMOSEXUALS OTHER DON’T KNOW REFUSE TO ANSWER
Skips ONLY ASK IF LIMFANAL=’1’ OR LIFMANAL=’1’
Comments S PREVENTION & VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
1 2 3 4
5 6 7 8
9.
LUCOST
Do you find water-based lubricants such as KY Jelly, Vendome or Assegai to be very affordable, somewhat affordable, or not affordable?
VERY AFFORDABLE SOMEWHAT AFFORDABLE NOT AFFORDABLE DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
KA Adapt lubricant brands to local context.
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 5 2 959
N/A 10.
Variable ID: LU LUOBTAIN
Lubricant Use
Response Options 1 2 3 4 7 8 1 2 7 8
Skips
Comments S Adapt lubricant brands to local context. S
Core or Secondary* M S
T S
P S
O S
How easy is it to obtain water-based lubricants such VERY EASY as KY Jelly, Vendome or Assegai? SOMEWHAT EASY NOT EASY CANNOT GET THEM DON’T KNOW REFUSE TO ANSWER
11.
LUBREAK
In the last 6 months, did you ever have a condom break during vaginal or anal sex?
YES NO DON’T KNOW REFUSE TO ANSWER
ONLY ASK IF LIMFVAG=’1’ OR LIFMVAG=’1’ OR LIMFANAL=’1’ OR LIFMANAL =’1’ ‘2’,’7’ OR ‘8’ →SKIP TO NEXT SECTION
RISK
S
S
S
S
S
12.
LUUSEBRK
Did you use lubricant the time that it broke?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
RISK
S
S
S
S
S
13.
LUFREE
In the last 12 months, have you been given “packets” of lubricant for free? For example, through an outreach service, drop-in center or health clinic.
YES NO DON’T KNOW REFUSE TO ANSWER
COVERAGE
C
C
C
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
260
19. LUBRICANT USE
Section
II N/A N/A
20. MICROBICIDES
This module is meant to assess awareness of and access to microbicides. Local knowledge should be used to modify the questions with local terminology. This module should only be used in settings where microbicides are available.
Variable ID: MI MI1MSG
Microbicides
Response Options
Skips
Comments S
Core or Secondary* M T P O
The next few questions are about gels that can prevent HIV infection when inserted into the vagina or anus. Some people call these gels “microbicides”.
1.
MIHEAR
YES Have you heard of a gel (product) that can help prevent HIV when you put it in your vagina or anus before sex? With sex we mean either vaginal sex or anal sex. With vaginal sex we mean NO a penis enters a vagina. With anal sex we mean a penis enters DON’T KNOW a person’s anus. REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8→ SKIP TO END OF CURRENT SECTION
KA
S
S
S
S
S
2.
MIEUSE
Have you ever used this kind of gel before vaginal or anal sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO END OF CURRENT SECTION
RISK & PREVENTION
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 6 2 161
3.
MI6USE
In the last 6 months, have you used this kind of gel before sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘1’,’7’ OR ‘8→SKIP TO MIDISCUS SKIP IF RCMAPA=’2’, ‘7’, OR ‘8’; RCFEPA=’2’, ‘7’, OR ‘8’; OR RCTAPA=’2’, ‘7’, ‘8’ SKIP TO END OF CURRENT SECTION
RISK & PREVENTION
S
S
S
S
S
4.
MINOTUSE
What is the main reason you have not used a microbicide?
HAVEN’T THOUGHT ABOUT IT DON’T LIKE IT TOO EXPENSIVE PARTNER OBJECTS CAN’T GET IT DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 7 8 1 2 7 8
PREVENTION
S
S
S
S
S
5.
MIDISCUS
Did you talk to your sex partner before starting to use a vaginal or anal microbicide?
YES NO DON’T KNOW REFUSE TO ANSWER
PREVENTION
S
S
S
S
S
6.
MITELL
The last time you used a microbicide, did you tell your sex partner that you used it?
YES NO DON’T KNOW REFUSE TO ANSWER
PREVENTION
S
S
S
S
S
262
20. MICROBICIDES
7.
MIGET
Where do you usually get your microbicide?
PHARMACY LOCATION A LOCATION B LOCATION C DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
PREVENTION & COVERAGE Use locally appropriate responses
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
2 6 2 363
Section
Il
21. ALCOHOL USE
AUDIT (Alcohol Use Disorders Identification Test) scale is a validated identification tool with a high sensitivity (92%) and specificity (94%), making it an accurate screening tool for alcohol abuse. The AUDIT scale includes 10 questions about the frequency and quantity of alcohol use, as well as binge drinking, dependence symptoms, and alcohol-related problems. AUDIT asks quantitative questions on drinking which may be hard to be answered in some settings (e.g. communal drinking out of a plastic bucket through straws), therefore a standard unit should be defined and be country/setting specific.
AUDIT-C (Consumption) is a subset derived from the first 3 questions of the full AUDIT and indicates whether an individual is potentially drinking at increased or high risk levels. The AUDIT-C tool can be used to obtain an initial understanding of alcohol dependence and if a respondent’s points are all from question 1 alone, then the full AUDIT can be used to confirm accuracy. CAGE is another validated tool to determine usage patterns which may reflect problems with alcohol. CAGE has 71% sensitivity and 90% specificity. It is a shorter instrument than AUDIT. CAGE asks qualitative questions on alcohol use. It consists of 4 simple yes/no questions and does not require information such as the amount or frequency of drinking. N/A Variable ID: AL N/A AL1MSG
Alcohol Use Now some questions about drinking alcohol during this past year.
Response Options
Skips
Comments S
Core or Secondary* M T P O
1.
ALFRQ
How often do you have a drink containing alcohol?
NEVER MONTHLY OR LESS 2-4 TIMES A MONTH 2-3 TIMES A WEEK 4 OR MORE TIMES A WEEK REFUSE TO ANSWER
0 1 2 3 4 8
‘0’→SKIP TO END OF CURRENT SECTION
AUDIT- C/AUDIT RISK
C
C
C
C
C
264
21. ALCOHOL USE
N/A Variable ID: AL 2. ALDAY
Alcohol Use How many drinks containing alcohol do you have on a typical day when you are drinking? 1 OR 2 3 OR 4 5 OR 6 7-9
Response Options 0 1 2 3 4 8
Skips
Comments S AUDIT- C/AUDIT RISK C
Core or Secondary* M C
T C
P C
O C
10 OR MORE REFUSE TO ANSWER
3.
ALBNGE
On one occasion, how often do you have six or more drinks?
NEVER LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
0 1 2 3 4 8
IF ALDAY=’0’ AND ALBINGE = ‘0’→SKIP TO END OF CURRENT SECTION
AUDIT- C/AUDIT RISK
C
C
C
C
C
4.
ALCTRL
During the last year, how often have you found that you were not able to stop drinking once you had started?
NEVER LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
0 1 2 3 4 8
AUDIT RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 6 2 565
N/A Variable ID: AL 5. ALFAIL
Alcohol Use
Response Options 0 1 2 3 4 8
Skips
Comments S AUDIT RISK S
Core or Secondary* M S
T S
P S
O S
During the last year, how often have you failed to do what NEVER was normally expected of you because of drinking? LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
6.
ALMORN
During the last year, how often have you needed a first drink in the morning to get yourself going after a heavy drinking session?
NEVER LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
0 1 2 3 4 8
AUDIT RISK
S
S
S
S
S
7.
ALGUILT
During the last year, how often have you had a feeling of guilt or remorse after drinking?
NEVER LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
0 1 2 3 4 8
AUDIT RISK
S
S
S
S
S
266
21. ALCOHOL USE
N/A Variable ID: AL 8. ALMEM
Alcohol Use During the last year, how often have you been unable to remember what happened the night before because of your drinking? NEVER
Response Options 0 1 2 3 4 8 0 1
Skips
Comments S AUDIT RISK S
Core or Secondary* M S
T S
P S
O S
LESS THAN MONTHLY MONTHLY WEEKLY DAILY OR ALMOST DAILY REFUSE TO ANSWER
9.
ALINJUR
Have you or someone else been injured because of your drinking?
NO YES, BUT NOT IN THE LAST YEAR
AUDIT RISK
S
S
S
S
S
YES, DURING THE LAST YEAR 3 REFUSE TO ANSWER 10. ALCNCRN Has a relative, friend, doctor, or other health care worker been concerned about your drinking or suggested you cut down? NO YES, BUT NOT IN THE LAST YEAR 8 0 1 AUDIT RISK S S S S S
YES, DURING THE LAST YEAR 3 REFUSE TO ANSWER *: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
8
Scoring of Audit-C: The AUDIT-C is scored on a scale of 0-12. Generally, the higher the score, the more likely it is that the patient’s drinking is affecting his or her safety. Each AUDIT-C question has 5 answer choices. Points given are: 0 points = Never 1 point = Less than monthly 2 points = Monthly
SECTION 2: QUESTIONNAIRE MODULES
2 6 2 767
3 points = Weekly 4 points = Daily or almost daily • In men, a score of 4 or more is considered positive, optimal for identifying hazardous drinking or active alcohol use disorders. • In women, a score of 3 or more is considered positive (same as above). • However, when the points are all from Question #1 alone (#2 & #3 are zero), it can be assumed that the patient is drinking below recommended limits and is suggested that the provider review the patient’s alcohol intake over the past few months to confirm accuracy. • Generally, the higher the score, the more likely it is that the patient’s drinking is affecting his or her safety. Scoring of Audit: Scores for questions 1 through 8 range from 0 to 4: Never= 0 points Less than monthly = 1 point Monthly = 2 points Weekly = 3 points Daily or almost dail y= 4 points Total score interpretation: A score of 8 or more is associated with harmful or hazardous drinking. A score of 13 or more in women, and 15 or more in men, is likely to indicate alcohol dependence. References: Audit, the Alcohol Use Disorders Identification Test. http://apps.who.int/iris/bitstream/10665/67205/1/WHO_MSD_MSB_01.6a.pdf (Accessed 14 July, 2016) OR
N/A 1.
AL ALCUTDN
Alcohol Use Have you ever felt you should cut down on your drinking? NO YES
Response Options 0 1 8 0 1 8
Skips
Comments S CAGE C
Core or Secondary* M C
T C
P C
O C
REFUSE TO ANSWER 2. ALCRITC Have people annoyed you by criticizing your drinking? NO YES REFUSE TO ANSWER
CAGE
C
C
C
C
C
268
21. ALCOHOL USE
3.
ALFTBAD
Have you ever felt bad or guilty about your drinking?
NO YES REFUSE TO ANSWER
0 1 8 0 1 8
CAGE
C
C
C
C
C
4.
ALEYEOP
Have you ever had a drink first thing in the morning to steady your nerves or to get rid of a hangover (eye opener)?
NO YES REFUSE TO ANSWER
CAGE
C
C
C
C
C
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
Scoring Of Cage: Item responses on the CAGE are scored 0 or 1, with a higher score, an indication of alcohol problems. A total score of 2 or greater is considered clinically significant.
References: CAGE instrument: http://www.integration.samhsa.gov/clinical-practice/sbirt/CAGE_questionaire.pdf
SECTION 2: QUESTIONNAIRE MODULES
2 6 2 969
Section
Il
22. NON-INJECTION DRUG USE
The questions in this section focus on non-injection drug use. All participants should be asked the core questions from this module. The responses from these questions are useful in determining the extent of non-injection drug use among the population being surveyed. The information can also be useful to assess associations between non-injection drug use and sexual risks. While the questions only list certain types of drugs, questions can be added to collect information on use of other non-prescribed injected drugs, given the country context. Similarly, in order to make the questions more suitable to the local context, appropriate local drug names should be used.
N/A N/A
Variable ID: DU DU1MSG
Non-Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M T P O
Thank you. Now some questions about drugs people may use to feel high. These include illicit drugs such as [insert drug names from local context].
1.
DUEVER
Did you ever use any drugs that you sniff, swallow, or inhale other than those prescribed for you?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO END OF CURRENT SECTION
RISK
C
C
C
C
C
2.
DU6MO
In the last 6 months, did you use any drugs that you sniff, swallow, or inhale, other than thoseprescribed for you?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO END OF CURRENT SECTION
RISK
C
C
C
C
C
270
22. NON-INJECTION DRUG USE
N/A N/A
Variable ID: DU DU1MSG
Non-Injection Drug Use
Response Options
Skips
Comments S Questions should be asked for each drug based on local context.
Core or Secondary* M T P O
Now I am going to ask you some questions about your use of non-injection drugs that you have smoked, inhaled, or snorted during the last 6 months. For each of these drugs, please mention how often you smoked, inhaled, or snorted any of these drugs in the last 6 months.
3.
DUMAR
In the last 6 months, did you use marijuana?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO DUHER
RISK
S
S
S
S
S
4.
DUMARF
In the last 6 months, how often did you use marijuana?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 1 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
5.
DUHER
In the last 6 months, did you use heroin?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO DUCOC
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 7 2 171
N/A 6.
Variable ID: DU DUHERF
Non-Injection Drug Use In the last 6 months, how often did you use heroin?
Response Options LESS THAN ONCE A MONTH TO 4 TIMES A MONTH TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
Skips RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 7 8
7.
DUCOC
In the last 6 months, did you use cocaine? YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO DUMET
RISK
S
S
S
S
S
8.
DUCOCF
In the last 6 months, how often did you use cocaine?
LESS THAN ONCE A MONTH TO 4 TIMES A MONTH TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
272
22. NON-INJECTION DRUG USE
N/A 9.
Variable ID: DU DUMET
Non-Injection Drug Use In the last 6 months, how often did you use crystal meth? YES NO
Response Options 1 2 7 8
Skips ‘2’,’7’ OR ‘8’→SKIP TO DUCRK RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
10.
DUMETF
In the last 6 months, how often did you use crystal meth
LESS THAN ONCE A MONTH TO 4 TIMES A MONTH TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
11.
DUCRK
In the last 6 months, did you use crack?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO DUSBL
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 7 2 373
N/A 12.
Variable ID: DU DUCRKF
Non-Injection Drug Use In the last 6 months, how often did you use crack?
Response Options LESS THAN ONCE A MONTH TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
Skips RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 7 8
13.
DUSBL
In the last 6 months, did you use speedball (cocaine and heroin mixed together)?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO DUSYN
RISK
S
S
S
S
S
14.
DUSBLF
In the last 6 months, how often did you use speedball (cocaine and heroin mixed together)?
LESS THAN ONCE A MONTH TO 4 TIMES A MONTH TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
274
22. NON-INJECTION DRUG USE
N/A 15.
Variable ID: DU DUSYN
Non-Injection Drug Use In the last 6 months, did you use synthetic YES opioid? NO
Response Options 1 2 7 8
Skips ‘2’,’7’ OR ‘8’→SKIP TO DUOTR RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
16.
DUSYNF
In the last 6 months, how often did you use synthetic opioid?
LESS THAN ONCE A MONTH TO 4 TIMES A MONTH TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
17.
DUOTR
In the last 6 months, did you use any other drug that I did not mention?
YES NO OTHER PLEASE SPECIFY _________ DON’T KNOW REFUSE TO ANSWER
1 2 3
‘2’,’7’ OR ‘8’→SKIP TO DUMAIN
RISK
S
S
S
S
S
7 8
SECTION 2: QUESTIONNAIRE MODULES
2 7 2 575
N/A 18.
Variable ID: DU DUOTRF
Non-Injection Drug Use In the last 6 months, how often did you use any other drug that I did not mention?
Response Options LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 1 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
Skips RISK
Comments S S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 7 8
19.
DUMAIN
In the last 6 months, which drug is your primary drug of choice?
HEROIN COCAINE CRACK CRYSTAL METH SYNTHETIC OPIOID OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
276
22. NON-INJECTION DRUG USE
Section
II
23. INJECTION DRUG USE
Injecting drugs with used (contaminated) syringes or needles may lead to infection with HIV, Hepatitis B or Hepatitis C virus. The responses from these questions are useful in determining the extent of injection drug use among the population being surveyed. The information also helps to determine the extent of sharing of injection paraphernalia (equipment) and indirectly measures access to and use of sterile needle/syringes by people who inject drugs (PWID). These questions only list certain types of drugs. Based on the country context, questions can be added to collect information on the use of other non-prescribed drugs that are injected. In order to make the questions more suitable to the local context, appropriate local drug names should be used.
N/A Variable ID: ID N/A ID1MSG
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
Now some questions about using a needle to inject drugs to get high. By injecting I mean sticking yourself with a needle
S
1.
IDEVER
Have you ever injected any drugs to get high?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
2’, ‘7’ OR ‘8’ →SKIP TO END OF CURRENT SECTION
RISK Use terminology from local context for injecting into vein, under the skin or into the muscle. RISK/VULNERABILITY
C
C
C
N/A
C
2.
IDOLD
Think back the very first time you injected drugs, other than # AGE OLD [ | those prescribed for you. How old were you when you first DON’T KNOW 97 injected drugs?
]
S
S
S
C
S
REFUSE TO ANSWER 98
SECTION 2: QUESTIONNAIRE MODULES
2 7 2 777
N/A Variable ID: ID 3. IDFDR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
The first time when you injected, what drug did you inject?
HEROIN COCAINE OPIUM/MORPHINE MIXTURE OF HEROIN & COCAINE METHAMPHETAMINE (CRYSTAL) CRACK/LOCAL DRUG DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8 1 2 3 4
RISK Use drugs and terminology from local context.
S
4.
IDREL
The first time you injected, what was your relationship to the person who injected you?
I INJECTED MYSELF SOMEONE I HAD SEX WITH A RELATIVE FRIEND DEALER/GALLERY OPERATOR/OTHER DRUG USER DON’T KNOW REFUSE TO ANSWER
RISK/VULNERABILITY
S
S
S
S
S
5 7 8 1 2 7 8 RISK S S S S S
5.
IDLENT
That first time you injected, did you share a needle or syringe, i.e. use a needle or syringe that someone else used before you?
YES NO DON’T KNOW REFUSE TO ANSWER
278
23. INJECTION DRUG USE
N/A Variable ID: ID 6. IDDRGET
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
The very first time you injected, how did you get the drug?
BOUGHT FROM SOMEONE 1 TRADED SEX FOR IT GIVEN TO ME FREE I WAS A DRUG DEALER DON’T KNOW REFUSE TO ANSWER 2 3 4 7 8
RISK
S
N/A
ID2MSG
In your lifetime to the best of your knowledge, did you ever inject (share needles or syringe) with any of the following? Anyone who later found out he/she had HIV? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 1 2 7 8 1 2 7 8 ‘2’, ‘7’ OR ‘8’→ SKIP TO END OF CURRENT SECTION RISK Should be considered N/A for PWID depending on eligibility criteria. RISK RISK
S
S
S
C
S
7.
IDIHIV
S
S
S
S
S
8.
IDIHEP
Anyone who had Hepatitis B and/or C?
YES NO DON’T KNOW REFUSE TO ANSWER
S
S
S
S
S
9.
ID6MOS
In the last 6 months, have you injected any illicit drugs or prescription drugs in a way other than instructed by a health care provider?
YES NO DON’T KNOW REFUSE TO ANSWER
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
2 7 2 979
N/A Variable ID: ID 10. ID6FRQ
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
Now we want you to think about your overall injecting in the LESS THAN ONCE A MONTH last 6 months. In the last 6 months, how often did you inject? 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
N/A
ID3MSG
Thank you. Now we are going to ask you some questions about specific drugs that you have injected to get high in the last 6 months.
Each 6 month question should be asked for each local drug. ‘2’, ‘7’ OR ‘8’ →SKIP TO ID6COC RISK
S
S
S
C
S
11.
ID6HRN
In the last 6 months, did you inject heroin?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
S
S
S
C
S
280
23. INJECTION DRUG USE
N/A Variable ID: ID 12. ID6HRFR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, how often did you inject heroin?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
13.
ID6COC
In the last 6 months, did you inject powdered cocaine?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO ID6MIX
RISK
S
S
S
C
S
14.
ID6COFR
In the last 6 months, how often did you inject powdered cocaine?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 8 2 181
N/A Variable ID: ID 15. ID6MIX
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M T P O
In the last 6 months, did you use any drug combinations (this includes drugs mixed together such as cocaine+heroin, ATS+heroin or other combination)?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO ID6CRCK
16.
ID6MIXO
In the last 6 months, how often did you use this particular LESS THAN ONCE A drug combination (e.g. cocaine+heroin, ATS+heroin or other MONTH combination)? 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 7 8 ‘2’, ‘7’ OR ‘8’ →SKIP TO ID6METH RISK S S S C S
17.
ID6CRCK
In the last 6 months, did you inject crack?
YES NO DON’T KNOW REFUSE TO ANSWER
282
23. INJECTION DRUG USE
N/A Variable ID: ID 18. ID6CRFR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, how often did you inject crack?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
19.
ID6METH In the last 6 months, did you inject methamphetamine (crystal)?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’, ‘7’ OR ‘8’ →SKIP TO ID6SBALL
RISK
S
S
S
C
S
20.
ID6MTFR
In the last 6 months, how often did you inject methamphetamine (crystal)?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 8 2 383
N/A Variable ID: ID 21. ID6SBALL
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, did you inject speedball – heroin and cocaine combined?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO ID6OTHR
RISK
S
22.
ID6SBFR
In the last 6 months, how often did you inject speedball?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
C
S
23.
ID6OTHR
In the last 6 months, did you inject any other drugs that I have not mentioned?
YES PLEASE SPECIFY_________ NO DON’T KNOW REFUSE TO ANSWER
1
‘2’,’7’ OR ‘8’→SKIP TO IDNLOC
RISK
S
S
S
C
S
2 7 8
284
23. INJECTION DRUG USE
N/A Variable ID: ID 24. ID6OTFR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, how often did you inject any other drugs?
LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
25.
IDSYN
In the last 6 months, did you inject synthetic opioid?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO IDMAIN
RISK
S
S
S
S
S
26.
IDSYNF
In the last 6 months, how often did you inject synthetic opioid? LESS THAN ONCE A MONTH 1 TO 4 TIMES A MONTH 2 TO 7 TIMES WK, ONCE A DAY 2 TO 3 TIMES A DAY 5 OR MORE TIMES A DAY DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 8 2 585
N/A Variable ID: ID 27. IDMAIN
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
In the last 6 months, which drug is your primary drug of choice?
HEROIN COCAINE CRACK CRYSTAL METH SYNTHETIC OPIOID OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
RISK
S
286
23. INJECTION DRUG USE
N/A Variable ID: ID 28. IDNLOC
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months when you injected, where did you get your needles/syringes from? CHECK ALL THAT APPLY.
PHARMACY/CHEMIST DRUG STORE/OTHER STORE DOCTOR’S OFFICE, CLINIC OR HOSPITAL, OTHER HEALTH AGENCY, OR HIV PREVENTION PROGRAM
COVERAGE A Use locations from local context.
S
B
MARKET PLACE OR STREET C VENDOR DRUG WORKER OR AGENCY/OUTREACH WORKER OR STREET UNIT SEX PARTNER, FRIEND, ACQUAINTANCE, RELATIVE DRUG DEALER OR OTHER DRUG USERS NEEDLE/SYRINGE EXCHANGE PROGRAM OTHER DON’T KNOW REFUSE TO ANSWER 29. IDNSTE Are sterile needles and syringes available when you need them? YES NO DON’T KNOW REFUSE TO ANSWER
D E F G X Y Z 1 2 7 8 S S S S C
SECTION 2: QUESTIONNAIRE MODULES
2 8 2 787
N/A Variable ID: ID 30. IDNEWN
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months when
NEVER
1 2 3 4 5 7 8 1 2 3 4
RISK & PREVENTION
S
you injected, how often did you use a new, sterile needle? RARELY By new, sterile needle, we mean a needle never used before HALF OF THE TIME by anyone, even you MOST OF THE TIME ALWAYS DON’T KNOW REFUSE TO ANSWER 31. IDBUYN Why do you not always use a new needle or syringe? NOT AVAILABLE DIFFICULT TO FIND EXPENSIVE PEER PRESSURED TO SHARE BELIEVE THE PERSON SHARING WITH HAS THE SAME HIV STATUS I ALWAYS USE A NEW NEEDLE OR SYRINGE DON’T KNOW REFUSE TO ANSWER ID4MSG Next, I’m going to ask you about your injecting behaviors in the last 6 months.
S
S
S
C
S
5 6 7 8 S S S C S
288
23. INJECTION DRUG USE
N/A Variable ID: ID 32. IDINJT
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, how often did you use needles that someone else had already injected with?
NEVER RARELY HALF OF THE TIME MOST OF THE TIME ALWAYS DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
‘1’,’7’,’8’→ SKIP RISK TO IDCOOK
S
33.
IDNIMN
In the last 6 months, with how many people did you use a needle after they injected with it?
# PEOPLE [
|
]
RISK
S
S
S
C
S
DON’T KNOW 97 REFUSE TO ANSWER 98
34.
IDCOOK
In the last 6 months when you injected, how often did you use cookers, cottons, or tourniquets of water that someone else had already used?
NEVER RARELY HALF OF THE TIME MOST OF THE TIME ALWAYS DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
‘1’,’7’,’8’→ SKIP RISK TO IDSYR
S
S
S
C
S
35.
IDCKMN
In the last 6 months, with how many people did you use the # PEOPLE [ | ] same cooker, cotton, or water (water for rinsing needles or DON’T KNOW preparing drugs)? REFUSE TO ANSWER
RISK 7 8
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 8 2 989
N/A Variable ID: ID 36. IDSYR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
In the last 6 months when you injected, how often did you use drugs that had been divided with a syringe that someone else had already injected with?
NEVER RARELY HALF OF THE TIME MOST OF THE TIME ALWAYS DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 7 8
‘1’,’7’,’8’→ SKIP RISK TO IDPOL
S
37.
IDPOL
In the last 6 months, did the police or other authorities ever YES confiscate any injecting equipment (needle, syringe, etc.) NO from you? DON’T KNOW REFUSE TO ANSWER
VULNERABI LITY
S
S
S
C
S
ID5MSG
Now I’d like you to think about the last time you injected with someone. By “injecting with someone,” I mean you shared drugs or drug injecting equipment, or both with at least one other person.
S
S
S
C
S
38.
IDLAST
When was the last time you injected with one or more people?
I NEVER INJECTED WITH SOMEONE LESS THAN 1 MONTH AGO 1 OR MORE MONTHS AGO BUT LESS THAN 12 MONTHS AGO 1 OR MORE YEARS AGO DON’T KNOW REFUSE TO ANSWER
1 2
‘1’, 7 or 8→SKIP to IDSLF ‘2’ → SKIP to IDLASTD ‘3’ → SKIP to IDLASTM ‘4’ → SKIP to IDLASTY
RISK
S
S
S
C
S
3 4 7 8
290
23. INJECTION DRUG USE
N/A Variable ID: ID 39. IDLASTD
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
How many DAYS ago did you last inject with someone or with more than one person? By “injecting with someone”, I mean you shared drugs or drug injecting equipment, or both with at least one other person.
# DAYS [
|
]
RISK
S
DON’T KNOW 97 REFUSE TO ANSWER 98 MINIMUM: 1 MAXIMUM:30
40.
IDLASTM
How many MONTHS ago did you last inject with someone or with more than one person? By “injecting with someone”, I mean you shared drugs or drug injecting equipment, or both with at least one other person.
# MONTHS [
|
]
RISK
S
S
S
C
S
DON’T KNOW 97 REFUSE TO ANSWER 98 MINIMUM: 1 MAXIMUM: 11
41.
IDLASTY
How many YEARS ago did you last inject with someone or with more than one person? By “injecting with someone”, I mean you shared drugs or drug injecting equipment, or both with at least one other person.
# YEARS [
|
]
RISK
S
S
S
C
S
DON’T KNOW 97 REFUSE TO ANSWER 98 MINIMUM: 1 MAXIMUM: IDOLD (age first injected)
42.
IDLASTPL
Where did you last inject with someone?
HOME STREET SUPERVISED INJECTION SITE PRISON OTHER DON’T KNOW
1 2 3 4 5 97
RISK
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 9 2 191
N/A Variable ID: ID 43. IDREL
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
The last time you injected with someone, what was your relationship with the person or people? CHECK ALL THAT APPLY.
SEX PARTNER
A
RISK ADAPT RESPONSES TO LOCAL CONTEXT.
S
FRIEND OR ACQUANTANCE B RELATIVE NEEDLE OR DRUG DEALER STRANGER OTHER DON’T KNOW REFUSE TO ANSWER C D E X Y Z 1 2 7 8 1 2 7 8 1 2 7 8
44.
IDNDSTR
The last time you injected with another person did you use a new sterile needle to inject? By new, sterile needle, we mean a needle never used before by anyone, even you.
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
C
S
45.
IDLSTND
The last time you injected with another person, did you share a needle with that person (i.e. use it after they had used it)?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
C
S
46.
IDWATR
The last time you injected with someone, did you use a cooker, cotton, or water that that that person had already used?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
C
S
292
23. INJECTION DRUG USE
N/A Variable ID: ID 47. IDDRSR
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
The last time you injected with someone, did you use drugs that had been divided with a syringe that a person had already injected with?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
RISK
S
48.
IDSLF
The last time you injected drugs by yourself, did you use a new sterile needle to inject? By new, sterile needle, we mean needle never been used before by anyone, even you.
YES NO DON’T KNOW REFUSE TO ANSWER
S
S
S
C
S
ID6MSG
Next, we are going to ask you about drug treatment programs. These include out-patient, inpatient, residential, detox, and methadone maintenance treatment programs Are you aware of any drug treatment program that may be available for you? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 1 2 7 8 1 2 7 8 ‘1’→ SKIP TO ID6METH COVERAGE COVERAGE ‘2’,’7’ OR ‘8’→ SKIP TO IDTRY COVERAGE
S
S
S
C
S
49.
IDAWR
S
S
S
C
S
50.
IDDRTR
Have you ever been in a drug treatment program?
YES NO DON’T KNOW REFUSE TO ANSWER
S
S
S
C
S
51.
ID6TRT
In the last 6 months, have you been in a drug treatment program?
YES NO DON’T KNOW REFUSE TO ANSWER
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 9 2 393
N/A Variable ID: ID 52. ID6TRYT
Injection Drug Use
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P C
O S
In the last 6 months, did you try to get into a drug treatment program but were unable to?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 3 7 8 1 2 7 8 1 2 7 8
‘2’,’7’ OR ‘8’ →SKIP TO IDTRY
COVERAGE
S
53.
ID6MEDS In the last 6 months, did you receive medication such as methadone or buprenorphine for your drug dependency?
YES NO - BUT I WOULD LIKE TO NO - I DON’T KNOW ABOUT SUCH TREATMENT DON’T KNOW REFUSE TO ANSWER
‘1’→ SKIP TO IDSEX
COVERAGE & PREVENTION
S
S
S
C
S
54.
IDTRY
In the last 6 months, did you try to reduce or give up drug use?
YES NO DON’T KNOW REFUSE TO ANSWER
PREVENTION
S
S
S
C
S
55.
IDSEX
In the last 6 months, did you sell sex for drugs?
YES NO DON’T KNOW REFUSE TO ANSWER
RISK
S
S
S
C
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
294
23. INJECTION DRUG USE
Section
II
24. DRUG OVERDOSE
Overdose is a risk when using many psychoactive drugs – but is much more common for some drugs (such as opioids) than others (such as cannabis and nicotine). It is especially common for people who use combinations of substances and those who inject drugs. The questions in this section focus on overdose experiences, if any, of the participant, if the participant had seen or witnessed another person go thru overdose, or s/he herself/himself experienced overdose. Information collected through this module can help understand the extent of overdose, if any, and possibly help in the development of interventions to minimize overdose, as in many cases, overdose can lead mortality.
N/A N/A
Variable ID: OD OD1MSG
Overdose
Response Options
Skips
Comments S
Core or Secondary* M T P O
Now we are going to ask you some questions about overdosing on drugs. Overdosing means that someone lost consciousness or stopped breathing as a result of taking drugs.
1.
ODEVR
Have you ever overdosed on narcotics to the point where you lost consciousness?
YES NO DON’T REMEMBER REFUSE TO ANSWER
1 2 7 8
‘2’, ‘7’ OR ‘8’ →SKIP TO ODTRT
RISK Insert local term for “passed out”.
S
S
S
C
S
SECTION 2: QUESTIONNAIRE MODULES
2 9 2 595
N/A 2.
Variable ID: OD ODTIM
Overdose How many times has this happened? NONE
Response Options 1 2 3 4 7 8
Skips
Comments S RISK S
Core or Secondary* M S
T S
P C
O S
1 TO 5 TO 10 MORE THAN 10 DON’T KNOW REFUSE TO ANSWER
3.
ODLSM
Please think about the last time you overdosed on narcotics. How long ago did that occur?
WITHIN 7 DAYS 1 BETWEEN 7 AND 30 DAYS BETWEEN 1 AND 6 MONTHS BETWEEN 6 AND 12 MONTHS OVER 12 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
RISK
S
S
S
C
S
4.
ODOSNAR
What narcotics or other substances did you use (injected and/or non- injected) on that occasion?
SPECIFY: _______ DON’T KNOW 9997 REFUSE TO ANSWER 9998
RISK ADAPT RESPONSES TO LOCAL CONTEXT
S
S
S
S
S
296
24. DRUG OVERDOSE
N/A 5.
Variable ID: OD ODSALN
Overdose Were you alone the last time you overdosed? YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 7 8 1 2 7 8 1 2 7 8
Skips
Comments S RISK S
Core or Secondary* M S
T S
P S
O S
DON’T REMEMBER REFUSE TO ANSWER 6. ODSHLP Did you receive help from anyone the last time you overdosed? YES NO DON’T REMEMBER REFUSE TO ANSWER 7. ODTRT If you were to seek medical help for an overdose of narcotics drugs, do you believe that you would be treated well? YES NO DON’T REMEMBER REFUSE TO ANSWER 8. ODLAW If you were to seek medical help for an overdose of narcotics drugs, do you believe that you would be reported to law enforcement authorities? YES NO DON’T REMEMBER REFUSE TO ANSWER 9. ODSPR Have you ever seen another person overdose on narcotic drugs to the point where they lost consciousness or stopped breathing? YES NO DON’T REMEMBER REFUSE TO ANSWER
COVERAGE
S
S
S
S
S
KA
S
S
S
S
S
KA
S
S
S
S
S
‘2’ OR ‘7’ OR ‘8’ →SKIP TO END OF CURRENT SECTION
RISK
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
2 9 2 797
N/A 10.
Variable ID: OD ODTM
Overdose How many times has that happened? ONCE
Response Options 1 2 3 4 7 8 1 2 3 4 5 7 8
Skips
Comments S RISK S
Core or Secondary* M S
T S
P S
O S
2 TO 5 TIMES TO 10 TIMES MORE THAN 10 TIMES DON’T REMEMBER REFUSE TO ANSWER 11. ODLST Please think about the last time that happened. How long ago did that occur? WITHIN 7 DAYS BETWEEN 7 AND 30 DAYS BETWEEN 30 DAYS AND 6 MONTHS BETWEEN 6 MONTHS AND 12 MONTHS OVER 12 MONTHS AGO DON’T REMEMBER REFUSE TO ANSWER
RISK
S
S
S
S
S
298
24. DRUG OVERDOSE
N/A 12.
Variable ID: OD ODDO
Overdose What did you and the others present do? CHECK ALL THAT APPLY.
Response Options DID NOTHING A
Skips IF ‘A’, ‘Y’, OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
Comments S PREVENTION & COVERAGE ADAPT RESPONSES TO LOCAL CONTEXT. S
Core or Secondary* M S
T S
P S
O S
DIDN’T KNOW WHAT TO DO B MOUTH TO MOUTH RESUSCITATION USED ICE, THREW COLD WATER ON OR SHOUTED AT THE PERSON INJECTED ANOTHER DRUG (e.g. NARCAN/ NALOX/NALTREXONE/ OTHER) TOOK THE PERSON TO HOSPITAL/MEDICAL FACILITY CALLED AMBULANCE/ MEDICAL PERSONNEL LEFT THE PLACE IMMEDIATELY HIT/SLAPPED/PINCHED THE PERSON DRAWING BLOOD FROM THE PERSON TO GET RID OF THE DRUG OTHER DON’T KNOW REFUSE TO ANSWER C
D
E
F G H I
J X Y Z
SECTION 2: QUESTIONNAIRE MODULES
2 9 2 999
N/A 13.
Variable ID: OD ODDTH
Overdose How many people have you known who died of a drug overdose? NONE
Response Options 1 2 3 4 7 8 1 2 7 8 1 2 7 8 1 2 7 8
Skips
Comments S RISK S
Core or Secondary* M S
T S
P S
O S
1 TO 5 TO 10 MORE THAN 10 DON’T KNOW REFUSE TO ANSWER
14.
ODEDUC
Have you ever received education or training on how to revive someone after an overdose?
YES NO DON’T KNOW REFUSE TO ANSWER
2’ OR ‘7’ OR ‘8’ →SKIP TO END OF CURRENT SECTION
COVERAGE
S
S
C
S
S
15.
ODNLX
Have you ever received education or training on naloxone as an antidote to opioids?
YES NO DON’T KNOW REFUSE TO ANSWER
COVERAGE
S
S
S
S
S
16.
ODNAV
Have you been able to get naloxone when you wanted?
YES NO DON’T KNOW REFUSE TO ANSWER
COVERAGE
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
300
24. DRUG OVERDOSE
Section
II
25. SEXUALLY TRANSMITTED INFECTIONS
This module measures characteristics related to self-reported sexually transmitted infections (STI). The terms sexually transmitted infections and sexually transmitted disease are used interchangeably. However, the term STI has been used more recently because people may be infected but may not show symptoms or the infection may not turn into a disease. The first questions ask about symptoms associated with STI. If a participant responds “yes” to one of the symptoms, we assume that person had an STI in the last 12 months. N/A N/A
Variable ID: ST STD1MSG
Sexually Transmitted Infections Thank you. Now we will ask a few questions about your sexual health.
Response Options
Skips ONLY ASK IF KEYPOP= ‘1’,, ‘2’, or 3’
Comments S
Core or Secondary* M T P O
1.
STABNM
In the last 12 months, have you had an abnormal discharge from your penis?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
MALES/TG WOMEN ONLY.
RISK
C
C
C
C
C
2.
STULCM
In the last 12 months, have you had an ulcer or sore on or near YES your penis? NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
MALES/TG WOMEN ONLY.
RISK
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
3 0 3 101
N/A 3.
Variable ID: ST STPAINUR
Sexually Transmitted Infections In the last 12 months, have you had pain on urination? YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 7 8 1 2 7 8 1 2 7 8
Skips MALES/TG WOMEN ONLY.
Comments S RISK C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 4. STULCPM In the last 12 months, have you found an ulcer or sore on or near your partner’s penis or vagina? YES NO DON’T KNOW REFUSE TO ANSWER 5. STABNF During the last 12 months, have had an abnormal discharge from your vagina or experienced pelvic pain? This may include an unusual smell, color, or texture. YES NO DON’T KNOW REFUSE TO ANSWER 6. STULCF In the last 12 months, have you had an ulcer or sore on or near YES your vagina? NO DON’T KNOW REFUSE TO ANSWER 7. STULCPF In the last 12 months, have you found an ulcer or sore on or near your partner’s penis? YES NO DON’T KNOW REFUSE TO ANSWER
MALES/TG WOMEN ONLY.
RISK
S
S
S
S
S
FEMALES ONLY.
RISK
S
N/A
N/A
S
S
FEMALES ONLY.
RISK
C
N/A
N/A
C
C
FEMALES ONLY.
RISK
S
S
S
S
S
302
25. SEXUALLY TRANSMITTED INFECTIONS
N/A 8.
Variable ID: ST STPROV
Sexually Transmitted Infections Did you see a healthcare provider because of these problems? YES NO
Response Options 1 2 7 8
Skips ‘2’,’7’,’8’→ STDIAG SKIP IF STABNM=’2’, ‘7’, OR ‘8’ AND STULCM=’2’, ‘7’, OR ‘8’ AND STPAINUR=’2’, ‘7’, OR ‘8’ AND STULCPM=’2’, ‘7’, OR ‘8’ AND STABNF=’2’, ‘7’, OR ‘8’ AND STULCF=’2’, ‘7’, OR ‘8’ AND STULCPF=’2’, ‘7’, OR ‘8’
Comments S PREVENTION C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER
9.
STLONG
How long did it take you to go see the healthcare provider since the appearance of these symptoms?
LESS THAN ONE WEEK MORE THAN ONE WEEK LESS THAN ONE MONTH MORE THAN ONE MONTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
PREVENTION
S
S
S
S
S
10.
STDIAG
In the last 12 months, did a healthcare provider tell you that you had a sexually transmitted disease, other than HIV?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
PREVENTION
C
C
C
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 0 303
N/A 11.
Variable ID: ST STPHARM
Sexually Transmitted Infections Did you get treatment for these problems? YES NO
Response Options 1 2 7 8
Skips ‘2’,’7’,’8’→SKIP TO STHPV SKIP IF STABNM=’2’, ‘7’, OR ‘8’ AND STULCM=’2’, ‘7’, OR ‘8’ AND STPAINUR=’2’, ‘7’, OR ‘8’ AND STULCPM=’2’, ‘7’, OR ‘8’ AND STABNF=’2’, ‘7’, OR ‘8’ AND STULCF=’2’, ‘7’, OR ‘8’ AND STULCPF=’2’, ‘7’, OR ‘8’
Comments S PREVENTION C
Core or Secondary* M C
T C
P S
O S
DON’T KNOW REFUSE TO ANSWER
12.
STFACI
Where did you go to get treatment? Did you go to a public facility, a private facility or a pharmacy?
PUBLIC CLINIC/HOSPITAL PRIVATE CLINIC/HOSPITAL PHARMACY OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 6 7 8
SKIP IF STABNM=’2’, ‘7’, OR ‘8’ AND STULCM=’2’, ‘7’, OR ‘8’ AND STPAINUR=’2’, ‘7’, OR ‘8’ AND STULCPM=’2’, ‘7’, OR ‘8’ AND STABNF=’2’, ‘7’, OR ‘8’ AND STULCF=’2’, ‘7’, OR ‘8’ AND STULCPF=’2’, ‘7’, OR ‘8’
PREVENTION
C
C
C
C
C
304
25. SEXUALLY TRANSMITTED INFECTIONS
N/A 13.
Variable ID: ST STHPV
Sexually Transmitted Infections There is a vaccine to protect people against human papilloma virus. This virus can cause cancer in the cervix, penis or anus. Did you get this vaccination? YES NO
Response Options 1 2 7 8 1 2 7 8
Skips
Comments S COVERAGE S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER 14. STCHE Do you know where you can go for a STI/sexual health check-up? YES NO DON’T KNOW REFUSE TO ANSWER
SKIP IF STABNM=’2’, ‘7’, OR ‘8’ AND STULCM=’2’, ‘7’, OR ‘8’ AND STPAINUR=’2’, ‘7’, OR ‘8’ AND STULCPM=’2’, ‘7’, OR ‘8’ AND STABNF=’2’, ‘7’, OR ‘8’ AND STULCF=’2’, ‘7’, OR ‘8’ AND STULCPF=’2’, ‘7’, OR ‘8’
C
C
C
C
C
15.
STCOM
Would you be comfortable using this service if you want to have an STI/sexual health checkup?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
C
C
C
C
C
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
3 0 3 505
Section
Il N/A
26. HEPATITIS B AND C
This module is intended to capture information on participants’ awareness of Hepatitis (B and C) and also whether the participants have been tested for Hepatitis and received treatment.
Variable ID: HP HP1MSG
HBV And HCV
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A
Next, we’d like to ask you some questions about your health including hepatitis, meaning inflammation of the liver. Have you ever received the Hepatitis B vaccine? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 1 2 3 7 8 COVERAGE S S S S S ‘2’,’7’ OR ‘8’→ SKIP TO HPBTEST COVERAGE C C C C S
1.
HPVAC
2.
HPSHOT
How many shots for Hepatitis B vaccine did you receive?
ONE TWO THREE DON’T KNOW REFUSE TO ANSWER
306
26. HEPATITIS B AND C
N/A
Variable ID: HP HPBTEST
HBV And HCV
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P C
O S
3.
Have you ever been tested for Hepatitis B?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO HPCTEST
COVERAGE
C
4.
HPBTME
When did you have your most recent Hepatitis B test?
LESS THAN 12 MONTHS AGO MORE THAN 12 MONTHS AND LESS THAN FIV YEARS AGO MORE THAN FIVE YEARS AGO DON’T KNOW REFUSE TO ANSWER
1
COVERAGE
C
C
C
C
S
2 3 7 8
5.
HPRSL
What was the result of your last Hepatitis B test?
POSITIVE NEGATIVE DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP TO HPCTEST
RESULT
S
S
S
S
S
6.
HPTREAT
Have you ever taken medicine to treat your Hepatitis B infection?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 0 3 707
N/A
Variable ID: HP HPCTEST
HBV And HCV
Response Options
Skips
Comments S
Core or Secondary* M C
T C
P C
O S
7.
Have you ever been tested for Hepatitis C?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→SKIP COVERAGE TO END OF CURRENT SECTION
C
8.
HPCTME
When did you have your most recent Hepatitis C test?
LESS THAN 12 MONTHS AGO MORE THAN 12 MONTHS AND LESS THAN FIVE YEARS AGO MORE THAN FIVE YEARS AGO DON’T KNOW REFUSE TO ANSWER
1
COVERAGE
C
C
C
C
S
2 3 7 8
9.
HPRSL
What was the result of your last Hepatitis C test?
POSITIVE NEGATIVE DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→ SKIP RESULT TO END OF CURRENT SECTION
S
S
S
C
S
10.
HPMED
Have you ever taken medicine to treat your Hepatitis C infection?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
S
S
S
C
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
308
26. HEPATITIS B AND C
Section
II
27. SERVICES UPTAKE
A key goal of HIV prevention and care is to identify HIV-infected persons and link them to services for care, treatment and prevention services. This module measures the uptake of HIV-related services from community outreach to linkage to care, ART initiation and retention.
Sex workers, men who have sex with men, transgender persons and people who inject drugs may not always identify themselves as such when utilizing health services. Surveys represent an additional data source to help describe the uptake of services by key affected populations. Surveys have the advantage of measuring service uptake at the population (rather than clinic) level, sampling both respondents who do and do not access services. Surveys can thus characterize people who do not access services and describe barriers to service uptake. The many strengths of this module may be limited by small sample sizes, low HIV prevalence, and/or poor uptake of services that can lead to small numbers in analysis. Further, respondents with poor health literacy may find some questions challenging to answer. While this module is extensive, the number of questions considered “core” is small, varying by key population group and skip patterns. Depending on the final questionnaire structure, some questions listed here may be better placed within other modules. Some questions or answer options may need to be altered to reflect national guidelines or to program goals. Certain terms (e.g., “STD” or “cotrimoxazole”) may need to be replaced with better known local terms. For medication-related questions (e.g., ARVs) consider showing participants photos to help identify these medications. Variable ID: CS OR N/A CS1MSG
Services Uptake Outreach
Response Options
Skips (SUBSECTION)
Comments S
Core or Secondary* M T P O
We will now ask you about services that peer educators or outreach workers may have given you. A peer educator is someone like you who has been trained in HIV. An outreach worker is someone employed by an organization, government or private agency, who might provide these same services to people like you.
SECTION 2: QUESTIONNAIRE MODULES
3 0 3 909
Variable ID: CS 1. CSOREV
Services Uptake Has a peer educator or outreach worker ever talked to you YES about HIV? NO
Response Options 1 2 7 8
Skips
Comments S C
Core or Secondary* M C
T C
P C
O C
‘2’,’7’ OR COVERAGE ‘8’→SKIP TO NEXT SUBSECTION
DON’T KNOW REFUSE TO ANSWER
2.
CSORTI
How long ago did a peer educator or outreach worker talk to you about HIV?
IN THE LAST 30 DAYS IN THE LAST 3 MONTHS IN THE LAST YEAR LONGER THAN A YEAR AGO DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8 A B C D E X Y Z IF ‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
COVERAGE
C
C
C
C
C
3.
CSORPR
The last time you met a peer educator or outreach worker, NOTHING what items or things did you receive? CONDOMS CHECK ALL THAT APPLY. LUBRICANTS PAMPHLET OR BROCHURE MEDICINES OTHER DON’T KNOW REFUSE TO ANSWER
COVERAGE & PREVENTION
S
S
S
N/A
S
310
27. SERVICES UPTAKE
Variable ID: CS 4. CSORID
Services Uptake
Response Options A B C D E X Y Z
Skips IF ‘A, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
Comments S COVERAGE & PREVENTION N/A
Core or Secondary* M N/A
T N/A
P S
O N/A
The last time you met a peer educator or outreach worker, NOTHING what items or things did you receive? NEEDLES, SYRINGES, CHECK ALL THAT APPLY. SAFE INJECTION KITS PAMPHLET OR BROCHURE CONDOMS MEDICINES OTHER DON’T KNOW REFUSE TO ANSWER
5.
CSORSV
The last time you met a peer educator or outreach worker, NOTHING what services did you receive? HIV TESTING CHECK ALL THAT APPLY. TRAINING ON CONDOM USE COUNSELING ON RISK REFERRAL OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
IF ‘A, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
COVERAGE & PREVENTION
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 1 3 111
Variable ID: CS A. N/A
Services Uptake Counseling and Testing
Response Options
Skips (SUBSECTION)
Comments S
Core or Secondary* M T P O
CT CS2MSG
We will now ask you a few questions about HIV testing. Remember that your responses will be kept confidential and will not be shared with anyone.
6.
CSCRET
Do you know where you can go If You want to receive an HIV test?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
C
C
C
C
C
7.
CSCTCO
Would you feel comfortable with using this service if you wanted to be tested for HIV?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
C
C
C
C
C
8.
CSCTEV
Have you ever tested for HIV?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
’1’→SKIP TO CSCTPL1
COVERAGE
C
C
C
C
C
312
27. SERVICES UPTAKE
Variable ID: CS 9. CSCTNV
Services Uptake What is the main reason you have never tested for HIV?
Response Options I FEEL I AM NOT AT RISK FOR HIV FEAR OF POSITIVE RESULT NO MONEY TO GET TESTED NO TIME TO GET TESTED STIGMA BY HEALTH CARE WORKERS OTHER DON’T KNOW REFUSE TO ANSWER
Skips SKIP TO END OF CURRENT SECTION
Comments S C
Core or Secondary* M C
T C
P C
O C
1 2 3 4 5 6 7 8
PREVENTION
10.
CSCTPL1
Please think about to the last time you tested for HIV. Did you go somewhere, did someone come to you or did you test yourself?
I WENT THERE THEY CAME TO ME I TESTED MYSELF DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
‘2’→SKIP TO CSCTPL3; ‘3’→ SKIP TO CSCTRS1
COVERAGE & PREVENTION
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
3 1 313
Variable ID: CS 11. CSCTPL2
Services Uptake Where did you have your last HIV test?
Response Options TESTING AND COUNSELING CENTER HEALTH CLINIC, HOSPITAL, OR SIMILAR OUTREACH/MOBILE TESTING OTHER DON’T KNOW REFUSE TO ANSWER
Skips ‘3’ → SKIP TO CSCTPL3 ‘6’ → SKIP TO CSCTPL4 ‘1’, ‘2’, ‘7’, ‘8’ → SKIP TO CSCTWHO
Comments S COVERAGE C
Core or Secondary* M C
T C
P C
O C
1
2 3 6 7 8 1 2 3 6 7 8 1 2 3 6 7 8
12.
CSCTPL3
Where did you get tested?
IN MY HOME AT MY WORK WHERE I HANG OUT OTHER DON’T KNOW REFUSE TO ANSWER
‘1’, ‘2’, ‘3’, ‘7’, ‘8’ →SKIP TO CSCTWHO
COVERAGE
C
C
C
C
C
13.
CSCTPL4
You said “Other”. Where was that?
SCHOOL PRISON CHURCH OTHER DON’T KNOW REFUSE TO ANSWER
COVERAGE
C
C
C
C
C
314
27. SERVICES UPTAKE
Variable ID: CS 14. CSCTWHO
Services Uptake Who tested you? Choose the best answer.
Response Options GOVERNMENT WORKER NGO/CBO NAME A NGO/CBO NAME B NGO/CBO NAME C OTHER PRIVATE OR COMMERCIAL PLACE OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S COVERAGE ADAPT RESPONSES BASED ON LOCAL CONTEXT. S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 6 7 8
15.
CSCTRS1
Why did you last test for HIV?
HEALTH CARE OR OUTREACH OFFERED TEST I JUST WANTED TO KNOW FELT AT RISK OR SICK
1 2 3
’2’, ’3’, ’4’, ‘7’, ‘8’ → SKIP TO CSCTYY OR CSCTTI
PREVENTION
S
S
S
S
S
GOT A NEW PARTNER 4 OTHER DON’T KNOW REFUSE TO ANSWER 6 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 1 3 515
Variable ID: CS 16. CSCTRS2
Services Uptake You said “Other”. Which of the following best fit your reason to test?
Response Options EMPLOYER ASKED ME TO TEST PRE-MARITAL TESTING PARTNER ASKED ME TO TEST OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1 2 3 6 7 8
17.
CSCTYY
In what year was your last HIV test?
| | | | | MIN: 1985 MAX: CURRENT YEAR DON’T KNOW REFUSE TO ANSWER 7 8
OPTION A
COVERAGE & PREVENTION CHOOSE EITHER OPTION A (CSCTYY AND CSCTMM) OR OPTION B (CSCTTI)
S
S
S
S
S
18.
CSCTMM
In which month did you take your last HIV test?
| | | RANGE: 1-12 IF CSCTYY = CURRENT YEAR, THEN MAX= CURRENT MONTH DON’T KNOW REFUSE TO ANSWER 7 8
OPTION A SKIP IF CURRENT YEAR = CSCTYY >1
COVERAGE CHOOSE EITHER OPTION A (CSCTYY AND CSCTMM) OR OPTION B (CSCTTI)
S
S
S
S
S
316
27. SERVICES UPTAKE
Variable ID: CS 19. CSCTTI
Services Uptake When did you last test for HIV?
Response Options IN THE LAST 6 MONTHS BETWEEN 7-12 MONTHS AGO MORE THAN 12 MONTHS AGO DON’T KNOW REFUSE TO ANSWER
Skips OPTION B
Comments S COVERAGE & PREVENTION CHOOSE EITHER OPTION A (CSCTYY AND CSCTMM) OR OPTION B (CSCTTI) C
Core or Secondary* M C
T C
P C
O C
1 2 3 7 8 1 2
20.
CSCTRS
What was the result of your last test?
POSITIVE NEGATIVE UNCLEAR / NEITHER POSITIVE OR NEGATIVE DID NOT RECEIVE RESULT DON’T KNOW REFUSE TO ANSWER
‘1’→CSCFP
PREVENTION
C
C
C
C
C
3 4 7 8 SKIP TO CSDSMM KA & VULNERABILITY S S S S
21.
CSCTINF
How likely do you think it is that you will get infected with HIV in the next year?
EXTREMELY UNLIKELY 1 SOMEWHAT UNLIKELY SOMEWHAT LIKELY EXTREMELY LIKELY DON’T KNOW REFUSE TO ANSWER 2 3 4 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 1 3 717
Variable ID: CS 22. CSCFP
Services Uptake Was the last test your first positive test result? This will be the very first HIV positive test result you have received. YES NO
Response Options 1 2 7 8
Skips ‘1’→SKIP TO CSCFPY ’7’ OR ‘8’→SKIP TO CSCTTL
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
23.
CSCFPY
What year did you first test positive?
|
|
|
|
|
OTHER
S
S
S
S
S
MIN: 1900 MAX: CURRENT YEAR DON’T KNOW REFUSE TO ANSWER 24. CSCFPM What month did you first test positive? | | | 7 8 OTHER S S S S S
RANGE: 1-12 IF CSCTYY = CURRENT YEAR, THEN MAXIMUM = CURRENT MONTH DON’T KNOW REFUSE TO ANSWER 25. CSCFPY Before your first positive test, did you ever have a negative YES HIV test? NO DON’T KNOW REFUSE TO ANSWER 7 8 1 2 7 8 ‘2’,’7’,’8’ → SKIP TO CSDSTL OTHER S S S S S
318
27. SERVICES UPTAKE
Variable ID: CS 26. CSCFPY
Services Uptake What year did you last test HIV negative? |
Response Options | | | |
Skips
Comments S OTHER Probing the date of the last HIV neg test provides information for estimating HIV incidence in a casebased HIV incidence surveillance system S
Core or Secondary* M S
T S
P S
O S
MINIMUM: 1985 MAXIMUM: CURRENT YEAR DON’T KNOW REFUSE TO ANSWER 7 8
27.
CSCFPM
What month did you last test HIV negative?
|
|
|
OTHER
S
S
S
S
S
RANGE: 1-12 IF CSCTYY = CURRENT YEAR, THEN MAXIMUM = CURRENT MONTH DON’T KNOW REFUSE TO ANSWER 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 1 3 919
Variable ID: CS B. 28.
Services Uptake Disclosure Of the following people, who have you told that you are HIV positive? CHECK ALL THAT APPLY.
Response Options
Skips (SUBSECTION)
Comments S
Core or Secondary* M T P O
DS CSDSTL
NO ONE SPOUSE/SEX PARTNER DOCTOR FRIEND FAMILY MEMBER OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
ONLY ASK IF CSCTRS=’1’ ‘A’ → CSDSMM
PREVENTION & VULNERABILITY
C
C
C
C
C
29.
CSDSLK
Did the person who tested you do any of the following? CHECK ALL THAT APPLY.
NOTHING COUNSEL YOU ON HIV CARE REFER YOU TO A CARE SERVICE ACCOMPANY YOU TO A CARE SERVICE DON’T KNOW REFUSE TO ANSWER
A B C D Y Z
ONLY ASK IF CSCTRS=’1’
PREVENTION
S
S
S
S
S
320
27. SERVICES UPTAKE
Variable ID: CS 30. CSDSMM
Services Uptake If you have a main sex partner, did you ever test together with that partner?
Response Options I DON’T HAVE A MAIN SEX PARTNER YES – WE TESTED TOGETHER NO – WE NEVER TESTED TOGETHER NO—PARTNER WAS TESTED BUT WITHOUT ME DON’T KNOW REFUSE TO ANSWER
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1 2 3
4 7 8
31.
CSCTMSM
The last time you tested for HIV, did the you tell the HIV testing counselor that you have sex with men?
YES NO NEVER TESTED FOR HIV DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
PREVENTION & VULNERABILITY
N/A
S
N/A
N/A
N/A
32.
CSDSSW
The last time you tested for HIV, did you let the HIV testing YES counselor know that you exchange sex for money? NO NEVER TESTED FOR HIV DON’T KNOW
1 2 3 7
PREVENTION & VULNERABILITY
S
N/A
N/A
N/A
N/A
REFUSE TO ANSWER 8
SECTION 2: QUESTIONNAIRE MODULES
3 2 3 121
Variable ID: CS 33. CSDSID
Services Uptake The last time you tested for HIV, did you let the HIV testing YES counselor know that you inject drugs? NO
Response Options 1 2 3 7 8
Skips
Comments S PREVENTION & VULNERABILITY N/A
Core or Secondary* M N/A
T N/A
P S
O N/A
NEVER TESTED FOR HIV DON’T KNOW REFUSE TO ANSWER
34.
CSDSQU1
Which of the following best describes the counseling you received?
RESPECTFUL, CARING, UNDERSTANDING DISRESPECTFUL, UNCARING, STIGMATIZING, UNCOMFORTABLE NEITHER RESPECTFUL NOR DISRESPECTFUL DON’T KNOW REFUSE TO ANSWER
PREVENTION 1
S
S
S
S
S
2
3 7 8
35.
CSDSQU2
Did the counselor address most of your concerns, or questions?
YES NO SOMEWHAT DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
PREVENTION
S
S
S
S
S
322
27. SERVICES UPTAKE
Variable ID: CS C. NA
Services Uptake Care and treatment
Response Options
Skips
Comments S
Core or Secondary* M T P O
Thank you. You told us you tested positive for HIV. Now we will ask you about your participation in a support group. PS CSPSEV Support Group Have you ever attended a support group for people living with HIV? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8
ONLY ASK IF CSCTRS=’1’
NA
36.
‘2’,’7’,’8’→SKIP TO CSCANW
COVERAGE
S
S
S
S
S
37.
CSPSFQ
In the last 12 months, how many times did you attend a support group? TYPE ‘0’ IF NONE.
|
|
|
PREVENTION
S
S
S
S
S
DON’T KNOW 97 REFUSE TO ANSWER 98
SECTION 2: QUESTIONNAIRE MODULES
3 2 323
Variable ID: CS 38. CSPSTY
Services Uptake Which of the following do you receive from the support group for your HIV infection? CHECK ALL THAT APPLY.
Response Options NOTHING COUNSELING/ HEALTH LIVING MESSAGES REMINDED OF IMPORTANCE OF TAKING ARV REGULARLY REMINDED TO KEEP HIV APPOINTMENTS REFILLS OR PICKING UP ARV MEDICATIONS PSYCHOSOCIAL SUPPORT LIVELIHOOD/ MATERIAL SUPPORT OTHER DON’T KNOW REFUSE TO ANSWER A
Skips ‘A’ → CSLKEV
Comments S COVERAGE. Adapt responses based on local context. S
Core or Secondary* M S
T S
P S
O S
B
C D
E F G X Y Z
LK NA CS3MSG
Linkage to care You told us you tested positive for HIV. Many people living with HIV register with a health care provider to get care. By “care” we mean someone goes for check-ups to a health care provider or gets ARVs for their HIV infection. The next few questions are about the first time you saw a provider for your HIV. ONLY ASK IF CSCTRS=’1’
324
27. SERVICES UPTAKE
Variable ID: CS 39. CSLKEV
Services Uptake After learning you had HIV, have you ever received HIV medical care from a health care provider? YES NO
Response Options 1 2 7 8 1
Skips ‘1’→SKIP TO CSLKSTYR
Comments S PREVENTION C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 40. CSlknv What is the main reason you have never received HIV medical care from a healthcare provider? FEEL HEALTHY STIGMA, DON’T WANT OTHERS TO KNOW COST/DISTANCE TO CLINIC POOR ATTITUDE OF HEALTH CARE WORKERS WAITING TIME OR CLINIC HOURS NOT GOOD OTHER DON’T KNOW REFUSE TO ANSWER
PREVENTION
C
C
C
C
C
2 3
4
5 6 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 2 3 525
Variable ID: CS 41. CSLKSTYR
Services Uptake In which year did you start taking ARVs? |
Response Options | | | |
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
MIN: CANNOT BE BEFORE YEAR TESTED MAX: CURRENT YEAR DON’T KNOW REFUSE TO ANSWER 7 8
42.
CSLKSTMT
In which month did you start taking ARVs?
|
|
|
OTHER
S
S
S
S
S
RANGE: 1-12 IF RXARVY = CURRENT YEAR, THEN MAXIMUM = CURRENT MONTH DON’T KNOW REFUSE TO ANSWER 7 8
43.
CSLKTY1
What kind of clinic did you first go to discuss or receive care for your HIV?
PUBLIC OR GOVERNMENT PRIVATE DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
S
S
S
S
S
326
27. SERVICES UPTAKE
Variable ID: CS 44. CSLKTY3
Services Uptake Was this a mobile clinic, or did you go to a fixed building/ office?
Response Options MOBILE CLINIC FIXED CLINIC OTHER DON’T KNOW REFUSE TO ANSWER 1 2 6 7 8
Skips
Comments S COVERAGE S
Core or Secondary* M S
T S
P S
O S
45.
CSLKFEE
Was the care you received from that first health care provider free of charge?
YES – FREE OF CHARGE NO – I HAVE TO OR HAD TO PAY DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
S
S
S
S
S
CA NA CS4MSG
Care Thank you. The next few questions are about the HIV care you may receive now.
46.
CSCANW
Do you currently get health care or check-ups for your HIV?
YES, STILL IN CARE NO, I STOPPED RECEIVING CARE/ GOING TO THE CLINIC DON’T KNOW REFUSE TO ANSWER
1
‘1’→SKIP TO CSCALST
COVERAGE
C
C
C
C
C
2 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 2 3 727
Variable ID: CS 47. CSCARS
Services Uptake What is the main reason you no longer receive HIV care from a health care provider?
Response Options FEEL HEALTHY STIGMA, DON’T WANT OTHERS TO KNOW COST OR TRANSPORTATION PROBLEMS POOR ATTITUDE OF HEALTH CARE WORKERS WAITING TIME OR CLINIC HOURS NOT GOOD OTHER DON’T KNOW REFUSE TO ANSWER 1
Skips SKIP TO END OF CURRENT SECTION.
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
2
3
4
5 6 7 8
48.
CSCALST
When is the next time you will receive HIV care from a health care provider?
LESS THAN ONE MONTH BETWEEN 1 AND 3 MONTHS BETWEEN 4 AND 6 MONTHS LONGER THAN 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
PREVENTION
C
C
C
C
C
328
27. SERVICES UPTAKE
Variable ID: CS NA 49. CX CSCXIN Cotrimoxazole
Services Uptake
Response Options
Skips
Comments S
Core or Secondary* M T P O
Now some questions about Cotrimoxazole or Cotrim. Cotrim is a medicine recommended for people with HIV, even if they have not started treatment. It helps prevent certain infections but it does not treat HIV.
50.
CSCXEV
Have you ever taken cotrimoxazole?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 3
‘1’→SKIP TO CSCXNW
COVERAGE
C
C
C
C
C
51.
CSCXNV
What is the main reason you have never taken cotrimoxazole?
FEEL HEALTHY FEAR OF STIGMA NOT PRESCRIBED
SKIP TO END OF CURRENT SUBSECTION
PREVENTION ADAPT RESPONSES TO LOCAL CONTEXT.
S
S
S
S
S
NO TIME/CLINIC TOO FAR 4 COST OTHER DON’T KNOW REFUSE TO ANSWER 52. CSCXNW Are you currently taking cotrimoxazole? YES NO DON’T KNOW REFUSE TO ANSWER 5 6 7 8 1 2 7 8 COVERAGE S S S S S
SECTION 2: QUESTIONNAIRE MODULES
3 2 3 929
Variable ID: CS D. N/A
Services Uptake CD4+ T cell count
Response Options
Skips
Comments S
Core or Secondary* M T P O
CD CS5MSG
We will now ask you a few questions about your “CD4” or “T-cell” count. The CD4 count tells how sick you are with HIV and if you need to take ARVs or other HIV medications. Have you ever had your CD4 tested? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 ‘2’,’7’,’8’ → SKIP TO NEXT SUBSECTION
COVERAGE
53.
CSCDEV
COVERAGE
C
C
C
C
C
54.
CSCDEV
When did your care provider last test your CD4 count?
IN THE LAST 6 MONTHS BETWEEN 7 AND 12 MONTHS AGO MORE THAN 12 MONTHS AGO DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
PREVENTION Tailor time periods to national or program guidelines
C
C
C
C
C
330
27. SERVICES UPTAKE
Variable ID: CS 55. CSCDRS1
Services Uptake What was the result of your last CD4 count?
Response Options DID NOT RECEIVE RESULT MORE THAN 500 BETWEEN 350 AND 500 BETWEEN 200 AND 349 BELOW 200 DON’T KNOW REFUSE TO ANSWER
Skips OPTION A
Comments S PREVENTION CHOOSE EITHER OPTION A (CSCDRS1) OR OPTION B (CSCDRS2) S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 7 8
56.
CSCDRS2
What was your last CD4 count?
|__ |__| __| __| DON’T KNOW 997 REFUSE TO ANSWER 998
OPTION B
PREVENTION CHOOSE EITHER OPTION A (CSCDRS1) OR OPTION B (CSCDRS2)
S
S
S
S
S
E.
TR
Anti-retroviral treatment
ONLY ASK IF CSCTRS=’1’
ART initiation NA CS6MSG Thank you. Now some questions on anti-retroviral treatment, also called ARVs, to treat HIV. Term ‘ARV’ should be changed based on local context.
SECTION 2: QUESTIONNAIRE MODULES
3 3 3 131
Variable ID: CS 57. CSTREV
Services Uptake Have you ever taken ARVs, that is, antiretroviral medication, to treat your HIV infection? YES NO
Response Options 1 2 7 8
Skips ‘1’→ SKIP TO csTRYY
Comments S COVERAGE, PREVENTION C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 58. CSTRNV1 What is the main reason you never started taking ARVs? MY CD4 COUNT IS STILL HIGH MY CD4 COUNT IS UNKNOWN I AM ON A WAITING LIST TO START HEALTH CARE PROVIDER TOLD ME ITS TOO EARLY TO START OTHER REASON DON’T KNOW REFUSE TO ANSWER
1 2 3
‘1’, ‘2’, ‘3’, ‘4’, ‘7’, OR ‘8’→ SKIP TO CSVLEV
PREVENTION Adapt responses based on local context.
C
C
C
C
C
4 6 7 8
332
27. SERVICES UPTAKE
Variable ID: CS 59. CSTRNV2
Services Uptake You said other. Which of the following is the main reason you never started taking ARVs?
Response Options NO CLINIC NEAR ME OFFERS IT HEALTH CARE PROVIDER REFUSED TO GIVE ME ARVs FEAR OF STIGMA OR THAT OTHERS FIND OUT COSTS TOO MUCH OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S COVERAGE, PREVENTION S
Core or Secondary* M S
T S
P S
O S
1
2
3 4 6 7 8 OTHER S S S S S
60.
CSTRYY
In which year did you start taking ARVs?
|
|
|
|
|
MIN: CANNOT BE BEFORE YEAR TESTED MAX: CURRENT YEAR DON’T KNOW REFUSE TO ANSWER 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 3 333
Variable ID: CS 61. CSTRMM
Services Uptake In which month did you start taking ARVs? |
Response Options | |
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
RANGE: 1-12 IF RXARVY = CURRENT YEAR, THEN MAXIMUM = CURRENT MONTH DON’T KNOW REFUSE TO ANSWER ART retention CSTRCURR Are you currently taking ARVs? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 ‘1’→CSTRSV COVERAGE, PREVENTION C C C C C 7 8
334
27. SERVICES UPTAKE
Variable ID: CS 62. CSTRSP2
Services Uptake
Response Options
Skips ‘1’, ‘2’, ‘3’, ‘4’, ‘7’, ‘8’→ CSVLEV
Comments S PREVENTION Adapt responses based on local context. C
Core or Secondary* M C
T C
P C
O C
Can you tell me the main reason why you are not currently CLINIC TOO FAR, taking ARVs? TRANSPORT COSTS TOO HIGH
1
COSTS AT CLINIC TOO 2 HIGH POOR SERVICE AT CLINIC CLINIC RAN OUT OF ARVs OTHER DON’T KNOW REFUSE TO ANSWER 63. CSTRSP3 You said other. Can you tell me the main reason why you are not currently taking ARVs? FEEL HEALTHY TOO MANY SIDE EFFECTS RELIGIOUS REASONS OTHER DON’T KNOW REFUSE TO ANSWER 3 4 6 7 8 1 2 3 6 7 8
PREVENTION Adapt responses based on local context.
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
3 3 3 535
Variable ID: CS 64. CSTRSV
Services Uptake While taking ARVs, do you or did you use any of the following services? READ RESPONSES ALOUD. CHECK ALL THAT APPLY.
Response Options MOBILE PHONE TEXT REMINDERS TREATMENT SUPPORT GROUP FOOD OR MONEY SUPPORT OUTREACH WORKER OR PEER EDUCATOR NONE OF THESE DON’T KNOW REFUSE TO ANSWER
Skips
Comments S S
Core or Secondary* M S
T S
P S
O S
A B C D E Y Z
COVERAGE
F.
VL
Viral load
ONLY ASK IF CSTREV=’1’ YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 PREVENTION S S S S S ‘2’, ‘7’, OR ‘8’ →SKIP TO END OF CURRENT SUBSECTION COVERAGE C C C C C
65.
CSVLEV
Some people with HIV get their viral load measured. this a test that measures how much HIV is in your blood. Did you ever have a viral load test?
66.
CSVLTI
When did you last have a viral load test?
IN THE LAST 12 MONTHS MORE THAN A YEAR AGO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
336
27. SERVICES UPTAKE
Variable ID: CS 67. CSVLRS1
Services Uptake Did you learn the result of your most recent viral load test? YES NO
Response Options 1 2 7 8 1 2 3 7 8 1
Skips ‘2’, ‘7’, OR ‘8’ →SKIP to NEXT SUBSECTION
Comments S COVERAGE S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER 68. CSVLRS1 SUPPRESSED Viral load results can be “Suppressed” or “Not suppressed”? “Suppressed” means there is very low or no virus in your blood. “Not suppressed” means there is still NOT SUPPRESSED plenty of virus in your blood. Was you last viral load test THEY DID NOT TELL “Suppressed” or “Not suppressed”? ME THE RESULT DON’T KNOW REFUSE TO ANSWER 69. CSVLRS2 How much virus was in your blood when you last tested your viral load? LESS THAN 1,000
‘2’, ‘3’, ‘7’, OR ‘8’ →SKIP TO END OF CURRENT SUBSECTIOn
PREVENTION This language should be adapted to the local context
S
S
S
S
S
OTHER
S
S
S
S
S
BETWEEN 1,0011001 2 AND 10,000 BETWEEN 10,0011001 AND 100,000 MORE THAN 100,0011001 DON’T KNOW REFUSE TO ANSWER
3 4 5 8
TB NA CS8MSG
TUBERCULOSIS The next few questions are about tuberculosis or TB. All people with HIV should be asked about TB.
SECTION 2: QUESTIONNAIRE MODULES
3 3 3 737
Variable ID: CS 70. CSTBSC1
Services Uptake In the last 12 months, during any of your visits to the HIV clinic, were you asked if you had the following TB symptoms: night sweats, cough, fever, and weight loss? YES
Response Options 1 2 3 7 8 1 2 3 7 8
Skips ‘2’,’3’,’7’,’8’ → SKIP TO END OF CURRENT SUBSECTION
Comments S COVERAGE S
Core or Secondary* M S
T S
P S
O S
NO, I WAS NOT ASKED NO, I AM CURRENTLY ON TB TREATMENT DON’T KNOW REFUSE TO ANSWER
71.
CSTBSC2
In the last 12 months, how often were you asked these questions?
AT EVERY VISIT ON SOME VISITS ON ONE VISIT DON’T KNOW REFUSE TO ANSWER
COVERAGE
S
S
S
S
S
72.
CSTBSC3
In the last 12 months, have you experienced any of these symptoms?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 A B C Y Z
‘2’,’7’,’8’→ SKIP TO COVERAGE END OF CURRENT SUBSECTION
S
S
S
S
S
73.
CSTBTT
In the last 12 months, did you receive a chest x-ray or sputum test to look for TB?
NONE OF THESE
CHEST X-RAY A sputum test is when the patient has to cough and collect SPUTUM TEST the sample in a cup. CHECK ALL THAT APPLY. DON’T KNOW REFUSE TO ANSWER
‘A’→ SKIP TO END OF CURRENT SUBSECTION
COVERAGE
S
S
S
S
S
338
27. SERVICES UPTAKE
Variable ID: CS 74. CSTBRS
Services Uptake What was the result of your last TB test?
Response Options TB-NEGATIVE TB-POSITIVE DON’T KNOW REFUSE TO ANSWER 1 2 7 8
Skips ‘1’,’7’,’8’→ CS8MSG
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
75.
CSTBTR1
The last time you had TB, did you get treatment for TB?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
COVERAGE
S
S
S
S
S
CC NA CS8MSG
CERVICAL CANCER SCREENING Now I’m going to ask you about tests a health care provider can do to check for cervical cancer. The cervix connects the uterus to the vagina. The tests a health care provider can do to check for cervical cancer are called a Pap smear, HPV test and VIA test. For a Pap smear and HPV test, a health care provider puts a small stick inside the vagina to wipe the cervix and sends the sample to the laboratory. For a VIA test, a healthcare worker puts vinegar on the cervix and looks to see if the cervix changes color.
76.
CSCCEV
Have you ever been tested for cervical cancer?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
FOR WOMEN ONLY.
COVERAGE
S
N/A
N/A
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 3 3 939
Variable ID: CS 77. CSCCAB
Services Uptake At your last cervical cancer test, what was the result?
Response Options NORMAL/NEGATIVE ABNORMAL/ SUSPECT CANCER/ POSITIVE UNCLEAR/ INCONCLUSIVE DID NOT RECEIVE RESULTS DON’T KNOW REFUSE TO ANSWER 1
Skips FOR WOMEN ONLY. ‘1’,’3’,’4’,’7’,’8’→ SKIP TO END OF CURRENT SECTION
Comments S COVERAGE S
Core or Secondary* M N/A
T N/A
P S
O S
2 3 4 7 8
78.
CSCCREF
Were you referred or treated for cervical abnormality?
REFERRED TREATED BOTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
FOR WOMEN ONLY.
COVERAGE RESPONSES BASED ON LOCAL CONTEXT.
S
N/A
N/A
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
340
27. SERVICES UPTAKE
Section
II NA N/A
28. PEP and PrEP
These modules are meant to assess knowledge and use of pre-exposure prophylaxis (PrEP) and post-exposure prophylaxis (PEP). These questions should only be asked in settings where PrEP and PEP are available.
Variable ID: PP PP1MSG
PEP and PREP
Response Options
Skips IF CSCTRS=’1’→ SKIP TO END OF CURRENT SECTION
Comments S
Core or Secondary* M T P O
Thank you. Now we will ask some questions about pre-exposure prophylaxis (PrEP). PrEP is a medicine that can prevent HIV. It is taken by HIV-negative people.
1.
PPRKNOW
Have you heard of PrEP?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→ SKIP TO PPRCONS
KA
C
C
C
C
C
2.
PPRTAKE
Have you ever taken PrEP?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘1’→ SKIP TO PPR6TAKE
COVERAGE
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
3 4 3 141
NA 3.
Variable ID: PP PPRNEVER
PEP and PREP What is the main reason you have never taken PrEP?
Response Options EMBARRASSED TO TALK ABOUT IT WITH DOCTOR/ NURSE DON’T FEEL AT RISK FOR HIV NOT AVAILABLE WHERE I LIVE DON’T KNOW WHERE TO GET IT DON’T WANT IT AFRAID OF SIDE EFFECTS DON’T WANT OTHERS TO KNOW OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S COVERAGE & PREVENTION C
Core or Secondary* M C
T C
P C
O C
1 2 3 4 5 6 7 8 97 98 1 2 7 8 1 2 7 8 ‘1’→ SKIP TO PPRWHERE ‘2’, ‘7’, OR ‘8’→ SKIP TO PP2MSG
4.
PPRCONS
PrEP has similar side effects to other drugs used to treat HIV. Would you take PrEP to help prevent HIV?
YES NO DON’T KNOW REFUSE TO ANSWER
KA
S
S
S
S
S
5.
PPR6TAKE
In the last 6 months, have you taken PrEP?
YES NO DON’T KNOW REFUSE TO ANSWER
COVERAGE
C
C
C
C
C
342
28. PEP AND PREP
NA 6.
Variable ID: PP PPRMAIN
PEP and PREP What is the main reason you stopped taking PrEP?
Response Options I TRUST MY PARTNERS CAN’T GET PrEP ANYMORE HAD SIDE EFFECTS DON’T WANT OTHERS TO KNOW TESTED HIV-POSITIVE OTHER DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Skips SKIP TO PP2MSG
Comments S COVERAGE & PREVENTION C
Core or Secondary* M C
T C
P C
O C
7.
PPRWHERE
From where do you usually get PrEP?
GOVERNMENT CLINIC COMMUNITY-BASED NGO PHARMACY FRIENDS STRANGERS OTHER DON’T KNOW REFUSE TO ANSWER
COVERAGE & PREVENTION
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 4 343
NA 8.
Variable ID: PP PPRDAILY
PEP and PREP Do you take PrEP daily as prescribed or non-daily as prescribed?
Response Options DAILY AS PRESCRIBED NON-DAILY AS PRESCRIBED OTHER DON’T KNOW REFUSE TO ANSWER 1 2 3 7 8
Skips
Comments S COVERAGE C
Core or Secondary* M C
T C
P C
O C
9.
PPRLAST
When was the last time you took PrEP?
YESTERDAY OR TODAY 2-3 DAYS AGO 4-7 DAYS AGO 1-2 WEEKS AGO MORE THAN 2 WEEKS AGO DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
COVERAGE & PREVENTION
C
C
C
C
C
N/A
PP2MSG
Thank you. Now some questions about post-exposure prophylaxis (PEP). PEP is a medicine that HIV-negative people take for one month after they had contact with HIV, so that they do not get HIV.
10.
PPEKNOW
Have you heard of PEP?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’ OR ‘8’→ SKIP TO PPECONS
KA
C
C
C
C
C
344
28. PEP AND PREP
NA 11.
Variable ID: PP PPETAKE
PEP and PREP Have you ever taken PEP? YES NO
Response Options 1 2 7 8 1 2 7 8
Skips ‘2’,’7’ OR ‘8’→ SKIP TO PPECONS
Comments S COVERAGE C
Core or Secondary* M C
T C
P C
O C
DON’T KNOW REFUSE TO ANSWER 10. PPE6TAKE In the last 6 months, have you taken PEP? YES NO DON’T KNOW REFUSE TO ANSWER
‘2’,’7’ OR ‘8’→ SKIP TO PPECONS
COVERAGE
C
C
C
C
C
11.
PPEREAS
Why did you take PEP?
I HAD UNPROTECTED SEX I WAS RAPED/FORCED TO HAVE SEX I SHARED NEEDLES DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
PREVENTION
C
C
C
C
C
12.
PPECONS
If you think you came into contact with HIV, would you go to a health care worker and ask for this medicine?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
KA
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SECTION 2: QUESTIONNAIRE MODULES
3 4 3 545
Section
Il
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
Stigma is a barrier to HIV prevention and treatment. The first part of the module assesses stigma towards being a member of a key population or engaging in high risk behavior*. The second part assesses stigma towards people living with HIV/AIDS.
N/A
Variable ID: ST ST1MSG
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M T P O
N/A
Thank you for your answers. Now we will ask you questions about ways that other people treat you.
1.
STGSHMP
I am ashamed to be an injection drug user.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
N/A
N/A
N/A
S
N/A
346
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGMMEETP
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P S
O N/A
2.
I am not ashamed to say I am an injection drug user in a gathering with other people who inject drugs.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 3 4 5 7 8
VULNERABILITY
N/A
3.
STGHLCRP
I am not ashamed to say I am an injection drug user when I meet with a social or health worker who works in my community.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
S
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 4 3 747
N/A
Variable ID: ST STGTLDP
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P C
O N/A
4.
Who have you told that you inject drugs? CHECK ALL THAT APPLY.
NO ONE PARTNER/SPOUSE FAMILY FRIENDS/ACQUAINTANCES WHO INJECT DRUGS FRIENDS/ACQUAINTANCES WHO DO NOT INJECT DRUGS HEALTHCARE PROVIDERS OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E F X Y Z 1 2 3 7 8 1 2 3 7 8
IF ‘NO ONE’, ‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY ADAPT RESPONSES TO LOCAL CONTEXT.
N/A
5.
STGEXCLP
Have you ever felt excluded from family activities because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
C
N/A
6.
STGFMMP
Have you ever felt that family members have made unfair comments or gossiped about you because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
S
N/A
348
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGHEALP
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P C
O N/A
7.
Have you ever felt afraid to seek health services because you worried someone may learn you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
8.
STGHIDEP
Have you ever avoided seeking health services because you worried someone may learn you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
N/A
N/A
C
N/A
9.
STGVHRSP
Has someone ever scolded you because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
N/A
N/A
C
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 4 3 949
N/A
Variable ID: ST STGBLMP
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P C
O N/A
10.
Has someone ever blackmailed you because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 3 7 8 1 2 3 7 8 1 2 7 8 ‘1’ → SKIP TO STVGTBLP
VULNERABILITY
N/A
11.
STGPHRSP**
Has someone ever physically harassed or hurt you because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTH DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
C
N/A
12.
STGPFSXP**
Has someone ever forced you to have any type of sex when you did not want to? By “forced”, I mean physically forced or coerced to have any type of sex, including penetration with an object, when you did not want to?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTH DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
C
N/A
13.
STGFSRP**
Do you believe any of your experiences of forced sex were related to the fact that you inject drugs?
NO YES DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
N/A
C
N/A
350
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGARSTP
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P C
O N/A
14.
Have you ever been arrested because you inject drugs?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
15.
STGSHMM
I am ashamed to be a man who has sex with men.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
N/A
S
N/A
N/A
N/A
16.
STGMMEETM
I am not ashamed to say I am a man who has sex with men in a gathering with other people who are men who have sex with men
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
N/A
S
N/A
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 5 3 151
N/A
Variable ID: ST STGHLCRM
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M S
T N/A
P N/A
O N/A
17.
I am not ashamed to say I am a man who has sex with men when I meet with a social or health worker who works in my community.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
N/A
18.
STGTLDM
Who have you told that you have sex with men? CHECK ALL THAT APPLY.
NO ONE PARTNER/SPOUSE FAMILY MALE FRIENDS/ ACQUAINTANCES WHO HAVE SEX WITH OTHER MEN FRIENDS/ACQUAINTANCES WHO MEN WHO HAVE SEX WITH MEN HEALTHCARE PROVIDERS OTHER DON’T KNOW REFUSE TO ANSWER
A B C
IF ‘A’, ‘Y’ OR ‘Z’ NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY ADAPT RESPONSES TO LOCAL CONTEXT.
N/A
C
N/A
N/A
N/A
D
E F X Y Z
352
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGEXCLM
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M C
T N/A
P N/A
O N/A
19.
Have you ever felt excluded from family activities because you have sex with men
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
20.
STGFMMM
Have you ever felt that family members have made unfair comments or gossiped about you because you have sex with men
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
S
N/A
N/A
N/A
21.
STGHEALM
Have you ever felt afraid to seek health services because you worried someone may learn you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
C
N/A
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 5 353
N/A
Variable ID: ST STGHIDEM
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M C
T N/A
P N/A
O N/A
22.
Have you ever avoided seeking health services because you worried someone may learn you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
23.
STGVHRSM
Has someone ever scolded you because you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
C
N/A
N/A
N/A
24.
STGBLMM
Has someone ever blackmailed you because you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
C
N/A
N/A
N/A
354
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGPHRSM**
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M C
T N/A
P N/A
O N/A
25.
Has someone ever physically harassed or hurt you because you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTH DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 3 7 8 1 2 7 8 1 2 3 7 8 ‘1’ → SKIP TO STGARSTM
VULNERABILITY
N/A
26.
STGPFSXM**
Has someone ever forced you to have any type of sex when you did not want to? By “forced”, I mean physically forced or coerced to have any type of sex, including penetration with an object, when you did not want to?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTH DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
C
N/A
N/A
N/A
27.
STGFSRM**
Do you believe any of your experiences of forced sex were related to the fact that you have sex with men?
NO YES DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
C
N/A
N/A
N/A
28.
STGARSTM
Have you ever been arrested because you have sex with men?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
C
N/A
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 5 3 555
N/A
Variable ID: ST STGSHMT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T S
P N/A
O N/A
29.
I am ashamed to be transgender.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 1 2 3 4 5 7 8 1 2 3 4 5 7 8
VULNERABILITY
N/A
30.
STGMMEETT
I am not ashamed to say I am transgender in a gathering with other people who are transgender.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
S
N/A
N/A
31.
STGHLCRT
I am not ashamed to say I am transgender when I meet with a social or health worker who works in my community.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
S
N/A
N/A
356
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGTLDT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T C
P N/A
O N/A
32.
Who have you told that you are transgender? CHECK ALL THAT APPLY.
NO ONE PARTNER/SPOUSE FAMILY FRIENDS/ACQUAINTANCES WHO ARE TRANSGENDER FRIENDS/ACQUAINTANCES WHO ARE NOT TRANSGENDER HEALTHCARE PROVIDERS OTHER DON’T KNOW REFUSE TO ANSWER
A B C D
‘A’, ‘Y’ OR ‘Z’, NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY ADAPT RESPONSES TO LOCAL CONTEXT.
N/A
E F X Y Z 1 2 3 7 8 1 2 3 7 8 VULNERABILITY N/A N/A S N/A N/A VULNERABILITY N/A N/A C N/A N/A
33.
STGEXCLT
Have you ever felt excluded from family activities because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
34.
STGFMMT
Have you ever felt that family members have made unfair comments or gossiped about you because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
SECTION 2: QUESTIONNAIRE MODULES
3 5 3 757
N/A
Variable ID: ST STGHEALT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T C
P N/A
O
35.
Have you ever felt afraid to seek health services because you worried someone may learn you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
N/A
36.
STGHIDET
Have you ever avoided seeking health services because you worried someone may learn you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
N/A
C
N/A
N/A
37.
STGVHRST
Has someone ever scolded you because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
N/A
C
N/A
N/A
358
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGBLMT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T C
P N/A
O N/A
38.
Has someone ever blackmailed you because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 3 7 8 1 2 3 7 8 1 2 7 8 ‘1’ → SKIP TO STGARSTT
VULNERABILITY
N/A
39.
STGPHRST**
Has someone ever physically harassed or hurt you because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
C
N/A
N/A
40.
STGPFSXT**
Has someone ever forced you to have any type of sex when you did not want to? By “forced”, I mean physically forced or coerced to have any type of sex, including penetration with an object, when you did not want to.
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
C
N/A
N/A
41.
STGFSRT**
Do you believe any of your experiences of forced sex were related to the fact that you are transgender?
NO YES DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
C
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 5 3 959
N/A
Variable ID: ST STGARSTT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M C
T N/A
P N/A
O N/A
42.
Have you ever been arrested because you are transgender?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
N/A
43.
STGSHMS
I am ashamed to sell sex?
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
S
N/A
N/A
N/A
N/A
44.
STGMMEETS
I am not ashamed to say I sell sex in a gathering with other people who sell sex.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
S
N/A
N/A
N/A
N/A
360
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGHLCRS
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O N/A
45.
I am not ashamed to say I sell sex when I meet with a social or health worker who works in my community.
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8 A B C D IF ‘A’, ‘Y’ OR ‘Z’ NO OTHER CATEGORY CAN BE CHOSEN AS WELL.
VULNERABILITY
S
46.
STGTLDS
Who have you told that you sell sex? CHECK ALL THAT APPLY.
NO ONE PARTNER/SPOUSE FAMILY FRIENDS/ACQUAINTANCES WHO SELL SEX FRIENDS/ACQUAINTANCES WHO ARE NOT DO NOT SELL SEX HEALTHCARE PROVIDERS OTHER DON’T KNOW REFUSE TO ANSWER
VULNERABILITY ADAPT RESPONSES TO LOCAL CONTEXT.
C
N/A
N/A
N/A
N/A
E F X Y Z
SECTION 2: QUESTIONNAIRE MODULES
3 6 3 161
N/A
Variable ID: ST STGEXCLS
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O N/A
47.
Have you ever felt excluded from family activities because you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
C
48.
STGFMMS
Have you ever felt that family members have made unfair comments or gossiped about you because you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
S
N/A
N/A
N/A
N/A
49.
STGHEALS
Have you ever felt afraid to seek health services because you worried someone may learn you sell sex
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
C
N/A
N/A
N/A
N/A
362
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STGHIDES
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O N/A
50.
Have you ever avoided seeking health services because you worried someone may learn you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
C
51.
STGVHRSS
Has someone ever scolded you because you are sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
C
N/A
N/A
N/A
N/A
52.
STGBLMS
Has someone ever blackmailed you because you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
VULNERABILITY
C
N/A
N/A
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 6 363
N/A
Variable ID: ST STGPHRSS**
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M N/A
T N/A
P N/A
O N/A
53.
Has someone ever physically harassed or hurt you because you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 3 7 8 1 2 7 8 1 2 3 7 8 ‘1’ → SKIP TO STGARSTS
VULNERABILITY
C
54.
STGPFSXS**
Has someone ever forced you to have any type of sex when you did not want to? By “forced”, I mean physically forced or coerced to have any type of sex, including penetration with an object, when you did not want to.
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
C
N/A
N/A
N/A
N/A
55.
STGFSRS**
Do you believe any of your experiences of forced sex were related to the fact that you sell sex?
NO YES DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
C
N/A
N/A
N/A
N/A
56.
STGARSTS
Have you ever been arrested because you sell sex?
NO YES, IN THE LAST 6 MONTHS YES, BUT NOT IN THE LAST 6 MONTHS DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
C
N/A
N/A
N/A
N/A
364
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST SWPOGIVE
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
57.
In the past 12 months, how many times have you TIMES: | | | | given something (in cash or kind) to the police to avoid DON’T KNOW 997 trouble with the? REFUSE TO ANSWER 998 MAX: 300
‘0’, ‘997’, ‘998’ → VULNERABILITY SKIP TO END OF CURRENT SECTION
C
58
SWPOITM
What did you give to the police to avoid trouble with them? CHECK ALL THAT APPLY
MONEY SEX OTHER REFUSE TO ANSWER
A B X Z
VULNERABILITY ADAPT ANSWERS TO LOCAL CONTEXT
C
S
S
S
S
N/A 59.
ST2MSG STVGTBL
Next we ask some questions about people with HIV. Would you buy fresh vegetables from a shop keeper or vendor if you knew the person had HIV YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 1 2 7 8 1 2 7 8 KA S S S S S KA S S S S S KA S S S S S
60.
STSCHL
Do you think children living with HIV should be allowed to attend school?
YES NO DON’T KNOW REFUSE TO ANSWER
61.
STFRSLV
Do you fear that you could get HIV if you come into contact with the saliva of a person with HIV?
YES NO DON’T KNOW REFUSE TO ANSWER
SECTION 2: QUESTIONNAIRE MODULES
3 6 3 565
N/A
Variable ID: ST STFMSHM
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
62.
Would you be ashamed if someone in your family had HIV?
AGREE DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 7 8
KA
S
63.
STDSCRM
Do you agree that most people in [insert survey area] would discriminate against someone with HIV?
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
KA
S
S
S
S
S
64.
STRGHTS
Do you agree that most people in [insert survey area] would support the rights of a person with HIV to live and work wherever they wanted to?
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
KA
S
S
S
S
S
366
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
N/A
Variable ID: ST STDSRV
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
65.
Do you agree that most people in [insert survey area] think that people who got HIV through sex or drug use have gotten what they deserve?
STRONGLY AGREE AGREE NEUTRAL DISAGREE STRONGLY DISAGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
KA
S
66.
STFRTST
In your opinion, are people hesitant to test for HIV out of fear how people will react if they test positive?
YES NO NOT SURE DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
KA
S
S
S
S
S
67.
STTLK
Do people talk badly about people living with or thought to be living with HIV?
YES NO IT DEPENDS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
KA
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 6 3 767
N/A
Variable ID: ST STRSPCT
Shame, Stigma, Harassment & Discrimination
Response Options
Skips
Comments S
Core or Secondary* M S
T S
P S
O S
68.
Do people living with or thought to be living with HIV lose respect or standing?
YES NO IT DEPENDS DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8
KA
S
#: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
* Obtained through personal communication with Stefan Baral, Johns Hopkins University, US **Similar questions are available in Questionnaire Modules for Physical Violence and Sexual Violence.
References: Stangl, Anne. MERG Approved Indicators: Stigma & Discrimination in the General Population. 19 March 2014. OGAC meeting on stigma measurement.
368
29. SHAME, STIGMA, HARASSMENT & DISCRIMINATION
Section
II
30. PHYSICAL VIOLENCE
Members of key populations experience increased vulnerability to physical violence due to the stigmatized nature of their behaviors, and the perceived lack of adherence to existing social and/or gender norms. Several studies have noted the connection between experience of physical violence and HIV.
The survey data instruments that are the base for this module have been implemented in many countries. A formative assessment with these variables in mind is necessary in order to appropriately contextualize the prompts and answer choices. The UN Statistical Commission recommends the monitoring of violence using 3 foci within a lifetime and the last 12 months: severity, relationship to perpetrator and frequency1. The core questions listed here focus on frequency, while the secondary questions add the severity and relationship to perpetrator foci. These secondary questions can be used when a country portfolio has an expanded focus on violence and GBV; and/or for trend analysis with previously implemented population based surveys (e.g. DHS, WHO MultiCountry Study, VACS, etc.). When using secondary questions in analysis, it is recommended that both the LAST and FIRST incidents be included, however you may choose to ask about one or the other. With questions of such a sensitive nature, it is important to have a process in place for referral to psychosocial, medical and legal services, should the survey respondent express an interest in accessing services or experience duress as a result of recalling traumatic events.
SECTION 2: QUESTIONNAIRE MODULES
3 6 3 969
Variable ID: PV N/A PVMSG
Physical Violence
Response Options
Skips
Comments S
Core or Secondary* M T P O
Thank you. Now some questions about how other people have treated you. Please remember that your answers are confidential.
1.
PVEVERFRS
In your lifetime, how many times has anyone ever physically hurt you such as hit or choked you, or threatened you with a knife or other weapon?
NEVER ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
‘1’,’7’,’8’→ SKIP VULNERABILITY TO PVPERP
C
N/A
N/A
N/A
N/A
5.
PVFAGE
The first time someone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon, how old were you?
YEARS: [ MIN: 1
|
]
VULNERABILITY
S
S
S
S
S
MAX: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
370
30. PHYSICAL VIOLENCE
Variable ID: PV 6. PVFBY
Physical Violence The first time this happened, what was your relationship to the person who did this to you? If it was more than one person, what was your relationship with the person who was the first person who did this to you?
Response Options PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
7.
PVFBYP
The first time this happened, what type of partner was this person?
SPOUSE OR LIVE-IN PARTNER 1 BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER 2
ONLY ASK IF PVFBY=’1’, ‘2’
VULNERABILITY
S
S
S
S
S
3 6 7 8 1 2 7 8 VULNERABILITY C N/A N/A N/A N/A
8.
PVFBEHS*
Any of these times this happened, do you think it was because you sell sex?
YES NO DON’T KNOW REFUSE TO ANSWER
SECTION 2: QUESTIONNAIRE MODULES
3 7 3 171
Variable ID: PV 9. PVFBEHM*
Physical Violence Any of these times this happened, do you think it was because you have sex with men? YES NO
Response Options 1 2 7 8 1 2 7 8 1 2 7 8
Skips
Comments S VULNERABILITY N/A
Core or Secondary* M C
T N/A
P N/A
O N/A
DON’T KNOW REFUSE TO ANSWER 10. PVFBEHT* Any of these times this happened, do you think it was because you are transgender? YES NO DON’T KNOW REFUSE TO ANSWER 11. PVFBEHP* Any of these times this happened, do you think it was because you inject drugs? YES NO DON’T KNOW REFUSE TO ANSWER 12. PVYRFRQS In the last 12 months, how many times has anyone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon? THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
C
N/A
N/A
VULNERABILITY
N/A
N/A
N/A
C
N/A
1 2 3 4 5 7 8
‘1’,’7’,OR ‘8’→SKIP TO PVPERP
VULNERABILITY
C
N/A
N/A
N/A
N/A
372
30. PHYSICAL VIOLENCE
Variable ID: PV 13. PVYRFRQM
Physical Violence In the last 12 months, how many times has anyone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon?
Response Options THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
Skips ‘1’,’7’,OR ‘8’→SKIP TO PVPERP
Comments S VULNERABILITY N/A
Core or Secondary* M C
T N/A
P N/A
O N/A
1 2 3 4 5 7 8
14.
PVYRFRQT
In the last 12 months, how many times has anyone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon?
THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
‘1’,’7’,OR ‘8’→SKIP TO PVPERP
VULNERABILITY
N/A
N/A
C
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 7 373
Variable ID: PV 15. PVYRFRQP
Physical Violence In the last 12 months, how many times has anyone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon?
Response Options THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
Skips ‘1’,’7’,OR ‘8’→SKIP TO PVPERP
Comments S VULNERABILITY N/A
Core or Secondary* M N/A
T N/A
P C
O N/A
1 2 3 4 5 7 8 1 2 7 8 1 2 7 8 1 2 7 8
16.
PVYRBEHS*
In any of the times in the last 12 months when these happened, do you think it was because you sell sex?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
C
N/A
N/A
N/A
N/A
17.
PVYRBEHM* In any of the times in the last 12 months when these happened, do you think it was because you have sex with men?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
C
N/A
N/A
N/A
18.
PVYRBEHT*
In any of the times in the last 12 months when these happened, do you think it was because you are transgender?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
N/A
N/A
C
N/A
N/A
374
30. PHYSICAL VIOLENCE
Variable ID: PV 19. PVYRBEHP*
Physical Violence In any of the times in the last 12 months when these happened, do you think it was because you Inject drugs? YES NO
Response Options 1 2 7 8
Skips
Comments S VULNERABILITY N/A
Core or Secondary* M N/A
T N/A
P C
O N/A
DON’T KNOW REFUSE TO ANSWER N/A PVMSG2
The next questions will be asking you about the most recent time someone has physically hurt you such as hit or choked you, or threatened you with a knife or other The last time this has happened, what was your relationship to the person who did any of these things to you? If it was more than one person, what was your relationship with the person who started the violence in the most recent time this happened? CHECK ALL THAT APPLY. PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z ‘A’, ‘C’, ‘D’, ‘E’, ‘X’, ‘Y’, ‘Z’ → SKIP TO PVRSEV VULNERABILITY S S S S S
20.
PVRBY
SECTION 2: QUESTIONNAIRE MODULES
3 7 3 575
Variable ID: PV 21. PVRBYP
Physical Violence The last time this happened, what type of partner was this person?
Response Options SPOUSE OR LIVE-IN PARTNER 1 BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER 2
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
3 6 7 8 A B C D E X Y Z 1 2 7 8 VULNERABILITY S S S S S VULNERABILITY S S S S S
22.
PVRSEV
The last time this has happened, did any of the following happen to you as a result of someone having physically hurt you such as hit or choked you, or threatened you with a knife or other weapon? CHECK ALL THAT APPLY
CUTS, BRUISES OR ACHES EYE INJURIES, SPRAINS, DISLOCATIONS, OR BURNS DEEP WOUNDS, BROKEN BONES, BROKEN TEETH OTHER SERIOUS INJURY NONE OF THESE OTHER DON’T KNOW REFUSE TO ANSWER
23.
PVRBEH
The last time this has happened, do you think it was because you [insert behavior]?
YES NO DON’T KNOW REFUSE TO ANSWER
376
30. PHYSICAL VIOLENCE
Variable ID: PV PV MSG3
Physical Violence
Response Options
Skips
Comments S S
Core or Secondary* M S
T S
P S
O S
For the next few questions, please think about any experiences you have had when someone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon. As a result of someone doing any of those things to you, did you ever try to seek professional help or services from any of the following? CHECK ALL THAT APPLY I DID NOT TRY TO SEEK HELP HEALTHCARE PROFESSIONAL POLICE OR OTHER SECURITY PERSONNEL SOCIAL WORKER, COUNSELOR OR NGO RELIGIOUS LEADER OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z IF ‘A’, ‘Y’, OR ‘Z’, NO OTHER OPTION CAN BE CHOSEN AS WELL. SKIP TO PVSRVWHY PREVENTION
24.
PVSRVW
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 7 3 777
Variable ID: PV 25. PVSRVP
Physical Violence
Response Options
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
When trying to seek help for what someone had done to you, DOCTOR, NURSE, OTHER were you ever refused services by any of the following? HEALTHCARE PROFESSIONAL CHECK ALL THAT APPLY. POLICE OR OTHER SECURITY PERSONNEL LAWYER, JUDGE OR OTHER LEGAL PROFESSIONAL SOCIAL WORKER OR COUNSELOR AN NGO OR OTHER LOCALORGANIZATION OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
378
30. PHYSICAL VIOLENCE
Variable ID: PV 26. PVSRVWHY
Physical Violence What was the main reason I that you did not try to seek professional help or services?
Response Options DID NOT KNOW WHAT SERVICES WERE AVAILABLE/ WHERE TO GO THE SERVICES I WANTED/ NEEDED WERE NOT AVAILABLE I COULD NOT AFFORD SERVICES I WAS UNCOMFORTABLE ACCESSING SERVICES I DID NOT FEEL THAT I NEEDED SERVICES OTHER DON’T KNOW REFUSE TO ANSWER
Skips ‘1’, ‘2’, ‘3’, ‘5’, ‘6’, ‘7’, ‘8’ → SKIP TO PVTELL
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1
2 3 4 5 6 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 7 3 979
Variable ID: PV 27. PVSRVUNC
Physical Violence You said that you were uncomfortable accessing services. Why were you uncomfortable? CHECK ALL THAT APPLY.
Response Options I WAS WORRIED THAT THE SERVICE PROVIDER WOULD TREAT ME DIFFERENTLY/ POORLY IF THEY FOUND OUT I AM [insert population] A I WAS AFRAID THAT MY FAMILY/PARTNER/HUSBAND/ WIFE WOULD FIND OUT I AM [insert population] B I WAS AFRAID OF WHAT THE AUTHORITIES/POLICE WOULD DO TO ME IF THEY FOUND OUT I AM MSM/ FSW/TG/PWID/OVP I WAS WORRIED THAT OTHERS WOULD FIND OUT I AM [insert population] OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
C
D X Y Z A B C D E X Y Z VULNERABILITY S S S S S
28.
PVTELL
Who else did you tell about any of these experiences? CHECK ALL THAT APPLY.
PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER
380
30. PHYSICAL VIOLENCE
Variable ID: PV 29. PVPERP
Physical Violence In your lifetime, have you ever physically hurt someone else such as hit or choked, or threatened with a knife or other weapon? YES NO
Response Options 1 2 7 8 1 2 3 4 7 8
Skips ‘2’,’7’,’8’→ SKIP TO END OF CURRENT SECTION
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW/REMEMBER REFUSE TO ANSWER 30. PVPRPFR In your lifetime, how many times have you done this to someone? ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER 31. PVPRPYR In the last 12 months, how many times have you done this to someone? THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
OTHER
S
S
S
S
S
1 2 3 4 5 7 8
OTHER
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 8 3 181
Variable ID: PV 32. PVPRPWH
Physical Violence Please think about any time that you have ever done this to someone. Who did you did this to? CHECK ALL THAT APPLY.
Response Options SPOUSE/PARTNER/ BOYFRIEND/GIRLFRIEND FRIEND/ACQUAINTANCE FAMILY/RELATIVE FRIEND/ACQUAINTANCE STRANGER OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
A B C D E X Y Z
#: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
*Similar questions can be found in Questionnaire module for Shame, Stigma, Harassment & Discrimination. 1 UN Statistical Commission recommendations: http://unstats.un.org/unsd/demographic/meetings/vaw/docs/ConclusionsAndRecommendations.pdf Additional Sources of Violence Indicators: UN General Assembly Human Rights Council (2008) A/HCR/7/6 http://www2.ohchr.org/english/bodies/hrcouncil/7session/reports.htm MEASURE Evaluation (2008). Violence against Women and girls. A compendium of Monitoring and Evaluation Indicators. http://www.cpc.unc.edu/measure/publications/pdf/ms-08-30.pdf
382
30. PHYSICAL VIOLENCE
Section
II
31. SEXUAL VIOLENCE: COERCION AND RAPE
The survey data instruments that are the base for this module have been implemented in many countries. A formative assessment with these variables in mind is necessary in order to appropriately contextualize answer choices. Survey inclusion of the PEP/PrEP module would add additional information on sexual violence service access to that captured by this module. The UN Statistical Commission recommends the monitoring of violence using 3 foci within a lifetime and the last 12 months: severity, relationship to perpetrator and frequency1. The core questions listed here focus on frequency, while the secondary questions add the severity and relationship to perpetrator foci, as well as information on service access and delivery. These secondary questions are intended to gather information on the different aspects of sexual violence that may be associated with HIV risk and transmission in these populations (e.g. multiple perpetrator incidents, age-disparity and agency). With questions of such a sensitive nature, it is very important to have a process in place for referral to psychosocial, medical and legal services, should the survey respondent express an interest in accessing services or experience duress as a result of recalling traumatic events. NO. N/A
Variable ID: SV SVMSGS
Sexual Violence: Coercion and Rape
Response Options
Skips
Comments S
Core or Secondary* M T P O
Thank you. Now some questions about unwanted sex. The next few questions are about when someone tricked you, lied to you or threatened you to have sex with them. These are not questions about when someone may have physically forced you to have sex. Please remember that your answers are confidential. If you would like to speak to a health care worker or social worker or anyone else about your experiences, please let the Interview Facilitator know.
SECTION 2: QUESTIONNAIRE MODULES
3 8 383
NO. 1.
Variable ID: SV SVCEVFRS
Sexual Violence: Coercion and Rape In your lifetime, how many times has anyone ever tricked you, lied to you, or threatened you in order to make you have sex with them when you didn’t want to?
Response Options NEVER ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips ‘1’,’7’,’8’→ SKIP TO SVREVFR
Comments S VULNERABILITY C
Core or Secondary* M N/A
T N/A
P N/A
O N/A
2.
SVCAGE
The first time someone tricked you, lied to you, or YEARS: | | | threatened you in order to make you have sex when you MIN: 1 didn’t want to, how old were you? MAX:CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
VULNERABILITY
S
S
S
S
S
3.
SVCFBY
The first time this has happened, what was your relationship to the person who did this to you? If it was more than one person, what was your relationship with the person you remember best. CHECK ALL THAT APPLY
PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORI TY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
‘A’, ‘C’, ‘D’, ‘E’, ‘X’, ‘Y’, ‘Z’ → SKIP TO SVCFCDM
VULNERABILITY
S
S
S
S
S
384
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 4.
Variable ID: SV SVCFBYP
Sexual Violence: Coercion and Rape The first time this happened, what type of partner was this person?
Response Options SPOUSE OR LIVE-IN PARTNER BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/ GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
1 2
3 6 7 8
5.
SVCFCDM
The first time this has happened, was a condom used? If more than one person did this to you, answer about the person you remember best.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
S
6.
SVCFBEH*
The first time this happened, do you think it was because YES you <insert behavior>? NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 8 3 585
NO. 7.
Variable ID: SV SVCYRF
Sexual Violence: Coercion and Rape In the last 12 months, how many times has someone tricked you, lied to you, or threatened you in order to make you have sex when you didn’t want to?
Response Options THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
Skips ‘1’,’7’,’8’→ SKIP TO SVREVFR
Comments S VULNERABILITY C
Core or Secondary* M C
T C
P C
O S
1 2 3 4 5 7 8
8.
SVCYRBEHS*
In the last 12 months when these happened, do you think it was because you <insert risk behavior>?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
C
C
C
C
S
SVMSG2
The next questions will be asking you about the most recent time someone tricked you, lied to you, or threatened you in order to make you have sex when you didn’t want to.
386
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 9.
Variable ID: SV SVCLBY
Sexual Violence: Coercion and Rape The last time this has happened, what was your relationship to the person who did this to you? If it was more than one person, what was your relationship with the person you remember best. CHECK ALL THAT APPLY.
Response Options PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z
Skips ‘A’, ‘C’, ‘D’, ‘E’, ‘X’, ‘Y’, ‘Z’ → SKIP TO SVCLCDM
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
10.
SVCFBYP
The last time this happened, what type of partner was this person?
SPOUSE OR LIVE-IN PARTNER 1 BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/ GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER
1 2
VULNERABILITY
S
S
S
S
S
3 6 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 8 3 787
NO. 11.
Variable ID: SV SVCLCDM
Sexual Violence: Coercion and Rape The last time this has happened, was a condom used? If more than one person did this to you, answer about the person you remember best. YES NO
Response Options 1 2 7 8
Skips
Comments S RISK & VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
DON’T KNOW REFUSE TO ANSWER
12.
SVCLBEH*
The last time this has happened, do you think it was because you <insert behavior>?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
S
N/A
SVMSG3
Sometimes, people will also use physical force to make a person have sex with them. We would like to ask you some questions about this happening to you. NEVER ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8 ‘1’,’7’,’8’ → SKIP TO SVCPRP RISK & VULNERABILITY C C C C S
13.
SVREVFRS
In your lifetime, how many times has anyone ever physically forced you to have sex when you didn’t want to?
388
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 14.
Variable ID: SV SVRAGE
Sexual Violence: Coercion and Rape
Response Options YEARS: | | | MINIMUM: 1 MAXIMUM: CURRENT AGE DON’T KNOW 97 REFUSE TO ANSWER 98
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
The first time someone physically forced you to have sex when you didn’t want to, how old were you?
15.
SVRFBY
What was your relationship to the person who physically forced you to have sex the first time? If it was more than one person, what was your relationship with the person you remember best. CHECK ALL THAT APPLY
PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER
A B C D E X Y Z
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 8 3 989
NO. 16.
Variable ID: SV SVRFBYP
Sexual Violence: Coercion and Rape The first time this happened, what type of partner was this person?
Response Options SPOUSE OR LIVE-IN PARTNER BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/ GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
1 2
3 6 7 8 1 2 7 8 1 2 7 8 1 2 3 4 9998 VULNERABILITY S S S S S VULNERABIL ITY S S S S S RISK & VULNERABILITY S S S S S
17.
SVRFCDM
The first time this had happened, was a condom used? If more than one person did this to you, answer about the person you remember best.
YES NO DON’T KNOW REFUSE TO ANSWER
18.
SVRFBEH*
The first time someone physically forced you to have sex, do you think it was because you <insert risk behavior>]?
YES NO DON’T KNOW REFUSE TO ANSWER
19.
SVRFREP
How many times has this same person (people) physically forced you to have sex?
ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES REFUSE TO ANSWER
390
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 20.
Variable ID: SV SVCYR12
Sexual Violence: Coercion and Rape In the last 12 months, how many times has this happened?
Response Options THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
Skips ‘1’,’7’,’8’→ SKIP TO SVRPRP
Comments S RISK & VULNERABILITY C
Core or Secondary* M C
T C
P C
O S
1 2 3 4 5 7 8
21.
SVCYRFRQS*
In any of the times in the last 12 months when these happened, do you think it was because you sell sex?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
C
N/A
N/A
N/A
N/A
23.
SVCYRFRQM*
In any of the times in the last 12 months when these happened, do you think it was because you have sex with men?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
N/A
C
N/A
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
3 9 3 191
NO. 24.
Variable ID: SV SVCYRFRQT*
Sexual Violence: Coercion and Rape In any of the times in the last 12 months when these happened, do you think it was because you are transgender? YES NO
Response Options 1 2 7 8
Skips
Comments S VULNERABILITY N/A
Core or Secondary* M N/A
T C
P N/A
O N/A
DON’T KNOW REFUSE TO ANSWER
24.
SVCYRFRQP*
In any of the times in the last 12 months when these happened, do you think it was because you inject drugs?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
N/A
N/A
N/A
C
N/A
SVMSG4
The next questions will be asking you about the most recent time someone physically forced you to have sex. The last time this had happened, what was your relationship to the person who did this to you? If it was more than one person, what was your relationship with the person you remember best. CHECK ALL THAT APPLY. PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORITY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z VULNERABILITY S S S S
25.
SVRLBY
392
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 26.
Variable ID: SV SVRLBYP
Sexual Violence: Coercion and Rape The last time this had happened, what type of partner was this person?
Response Options SPOUSE OR LIVE-IN PARTNER BOYFRIEND/GIRLFRIEND SOMEONE THAT I HAVE SEX WITH BUT DO NOT CALL A BOYFRIEND/ GIRLFRIEND OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
1 2
3 6 7 8
27.
SVRLCDM
The last time this had happened, was a condom used? If more than one person did this to you, answer about the person you remember best.
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
RISK & VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 9 393
NO. 28.
Variable ID: SV SVRLSEV
Sexual Violence: Coercion and Rape The last time this had happened, did any of the following happen to you as a result of the time someone physically forced you to have sex when you didn’t want to? CHECK ALL THAT APPLY.
Response Options CUTS, BRUISES OR ACHES A
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
EYE INJURIES, SPRAINS, DISLOCATIONS, OR BURNS B DEEP WOUNDS, BROKEN BONES, BROKEN TEETH OTHER SERIOUS INJURY NONE OF THESE REFUSE TO ANSWER C D E Z
29.
SVRLBEH*
The last time this has happened, do you think it was because you <insert risk behavior>?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
S
30.
SVRLREP
How many times has this same person (people) physically forced you to have sex?
ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES REFUSE TO ANSWER
1 2 3 4 9
RISK & VULNERABILITY
S
S
S
S
S
394
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO.
Variable ID: SV SVMSG5
Sexual Violence: Coercion and Rape
Response Options
Skips
Comments S
Core or Secondary* M T P O
For the next few questions, we will be asking you about any time someone has ever physically forced OR tricked you, lied to you, threatened you in order to make you have sex when you didn’t want to. As a result of someone doing any of those things to you, did you try to seek professional help or services for any of these incidents from any of the following? CHECK ALL THAT APPLY. I DID NOT TRY TO SEEK HELP HEALTHCARE PROFESSIONAL POLICE OR OTHER SECURITY PERSONNEL SOCIAL WORKER, COUNSELOR OR NGO RELIGIOUS LEADER OTHER DON’T KNOW REFUSE TO ANSWER IF ‘A’, ‘Y’, ‘Z’, NO OTHER OPTION CAN BE CHOSEN AS WELL. SKIP TO SVSRVWHY PREVENTION S S S S S
31.
SVSERV
A B C D E X Y Z
SECTION 2: QUESTIONNAIRE MODULES
3 9 3 595
NO. 32.
Variable ID: SV SVSRVREF
Sexual Violence: Coercion and Rape When trying to seek help for what someone had done to you, were you ever refused services by any of the following? Please choose all that ever refused to provide you with services.
Response Options DOCTOR, NURSE, OTHER HEALTHCARE PROFESSIONAL POLICE OR OTHER SECURITY PERSONNEL LAWYER, JUDGE OR OTHER LEGAL PROFESSIONAL SOCIAL WORKER OR COUNSELOR
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1 2
3 4
AN NGO OR OTHER LOCAL ORGANIZATION 5 OTHER DON’T KNOW REFUSE TO ANSWER 6 7 8
396
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 33.
Variable ID: SV SVSRVWHY
Sexual Violence: Coercion and Rape What was the main reason that you did not try to seek professional help or services?
Response Options I DID NOT KNOW WHAT SERVICES WERE AVAILABLE/WHERE TO GO THE SERVICES I WANTED/ NEEDED WERE NOT AVAILABLE I COULD NOT AFFORD SERVICES I WAS UNCOMFORTABLE ACCESSING SERVICES BECAUSE I AM FSW/ MSM/TG/PWID/OVP I DID NOT FEEL THAT I NEEDED SERVICES OTHER DON’T KNOW REFUSE TO ANSWER
Skips IF ‘1’, ‘2’, ‘3’, ‘5’, ‘6’, ‘7’, ‘8’ → SKIP TO SVTELL
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1
2 3
4 5 6 7 8
SECTION 2: QUESTIONNAIRE MODULES
3 9 3 797
NO. 34.
Variable ID: SV SVSRVF
Sexual Violence: Coercion and Rape You said that you were uncomfortable accessing services. Why were you uncomfortable? If none of the answer choices fit, please choose OTHER.
Response Options I WAS WORRIED THAT THE SERVICE PROVIDER WOULD TREAT ME DIFFERENTLY/POORLY IF THEY FOUND OUT I AM MSM/FSW/TG/PWID/OVP
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
1
I WAS AFRAID THAT MY FAMILY/PARTNER/HUSBAN D/WIFE WOULD FIND OUT I AM MSM/FSW/TG/ 2 PWID/OVP I WAS AFRAID OF WHAT THE AUTHORITIES/POLICE WOULD DO TO ME IF THEY FOUND OUT I AM MSM/FSW/TG/PWID/OVP I WAS WORRIED THAT OTHERS WOULD FIND OUT I AM MSM/FSW/TG/ PWID/OVP OTHER REASON DON’T KNOW REFUSE TO ANSWER
3
4 6 7 8
398
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 35.
Variable ID: SV SVTELL
Sexual Violence: Coercion and Rape Who else did you tell about any of these experiences? CHECK ALL THAT APPLY
Response Options PAYING SEX PARTNER NON-PAYING SEX PARTNER POLICE/MILITARY/ AUTHORTY FIGURE RELATIVE FRIEND/ACQUAINTANCE OTHER DON’T KNOW REFUSE TO ANSWER A B C D E X Y Z
Skips
Comments S PREVENTION S
Core or Secondary* M S
T S
P S
O S
36.
SVCPRP
In your lifetime, have you ever tricked, lied to or threatened someone in order to make them have sex with you when they didn’t want to?
YES NO I DON’T KNOW/DON’T REMEMBER DON’T KNOW REFUSE TO ANSWER
1 2 3 7 8 1 2 3 4 7 8
‘2’,’7’,’8’ → SKIP TO SVRPRP
OTHER
C
C
C
C
C
37.
SVCPRPFR
In your lifetime, how many times have you done this to someone?
ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
OTHER
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
3 9 3 999
NO. 38.
Variable ID: SV SVCPRPYR
Sexual Violence: Coercion and Rape In the last 12 months, how many times have you done this to someone?
Response Options THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
1 2 3 4 5 7 8
39.
SVCPRPWH
Please think about the most recent time that you have tricked, lied to, or threatened in order to make someone have sex with you when they didn’t want to. Who was the person you did this to?
SPOUSE/PARTNER/ BOYFRIEND/GIRLFRIEND FRIEND/ACQUAINTANCE FAMILY/RELATIVE FRIEND/ACQUAINTANCE STRANGER OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
OTHER
S
S
S
S
S
40.
SVRPRP
In your lifetime, have you ever physically forced someone to have sex with you when they didn’t want to?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
‘2’,’7’,’8’ → SKIP TO END OF CURRENT SECTION
OTHER
C
C
C
C
C
400
31. SEXUAL VIOLENCE: COERCION AND RAPE
NO. 41.
Variable ID: SV SVRPRPFR
Sexual Violence: Coercion and Rape In your lifetime, wow many times have you done this to someone?
Response Options ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER 1 2 3 4 7 8
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
42.
SVRPRPYR
In the last 12 months, how many times have you done this to someone?
THIS HAS NOT HAPPENED IN THE LAST 12 MONTHS ONCE 2 – 5 TIMES 6 – 10 TIMES MORE THAN 10 TIMES DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
OTHER
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 0 4 101
NO. 43.
Variable ID: SV SVRPRPWH
Sexual Violence: Coercion and Rape
Response Options SPOUSE/PARTNER/ BOYFRIEND/GIRLFRIEND FRIEND/ACQUAINTANCE FAMILY/RELATIVE FRIEND/ACQUAINTANCE STRANGER OTHER DON’T KNOW REFUSE TO ANSWER
Skips
Comments S OTHER S
Core or Secondary* M S
T S
P S
O S
Please think about the most recent time that you have physically forced someone to have sex with you when they didn’t want to. Who was the person you did this to?
1 2 3 4 5 6 7 8
#: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
*Similar questions can be found in the questionnaire module Shame, Stigma, Harassment & Discrimination. 1
http://unstats.un.org/unsd/demographic/meetings/vaw/docs/ConclusionsAndRecommendations.pdf
Additional Sources of Violence Indicators: UN General Assembly Human rights Council (2008) A/HRC/7/6: http://www.ohchr.org/Documents/Press/OHCHR_Report_2008.pdf MEASURE Evaluation (2008). Violence against Women and girls. A compendium of Monitoring and Evaluation Indicators. http://www.cpc.unc.edu/measure/publications/pdf/ms-08-30.
402
31. SEXUAL VIOLENCE: COERCION AND RAPE
Section
Il N/A N/A
32. INTERNALIZED HOMOPHOBIA
Internalized homophobia is the personal acceptance and endorsement of sexual stigma as part of a person’s value system and self-concept. Researchers should adapt the terminology ‘internalized homophobia’ to the way the person self-identified in demographic or group characteristics modules. The module consists of the 5-item revised internalized homophobia (IHP-R) scale, recommended by the 3rd edition of Diagnosis and Statistics Manual (DSM) by the American Psychiatric Association. This module should be used for surveys among MSM and TG, but it can also be used on lesbian or bisexual respondents.
Variable ID: IH IH1MSG
Internalize Homophobia
Response Options
Skips
Comments S
Core or Secondary* M T P O
Next we will ask you some questions on how you feel about your sexual attraction to men.
1.
IHATTRAC
I have tried to stop being attracted to men in general.
STRONGLY AGREE SOMEWHAT AGREE NEITHER AGREE NOR DISAGREE SOMEWHAT DISAGREE STRONGLY DISAGREE REFUSE TO ANSWER
1 2 3 4 5 8
VULNERABILITY
N/A
S
S
N/A
N/A
SECTION 2: QUESTIONNAIRE MODULES
4 0 4 303
N/A 2.
Variable ID: IH IHHETER
Internalize Homophobia If someone offered me the chance to be completely heterosexual, I would accept the chance.
Response Options STRONGLY AGREE SOMEWHAT AGREE NEITHER AGREE NOR DISAGREE SOMEWHAT DISAGREE STRONGLY DISAGREE REFUSE TO ANSWER 1 2 3 4 5 8
Skips
Comments S VULNERABILITY N/A
Core or Secondary* M S
T S
P N/A
O N/A
3.
IHNOTGAY
I wish I weren’t gay/bisexual.
STRONGLY AGREE SOMEWHAT AGREE NEITHER AGREE NOR DISAGREE SOMEWHAT DISAGREE STRONGLY DISAGREE REFUSE TO ANSWER
1 2 3 4 5 8
VULNERABILITY
N/A
S
S
N/A
N/A
4.
IHSHORT
I feel that being gay/bisexual is a personal shortcoming for me.
STRONGLY AGREE SOMEWHAT AGREE NEITHER AGREE NOR DISAGREE SOMEWHAT DISAGREE STRONGLY DISAGREE REFUSE TO ANSWER
1 2 3 4 5 8
VULNERABILITY
N/A
S
S
N/A
N/A
404
32. INTERNALIZED HOMOPHOBIA
N/A 5.
Variable ID: IH IHHELP
Internalize Homophobia I would like to get professional help in order to change my sexual orientation from gay/bisexual to straight.
Response Options STRONGLY AGREE SOMEWHAT AGREE NEITHER AGREE NOR DISAGREE SOMEWHAT DISAGREE STRONGLY DISAGREE REFUSE TO ANSWER 1 2 3 4 5 8
Skips
Comments S VULNERABILITY N/A
Core or Secondary* M S
T S
P N/A
O N/A
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
SCORING: Sum responses and divide by the total number of items, thereby maintaining the 1-5 metric for ease of interpretation. Higher scores indicate more negative self-attitudes. References: 1. Assessing Internalized Sexual Stigma (“Internalized Homophobia”) in Sexual Minority Adults. Accessed at: http://psychology.ucdavis.edu/rainbow/html/ihpitems.html 2. American Psychiatric Association (1980). Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). Washington, DC: Author. 3. Herek, G. M., Cogan, J. C., Gillis, J. R., & Glunt, E. K. (1998). Correlates of internalized homophobia in a community sample of lesbians and gay men. Journal of the Gay and Lesbian Medical Association, 2, 17-25. 4. Herek, G. M., & Glunt, E. K. (1995). Identity and community among gay and bisexual men in the AIDS era: Preliminary findings from the Sacramento Men’s Health Study. In G. M. Herek, & B. Greene (Eds.), AIDS, identity, and community: The HIV epidemic and lesbians and gay men (pp. 55-84). Thousand Oaks, CA: Sage. 5. Meyer, I. H. (1995). Minority stress and mental health in gay men. Journal of Health and Social Behavior, 36, 38-56
SECTION 2: QUESTIONNAIRE MODULES
4 0 4 505
Section
II
33. DEPRESSION
This module evaluates depression using the Patient Health Questionnaire (PHQ). The PHQ-2 and PHQ-9 are validated tools used to assess and monitor depression severity. When evaluating depression either the PHQ-2 with two questions or the more extensive PHQ-9, nine questions may be used. The PHQ-9 includes nine items and is calculated by assigning scores of 0, 1, 2, and 3, to the response categories of: not at all, several days, more than half the day, and nearly every day. The PHQ-9 total score ranges from 0 to 27. PHQ-9 score of 10 or more has a sensitivity of 88% and a specificity of 88% for major depression. The scores of 5, 10, 15, and 20 denote cut-points for mild, moderate, moderately severe and severe depression, respectively. The PHQ-2, comprised of the first two items of the PHQ-9, inquires about the degree to which an individual has experienced depressed mood and anhedonia for the past two weeks. Its purpose is to screen for depression, not to establish a final diagnosis or monitor depression. PHQ-2 scores range from 0-6, with a score of 3 or more has a sensitivity of 83% and specificity of 92% for major depression. If a brief measure is desirable for use or as part of a comprehensive health questionnaire, researchers may want to use PHQ-2 instead of PHQ-9. Patients who screen positive using PHQ-2 should then be further evaluated with the PHQ-9 to determine whether or not they meet criteria for a depressive disorder. Studies should have mental health referral protocols to assure access to mental health services for respondents with moderate to severe depression. N/A N/A
Variable ID: DP DP1MSG
Depression
Response Options
Skips
Comments
Core or Secondary* S M T P O
Next we will ask you some questions on depression. Over the last 2 weeks, how often have you been bothered by any of the following problems:
1.
DPINTRST
Little interest or pleasure in doing things
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
0 1 2 3 8
PHQ-2 PHQ-9 VULNERABILITY
S
S
S
S
S
406
33. DEPRESSION
N/A 2.
Variable ID: DP DPDOWN
Depression Feeling down, depressed, or hopeless
Response Options NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER 0 1 2 3 8 0 1 2 3 8 0 1 2 3 8 0 1 2 3 8
Skips IF DPINTRST=0 AND DPDOWN=0 → SKIP TO END OF CURRENT SECTION
Comments PHQ-2 PHQ-9 VULNERABILITY SKIP IS OPTIONAL.
Core or Secondary* S S
M S
T S
P S
O S
3.
DPSLEEP
Trouble falling or staying asleep, or sleeping too much
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
4.
DPTIRE
Feeling tired or having little energy
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
5.
DPEAT
Poor appetite or overeating
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 0 4 707
N/A 6.
Variable ID: DP DPFAIL
Depression Feeling bad about yourself or that you are a failure or have let yourself or your family down
Response Options NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER 0 1 2 3 8 0 1 2 3 8 0 1 2 3 8 0 1 2 3 8
Skips
Comments PHQ-9 VULNERABILITY
Core or Secondary* S S
M S
T S
P S
O S
7.
DPREAD
Trouble concentrating on things, such as reading the newspaper or watching television
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
8.
DPMOVE
Moving or speaking so slowly that other people could have noticed? Or the opposite—being so fidgety or restless that you have been moving around a lot more than usual
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
9.
DPHURT
Thoughts that you would be better off dead or of hurting yourself in some way
NOT AT ALL SEVERAL DAYS MORE THAN HALF THE DAYS NEARLY EVERY DAY REFUSE TO ANSWER
PHQ-9 VULNERABILITY
S
S
S
S
S
408
33. DEPRESSION
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: Kroenke K, Spitzer R.L., & Williams J B.W. The PHQ-9: Validity of a Brief Depression Severity Measure. J Gen Intern Med. 2001 September; 16(9): 606- 613. Accessed at: http://www.ncbi.nlm. nih.gov/pmc/articles/PMC1495268/pdf/jgi_01114.pdf Kroenke K, Spitzer R.L., & Williams J B.W. The Patient Health Questionnaire-2: Validity of a Two-Item Depression Screener. 2003. Medical Care. 41(11): 1284-1292. Accessed at: http://www.ihs. gov/crs/documents/phq_2_medical_care.pdf Spitzer R.L., Williams J B.W., Kroenke K., and colleagues developed with an educational grant from Pfizer Inc. No permission required to reproduce, translate, display or distribute. Accessed at: http://www.phqscreeners.com/pdfs/02_PHQ-9/English.pdf
SECTION 2: QUESTIONNAIRE MODULES
4 0 4 909
Section
Il
34. SUICIDE IDEATION AND ATTEMPTS
The suicide ideation and attempts module measures suicide-related thoughts and behaviors. Questions have been adapted from a validated instrument, the Suicide Behaviors Questionnaire-Revised (SBQ-R). SBQ-R is made up of four items each assessing a different dimension of the risk of suicide (or suicidality). The coding of this scale should not be changed because a score is calculated. If these questions are used, consent needs to state that if participants are deemed suicidal then participants will need to be referred to mental health professionals and services. Due to the sensitive nature of the questions included in this scale, a participant may Refused them. If this is the case, a score cannot be calculated for that participant. This module can be used in conjunction with other modules such as stigma (shame, stigma, harassment & discrimination), internalized homophobia, depression etc. to gain a thorough understanding of psychosocial measures.
N/A Variable ID: SU N/A SU1MSG
Suicide Ideation and Attempts
Response Options
SKIPS
COMMENTS S
Core or Secondary* M T P O
Next we will ask you some questions on thoughts of ending your life and actions taken to end your life. Please choose the number beside the statement or phrase that best applies to you.
410
34. SUICIDE IDEATION AND ATTEMPTS
N/A Variable ID: SU 1. SUIDEA
Suicide Ideation and Attempts Have you ever thought about or attempted to kill yourself? NEVER
Response Options 1 2
SKIPS
COMMENTS S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
IT WAS JUST A BRIEF PASSING THOUGHT I HAVE HAD A PLAN AT LEAST ONCE TO KILL MYSELF BUT DID NOT TRY TO DO IT I HAVE HAD A PLAN AT LEAST ONCE TO KILL MYSELF AND REALLY WANTED TO DIE
3
4
I HAVE ATTEMPTED TO KILL MYSELF, 5 BUT DID NOT WANT TO DIE I HAVE ATTEMPTED TO KILL MYSELF, AND REALLY HOPED TO DIE 6 REFUSE TO ANSWER 8
2.
SUFREQ
How often have you thought about killing yourself in the past year?
NEVER RARELY (1 TIME) SOMETIMES (2 TIMES) OFTEN (3-4 TIMES) VERY OFTEN (5 OR MORE TIMES) REFUSE TO ANSWER
1 2 3 4 5 8
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 1 4 111
N/A Variable ID: SU 3. SUTOLD
Suicide Ideation and Attempts Have you ever told someone you were going to commit suicide, or that you might do it? NO
Response Options 1 2 3 4 5 8
SKIPS
COMMENTS S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
YES, AT ONE TIME, BUT DID NOT REALLY WANT TO DIE YES, AT ONE TIME, AND REALLY WANTED TO DIE YES, MORE THAN ONCE, BUT DID NOT WANT TO DO IT YES, MORE THAN ONCE, AND REALLY WANTED TO REFUSE TO ANSWER
4.
SUTRY
How likely is it that you will attempt suicide someday?
NEVER NO CHANCE AT ALL RATHER UNLIKELY UNLIKELY LIKELY RATHER LIKELY VERY LIKELY REFUSE TO ANSWER
0 1 2 3 4 5 6 8
VULNERABILITY
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: The Suicide Behaviors Questionnaire-Revised (SBQ-R) http://www.integration.samhsa.gov/images/res/SBQ.pdf
412
34. SUICIDE IDEATION AND ATTEMPTS
PSYCHOMETRIC PROPERTIES Cutoff score Adult general population Adult Psychiatric Inpatients ≥7 ≥8
Sensitivity 97% 80%
Specificity 95% 91%
SBQ-R SCORING Item 1: taps into lifetime suicide ideation and/or suicide attempts Selected response 1 Selected response 2 Selected response 3 or 4 Selected response 5 or 6 Non-Suicidal subgroup Suicide Risk Ideation subgroup Suicide Plan subgroup Suicide Attempt subgroup Total pts.: 1 point 2 points 3 points 4 points
Item 2: assess the frequency of suicide ideation over the past 12 months Selected Response: Never Rarely (1 time) Sometimes (2 times) Often (3-4 times) Very Often (5 or more times) Total pts.: 1 point 2 points 3 points 4 points 5 points
SECTION 2: QUESTIONNAIRE MODULES
4 1 4 313
Item 3: taps into the threat of suicide attempt Selected response 1 Selected response 2 or 3 Selected response 4 or 5 Total pts.: 1 point 2 points 3 points
Item 4: evaluates self-reported likelihood of suicidal behavior in the future Selected Response: Never No chance at all Rather unlikely Unlikely Likely Rather Likely Very Likely Sum all the scores selected by the respondents. The total score should range from 3-18. Total pts.: 0 points 1 point 2 points 3 points 4 points 5 points 6 points
414
34. SUICIDE IDEATION AND ATTEMPTS
Section
II
35. SOCIAL SUPPORT
This module measures social support from friends, family and other persons the respondent feels close to. The module contains two validated instruments, the MSP-SS (Multidimensional Scale of Perceived Social Support) and the MOS-SS (Medical Outcomes Study: Social Support Survey). The MSP-SS, comprised of 12 items, measures perceived availability of social support from family, friends, and significant other. Each item is measured on a likert-scale and scores are summed with possible score range between 12 and 84, the higher the score the higher the perceived social support. MSP-SS scale has a high internal-consistency reliability of 0.91 (Cronbach’s coefficient alpha). The MOS-SS was developed for participants in the Medical Outcomes Study (MOS), a two-year longitudinal study of the process and outcomes of care for people with prevalent and treatable chronic conditions. It is comprised of 19 items and has an overall functional social support index as well as four separate social support subscales (emotional/ informational support, tangible support, affectionate support, positive social interaction). A higher score for an individual score or for overall support index indicates more support. The overall MOS-SS scale has a high internal-consistency reliability of 0.97. Researchers should use MSP-SS if their goal is to measure social support in terms of who provides respondents that support or if brevity is an issue. If interested in the types of support as well as an overall measure of support, then the MOS-SS should be used. N/A N/A
Variable ID: SS SS1MSG
Social Support
Response Options
Skips
Comments S
Core or Secondary* M T P O
Next some statements about help and support from your friends, family or special persons. Special persons may be any other person you feel close to, such as your sex partner or other person you feel close to. Please tell us whether you agree or disagree with the statements.
SECTION 2: QUESTIONNAIRE MODULES
4 1 4 515
N/A 1.
Variable ID: SS SSSPHER
Social Support There is a special person who is around when I am in need.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
2.
SSSPSHR
There is a special person with whom I can share my joys and sorrows.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
S
416
35. SOCIAL SUPPORT
N/A 3.
Variable ID: SS SSFMHLP
Social Support My family really tries to help me.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
4.
SSFMEM
I get the emotional help and support I need from my family.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 1 4 717
N/A 5.
Variable ID: SS SSSPCMF
Social Support I have a special person who is a real source of comfort to me.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
6.
SSFRHLP
My friends really try to help me.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
1 2 3 4 5 6 7 8
VULNERABILITY
S
S
S
S
S
418
35. SOCIAL SUPPORT
N/A 7.
Variable ID: SS SSFRCNT
Social Support I can count on my friends when things go wrong.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
8.
SSFMTLK
I can talk about my problems with my family.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
1 2 3 4 5 6 7 8
VULNERABILITY
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 1 4 919
N/A 9.
Variable ID: SS SSFRSHR
Social Support I have friends with whom I can share my joys and sorrows.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
10.
SSSPCAR
There is a special person in my life that cares about my feelings.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
1 2 3 4 5 6 7 8
VULNERABILITY
S
S
S
S
S
420
35. SOCIAL SUPPORT
N/A 11.
Variable ID: SS SSFMHLP
Social Support My family is willing to help me make decisions.
Response Options VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER 1 2 3 4 5 6 7 8
Skips
Comments S VULNERABILITY S
Core or Secondary* M S
T S
P S
O S
12.
SSFRTLK
I can talk about my problems with my friends.
VERY STRONGLY DISAGREE STRONGLY DISAGREE MILDLY DISAGREE NEUTRAL MILDLY AGREE STRONGLY AGREE VERY STRONGLY AGREE REFUSE TO ANSWER
1 2 3 4 5 6 7 8
VULNERABILITY
S
S
S
S
S
*S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary References
References Donald Sherbourne, D., & Stewart, A.L. (1991). The MOS Social Support Survey. Soc. Sci. Med, 32 (6), 705-714. Zimet, G.D., Dahlem, N.W., Zimet, S.G. & Farley, G.K. (1988). The Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment, 52, 30-41. Survey accessed at: http://www.yorku.ca/rokada/psyctest/socsupp.pdf
SECTION 2: QUESTIONNAIRE MODULES
4 2 4 121
Multidimensional Scale of Perceived Support Scoring Instructions Sum the score of all 12 items, with the possible score range for total is 7-84. Scoring Algorithm Total= 69-84 High Acuity Total= 49-68 Moderate Acuity Total= 12-48 Low Acuity OR
N/A N/A
SS SS2MSG
Social Support
Response Options
Skips N/A
Comments S
Core Or Secondary* M T P O
People sometime look to others for companionship, assistance, or other types of support. How often is each of the following kinds of support available to you if you need it?
N/A 1.
N/A SSLISTE
Emotional/Information Support Someone you can count on to listen to you whenyou need to talk. NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8 S S S S S
422
35. SOCIAL SUPPORT
N/A 2.
SS SSINFO
Social Support Someone to give you information to help you understand a situation.
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 3 4 5 7 8 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O S
3.
SSADVIC
Someone to give you good advice about a crisis.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
S
S
S
S
S
4.
SSCONFI
Someone to confide in or talk to about yourself or your problems.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 2 4 323
N/A 5.
SS SSWANTA
Social Support Someone whose advice you really want.
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 3 4 5 7 8 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O
6.
SSSHARE
Someone to share your most private worries and fears with.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
S
S
S
S
S
7.
SSSUGGE
Someone to turn to for suggestions about how to deal with a personal problem.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
S
S
S
S
S
424
35. SOCIAL SUPPORT
N/A 8.
SS SSPROBL
Social Support Someone who understands your problems.
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O S
N/A 9.
N/A SSCONFI
Tangible Support Someone to help you if you were confined to bed. NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER 1 2 3 4 5 7 1 2 3 4 5 7 8 S S S S S S S S S S
10.
SSDOCTO
Someone to take you to the doctor if you needed it.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
SECTION 2: QUESTIONNAIRE MODULES
4 2 4 525
N/A 11.
SS SSMEALS
Social Support Someone to prepare your meals if you were unable to do it yourself.
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O S
12.
SSCHORE
Someone to help with daily chores if you were sick.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
1 2 3 4 5 7 8
S
S
S
S
S
426
35. SOCIAL SUPPORT
N/A N/A 13.
SS N/A SSLOVEA
Social Support Affectionate Support Someone who shows you love and affection.
Response Options
Skips
Comments S
Core Or Secondary* M T P O
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
1 2 3 4 5 7 8
S
S
S
S
S
14.
SSWANTE
Someone to love and make you feel wanted.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
1 2 3 4 5 7 8
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 2 4 727
N/A 15.
SS SSHUGS Someone who hugs you.
Social Support
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O S
N/A 16.
N/A SSGOODT
Positive Social Interaction Someone to have a good time with. NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8 S S S S S
428
35. SOCIAL SUPPORT
N/A 17.
SS SSRELAX
Social Support Someone to get together for relaxation.
Response Options NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 3 4 5 7 8
Skips
Comments S S
Core Or Secondary* M S
T S
P S
O S
18.
SSENJOY
Someone to do something enjoyable with.
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
S
S
S
S
S
SECTION 2: QUESTIONNAIRE MODULES
4 2 4 929
N/A N/A 19.
SS N/A SSMIND
Social Support Additional Item Someone to do things with to help you get your mind off things.
Response Options
Skips
Comments S
Core Or Secondary* M T P O
NONE OF THE TIME A LITTLE OF THE TIME SOME OF THE TIME MOST OF THE TIME ALL OF THE TIME DON’T REMEMBER REFUSE TO ANSWER
1 2 3 4 5 7 8
S
S
S
S
S
*: S=Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
Medical Outcomes Study: Social Support Survey Scoring (MOS-SS) Instructions • To obtain a score for each subscale, calculate the average of the scores for each item in the subscale. • To obtain an overall support index, calculate the average of the scores for all 18 items in the four subscales, and the score for the one additional item. A higher score indicates more support. Reference: Sherbourne, C.D., & Stewart, A.L. (1991). The MOS Social Support Survey. Social Science Medicine, 52, 6, 705-714. Survey accessed at: http://www.rand.org/content/dam/rand/www/external/health/surveys_tools/mos/mos_socialsupport_survey.pdf
430
35. SOCIAL SUPPORT
Section
II
36. SOCIAL COHESION, SOCIAL PARTICIPATION AND COLLECTIVE AGENCY
Recent research among FSW and MSM has shown associations between higher levels of social cohesion, social participation and collective agency and better HIV prevention and care outcomes. This questionnaire module contains items to assess respondents’ social cohesion, social participation, and degree of collective agency. Social cohesion is defined here as the mutual assistance or support that respondents may garner from their connection and interaction with others in their KP group. Social participation items assess respondents’ involvement in specific activities with their KP community to build support networks and a sense of belonging. Collective agency is defined as the choice, control, and power that poor or marginalized groups have to act for themselves to claim their rights (whether civil, political, economic, social or cultural) and to hold others accountable for these rights.
N/A N/A
Variable ID: ST N/A
Social Cohesion
Response Options
Skips
Comments S Previously used in SW, MSM
Core Or Secondary* M T P O
The next several questions are about your social life and your relationships with other (sex workers/gay men or other men who have sex with men/transgender women/people who inject drugs). Please mark if you strongly disagree, disagree, are neutral, agree, or strongly agree with the statements.
1.
SCMONEY
You can count on other [insert key population group] if you need to borrow money.
STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
S
S
S
S
N/A
SECTION 2: QUESTIONNAIRE MODULES
4 3 4 131
N/A 2.
Variable ID: ST SCDOC
Social Cohesion You can count on other [insert key population group] to accompany you to the doctor or hospital.
Response Options STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 3 4 5 7 8 1 2 3 4 5 7 8
Skips
Comments S VULNERABILITY S
Core Or Secondary* M S
T S
P S
O N/A
3.
SCTALK
You can count on other [insert key population group] if you need to talk about your problems.
STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
N/A
4.
SCSTAY
You can count on other [insert key population group] if you need somewhere to stay.
STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
N/A
432
36. SOCIAL COHESION, SOCIAL PARTICIPATION AND COLLECTIVE AGENCY
N/A 5.
Variable ID: ST SCDEAL
Social Cohesion You can count on other [insert key population group] to help you deal with a violent or difficult [client/partner/person in your life].
Response Options STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments S VULNERABILITY S
Core Or Secondary* M S
T S
P S
O N/A
6.
SCINJ
You can count on other [insert key population group] to support your use of [condoms/ sterile injecting equipment].
STRONGLY DISAGREE DISAGREE NEUTRAL AGREE STRONGLY AGREE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
VULNERABILITY
S
S
S
S
N/A
N/A
N/A
SOCIAL PARTICIPATION The next several questions ask about social activities you may have taken part in over the past 12 months. Please mark how often you have done any of these social activities.
Ayala citation. This is an adapted subset of scale used in 2012 Global MSM Survey
SECTION 2: QUESTIONNAIRE MODULES
4 3 4 333
N/A 7.
Variable ID: ST SCORG
Social Cohesion During the past 12 months, how often have you participated in a community organization for [insert key population]?
Response Options NEVER DAILY WEEKLY MONTHLY LESS THAN MONTHLY DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8 1 2 3 4 5 7 8 1 2 3 4 5 7 8
Skips
Comments S VULNERABILITY S
Core Or Secondary* M S
T S
P S
O N/A
8.
SCGROUP
During the past 12 months, how often have you gone to a [insert key population group] support group?
NEVER DAILY WEEKLY MONTHLY LESS THAN MONTHLY DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
N/A
9.
SCSOCIAL
During the past 12 months, how often have you participated in social groups or activities for [insert key population]?
NEVER DAILY WEEKLY MONTHLY LESS THAN MONTHLY DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
N/A
434
36. SOCIAL COHESION, SOCIAL PARTICIPATION AND COLLECTIVE AGENCY
N/A 10.
Variable ID: ST SCREAD
Social Cohesion During the past 12 months, how often have you read [insert key population] materials online or in magazines?
Response Options NEVER DAILY WEEKLY MONTHLY LESS THAN MONTHLY DON’T KNOW REFUSE TO ANSWER 1 2 3 4 5 7 8
Skips
Comments S VULNERABILITY N/A
Core Or Secondary* M S
T S
P N/A
O N/A
N/A
COLLECTIVE AGENCY The next several questions ask about things you may have done to help others. Please indicate if you have done any of the following in the past 12 months.
Used in Avahan
11.
SCPOLI
In the past 12 months, have you negotiated with or stood up against police in order to help a fellow [insert key population group]?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8
VULNERABILITY
S
S
S
S
N/A
SECTION 2: QUESTIONNAIRE MODULES
4 3 4 535
N/A 12.
Variable ID: ST SCMAD
Social Cohesion In the past 12 months, have you negotiated with or stood up against a madam/broker/ pimp in order to help a fellow sex worker? YES NO
Response Options 1 2 7 8
Skips
Comments S VULNERABILITY Define madam/ broker/pimp based on local context. S
Core Or Secondary* M N/A
T N/A
P N/A
O N/A
DON’T KNOW REFUSE TO ANSWER
13.
SCGANG
In the past 12 months, have you negotiated with or stood up against a local gang member (thug, hooligan or gangster) in order to help a fellow [insert key population group]?
YES NO DON’T KNOW REFUSE TO ANSWER
1 2 7 8 1 2 7 8
VULNERABILITY Define local gang member based on local context.
S
S
S
S
N/A
14.
SCPART
In the past 12 months, have you negotiated with or stood up against clients/any other sexual partner in order to help a fellow [insert key population group]?
YES NO DON’T KNOW REFUSE TO ANSWER
VULNERABILITY
S
S
S
S
N/A
*: S =Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: • • • • Lippman SA, Donini A, Diaz J, Chinaglia M, Reingold A, et al. Social-environmental factors and protective sexual behavior among sex workers: the Encontros intervention in Brazil. AJPH. 2010; 100 Suppl 1: S216-223. Fonner VA, Kerrigan D, Mnisi Z, Ketende S, Kennedy CE, Baral S. Social cohesion, social participation, and HIV related risk among female sex workers in Swaziland. PLoS One. 2014 Jan 31;9(1):e87527 Ayala G, Makofane K, Santos G, et al. Access to Basic HIV-Related Services and PrEP Acceptability among Men Who Have sex with Men Worldwide: Barriers, Facilitators, and Implications for Combination Prevention. Journal of Sexually Transmitted Diseases, vol. 2013, doi:10.1155/2013/953123 Saggurti N, Mishra RM, Proddutoor L, Tucker S, Kovvali D, Parimi P, Wheeler T. Community collectivization and its association with consistent condom use and STI treatment seeking behaviors among female sex workers and high-risk men who have sex with men/transgenders in Andhra Pradesh, India. AIDS Care. 2013; 25 Suppl 1:S55-66. doi: 10.1080/09540121.2012.749334.
436
36. SOCIAL COHESION, SOCIAL PARTICIPATION AND COLLECTIVE AGENCY
Section
II
37. GAME OF CONTACTS
The Game of Contacts (GoC) is a technique to estimate (quantify) the social visibility of hard-to-count populations such as sex workers, people who inject drugs, or men who have sex with men. The primary purpose of the GoC is to provide an important adjustment factor for network scale-up surveys1 that estimate the population size of socially hidden populations. In the absence of a planned network scale-up survey there may be little or no utility to collect GoC data.
The GoC interview may mimic a game by using a board as a data collection form. The GoC yields a social visibility estimate between 0 (zero social visibility) and 1 (complete social visibility). As part of this interview, the respondent randomly selects one first name after the other from a set of 24 pre-arranged names. These names should represent the entire universe of people in the survey locale and include both frequent and infrequent names, selected from representative data bases such as DHS survey data, professional associations or other large unbiased data bases. For each name selected the respondent may have 0, 1, or more contacts he/she knows personally. The respondent is then probed whether the contact knows about the respondent’s defining behavior and whether the contact practices the same behavior. This will yield four different answer combinations. The interview continues until all 24 names are exhausted. The question about sharing a meal may serve as a proxy indicator for social closeness. More information is available through the referenced literature at the end of this chapter. N/A Variable ID: GC GC1
Game of Contacts
Response Options
Skips
Comments
Core or Secondary*
1.
How many people called “CONTACT NAME” do you know? Remember by “People you know” we mean people you know by name or sight. These should only be people who know you too. Type ‘0’ if you don’t know anyone called “CONTACT NAME”
[ | ] MIN: 0 MAX: 10 DON’T KNOW 97 REFUSE TO ANSWER 98
KA
S
Adapted from: 1
Network scale-up surveys are conducted among the general population with the purpose to estimate the size of hard-to-count populations. Because network scale-up surveys are not conducted in key populations themselves, they are not described in this guideline.
SECTION 2: QUESTIONNAIRE MODULES
4 3 4 737
N/A
Variable ID: GC GS1MSG
Game of Contacts
Response Options
Skips
Comments
Core or Secondary*
Now let’s talk about each of these “CONTACT NAME’s” you know. Let’s start with the one you know best. In the past 6 months, did you share a meal with “CONTACT NAME”?? YES NO DON’T KNOW REFUSE TO ANSWER 1 2 7 8 1 2 7 8 1 2 7 8 KA KA KA
S
2.
GC1AML
S
3.
GC1AKN
Does “CONTACT NAME” know that you [insert defining risk behavior] ?
YES NO DON’T KNOW REFUSE TO ANSWER
S
4.
GC1ADO
Does “CONTACT NAME” [insert defining risk behavior] himself/herself?
YES NO DON’T KNOW REFUSE TO ANSWER
S
Continue with the remaining names/cards, using the same question phrasing and correspondingly altered variable names. *: C=Core, S=Secondary question for all KP groups.
References: The Game of Contacts: Estimating the Social Visibility of Groups. Salganik MJ, Mello MB, Abdo AH, Bertoni N, Fazito D, Bastos FI. Soc Networks. 2011 Jan 1;33(1):70-78 Assessing network scale-up estimates for groups most at risk of HIV/AIDS: evidence from a multiple-method study of heavy drug users in Curitiba, Brazil. Salganik, MJ, Fazito D, Bertoni N, Abdo AH, Mello MB, Bastos FI. Am J Epidemiol. 2011 Nov 15:174(10): 1190-6. Estimating the Sizes of Populations at High Risk for HIV: A Comparison Study. Jing L, Qu C, Yu H, Wang T, Cui Y. PLoS One. 2014 Apr 22;9(4).
438
37. GAME OF CONTACTS
Section
II N/A
38. HIV KNOWLEDGE AND PERCEPTIONS
Correct knowledge of HIV and AIDS is critical for adoption of behaviors which lower HIV transmission risk. Also important is the rejection of misconceptions about modes of HIV transmission that can weaken one’s adoption of safer sexual behavior or reinforce the stigma faced by people living with HIV/AIDS.
Variable ID: HK HK1MSG
HIV Knowledge and Perceptions
Response Options
Skips
Comments S
Core / Secondary* M T P O
N/A
Thank you. Next we will ask you some questions about what you know about HIV.
1.
HKONEPAR
Can the risk of HIV transmission be reduced by having sex with only one uninfected partner who has no other partners?
YES NO DON’T REMEMBER REFUSE TO ANSWER
1 2 7 8
KA
C
C
C
C
C
2.
HKCONUSE Can a person reduce the risk of getting HIV by using a condom every time they have sex?
YES NO DON’T REMEMBER REFUSE TO ANSWER
1 2 7 8
KA
C
C
C
C
C
SECTION 2: QUESTIONNAIRE MODULES
4 3 4 939
N/A
Variable ID: HK HKHEALTH
HIV Knowledge and Perceptions
Response Options
Skips
Comments S
Core / Secondary* M C
T C
P C
O C
3.
Can a healthy-looking person have HIV?
YES NO DON’T REMEMBER REFUSE TO ANSWER
1 2 7 8 1 2 3 8 1 2 3 8 1 2 3 8 1 2 3 8
KA
C
4.
HKMOSBIT
Can a person get HIV from mosquito bites?
YES NO DON’T KNOW REFUSE TO ANSWER
KA Can be replaced by most common misconceptions in local context.
C
C
C
C
C
5.
HKSRFOOD
Can a person get HIV by sharing food with someone who is infected?
YES NO DON’T KNOW REFUSE TO ANSWER
KA Can be replaced by most common misconceptions in local context
C
C
C
C
C
6.
HKNEEDLE
Can a person get HIV by getting injections with a needle that was already used by someone else?
YES NO DON’T KNOW REFUSE TO ANSWER
KA
S
S
S
C
S
7.
HKINJ
Can people who inject drugs protect themselves from HIV YES by switching to that you swallow, sniff, or inhale? NO DON’T KNOW REFUSE TO ANSWER
KA
S
S
S
C
S
440
38. HIV KNOWLEDGE AND PERCEPTIONS
N/A
Variable ID: HK HKMORSK
HIV Knowledge and Perceptions
Response Options
Skips
Comments S
Core / Secondary* M C
T C
P C
O C
8.
If a condom is not used, what kind of sex puts you at the most risk for HIV?
ORAL SEX VAGINAL SEX ANAL SEX MUTUAL MASTURBATION DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 8 1 2 3 4 5 8 1 2 7 8 SKIP IF FEMALE. SKIP IF LIMMANAL= ’2’, ‘7’, OR ‘8’
KA
C
9.
HKANTYP
If a condom is not used, what kind of anal sex puts you at most risk for HIV?
ACTIVE (TOP) ANAL SEX PASSIVE (BOTTOM) ANAL SEX BOTH HAVE SAME RISK BOTH HAVE NO RISK DON’T KNOW REFUSE TO ANSWER
KA
S
C
C
S
S
10.
HKBEHA
Do you agree with this statement: “I am not as careful about HIV and sex now because there is better treatment for HIV”
AGREE DON’T AGREE DON’T KNOW REFUSE TO ANSWER
KA
S
S
S
S
S
*S =Sex Worker, M=MSM, T=Transgender, P=PWID, O=Other Vulnerable Population, C=Core, S=Secondary
References: UNAIDS. Global AIDS Response progress reporting 2014: Construction of Core Indicators for monitoring the 2011 United Nations Political Declaration on HIV and AIDS. http://www.unaids. org/en/media/unaids/contentassets/documents/document/2014/GARPR_2014_guidelines_en.pdf
SECTION 2: QUESTIONNAIRE MODULES
4 4 4 141
Section
II
39. QUESTIONNAIRE FEEDBACK/INTERVIEW STATUS
Unlike the other modules, most of these questions in this module are not asked of participants: they are administrative variables that are required to understand the context for the participant’s information. It can be useful to have this information on every data collection instrument, such as an eligibility screener or follow-up interviews.
The amount of time a respondent takes to finish the questionnaire can be important. For example, if it takes a very small amount of time to complete the survey, it can indicate that they are ‘clicking through’ all the questions in A-CASI. By the end of a face-to-face interview, it can be clear to an interviewer that the respondent is not part of the target population or that there is some other problem encountered. If a respondent fakes his way into the study in the eligibility screen, but does not admit to the behavior in the survey, an analyst would consider excluding this record from analysis. N/A Variable ID: IN IN1MSG
Questionnaire Feedback, Interview Status
Response Options
CORE (C) or SECONDARY (S)
DATA GENERATOR
N/A
1.
INFEED
Respondent feedback: Do you have any feedback or comments about the interview?)
<TEXT 256 CHAR>
S
Interviewer: paraphrase feedback A/CASI
2.
INEDATE
End Date
<DATE>
C
3.
INETIME
End Time
<TIME>
C
A/CASI
4.
INTIME
Number of minutes in survey (QPEtime-QPStime
C
A/CASI
442
39. QUESTIONNAIRE FEEDBACK/INTERVIEW STATUS
N/A
Variable ID: IN INVALID
Questionnaire Feedback, Interview Status
Response Options
CORE (C) or SECONDARY (S) S
DATA GENERATOR
5.
Interviewer confidence in validity of answers
CONFIDENT SOME DOUBTS NOT CONFIDENT AT ALL
1 2 3
Interviewer
6.
INICOMM
Interviewer comments
<TEXT>
S
Interviewer
7.
INCOMP
Did respondent complete survey?
YES NO
1 2 1 2
S
A/CASI or interviewer
8.
INELV
Eligibility verification (still admits to behavior in eligibility criteria during survey)
YES NO
S
A/CASI
SECTION 2: QUESTIONNAIRE MODULES
4 4 4 343
Section
II
40. RDS, PEER RECRUITMENT
In most surveys the non-response rate can easily be determined because the investigator directly recruits the participants. However, in respondent-driven sampling (RDS), measuring non-response is difficult because investigators depend on participants to refer prospective survey respondents. The investigators will have little knowledge of those who had been offered a coupon but rejected participation and whether these people differ from those who accepted a coupon but did not present at the survey office and those who accepted a coupon and did present at the survey office. In most surveys using RDS, participants are asked to return for a second visit to collect secondary incentives for successful peer recruitment and/or to receive results from biological testing. Investigators can take advantage of this second visit to administer a brief questionnaire to determine patterns in coupon distribution, acceptance, and refusal. Although this module is not essential, it is highly recommended for evaluating recruitment and to estimate non-response for an RDS survey. The questions from this module should be selected according to the needs of the survey. This module should be administered face-to-face. N/A Variable ID: PE PE1MSG
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary*
N/A
We will ask you now about your efforts to recruit people you know for this survey. We will start with how many people you thought about giving a coupon to, how many you actually offered one to, and how many accepted a coupon. To how many people did you consider giving a coupon? # PEOPLE [ | | ] DON’T KNOW 997 REFUSE TO ANSWER 998 MIN: 1 MAX:998 ‘0’ →SKIP TO PEREAS; ‘997’, ‘998’ → SKIP TO END OF CURRENT SECTION S
1.
PECONS
444
40. RDS, PEER RECRUITMENT
N/A
Variable ID: PE PEOFFER
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
2.
Of these [insert number from PECONS] people, how many did you offer a coupon?
# PEOPLE [ | | ] DON’T KNOW 997 REFUSE TO ANSWER 998 MIN: 1 MAX:998
IF > 0 →SKIP TO PEREFUSE; ‘97’, ‘98’ SKIP TO END OF CURRENT SECTION
3.
PEREAS
If you did not give or offer any coupons to people you know, what was the primary reason?
HAD NO TIME LOST INTEREST LOST COUPONS FELT EMBARASSED NOBODY WANTED THE COUPON OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
SKIP TO END OF CURRENT SECTION
S
4.
PEREFUS
Of these [insert number from PEOFFER] people, how many Refused the coupons you offered?
NUMBER [ || ] MAX: NUMBER FROM[PEOFFER] DON’T KNOW 997 REFUSE TO ANSWER 998
S
SECTION 2: QUESTIONNAIRE MODULES
4 4 4 545
N/A
Variable ID: PE PEACCEP
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
5.
Of these [insert number from PEOFFER] people, how many accepted the coupons you offered?
NUMBER [ | | ] MIN: 1 MAX: NUMBER FROM PEOFFER DON’T KNOW 997 REFUSE TO ANSWER 998
At N/A
PE2MSG
Interviewer add the responses from PEREFUS and PEACCEP.
IF PEREFUS + PEACCEP = PEOFFER→SKIP TO PED1 IF PEREFUS OR PEACCEP = ‘97’ OR ‘98’→SKIP TO END OF CURRENT SECTION
IF PEREFUS AND PEACCEP DO NOT ADD UP TO PEOFFER, ASK PEREFUSN AND PEACCEPN UNTIL THESE VALUES ADD UP TO PEOFFER.
S
6.
PEREFUSN
The numbers you entered do not add up to the total refused. Among those you offered coupons, how many refused the coupons you offered?
NUMBER [ | ] MAX: NUMBER FROM PEOFFER
ONLY ASK IF PEREFUS AND PEACCEP DO NOT ADD UP TO PEOFFER. ONLY ASK IF PEREFUS AND PEACCEP DO NOT ADD UP TO PEOFFER.
S
7.
PEACCEPN
The numbers you entered do not add up to the total accepted. Among those you offered coupons, how many people accepted?
NUMBER [ |
]
S
MAX: NUMBER FROM PEOFFER
446
40. RDS, PEER RECRUITMENT
N/A
Variable ID: PE PED1
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
8.
How many days did it take to give out the first coupon? If you gave it out on the same day you received it, put ‘0’ (zero).
NUMBER [ | ] MIN: 0 MAX: 30 DON’T KNOW 97 REFUSE TO ANSWER 98
ONLY ASK IF PEREFUS + PEACCEP = PEOFFER.
N/A
PE3MSG
We will now ask you questions about people you offered a coupon but who refused to accept one. Please think of the first person who refused a coupon.
IF PEREFUS = ‘0’,’9997’ OR ‘9998’ → SKIP TO END OF CURRENT SECTION
Ask the set of questions, PERELA1 until PEREAS1, for each person who has refused to accept a coupon from PEREFUS. Change the variable names to identify each person. For example, the variables for the second person who refused would be PERELA2, PEKNOW2, etc.
S
SECTION 2: QUESTIONNAIRE MODULES
4 4 4 747
N/A
Variable ID: PE PERELA1
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
9.
What is your relationship to this person?
STRANGER, SOMEONE YOU DON’T KNOW SOMEONE YOU KNOW, BUT NOT CLOSELY CLOSE FRIEND, SOMEONE YOU KNOW WELL A SEXUAL PARTNER (CURRENT OR FORMER) FAMILY MEMBER OF OTHER RELATIVE OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8 1 2 3 7 8
’1’→ SKIP TO PEAGE1
10.
PEKNOW1
Have you known this person for some weeks, months, or for years?
WEEKS MONTHS YEARS DON’T KNOW REFUSE TO ANSWER
S
11.
PESEEN1
In the last 30 days, how many times have you seen this person?
NUMBER [ | ] MIN: 0 MAX: 30 DON’T KNOW 97 REFUSE TO ANSWER 98
S
448
40. RDS, PEER RECRUITMENT
N/A
Variable ID: PE PESEX1
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
12.
What is the sex of this person?
FEMALE MALE DON’T KNOW REFUSE TO ANSWER
1 2 7 8
Only for populations that have more than one sex.
13.
PEAGE1
Approximately what age is this person?
18-24 25-34 35-44 45 OR OLDER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 7 8
S
14.
PEEDUC1
Can you estimate this person’s highest level of education?
NO EDUCATION PRIMARY SECONDARY COLLEGE/TECHNICAL UNIVERSITY/ POSTGRADUATE DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 7 8
S
SECTION 2: QUESTIONNAIRE MODULES
4 4 4 949
N/A
Variable ID: PE PEREAS1
RDS, Peer Recruitment
Response Options
Skips
Comments
Core or Secondary* S
15.
What was the main reason given for refusing to accept a coupon?
HAD NO TIME/TOO BUSY INCENTIVE TOO SMALL/ NOT INTERESTED ALREADY HAS A COUPON/ PARTICIPATED FEAR OF BEING IDENTIFIED FEAR HIV RESULTS OTHER DON’T KNOW REFUSE TO ANSWER
1 2 3 4 5 6 7 8
450
40. RDS, PEER RECRUITMENT
Supplemental Materials
Section
I
APPENDICES
Section
II
QUESTIONNAIRE MODULES
Section
III Section
INDICATORS
IV
GLOSSARY
Section III - Indicators Introduction What are indicators? How are data collected for indicators? How are indicators used? Indicator selection for these guidelines Indicators Sexual debut Condom use at last sex Consistent condom use Buying sex Selling sex Unprotected receptive anal intercourse Alcohol use Injecting drug use Received clean needles or syringes Sharing injecting equipment Sharing drug-preparation equipment Use of sterile injecting equipment Received medication for drug dependency Drug overdose Drug overdose prevention training Targeted information, education and communications Received condoms Never tested for HIV HIV test in last 12 months Received tuberculosis treatment Currently in care CD4 count Currently on antiretroviral therapy Prevention of mother-to-child transmission Hepatitis B virus testing Hepatitis C virus testing Discussed pre-exposure prophylaxis Taken pre-exposure prophylaxis Current use of modern family planning methods Antenatal clinic access and HIV testing Seeking care for sexually transmitted infection symptoms Sexually transmitted infection screening Received sexually transmitted infection treatment Received lubricants HIV knowledge and perceptions Incarceration Health-care stigma Discrimination and social exclusion Violence Condom use at last sex (sex workers only) 454 454 455 456 457 460 460 461 463 464 465 466 467 470 471 472 473 474 475 476 477 478 479 480 481 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 504
452
SECTION 3: INDICATORS
INTRODUCTION OF BIOMARKER INDICATORS Awareness of HIV-positive status Eligible but not on HIV treatment HIV treatment coverage Viral load suppression Unsuppressed viral load Population viral load HIV/sexually transmitted infection coinfection HIV/hepatitis coinfection HIV prevalence Sexually transmitted infection prevalence
505 506 507 508 509 510 511 512 513 514
SECTION 3: INDICATORS
453
Section
III
INDICATORS
What are indicators? Indicators are standardized measures that can be used to report on the burden of disease, health status or behavioural traits in a given population and describe or measure programme activities. Indicators can be used by ministries of health, donors, policy-makers, programme managers and service delivery organizations to quantify progress made towards a specific end-point. Many indicators are defined by normative bodies, such as governments, or international organizations, such as WHO, the Joint United Nations Programme on HIV/AIDS (UNAIDS), the United States President’s Emergency Plan for AIDS Relief (PEPFAR) or the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund). The aim is to inform strategic planning efforts and assess whether agreed-upon public health goals are being met. United Nations (UN) organizations (e.g. WHO and UNAIDS) have established an Indicator Working Group jointly with countries and partners to validate indicators in a country’s context. Standardized indicators facilitate comparison across populations and programmes, and can help donors, policy-makers and programme planners to determine whether activities are on course and to evaluate programme effectiveness. Figure. III.1 provides an overview of the different categories of indicators and their role in monitoring and evaluating HIV prevention services for key populations.1
Figure. III.1. Framework for monitoring and evaluating interventions for key populations
MONITORING Programme indicators • Assess national-level HIV/AIDS prevention interventions for key populations Aligned with a logic model’s inputs (resources, contributions, and investments) and outputs (direct products, including activities and services, that reach program participants)
EVALUATION Behavioural outcome indicators • Measure individual actions of key populations that directly affect biological outcomes Aligned with a outcomes (results or changes) in a logic model •
•
Impact indicators Capture the epidemiologic population-level impact measures Aligned with impacts (longterm changes in social, economic, or environmental conditions) in a logic
Determinants indicators (risk and protective factors) • Measure factors not directly to HIV infection but either contributr to risk taking and vulnerability or provide some protection against it Aligned with outcomes (results or changes) in a logic model
•
•
•
1
Adapted from WHO. Guide to monitoring and evaluating national HIV/AIDS prevention programmes for young people: addendum to the UNAIDS “National AIDS programmes: a guide to monitoring and evaluation”. Geneva: World Health Organization (WHO); 2004 (http://www.who.int/hiv/pub/me/en/me_prev_intro.pdf, accessed 8 September 2016).
454
INDICATORS
How are data collected for indicators? To report indicators, it is necessary to collect the required data elements (referred to below as “indicator components”) and compute the measurement according to the indicator’s definition. Multiple methods can be used to collect data to inform indicators, including biobehavioural surveys (BBS); health facility surveys; sentinel surveillance; and extraction of information from programme or facility records, and from logistics management information systems. Survey data are used to calculate the percentage of participants to which a certain characteristic applies. Table III.1 describes how to generate an indicator that assesses the percentage of men who have sex with men (MSM) reporting unprotected receptive anal intercourse (URAI) in the last 6 months. First, we would need to know the number of MSM who had anal sex in the last 6 months (denominator) and, of those, the number who had receptive anal sex and did not use a condom (numerator). The indicator is calculated by dividing the numerator by the denominator, which in this case yields 40% (400/1000 × 100). By default, every person in the numerator is also included in the denominator. Where sampling weights are available, weighted indicators are preferable because they facilitate population estimates, whereas indicators using unweighted data provide sample estimates (see Chapter C-1 for more information on creating sampling weights). Table III.1: Sample indicator calculation
Indicator: Percentage of MSM who engaged in URAI in the last 6 months = 40% Indicator components Related questions from questionnaire module 3
Numerator: Number of MSM who engaged Ask the 1000 who had anal sex with males in the last 6 months: In the last 6 months, did you have receptive or insertive anal sex? in URAI in the last 6 months = 400 INTERVIEW RESULTS: Receptive: 260 Insertive: 225 Both: 390 Don’t know: 100 Refuse to answer: 25 Ask the 650 who had receptive sex in the last 6 months: Did you use a condom at any time you had anal receptive sex in the last 6 months? Interview results: Yes: 200 No: 400 Don’t know: 20 Refuse to answer: 30 Denominator: Number of MSM who had anal sex in the last 6 months = 1000 In the last 6 months, did you have anal sex with any males? Interview results: Yes: 1000 No: 1500 Don’t know: 500 Refuse to answer: 100
SECTION 3: INDICATORS
455
Table III.2 describes how to generate an indicator that assesses the percentage of MSM reporting condom use at last sex within the last 6 months. First, we would need to know the number of MSM who had sex within the last 6 months (denominator) and, of those, the number who used a condom at last sex (numerator). The indicator is calculated by dividing the numerator by the denominator, which in this case yields 60% (600/1000 x 100). By default, every person in the numerator is also included in the denominator. As noted above, where sampling weights are available, weighted indicators are preferable. Table III.2: Sample indicator calculation
Indicator: Percentage of MSM respondents who used a condom at last anal sex = 60% Indicator components Numerator: Number of MSM who used a condom at last sex = 600 Related questions from questionnaire module 3 Ask of the 1000 who had anal sex with males in the last 6 months: Did you use a condom the last time you had sex? Interview results: Yes: 600 No: 300 Don’t know: 55 Refuse to answer: 45 Denominator: Number of MSM who had anal sex in the last 6 months = 1000 In the last 6 months, did you have anal sex with any males? Interview results: Yes: 1000 No: 1500 Don’t know: 500 Refuse to answer: 100
How are indicators used? Indicators serve several purposes, including reporting, monitoring, advocacy and programme evaluation, as discussed below. Reporting International donors and initiatives, such as UNAIDS, the Global Fund and PEPFAR, require reporting on a core set of indicators. The aim is to promote accountability; provide standardized data for comparison across sites, districts, province, states and countries; and allow for aggregation at the global level. Monitoring Stakeholders use indicators to monitor a programme’s or country’s progress towards agreed-upon goals. When indicators suggest that a situation is not improving, is improving but not at the rate expected, or is worsening, this information can influence decision-making to modify programmes or responses, or develop new ones. Data from indicators can also help to guide resource allocation. Once several data points have been collected, this information can determine whether trends are moving in a desirable direction. Collecting and reviewing indicator data can increase a country’s understanding of the behaviours and circumstances associated with HIV transmission among key populations, and thus can improve the design of appropriate prevention responses. Indicators can be used to estimate the prevalence of risk factors in the population or subpopulation of interest, and to determine where to focus efforts to address the population’s needs. For example, a large percentage of people who inject drugs (PWID) who report sharing injecting equipment the last time they injected drugs may signal
456
INDICATORS
a broad lack of knowledge about transmission of bloodborne infections or lack of access to reliable sources of sterile syringes. This information can be used to justify expanding access to community-based outreach programmes and high-coverage needle and syringe programmes to reduce the reuse of contaminated injecting equipment. Advocacy Indicator data can also be used to raise public awareness about an issue. Likewise, changes to the legal or policy environment may be inspired by survey results that indicate a high level of stigma and discrimination experienced by key populations. Such findings can push issues to the forefront of a country’s agenda and inform advocacy efforts, resource allocation and policy development. Programme evaluation Finally, the results of behavioural risk analyses can be used to evaluate the effects of control efforts, identifying successes or areas of weakness. A BBS among street-based sex workers (SW) that reveals an increase in unprotected sexual activity with regular clients after several years of condom promotion, for example, would alert HIV service providers to an increase in risk behaviour and possible weaknesses in the current condom promotion programme that might need to be addressed for the programme to be as effective as it was previously. Similarly, if a BBS indicates that less than half of HIV-infected MSM were aware of their seropositive status, service providers and policy-makers might try to increase access to HIV testing for MSM.
Indicator selection for these guidelines Most of the indicators defined in these guidelines are generated from self-reported data from the questionnaire modules; a subset of the indicators use biomarkers as the unit of measurement. All have the potential to influence service delivery or policy. The process for selecting indicators for these guidelines involved reviewing indicators specific to key populations that are required for reporting by global entities, and ensuring that the modules’ core questions would provide the necessary data. These guidelines propose additional indicators that are useful for monitoring the HIV epidemic and response among key populations. The indicators described here focus on several areas of measurement that can be monitored over time or compared across locations and populations, including access to a comprehensive package of HIV prevention services, access to care and treatment services for HIV and other relevant infections, experiences of stigma and abuse, behavioural risk characteristics and disease prevalence. Figure. III.2 depicts the relationships between these indicators and their utility in assessing progress towards reducing vulnerability, risk for HIV transmission and disease burden among key populations. If countries identify additional indicators they need, they may elect to add to the set of indicators included in these guidelines.
SECTION 3: INDICATORS
457
Figure. III.2. Indicators for monitoring and evaluating interventions for key populations
MONITORING Programme indicators • • • • • • • • • • • ANC access and HIV testing Current use of modern FP methods STI symptoms STI screening Received STI treatment CD4 count Currently in care Currently on ARVs PMTCT Eligible but not on treatment HIV treatment coverage
• • • • • • • • • • • •
Targeted IEC Received condoms Received lubricants HIV test in last 12 months Discussed PrEP Taken PrEP Received clean needles or syringes Received medication for drug dependency Hepatitis B testing Hepatitis C testing TB screening Received TB treatment
EVALUATION Behavioural outcome indicators • • • • • • • • Sexual debut Condom use at last sex Condom use at last sex by client type (SW only) Consistent condom use Buying sex Selling sex Unprotected receptive anal intercourse Injecting drug use in last 6 months Use of sterile injecting equipment Sharing injecting equipment Sharing drug preparation equipment
Impact indicators
• • • • • •
Determinants indicators (risk and protective factors) • • • • • • • • Alcohol use Drug overdose Drug overdose prevention training HIV knowledge and perceptions Discrimination and social exclusion Health care stigma Incarceration Violence
• • •
HIV prevelance STI prevalence HIV/STI coinfection HIV/hepatitis coinfection Viral load suppression Population viral load
ANC, antenatal clinic; ARV, antiretroviral; FP, family planning; IEC, information, education and communication; PMTCT, prevention of mother-to-child transmission; PrEP, pre-exposure prophylaxis; STI, sexually transmitted infection
The reference sheets in this section provide information for each indicator, including the numerator, denominator, rationale, data use and any similar PEPFAR, UNAIDS or Global Fund indicators. Each reference sheet includes the corresponding survey questions from which the indicator originates, to ensure that the information is collected in a standardized manner. Table III.3 provides a summary template of the indicator reference sheets. Note that the reference sheets for biomarker indicators are structured in a slightly different manner, with a briefer description and a focus on inclusion or exclusion criteria for the numerator and denominator.
458
INDICATORS
Table III.3: Indicator reference sheet template
Indicator title: States the percentage of the characteristic (e.g. access, experience, knowledge, behaviour or health status) being measured Numerator Describes the subset of individuals included in the calculation (represents the number of individuals possessing the characteristic of interest). Describes the sample of the target population being measured (represents the number of individuals for whom a characteristic could potentially apply). States the key populations for which the indicator applies (e.g. SW, MSM, TG, PWID, OVP). States the relevant subgroups that collected data can be separated into in order to more precisely understand and analyse the findings. Common subgroups include sex and age. Provides an indicator justification using the following outline: • • • • Corresponding questionnaire item(s) Related existing indicators A statement about what is being measured Contextual statements about the characteristic being measured and its importance Potential reasons for an undesirably high or low percentage or problematic trend Potential programmatic responses to address any “negative” findings.
Denominator
Applicable KP Recommended disaggregation Purpose/ rationale
• States the questionnaire module(s) from which the indicator calculation is derived, and the related question(s) and response categories that feed into the numerator, denominator and disaggregations. Lists any indicators used by UNAIDS, the Global Fund and PEPFAR that measure similar characteristics.
Global Fund, Global Fund to Fight AIDS, Tuberculosis and Malaria; OVP, other vulnerable populations; PEPFAR, United States President’s Emergency Plan for AIDS Relief
Not all indicators will be relevant or useful to every country. Indicator selection will depend on various factors, including characteristics of the local epidemic and the resources available for data collection. Furthermore, appropriate adjustments and modifications may be needed to adapt some indicators to different contexts (e.g. aligning recall periods to fit reporting needs). Additionally, indicators applicable to SW are focused on female SW; there may be additional considerations for countries conducting BBS among sex workers who are male or transgender persons (TG). When collected at regular intervals (WHO recommends conducting BBS with key populations every 3–4 years), these recommended indicators will generate valuable trend information that can be used with other data, such as programme data, for prevention resource allocation, strategic planning, policy development and evaluation (see Chapter C-1 for more information on trend analysis).
SECTION 3: INDICATORS
459
Sexual debut Percentage aged 15–24 years who had sexual intercourse before the age of 15 years Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who had anal or vaginal intercourse before the age of 15 years Number of participants aged 15–24 years who have had anal or vaginal intercourse SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (15–19 years, 20–24 years) This indicator measures the percentage of participants aged 15–24 years who initiated sexual activity before the age of 15 years. The longer someone waits to have sex, the longer that person avoids the risk of sexual acquisition of HIV. According to the GARPR, “there is also evidence to suggest that first having sex at a later age reduces susceptibility to infection per act of sex, at least for women”.a An increasing trend of early age at sexual debut may point to an increase in risky behaviour among young people. In such cases, countries may consider HIV prevention services that target younger audiences. This indicator is constructed from responses to the following questions from Module 14: Sexual history. For males aged 15–24 years, including MSM Numerator: • LIMF1AGE ≤15 [How old were you when you first had vaginal or anal sex with a female partner?] OR • LIMM1AGE ≤15 [How old were you when you first had anal sex with a male partner?] Denominator: • LIMFVAG = 1 (yes) [Have you ever had vaginal sex?] OR • LIMFANAL = 1 (yes) [Have you ever had anal sex with a woman?] OR • LIMMANAL = 1 (yes) [Have you ever had anal sex with a man?] For females aged 15–24 years Numerator: • LIFM1AGE ≤15 [How old were you when you first had vaginal or anal sex with a male partner?] Denominator: LIFMVAG = 1 (yes) [Have you ever had vaginal sex?] OR LIFMANAL = 1 (yes) [Have you ever had anal sex?] Related existing indicators • GARPR 1.2: Percentage of young women and men aged 15–24 who have had sexual intercourse before the age of 15
Corresponding questionnaire item(s)
GARPR, Global AIDS Response Progress Reporting
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016.
460
INDICATORS
Condom use at last sex Percentage reporting the use of a condom at last sex Numerator Denominator Applicable KP Recommended disaggregation Number of participants who used a condom at last sex Number of participants who had anal or vaginal sex in the last 6 months SW, MSM, TG, PWID, OVP • • • • Sex for SW, PWID, OVP (male, female) Sex of partner (male, female) Partner type (main, exchange, casual) Age (<25 years, 25+ years)
Purpose/ rationale
This indicator measures the extent to which condoms are being used by the target population. Condoms can substantially reduce the risk of sexually transmitting HIV. They are most effective when their use is correct and consistent. The prevalence of condom use during the most recent sexual act will generally reflect the prevalence of consistent condom use.a A low percentage or decreasing trend in condom use at last sex may indicate a lack of knowledge about the modes of HIV transmission, low perceptions of personal risk, lack of access to condoms, low self-efficacy for condom use and negotiation, low perceptions of condom efficacy, or perceived diminished sexual sensation.b Peerled and community-led outreach approaches may help to increase knowledge, develop skills, and empower KP to use condoms and lubricants consistently.c
This indicator is constructed from responses to the following questions from Module 15: Recall Corresponding sexual behaviour. questionnaire item(s) (if applicable) Numerator (with disaggregation): Sex of partner = Male Partner type = Main • RCMAMNRC = 1 (yes) [The last time you had sex with your main male partner, did you use a condom?] Partner type = Exchange (Sex work: gave money for sex) • RCMASWAC = 1 (yes) [The last time you had sex with a man you gave money for sex, did you use a condom?] Partner type = Exchange (Sex work: received money for sex) • RCMACLAC = 1 (yes) [The last time you had sex with a man who gave you money, did you use a condom?] Partner type = Exchange (Transactional sex: gave nonmonetary items for sex) • RCMATSAC = 1 (yes) [The last time you had sex with a man you gave gifts, goods or services for sex, did you use a condom?] Partner type = Exchange (Transactional sex: received nonmonetary items for sex) • RCMACTAC = 1 (yes) [The last time you had sex with a man who gave you gifts, goods or services, did you use a condom?] Partner type = Casual • RCMACSAC = 1 (yes) [The last time you had sex with a casual male partner, did you use a condom?]
SECTION 3: INDICATORS
461
Sex of partner = Female Partner type = Main • RCFEMNAC = 1 (yes) [The last time you had sex with a female main partner, did you use a condom?] Partner type = Exchange (Sex work: gave money for sex) • RCFESWAC = 1 (yes) [The last time you had sex with a female you gave money for sex, did you use a condom?] Partner type = Exchange (Sex work: received money for sex) • RCFECLAC = 1 (yes) [The last time you had sex with a female who gave you money for sex, did you use a condom?] Partner type = Exchange (Transactional sex: gave nonmonetary items for sex) • RCFETSAC = 1 (yes) [The last time you had sex with a female you gave gifts, goods and services for sex, did you use a condom?] Partner type = Exchange (Transactional sex: received nonmonetary items for sex) • RCFETSAC = 1 (yes) [The last time you had sex with a female who gave you gifts, goods or services for sex, did you use a condom?] Partner type = Casual • RCFECSAC = 1 (yes) [The last time you had sex with a casual female partner, did you use a condom?] Denominator: • RCMAPA = 1 (yes) [In the last 6 months, did you have sex with any males?] OR • RCFEPA = 1 (yes) [In the last 6 months, did you have sex with any females?] Related existing indicators • GARPR 1.8: Percentage of sex workers reporting the use of a condom with their most recent client • GARPR 1.12: Percentage of men reporting the use of a condom the last time they had anal sex with a male partner • GARPR 2.2: Percentage of people who inject drugs reporting the use of a condom the last time they had sexual intercourse • GF HIV O-5: Percentage of sex workers reporting the use of a condom with their most recent client • GF HIV O-4a: Percentage of men reporting the use of a condom the last time they had anal sex with a male partner • GF HIV O-4b: Percentage of transgender people who sell sex reporting the use of a condom with their most recent client • GF HIV O-7: Percentage of other vulnerable populations who report the use of a condom at last sexual intercourse
OVP, other vulnerable populations
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016. Sarkar NN. Barriers to condom use. Eur J Contracep Repr. 2008;13(2):114–122.
b c
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016).
462
INDICATORS
Consistent condom use Percentage reporting consistent condom use in the last 6 months Numerator Number of participants who always used condoms with any partner or a particular partner type in the last 6 months Total number of participants who had sex in the last 6 months SW, MSM, TG, PWID, OVP • Sex for PWID (male, female) • Sex of partner for PWID and OVP (male, female) • Partner type (main, casual, client, SW) This indicator measures the percentage of participants who always used condoms with any partner over the last 6 months. Condom use is an important measure of protection against HIV and other STIs, especially among individuals with multiple sexual partners. The maximum protective effect of condoms is attained when they are used consistently instead of intermittently. Consistent and correct condom use reduces sexual transmission of HIV and other STIs in both vaginal and anal sex by up to 94%1. A low percentage or decreasing trend in consistent condom use may indicate lack of resources to purchase condoms, lack of knowledge of the protective benefits of condom use, fear of stigma or discrimination for carrying or purchasing condoms, or clients’ unwillingness or coercion. WHO guidelines recommend that condoms should be available to KP through multiple outlets free of charge.a Programmes should offer information and skills-building in negotiating condom use and behavioural interventions that encourage consistent condom use.a This indicator is constructed from responses to the following questions from Module 15: Recall sexual behaviour. Note: By including additional secondary questions, it may be possible to disaggregate further for some target populations (e.g. female partners for MSM). Sample numerator (and disaggregation): • RCMAMNFQ = 1 (In the last 6 months, how often did you use condoms with your main male partners?) (Partner type = main male sex partners) • RCMACLFQ = 1 (In the last 6 months, how often did you use condoms with men who gave you money for sex?) (Partner type = male clients) • RCMACSFQ = 1 (In the last 6 months, how often did you use condoms with casual male partners?) (Partner type = casual male sex partners) Note: To construct this indicator for any partner (i.e. across all partner types), the numerator consists of the number of respondents who report always using condoms for all partner types they have sex with. Sample denominator: • RCMAMNFQ = 1, 2, 3, 4 or 5 (always, most, sometimes, rarely, never) • RCMACLFQ = 1, 2, 3, 4 or 5 (always, most, sometimes, rarely, never) • RCMACSFQ = 1, 2, 3, 4 or 5 (always, most, sometimes, rarely, never) Related existing indicators Not applicable
Denominator Applicable KP Recommended disaggregation
Purpose/ rationale
Corresponding questionnaire item(s)
OVP, other vulnerable populations
a
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016)
SECTION 3: INDICATORS
463
Buying sex Percentage who paid money in exchange for sex in the last 6 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who paid money in exchange for sex in the last 6 months Total number of participants MSM, TG, PWID, OVP • Sex for PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who paid money for sex in the last 6 months. High frequency of client turnover and unprotected sexual practices put SW and their clients at higher risk for getting and passing on HIV and STI.a A high percentage or increasing trend of buying sex may indicate greater demand for sex work and the potential for passing on HIV if sexual practices do not include condom use. In such cases, countries may target SW clients for health care and prevention interventions that promote safer sex practices. This indicator is constructed from responses to the following questions from Module 15: Recall sexual behaviour. Numerator: • RCMASWPA >0 [In the last 6 months, how many different men did you pay money in exchange for sex?] OR • RCFESWPA >0 [In the last 6 months, how many different women did you pay money for sex?] Denominator: • Total number of participants Not applicable
Corresponding questionnaire item(s)
Related existing indicators
OVP, other vulnerable populations
a WHO. Preventing HIV among sex workers in sub-Saharan Africa: a literature review. Geneva: World Health Organization (WHO); 2014 (http://apps.who.int/iris/bitstream/10665/44549/1/9789241501279_eng.pdf?ua=1, accessed 8 September 2016).
464
INDICATORS
Selling sex Percentage who received money in exchange for sex in the last 6 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who received money in exchange for sex in the last 6 months Total number of participants MSM, TG, PWID • Sex for PWID (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of MSM, TG and PWID who provided sex in exchange for money in the last 6 months. Because of the stigmatized and often criminalized setting in which sex work is practiced, MSM, TG and PWID who engage in sex work are at increased risk of HIV and STI transmission, violence and imprisonment, and are less likely to access needed services.a The potential for getting and passing on HIV and STI is higher when drug-using and sex work networks overlap.b,c,d,e Assessing the prevalence of sex work among these KP can inform specialized prevention, care and treatment services that address their individual risks. If there is a high percentage or increasing trend of sex work, countries may consider combination prevention interventions (evidence-based, mutually reinforcing biomedical, behavioural and structural interventions) that are tailored to local contexts and designed and implemented by SW communities.f This indicator is constructed from responses to the following questions from Module 15: Recall sexual behaviour. Numerator: • RCMACLPA >0 [In the last 6 months, how many different men paid you money in exchange for sex?)] OR • RCFECLPA >0 [In the last 6 months, how many different women gave you money for sex?] Denominator: • Total number of participants Related existing indicators Not applicable
Corresponding questionnaire item(s)
a WHO. Prevention and treatment of HIV and other sexually transmitted infections for sex workers in low- and middle-income countries: recommendations for a public health approach. 2012 (http:// apps.who.int/iris/bitstream/10665/77745/1/9789241504744_eng.pdf?ua=1, accessed 8 September 2016. b
Needle R, Kroeger K, Belani H, Achrekar A, Parry CD, Dewing S. Sex, drugs, and HIV: rapid assessment of HIV risk behaviors among street-based drug using sex workers in Durban, South Africa. Soc Sci Med. 2008;67(9):1447–1455. Medhi GK, Mahanta J, Paranjape RS, Adhikary R, Laskar N, Ngully P. Factors associated with HIV among female sex workers in a high HIV prevalent state of India. AIDS Care. 2012;24(3):369–376. Strathdee SA, Philbin MM, Semple SJ, Pu M, Orozovich P, Martinez G et al. Correlates of injection drug use among female sex workers in two Mexico-U.S. border cities. Drug Alcohol Depen. 2008;92(1-3):132–140.
c d
e
Tuan NA, Fylkesnes K, Thang BD, Hien NT, Long NT, Kinh NV et al. Human immunodeficiency virus (HIV) infection patterns and risk behaviours in different population groups and provinces in Viet Nam. B World Health Organ. 2007;85(1):35–41.
f Bekker LG, Johnson L, Cowan F, Overs C, Besada D, Hillier S et al. Combination HIV prevention for female sex workers: what is the evidence? Lancet. 2015;385(9962):72-87 (https://www.ncbi.nlm. nih.gov/pubmed/25059942, accessed 9 September 2016).
SECTION 3: INDICATORS
465
Unprotected receptive anal intercourse Percentage of MSM and TG who engaged in URAI in the last 6 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ Rationale Number of MSM and TG who engaged in URAI in the last 6 months Total number of MSM and TG who had anal sex in the last 6 months MSM, TG • Partner type (main, casual) • Age (<25 years, 25+ years) This indicator measures the percentage of MSM and TG who engaged in URAI in the last 6 months. URAI with ejaculation inside the anus carries more than 10 times the risk of HIV transmission compared to unprotected insertive anal intercourse.a A high percentage or increasing trend of URAI may indicate lack of knowledge around HIV risk reduction. In such cases, countries may consider support for consistent condom use. This indicator is constructed from responses to the following questions from the Module 15: Recall sexual behaviour. Numerator and disaggregation: • VARIABLE = 1 (receptive) OR 3 (both) [In the last 6 months, did you have receptive or insertive anal sex?] AND • VARIABLE = 2 (no) [Did you use a condom at any time you had anal receptive sex in the last 6 months?] Denominator: • Total number of MSM who had anal sex in the last 6 months AND • Total number of TG who had anal sex in the last 6 months Related existing indicators Not applicable
Corresponding questionnaire item(s)
URAI, unprotected receptive anal intercourse
1
Patel P, Borkowf CB, Brooks JT, Lasry A, Lansky A, Mermin J. Estimating per-act HIV transmission risk: a systematic review. AIDS. 2014;28(10):1509–1519.
466
INDICATORS
Alcohol use The following reference sheets describe three options for measuring alcohol use. They vary based on the instrument used: AUDIT, AUDIT-C or CAGE. Countries should select the indicator that corresponds with the validated tool that was used to assess alcohol use.
Percentage who are hazardous drinkers (AUDIT) Numerator Number of participants with a score of ≥8 on AUDIT scale based on alcoholic beverage intake in the past year Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • Degree of alcohol abuse: Harmful or hazardous drinking (score of ≥8); Alcohol dependence (score of ≥13 for women or ≥15 for men) This indicator measures the percentage of participants with hazardous drinking behaviours. Alcohol use can contribute to risky sexual behaviour. It can reduce sexual inhibitions, which may result in more sexual partners, forgetting to use a condom, or incorrect use of a condom. Alcohol use is associated with a greater risk of incident HIV infection.a A high percentage or increasing trend of hazardous drinking among target populations may indicate a social environment in which sex under the influence of alcohol occurs more frequently or is acceptable. In such cases, countries may consider HIV interventions that screen KP for alcohol use and refer those with alcohol dependence to treatment and counselling services. This indicator is constructed from responses to questions from Module 21: Alcohol use. See AUDIT section in Module 6 for questions and scoring information.
Denominator Applicable KP Recommended disaggregation
Purpose/ rationale
Corresponding questionnaire item(s) (if applicable) Related existing indicators
Not applicable
AUDIT, alcohol use disorders identification test; OVP, other vulnerable populations
a
Baliunas D, Rehm J, Irving H, Shuper P. Alcohol consumption and risk of incident human immunodeficiency virus infection: a meta-analysis. Int J Public Health. 2010;55(3):159–166.
SECTION 3: INDICATORS
467
Percentage who are hazardous drinkers or have active alcohol use disorders (AUDIT-C) Numerator Number of participants with a positive score (≥4 for men, ≥3 for women) on AUDIT-C based on alcoholic beverage intake in the past year Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants with hazardous drinking behaviours. Alcohol use can contribute to risky sexual behaviour. It can reduce sexual inhibitions, which may result in more sexual partners, forgetting to use a condom, or incorrect use of a condom. Alcohol use is associated with a greater risk of incident HIV infection.a A high percentage or increasing trend of hazardous drinking among target populations may indicate a social environment in which sex under the influence of alcohol occurs more frequently or is acceptable. In such cases, countries may consider HIV interventions that screen KP for alcohol use and refer those with alcohol dependence to treatment and counselling services. This indicator is constructed from responses to questions from Module 21: Alcohol use. See AUDIT-C section in Module 6 for questions and scoring information.
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s) (if applicable) Related existing indicators
Not applicable
AUDIT-C, alcohol use disorders identification test-consumption; OVP, other vulnerable populations
a
Baliunas D, Rehm J, Irving H, Shuper P. Alcohol consumption and risk of incident human immunodeficiency virus infection: a meta-analysis. Int J Public Health. 2010;55(3):159–166.
468
INDICATORS
Percentage with potential alcohol problems (CAGE) Numerator Number of participants with a total score of 2 or more on CAGE scale based on (the effect of) alcoholic beverage intake Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants with hazardous drinking behaviours. Alcohol use can contribute to risky sexual behaviour. It can reduce sexual inhibitions, which may result in more sexual partners, forgetting to use a condom, or incorrect use of a condom. Alcohol use is associated with a greater risk of incident HIV infection.a A high percentage or increasing trend of hazardous drinking among target populations may indicate a social environment in which sex under the influence of alcohol occurs more frequently or is acceptable. In such cases, countries may consider HIV interventions that screen KP for alcohol use and refer those with alcohol dependence to treatment and counselling services.
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding This indicator is constructed from responses to questions from Module 21: Alcohol use. See CAGE questionnaire section in Module 21 for questions and scoring information. item(s) (if applicable) Related existing indicators Not applicable
CAGE, an acronym of four questions in a questionnaire to screen for problem drinking and potential alcohol problem; OVP, other vulnerable populations
a
Baliunas D, Rehm J, Irving H, Shuper P. Alcohol consumption and risk of incident human immunodeficiency virus infection: a meta-analysis. Int J Public Health. 2010;55(3):159–166.
SECTION 3: INDICATORS
469
Injecting drug use in last 6 months Percentage who injected illicit or illegal drugs in the last 6 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who injected illicit or illegal drugs in the last 6 months Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who injected illicit or illegal drugs in the last 6 months. KP, who are already at a higher risk for getting HIV, are at greater risk of getting HIV, HBV and HCV when they also inject drugs. Being high on drugs affects one’s ability to make safe choices and lowers inhibitions, which may result in people engaging in risky sexual and drug behaviour. Injecting drug use is the main route of transmission for approximately 10% of HIV infections globally and 30% of infections outside of sub-Saharan Africa.a In cases where the percentage who inject drugs is high or increasing, countries may consider interventions tailored to KP that screen for injecting drug use and provide referrals for treatment, high-coverage needle and syringe services, and information, education and counselling.b
Corresponding This indicator is constructed from responses to the following question from Module 23: Injecting questionnaire drug use. item(s) (if applicable) Numerator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Denominator: • Total number of participants Related existing indicators Not applicable
HBV, hepatitis B virus; HCV, hepatitis C virus; OVP, other vulnerable populations
a
UNAIDS, UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). MacArthur GJ, van Velzen E, Palmateer N, Kimber J, Pharris A, Hope V et al. Interventions to prevent HIV and Hepatitis C in people who inject drugs: a review of reviews to assess evidence of effectiveness. Int J Drug Policy. 2014;25(1):34–52.
b
470
INDICATORS
Received clean needles or syringes Percentage of PWID who received clean (sterile) needles or syringes from an outreach worker or a needle and syringe programme in the last 6 months Numerator Number of PWID who received clean (sterile) needles or syringes from an outreach worker or a needle and syringe programme in the last 6 months Total number of PWID PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID who received any clean (sterile) needles or syringes from an outreach worker or needle and syringe programme in the last 6 months. Injecting drug use is the main route of transmission for approximately 10% of HIV infections globally and 30% of infections outside of sub-Saharan Africa. Needle and syringe programmes are one of nine interventions in the WHO, UNODC and UNAIDS comprehensive package for the prevention, treatment and care of HIV among people who inject drugs.a There is a wealth of scientific evidence supporting the efficacy of providing clean (sterile) needles and syringes in preventing the transmission of HIV and other bloodborne infections among PWID.b A low percentage or decreasing trend in the percentage who received clean needles or syringes may indicate lack of knowledge of HIV transmission or lack of access to needle and syringe programmes. In such cases, countries may consider interventions to improve access to safe injection equipment.c
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 23: Injecting questionnaire items) drug use. (if applicable) Numerator: • IDNLOC = 1 (yes for at least one of the following response options: doctor’s office, clinic or hospital, other health agency or HIV prevention programme; drug worker or agency/outreach worker or street unit; or needle/syringe exchange programme) [In the last 6 months when you injected, where did you get your needles/syringes from?] Denominator: • ID6MOS = 1 (yes) [(In the last 6 months, have you injected any illicit or illegal drugs?)] Related existing indicators • GF KP-4: Percentage of all PWID who were reached by an NSP over the specified reporting period
NSP, needle and syringe programme; UNODC, United Nations Office on Drugs and Crime
a
UNAIDS, UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/ contentassets/documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). Des Jarlais DC, Feelemyer JP, Modi SN, Abdul-Quader A, Hagan H. High coverage needle/syringe programs for people who inject drugs in low and middle income countries: a systematic review. BMC Public Health. 2013;13:53 (https://www.ncbi.nlm.nih.gov/pubmed/23332005, accessed 8 September 2016).
b
c MacArthur GJ, van Velzen E, Palmateer N, Kimber J, Pharris A, Hope V et al. Interventions to prevent HIV and Hepatitis C in people who inject drugs: a review of reviews to assess evidence of effectiveness. Int J Drug Policy. 2014;25(1):34–52.
SECTION 3: INDICATORS
471
Sharing injecting equipment Percentage of PWID active in the last 6 months who shared injecting equipment the last time they injected drugs Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of PWID who shared injecting equipment the last time they injected drugs Number of PWID injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID active in the last 6 months who shared injecting equipment the last time they injected drugs. Sharing contaminated injecting equipment increases the risk of HIV, HBV and HCV transmission. Safer injecting and sexual practices among people who inject drugs are essential because risk of HIV transmission from contaminated injecting equipment is extremely high.a A high percentage or increasing trend of sharing injecting equipment may indicate lack of knowledge about transmission of blood-borne infections or lack of access to reliable sources of new sterile syringes. In such cases, countries may consider interventions to improve injection safety through education and access to safe injection equipment.b
Corresponding This indicator is constructed from responses to the following questions from Module 23: Injecting questionnaire item(s) drug use. (if applicable) Numerator: • IDLSTND = 1 (yes) [The last time you injected with another person, did you share a needle with that person (i.e. use it after they had used it)?] Denominator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Related existing indicators • GARPR 2.3: Percentage of people who inject drugs reporting the use of sterile injecting equipment the last time they injected
HBV, hepatitis B virus; HCV, hepatitis C virus
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). MacArthur GJ, van Velzen E, Palmateer N, Kimber J, Pharris A, Hope V et al. Interventions to prevent HIV and Hepatitis C in people who inject drugs: a review of reviews to assess evidence of effectiveness. Int J Drug Policy. 2014;25(1):34–52.
b
472
INDICATORS
Sharing drug-preparation equipment Percentage of PWID active in the last 6 months who shared drug-preparation equipment the last time they injected drugs Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of PWID who shared drug-preparation equipment the last time they injected drugs Number of PWID who injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID active in the last six month who shared drugpreparation equipment the last time they injected drugs. HIV, HBV and HCV transmission can occur among PWID who share preparation equipment, including drug-preparation containers, filters, rinse water, cookers or cotton.a,b A high percentage or increasing trend of sharing drug-preparation equipment may indicate lack of knowledge around the risk of sharing drug-preparation equipment. It may also indicate lack of access to new preparation equipment. In such cases, countries may consider ways to improve injection safety through education and access to preparation equipment. This indicator is constructed from responses to the following questions from Module 23: Injecting drug use. Numerator: • IDWATR = 1 (yes) [The last time you injected with someone, did you use a cooker, cotton or water that that person had already used?] OR • IDDRSR = 1 (yes) [The last time you injected with someone, did you use drugs that had been divided with a syringe that a person had already injected with?] Denominator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Related existing indicators Not applicable
Corresponding questionnaire item(s) (if applicable)
HBV, hepatitis B virus; HCV, hepatitis C virus
a
Doerrbecker J, Behrendt P, Mateu-Gelabert P, Ciesek S, Riebesehl N, Wilhelm C et al. Transmission of hepatitis C virus among people who inject drugs: viral stability and association with drug preparation equipment. J Infect Dis. 2013;207(2):281–28. Pouget ER, Hagan H, Des Jarlais DC. Meta-analysis of hepatitis C seroconversion in relation to shared syringes and drug preparation equipment. Addiction. 2012;107(6):1057–1065.
b
SECTION 3: INDICATORS
473
Use of sterile injecting equipment Percentage of PWID active in the last 6 months who used sterile injecting equipment the last time they injected drugs Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of PWID who used sterile injecting equipment the last time they injected drugs Number of PWID who injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID active in the last 6 months who used sterile injecting equipment the last time they injected drugs. Safer injecting practices among people who inject drugs are essential because risk of HIV transmission from contaminated injecting equipment is extremely high.a A low percentage or decreasing trend of use of sterile injecting equipment may point to lack of knowledge of safer injecting practices or lack of access to sterile injecting equipment. Interventions such as high-coverage needle and syringe programmes may encourage the use of sterile injecting equipment and reduce HIV and HCV transmission.b This indicator is constructed from responses to the following questions from Module 23: Injecting drug use. Numerator: • IDNDSTR = 1 (yes) [The last time you injected with another person did you use a new sterile needle to inject?] OR • IDLSTND = 2 (no) [The last time you injected with another person, did you share a needle with that person (i.e. use it after they had used it)?] Denominator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Related existing indicators • GARPR 2.3: Percentage of people who inject drugs reporting the use of sterile injecting equipment the last time they injected • GF HIV O-6: Percentage of PWID reporting the use of sterile injecting equipment the last time they injected
Corresponding questionnaire item(s) (if applicable)
HCV, hepatitis C virus
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). MacArthur GJ, van Velzen E, Palmateer N, Kimber J, Pharris A, Hope V et al. Interventions to prevent HIV and Hepatitis C in people who inject drugs: a review of reviews to assess evidence of effectiveness. Int J Drug Policy. 2014;25(1):34–52.
b
474
INDICATORS
Received medication for drug dependency Percentage of PWID who received medication for drug dependency in the last 6 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of PWID who received medication for drug dependency in the last 6 months Number of PWID who injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID who received medication for drug dependency in the last 6 months. Treatment for drug dependency is called opiate or opioid substitution therapy (OST), sometimes referred to as medication-assisted treatment. OST is the most effective public health tool for reducing injecting drug use among opioid injectors. OST also helps with the treatment of other health conditions, including HIV, TB, HBV and HCV.a One of the most consistent findings in both high-income and resource-poor settings is that the more time injecting drug users spend on OST, the better the outcomes and the less likely they are to engage in high-risk behaviours.b A low percentage or decreasing trend of OST use may indicate a need for further studies to assess how a country can scale up OST availability or to determine the factors that are preventing PWID from receiving OST.c
Corresponding This indicator is constructed from responses to the following questions from Module 23: Injecting questionnaire drug use. item(s) (if applicable) Numerator: ID6TRT = 1 (yes) [Do you receive medication such as methadone or buprenorphine for your drug dependency?] Denominator: ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Related existing indicators PEPFAR MER KP_MAT: Percentage of PWID on medication-assisted therapy for at least 6 months GARPR 2.6: Number of people on opioid substitution therapy GF KP-6: Percentage of individuals receiving OST who received treatment for at least 6 months
HBV, hepatitis B virus; HCV, hepatitis C virus; OST, opiate or opioid substitution therapy
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016) Kermode M, Crofts N, Kumar MS, Dorabjee J, World Health O. Opioid substitution therapy in resource-poor settings. B World Health Organ. 2011;89(4):243–243.
b c
PEPFAR. Monitoring, evaluation, and reporting indicator reference guide: version 2.1. President’s Emergency Plan For AIDS Relief (PEPFAR); 2015 (http://www.pepfar.gov/documents/ organization/240108.pdf, accessed 8 September 2016).
SECTION 3: INDICATORS
475
Drug overdose Percentage of PWID active in the last 6 months who ever overdosed on narcotics to the point of losing consciousness Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of PWID who ever overdosed on narcotics to the point of losing consciousness Number of PWID who injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) • Time of overdose (over 12 months; within the last 12 months) This indicator measures the percentage of PWID active in the last 6 months who ever overdosed on narcotics to the point of losing consciousness. Drug overdose is often the cause of non-AIDS-related death among people with HIV and the leading cause of death for PWID.a People living with HIV who also inject drugs have higher rates of drug overdose compared to PWID who are HIV negative.b This indicator can help understand the extent of nonfatal overdose experiences. If there is a high percentage or increasing trend in recent drug overdose, countries may consider training healthcare providers to counsel HIV-infected patients who also inject drugs on how to reduce their risk of overdose and to prescribe naloxone to patients as well as to their peers, injecting partners and family.c
Corresponding This indicator is constructed from responses to the following questions from Module 24: Drug questionnaire overdose. item(s) (if applicable) Numerator: • ODEVR = 1 (yes) [Have you ever overdosed on narcotics to the point where you lost consciousness (passed out)?] Denominator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Disaggregation (time of overdose): • ODLSM = 5 (over 12 months) [Please think about the last time you overdosed on narcotics. How long ago did that occur?] • ODLSM = 1, 2, 3 or 4 (within the last 12 months) [Please think about the last time you overdosed on narcotics. How long ago did that occur?] Related existing indicators Not applicable
a b c
Green TC, McGowan SK, Yokell MA, Pouget ER, Rich JD. HIV infection and risk of overdose: a systematic review and meta-analysis. AIDS. 2012;26(4):403–417. Vlahov D, Tang AM, Lyles C, Rezza G, Thomas D, Cohn S et al. Increased frequency of overdose deaths among HIV-infected injection drug users. Addict Res. 2000;8(4):311–326.
Seal KH, Thawley R, Gee L, Bamberger J, Kral AH, Ciccarone D et al. Naloxone distribution and cardiopulmonary resuscitation training for injection drug users to prevent heroin overdose death: a pilot intervention study. 2005;82(2):303–311.
476
INDICATORS
Drug overdose prevention training Percentage of PWID active in the last 6 months who ever received education or training on how to revive someone after an overdose Numerator Number of PWID active in the last 6 months who ever received education or training on how to revive someone after an overdose Number of PWID who injected drugs in the last 6 months PWID • Sex (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of PWID active in the last 6 months who ever received education or training on how to revive someone after an overdose. Many PWID may witness an overdose but may not know how to identify when someone is overdosing and how to respond.a However, they can often act as first responders if properly trained.b Overdose prevention programmes teach overdose response techniques and distribute naloxone – a safe and highly effective opioid overdose antidote, which can help prevent death if PWID know how to quickly respond to an overdose among their peers.c In cases where overdose is high, countries may consider increasing training on overdose prevention.
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 24: Drug questionnaire overdose. item(s) (if applicable) Numerator: • ODEDUC = 1 (yes) [Have you ever received education or training on how to revive someone after an overdose?] Denominator: • ID6MOS = 1 (yes) [In the last 6 months, have you injected any illicit or illegal drugs?] Related existing indicators Not applicable
a b c
Green TC, McGowan SK, Yokell MA, Pouget ER, Rich JD. HIV infection and risk of overdose: a systematic review and meta-analysis. AIDS. 2012;26(4):403–417. Vlahov D, Tang AM, Lyles C, Rezza G, Thomas D, Cohn S et al. Increased frequency of overdose deaths among HIV-infected injection drug users. Addict Res. 2000;8(4):311–326.
Seal KH, Thawley R, Gee L, Bamberger J, Kral AH, Ciccarone D et al. Naloxone distribution and cardiopulmonary resuscitation training for injection drug users to prevent heroin overdose death: a pilot intervention study. 2005;82(2):303–311.
SECTION 3: INDICATORS
477
Targeted information, education and communications Percentage who received targeted IEC from a peer educator or outreach worker in the last 12 months Numerator Number of participants who received targeted IEC from a peer educator or outreach worker in the last 12 months Total number of participants SW, MSM, TG, PWID • Sex for SW, PWID (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who received targeted IEC from a peer educator or outreach worker in the last 12 months. PEPFAR recommends targeted IEC as a prevention intervention that can be included in a comprehensive programme for KP.a A low percentage or decreasing trend in this indicator may suggest challenges with reaching the target population, developing IEC material or recruiting, or training and supervising peer educators and outreach workers. In such cases, countries may consider involving KP in the development of IEC strategies for KP. This indicator is constructed from responses to the following question from Module 27: Services uptake. Numerator: • CSORTI = 3 (in the last year) [How long ago did a peer educator or outreach worker talk to you about HIV?] Denominator: • Total number of participants Related existing indicators PEPFAR MER: KP_PREV: Percentage of KP reached with individual and/or small group level HIV preventive interventions that are based on evidence and/or meet the minimum standards required
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
IEC, information, education and communication
a
PEPFAR. MER level 1 indicator package. President’s Emergency Plan For AIDS Relief (PEPFAR); 2013.
478
INDICATORS
Received condoms Percentage who received condoms through an outreach service, drop-in centre or sexual health clinic in the last 12 months Numerator Number of participants who received condoms through an outreach service, drop-in centre or sexual health clinic in the last 12 months Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who received condoms through an outreach service, drop-in centre or sexual health clinic in the last 12 months. Correct and consistent use of condoms can significantly reduce the risk of HIV transmission and acquisition as well as transmission of some other STIs.a,b,c Condom programming is one of several interventions in the WHO, UNODC and UNAIDS comprehensive package for the prevention, treatment and care of HIV among KP. A low percentage or decreasing trend may suggest challenges with condom procurement, lack of access to services that distribute free condoms, or lack of awareness of these services. In such cases, countries may consider additional venues for distributing condoms. This indicator is constructed from responses to the following questions from Module 27: Services uptake and Module 18: Condom use/accessibility. Numerator: From Module 10: IF NOT PWID: • CSORPR = 1 (condoms – yes) [What items did you receive the last time you met a peer educator or outreach worker?] IF PWID: • CSORID = 1 (condoms – yes) [What items did you receive the last time you met a peer educator or outreach worker?] OR From Module 15: FOR ALL: • COFREE = 1 (yes) [In the last 12 months, have you been given condoms for free? For example, through an outreach service, drop-in centre or health clinic] Denominator: • Total number of participants Related existing indicators Not applicable
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
OVP, other vulnerable populations; UNODC, United Nations Office on Drugs and Crime
a
Charania MR, Crepaz N, Guenther-Gray C, Henny K, Liau A, Willis LA et al. Efficacy of structural-level condom distribution interventions: a meta-analysis of U.S. and international studies, 1998–2007. AIDS Behav. 2011;15(7):1283–1297. UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). PEPFAR. Technical guidance on combination HIV prevention. President’s Emergency Plan For AIDS Relief (PEPFAR); 2011.
b
c
SECTION 3: INDICATORS
479
Never tested for HIV Percentage who have never received an HIV test Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who have never received an HIV test Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who have never been tested for HIV. Routine, voluntary HIV testing enables individual awareness of one’s HIV status, linking those identified as HIV positive to care and treatment services. KP face unique barriers to accessing HIV services, including perceived or real discrimination by service providers. WHO recommends that voluntary HIV testing and counselling (HTC) be routinely offered to all KP in both community and clinical settings.a A high percentage or increasing trend in the percentage who have never been tested for HIV may indicate structural, operational, logistical and social barriers, such as stigma, discrimination and punitive legislation, that limit access to existing testing services.b Such findings may indicate the need for national governments and programmes to identify and address gaps in HTC services to increase uptake of HTC among KP. This indicator is constructed from responses to the following question from Module 27: Services uptake. Numerator: • CSCTEV = 2 (no) [Have you ever been tested for HIV?] Denominator: • Total number of participants Related existing indicators Not applicable
Corresponding questionnaire item(s)
HTC, HIV testing and counselling; OVP, other vulnerable populations
a
WHO. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – Section 7.3: Monitoring response to ART and the diagnosis of treatment failure. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/hiv/pub/guidelines/arv2013/art/artmonitoring/en/index3.html, accessed 8 September 2016). WHO. Service delivery approaches to HIV testing and counselling (HTC): a strategic HTC programme framework. 2012 (http://apps.who.int/iris/bitstream/10665/75206/1/9789241593877_eng.pdf, accessed 8 September 2016).
b
480
INDICATORS
HIV test in last 12 months Percentage who received an HIV test in the last 12 months and know their results Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who received an HIV test in the last 12 months and know their results Total number of participants (excluding known positives who were tested more than 12 months ago) SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • Time of test (in last 12 months, in last 6 months) This indicator measures the percentage of participants who received an HIV test in the last 12 months and know their results. Knowledge of one’s status is a key factor in the decision to seek treatment and prevent transmission.a HIV testing with high-quality counselling can reduce HIV risk and STIs among HIV-negative men and women.b,c A low percentage or decreasing trend of participants who have been tested for HIV in the last 12 months may indicate low access to HTC services or fear of HIV-related stigma. Such findings may help countries to identify and address gaps in HTC services to increase uptake of HTC among KP. This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSCTTI = 1 (in the last 6 months) or 2 (6–12 months) [When did you last test for HIV?] AND • CSCTRS = 1 (positive), 2 (negative) or 3 (undetermined) [What was the result of your last test?] Denominator (total number of respondents excluding known positives who were tested more than 12 months ago): • Total number of participants MINUS those who meet the following 2 conditions: • CSCTRS = 1 (positive) [What was the result of your last test?] AND • CSCTTI = 3 (more than 12 months ago) [When did you last test for HIV?] Disaggregation (time of test): • CSCTTI = 1 (in the last 6 months) [When did you last test for HIV?] • CSCTTI = 2 (6–12 months) [When did you last test for HIV?] Related existing indicators • PEPFAR MER HTC_TST: Number of individuals who received HIV Testing and Counseling (HTC) services for HIV and received their test results • GARPR 1.9: Percentage of sex workers who have received an HIV test in the past 12 months and know their results (BSS) HTC, HIV testing and counselling; OVP, other vulnerable populations
Corresponding questionnaire item(s)
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016). PEPFAR. Technical guidance on combination HIV prevention. President’s Emergency Plan For AIDS Relief (PEPFAR); 2011. PEPFAR. Comprehensive HIV prevention for people who inject drugs: revised guidance. President’s Emergency Plan For AIDS Relief (PEPFAR); 2010.
b c
SECTION 3: INDICATORS
481
Tuberculosis screening Percentage of HIV-positive participants who received TB screening after their HIV diagnosis Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of HIV-positive participants who received TB screening after their HIV diagnosis Number of participants who had a positive HIV test result at their last HIV test SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of HIV-positive participants who received TB screening after their HIV diagnosis. TB is the leading cause of mortality among PLHIV. Screening for TB among PLHIV at initial and subsequent HIV care visits is recommended to identify TB suspects and link them to diagnosis and treatment.a WHO recommends that all HIV-positive patients be screened for TB.b A low percentage or decreasing trend in this indicator may suggest that not all PLHIV are being screened for TB due to changes in policy or services, such as a turnover in trained staff or decreased supervision visits. In such cases, countries may consider ways to improve the integrated delivery of TB and HIV services or the TB screening capacity at health facilities. This indicator should be interpreted along with the indicator “Received TB treatment”. Corresponding questionnaire item(s) This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSTBSC2 = 2 (After I tested HIV-positive) or 3 (Both before and after I tested HIV-positive) [When did they screen you for TB?] Denominator: • CSCTRS = 1 (Positive) [What was the result of your last (HIV) test?] Related existing indicators • PEPFAR MER TB_SCREENDX: Number of PLHIV in HIV clinical care who were screened for TB symptoms at the last clinical visit • GF TB/HIV-3: Percentage of HIV-positive patients who were screened for TB in HIV care and treatment settings
OVP, other vulnerable populations; PLHIV, people living with HIV
a b
PEPFAR. MER level 1 indicator package. President’s Emergency Plan For AIDS Relief (PEPFAR); 2013.
WHO. Systematic screening for active tuberculosis: principles and recommendations. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/tb/tbscreening/en/, accessed 8 September 2016).
482
INDICATORS
Received tuberculosis treatment Percentage of HIV-positive participants who screened positive for TB and received TB treatment Numerator Denominator Number of HIV-positive participants who screened positive for TB and received TB treatment Number of participants who had a positive HIV test result at their last HIV test and screened positive for TB SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of HIV-positive participants who screened positive for TB and who received TB treatment. TB is a leading cause of morbidity and mortality in people living with HIV, including those on ART.a All HIV-positive patients should be screened for TB. Patients who screen positive are suspected TB cases who should receive additional evaluation, diagnosis and treatment. A low percentage or decreasing trend in this indicator may suggest weak referral systems, staff who are untrained in coordinated care or unmotivated to take on additional activities, or lack of access to TB diagnosis and treatment services among HIV patients. In such cases, countries may consider ways to improve integrated delivery of TB and HIV services for KP at health facilities. This indicator should be interpreted along with the indicator “TB screening”. This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSTBTR1 = 1 (yes) [The last time you had TB, did you get treatment for TB?] Denominator: • CSTBRS = 2 (TB-positive) [What was the result of your last TB test?] AND • CSCTRS = 1 (positive) [What was the result of your last (HIV) test?] Related existing indicators • GARPR 5.1: Percentage of estimated HIV-positive incident TB cases that received treatment for both TB and HIV • GF TB/HIV-4: Percentage of new HIV-positive patients starting IPT during the reporting period
Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
IPT, isoniazid preventive therapy; OVP, other vulnerable populations
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016).
SECTION 3: INDICATORS
483
Currently in care Percentage of HIV-positive participants who received at least one of the following in the last 12 months: CD4 count or viral load Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of HIV-positive participants who received at least one of the following in the last 12 months: CD4 count or viral load Number of HIV-positive participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of self-reported HIV-positive participants who are currently in care, based on receipt of CD4 count or viral load in the last 12 months. For those not eligible for ART, care visits provide opportunities for screening, prevention and treatment of other conditions and comorbidities. For those on treatment, uninterrupted ART and routine monitoring are required for optimal treatment outcomes.a A low percentage or decreasing trend in the percentage who are currently in care may result from a lack of access to equipped facilities, stigma and discrimination from health-care providers, or insufficient resources to pay for testing.b Interventions addressing multiple levels of the health-care system and harnessing social support can improve retention in care.c Note that this indicator does not capture those who received a clinical assessment (WHO staging). Corresponding questionnaire item(s) This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSCDEV = 1 (in the last 6 months) or 2 (between 7 and 12 months ago) [When did your care provider last test your CD4 count?] OR • CSVLTI = 1 (in the last 12 months) [When did you last have a viral load test?] Denominator: • CSCTRS = 1 (positive) [What was the result of your last test?] Related existing indicators Not applicable
OVP, other vulnerable populations
a
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016); Section 9.3 Retention across the continuum of care. Mtetwa S, Busza J, Chidiya S, Mungofa S, Cowan F. “You are wasting our drugs”: health service barriers to HIV treatment for sex workers in Zimbabwe. BMC Public Health. 2013;13:698 (https://www. ncbi.nlm.nih.gov/pubmed/23898942, accessed 8 September 2016).
b
c WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016); Table 9.1 Factors related to the health system and people receiving ART affecting retention and adherence with possible interventions.
484
INDICATORS
CD4 count Percentage of HIV-positive participants who received a CD4 count measurement in the last 12 months Numerator Denominator Applicable KP Recommended disaggregation Number of self-reported HIV-positive participants who received a CD4 count in the last 12 months Number of self-reported HIV-positive participants SW, MSM, TG, PWID, OVP • • • • Sex for SW, PWID, OVP (male, female) Age (<25 years, 25+ years) Treatment status (currently taking ART, never taken ART) Timing of CD4 count (in the last 6 months, 7–12 months ago)
Purpose/ rationale
This indicator measures the percentage of self-reported HIV-positive participants who received a CD4 count in the last 12 months. CD4 count measurements are used to monitor the progress of HIV and to determine the need to start ART. WHO recommends CD4 counts every 6–12 months for those not on ART and every 6 months for those on ART.a Because all HIV-positive individuals should receive a periodic clinical assessment, this indicator can serve as a proxy for tracking progress in providing clinical care to HIV-positive KP.b For individuals newly diagnosed with HIV, CD4 count may indicate how early or delayed the HIV diagnosis was made.c A low percentage or decreasing trend in the percentage who received a CD4 count may result from a lack of access to equipped facilities, stigma and discrimination from health-care providers, or insufficient resources to pay for testing.d In such cases, countries my consider interventions addressing multiple levels of the health-care system and harnessing social support to improve retention in care.e Note that this indicator does not capture those who were previously on treatment but are currently not taking ART.
Corresponding questionnaire item(s)
This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSCDEV2 = 1 (in the last 6 months) or 2 (between 7 and 12 months ago) [When did your care provider last test your CD4 count?] Denominator: • CSCDEV2 = 1, 2, 3, 7, 8 (in the last 6 months, between 7 and 12 months ago, more than 12 months ago, don’t know, refuse to answer) Disaggregation (treatment status): • CSTRCURR = 1 (yes) [Are you currently taking ARVs?] • CSTREV = 2 (no) [Have you ever taken ARVs – that is, antiretroviral medication – to treat your HIV infection?] Disaggregation (timing of CD4 count): • CSCDEV2 = 1 (in the last 6 months) [When did your care provider last test your CD4 count?] • CSCDEV2 = 2 (between 7 and 12 months ago) [When did your care provider last test your CD4 count?]
Related existing indicators
Not applicable
OVP, other vulnerable populations
a
WHO. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – Table 7.3.1 Laboratory monitoring before and after initiating ART. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/hiv/pub/guidelines/arv2013/art/WHO_CG_7.3.1.pdf?ua=1, accessed 8 September 2016). PEPFAR. MER level 1 indicator package. President’s Emergency Plan For AIDS Relief (PEPFAR); 2013 (p. 12).
b c
MacCarthy S, Bangsberg DR, Fink G, Reich M, Gruskin S. Late presentation to HIV/AIDS testing, treatment or continued care: clarifying the use of CD4 evaluation in the consensus definition. HIV Med. 2014;15(3):130–134. d
Mtetwa S, Busza J, Chidiya S, Mungofa S, Cowan F. “You are wasting our drugs”: health service barriers to HIV treatment for sex workers in Zimbabwe. BMC Public Health. 2013;13:698 (https://www. ncbi.nlm.nih.gov/pubmed/23898942, accessed 8 September 2016).
e
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016); Table 9.1 Factors related to the health system and people receiving ART affecting retention and adherence with possible interventions.
SECTION 3: INDICATORS
485
Currently on antiretroviral therapy Percentage of HIV-positive participants eligible for treatment who are currently taking ART Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of HIV-positive treatment-eligible participants who are currently taking ART Total number of HIV-positive respondents who are eligible for treatment SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of HIV-positive participants eligible for treatment who are currently taking ART. This indicator can be used to assess progress in treatment coverage and helps monitor the HIV service cascade.a,b A low percentage or decreasing trend in treatment coverage may indicate challenges in linking KP to care and treatment or in procurement and distribution of ART, or stigma and discrimination by service providers. In such cases, countries may consider making treatment accessible to KP, training service providers and improving linkages. Note: Some surveys may test for ARV metabolites (numerator) and CD4 count (denominator), which allows for comparison of self-reported data with biological data. Please see the biomarker HIV treatment coverage indicator reference sheet for more information. Corresponding questionnaire item(s) This indicator is constructed from responses to the following questions from Module 27: Services uptake. Numerator: • CSTRCURR = 1 (yes) [Are you currently taking ARVs?] Denominator: • CSCTRS = 1 (positive) [What was the result of your last test?] AND • CSCDRS1 = (dependent on country eligibility for treatment) (2 = more than 500); (3 = between 350 and 500); (4 = between 200 and 349); or (5 = below 200). [What was the result of your last CD4 count?] • If CSCDRS1 = 1 (did not receive result) exclude OR • CSCDRS2 = |__ |__| | (dependent on country eligibility for treatment) [What was your last CD4 count?] Related existing indicators • GARPR 4.1: Percentage of eligible adults and children currently receiving antiretroviral therapy • GF TCS-1: Percentage of eligible adults and children currently receiving antiretroviral therapy among all adults and children living with HIV
OVP, other vulnerable populations
a b
PEPFAR. MER level 1 indicator package. President’s Emergency Plan For AIDS Relief (PEPFAR); 2013.
Mountain E, Mishra S, Vickerman P, Pickles M, Gilks C, Boily M-C. Antiretroviral therapy uptake, attrition, adherence and outcomes among HIV-infected female sex workers: a systematic review and meta-analysis. PLoS One. 2014;9(9):e105645.
486
INDICATORS
Prevention of mother-to-child transmission Percentage of HIV-positive female participants who received ARVs to reduce the risk of mother-tochild transmission during their last pregnancy that resulted in a birth Numerator Number of HIV-positive female participants who received ARVs to reduce the risk of mother-to-child transmission during their last pregnancy that resulted in a birth Total number of women who were HIV-positive at their last pregnancy that resulted in a birth (in the last 5 years) Female PWID, FSW, female OVP • Age (<25 years, 25+ years)
Denominator
Applicable KP Recommended disaggregation Purpose/ rationale
This indicator measures the percentage of HIV-positive female participants who received ARVs to reduce the risk of mother-to-child transmission during their last pregnancy that resulted in a birth.a A low percentage or decreasing trend for coverage of HIV-positive pregnant women who receive ARV prophylaxis or treatment may indicate poor access by the target population to ANC services; a weak PMTCT system; and inadequate procurement and distribution of sufficient ARVs for pregnant women. In such cases, countries may explore ways to scale up PMTCT programmes for KP and address any barriers to accessing PMTCT services, including combating stigma and discrimination and increasing the availability of community-based services to the KP. This indicator is constructed from responses to the following questions from Module 13: Reproductive health. Numerator: • RHHIVARV = 1 (yes) [Some HIV-positive women who are pregnant take ARV medicine to reduce the risk of passing HIV on to their baby. Did you take ARVs before giving birth?] Denominator: • RHEVBRTH = 1 (yes) [Have you ever given birth?] AND • RHHIVRES = 1 (positive) [During the last time you were pregnant, what was the result of the last HIV test?]
Corresponding questionnaire item(s)
Related existing indicators
GARPR 3.1: Percentage of HIV-positive pregnant women who receive antiretrovirals to reduce the risk of mother-to-child transmission GF PMTCT-2: Percentage of HIV-positive pregnant women who received antiretrovirals to reduce the risk of mother-to-child transmission
ANC, antenatal clinic; ARV, antiretroviral; OVP, other vulnerable populations; PMTCT, prevention of mother-to-child transmission
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016).
SECTION 3: INDICATORS
487
Hepatitis B virus testing Percentage who were tested for HBV in the last 12 months Numerator Denominator Applicable KP Recommended disaggregation Number of participants who were tested for HBV in the last 12 months Total number of participants, except those who tested positive for HBV more than 12 months ago SW, MSM, TG, PWID • Sex for SW, PWID (male, female) • Age (<25 years, 25+ years) • Vaccination history (received HBV vaccine; did not receive HBV vaccine) This indicator measures the percentage of participants who were tested for HBV in the last 12 months. HBV can be transmitted through several pathways, including sexual contact and injecting drug use. Vaccination programmes have dramatically reduced HBV prevalence, but it continues to be spread, particularly where vaccine programmes are not routinely or universally implemented.a Early diagnosis of HBV infection through blood tests is important because it can lead to care and treatment and decrease transmission. HBV diagnosis, treatment and vaccination are part of WHO’s comprehensive package of services for KP.b A low percentage or decreasing trend of those tested (among those who are at risk) may indicate a lack of coverage of or access to HBV testing. In such cases, countries may consider targeted outreach, provider education or a routine vaccination programme. This indicator is constructed from responses to the following questions from Module 26: Hepatitis B and C. Numerator: • HBTTME = 1 (less than 12 months ago) [When did you have your most recent hepatitis B test?] Denominator: • Total number of participants MINUS THE FOLLOWING: • HBTTME = 2 (more than 12 months ago) or 3 (more than 5 years ago) [When did you have your most recent hepatitis B test?] AND • HBRSL = 1 (positive) [What was the result of your last hepatitis B test?] Disaggregation (vaccination history): • HBVAC = 1 (yes) [Have you ever received the hepatitis B vaccine?] • HBVAC = 2 (no) [Have you ever received the hepatitis B vaccine?]
Purpose/ rationale
Corresponding questionnaire item(s) (if applicable)
Related existing indicators HBV, hepatitis B virus
Not applicable
a
Ott JJ, Stevens GA, Groeger J, Wiersma ST, Visscherd BR, Vanablee PA et al. Global epidemiology of hepatitis B virus infection: new estimates of age-specific HBsAg seroprevalence and endemicity. Vaccine. 2012;30(12):2212–2219. PEPFAR. Comprehensive HIV prevention for people who inject drugs: revised guidance. President’s Emergency Plan For AIDS Relief (PEPFAR); 2010.
b
488
INDICATORS
Hepatitis C virus testing Percentage of PWID who were ever tested for HCV Numerator Denominator Applicable KP Recommended disaggregation Number of PWID who were ever tested for HCV Total number of PWID PWID • Sex (male, female) • Age (<25 years, 25+ years) • Treatment history (took medicine to treat HCV infection, did not take medicine to treat HCV infection) • Time tested (less than 12 months ago, more than 12 months ago) • HIV status (positive, negative, unknown) This indicator measures the percentage of PWID who were ever tested for HCV. HCV infection is high among PWID.a WHO guidelines recommend that one-time testing should be performed for those with increased risk of HCV infection, including PWID and PLHIV.b Once identified, HCV-infected people can be offered treatment. In addition, knowing one’s status may promote safer injecting drug use practices to reduce transmission to others.c A low percentage or decreasing trend in the number of PWID who were tested for HCV may indicate a lack of knowledge of HCV or low coverage of or access to testing. In such cases, countries may consider increasing awareness of HCV risks and access to HCV prevention tools and diagnostic testing among PWID to help reach undiagnosed HCV-infected individuals.
Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 26: Hepatitis B questionnaire and C and Module 27: Services uptake. item(s) (if applicable) Numerator: • HCTEST = 1 (yes) [Have you ever been tested for hepatitis C?] Denominator: • Total number of PWID Disaggregation (treatment history) [Module 11]: • HCMED = 1 (yes) [Have you ever taken medicine to treat your hepatitis C infection?] • HCMED = 2 (no) [Have you ever taken medicine to treat your hepatitis C infection?] Disaggregation (time tested) [Module 11]: • HCTTME = 1 (less than 12 months ago) [When did you have your most recent hepatitis B test?] • HCTIME = 2 (more than 12 months ago) or 3 (more than 5 years ago) [When did you have your most recent hepatitis B test?] Disaggregation (HIV status) [Module 10]: • CSCTRS = 1 (positive) [What was the result of your last (HIV) test?] • CSCTRS = 2 (negative) [What was the result of your last (HIV) test?] • CSCTRS = 3 (unclear/neither positive or negative) or 4 (did not receive result) [What was the result of your last (HIV) test?] Related existing indicators Not applicable
ART, antiretroviral therapy; HCV, hepatitis C virus; OVP, other vulnerable populations
a
Nelson PK, Mathers BM, Cowie B, Hagan H, Des Jarlais D, Horyniak D et al. Global epidemiology of hepatitis B and hepatitis C in people who inject drugs: results of systematic reviews. Lancet. 2011;378(9791):571–583. WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016). HIV: science and stigma. Lancet. 2014;384(9939):207.
b
c
SECTION 3: INDICATORS
489
Discussed pre-exposure prophylaxis Percentage who discussed PrEP with a health-care provider in the last 12 months Numerator Denominator Number of participants who discussed PrEP with a health-care provider in the last 12 months Number of participants eligible for PrEP (i.e. those who are HIV negative as of their last HIV test) who visited a health-care provider in the last 12 months SW, MSM, TG, PWID Sex for SW, PWID (male, female)
Applicable KP Recommended disaggregation Purpose/ rationale
This indicator measures the percentage of participants who discussed PrEP with a health-care provider in the last 12 months. PrEP involves daily use of ART by uninfected individuals to prevent them from getting HIV. Clinical trials of daily oral PrEP have provided evidence of effectiveness among MSM, TG and PWID.a Global consensus supports the use of PrEP to reduce the risk of getting or transmitting HIV, regardless of population or setting.b A low percentage or decreasing trend in this indicator may suggest limited use of health services by KP; health-care providers’ lack of knowledge of PrEP guidelines; or health-care providers’ concerns and preconceived notions about adherence, drug resistance, or behavioural disinhibition/risk compensation. In such cases, countries may consider training providers in the use of PrEP. This indicator is constructed from responses to the following questions from Module 28: PEP and PrEP, Module 27: Services uptake, and respondent background. Numerator [Module 12]: • PPRDISC: When has a health-care provider last offered or discussed PrEP with you? (1 = yes) Denominator [Module 10 and respondent background]: • CSCTRS = 2 (negative) [What was the result of your last test?] AND • PPRDISC: When has a health-care provider last offered or discussed PrEP with you? (1, 2 or 3)
Corresponding questionnaire item(s)
Related existing indicators
Not applicable
ART, antiretroviral therapy; PEP, post-exposure prophylaxis; PrEP, pre-exposure prophylaxis
a
WHO. Guidance on oral pre-exposure prophylaxis (PrEP) for serodiscordant couples, men and transgender women who have sex with men at high risk of HIV: recommendations for use in the context of demonstration projects. Geneva: World Health Organization (WHO); 2012 (http://apps.who.int/iris/bitstream/10665/75188/1/9789241503884_eng.pdf, accessed 8 September 2016). WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016).
2
490
INDICATORS
Taken pre-exposure prophylaxis Percentage who took PrEP in the last 12 months Numerator Denominator Number of participants who took PrEP in the last 12 months Number of participants who are eligible for PrEP (i.e. those who are HIV negative as of their last HIV test) SW, MSM, TG, PWID • Sex for SW, PWID (male, female) This indicator measures the percentage of participants who took PrEP in the last 12 months. PrEP involves daily use of ART by uninfected individuals to prevent them from getting HIV. Clinical trials of daily oral PrEP have provided evidence of effectiveness among MSM, TG and PWIDa. Global consensus supports the use of PrEP to reduce the risk of getting or transmitting HIV, regardless of population or settingb. A low percentage or decreasing trend in this indicator may suggest logistical and financial challenges to successful rollout of PrEP, including inefficient drug delivery and lack of human resources and community education. It may also be due to lack of knowledge among providers and KP. In such cases, countries should integrate PrEP into existing comprehensive HIV prevention servicesc. As additional rounds of BBS occur, it will be possible to examine trends in adoption of PrEP within countries and the target population. This indicator may inform modifications to PrEP guidelines, outreach strategies, patient education and behavioural interventions. This indicator is constructed from responses to the following questions from Module 28: PEP and PrEP and Module 27: Services uptake. Numerator [Module 12]: • PPRTAKE2 = 1, 2 or 3 (last sex, last month, last 12 months) [Which answer describes best when you last have taken PrEP?] Denominator [Module 10]: • CSCTRS = 2 (negative) [What was the result of your last test?] Related existing indicators PEPFAR MER PrEP_NEW: Number of adults and adolescents who have received antiretroviral preexposure prophylaxis (PrEP) in the reporting period to prevent HIV infection
Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
ART, antiretroviral therapy; PEP, post-exposure prophylaxis; PrEP, pre-exposure prophylaxis
a
WHO. Guidance on oral pre-exposure prophylaxis (PrEP) for serodiscordant couples, men and transgender women who have sex with men at high risk of HIV: recommendations for use in the context of demonstration projects. Geneva: World Health Organization (WHO); 2012 (http://apps.who.int/iris/bitstream/10665/75188/1/9789241503884_eng.pdf, accessed 8 September 2016). WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016). Underhill K, Operario D, Mimiaga MJ, Skeer MR, Mayer KH. Implementation science of pre-exposure prophylaxis: preparing for public use. Curr HIV/AIDS Rep. 2010;7(4):210–219.
b
c
SECTION 3: INDICATORS
491
Current use of modern family planning methods Percentage of female participants currently using a modern family planning method Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of female participants who are currently using a modern family planning method Number of female participants not seeking to get pregnant FSW, female PWID • Age (<25 years, 25+ years) • Type of contraceptive method (pill, injection, norplant, IUD or other) • HIV status (positive, negative, unknown) This indicator measures the percentage of female participants who are currently using a modern method of family planning among those who are not seeking to get pregnant. Contraceptives prevent unintended pregnancies, reduce the number of abortions, and lower the incidence of death and disability related to complications of pregnancy and childbirth.a Access to modern family planning methods allows women to choose more effective methods and prevent unintended pregnancies. A low percentage or decreasing trend in this indicator may suggest that participants do not have access to modern methods or do not feel like they have access to family planning facilities because of experienced discrimination and stigma.b It may also indicate procurement challenges or insufficiently trained health-care providers. In such cases, countries may need to address operational policies and practices as well as stigma and discrimination to help FSW and female PWID access modern methods. This indicator is constructed from responses to the following questions from Module 13: Reproductive health and Module 27: Services uptake. Numerator [Module 13]: • RHTYPNOW = 1 (pill), 2 (injection), 3 (norplant), 4 (IUD) [Which method do you mainly use?] OR • RHTYPOTH = 1 (sterilization), 2 (patch), 3 (vaginal ring), 4 (condom) [You said “other”. Which of the following family planning methods do you mainly use?] Denominator [Module 13]: • RHTRYPRG = 2 (no) [Are you currently trying to get pregnant?] Disaggregation (type of contraceptive method) [Module 13]: • RHTYPNOW = 1 (pill), 2 (injection), 3 (norplant) or 4 (IUD) [Which method do you mainly use?] • RHTYPOTH = 1 (sterilization), 2 (patch), 3 (vaginal ring) and 4 (condoms) [You said “other”. Which of the following family planning methods do you mainly use?] Disaggregation (HIV status) [Module 10]: • CSCTRS = 1 (positive) [What was the result of your last (HIV) test?] • CSCTRS = 2 (negative) [What was the result of your last (HIV) test?] • CSCTRS = 3 (unclear/neither positive or negative) or 4 (did not receive result) [What was the result of your last (HIV) test?] Related existing indicators IUD, intrauterine device
Corresponding questionnaire item(s) (if applicable)
Not applicable
a b
Singh S, Darroch JE. Adding it up: costs and benefits of contraceptive services estimates for 2012. Guttmacher Institute: UNFPA. 2012.
FHI360/PTA/USAID. Integrating family planning into HIV programs: evidence-based practices. fhi360; 2013 (https://www.fhi360.org/sites/default/files/media/documents/fp-hiv-evidence%20 based%20practices%202013.pdf, accessed 8 September 2013).
492
INDICATORS
Antenatal clinic access and HIV testing Percentage who gave birth in the last 5 years who were tested for HIV in an ANC during their most recent pregnancy Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of female participants who gave birth in the last 5 years who were tested for HIV in an ANC during their most recent pregnancy Number of female participants who gave birth in the last 5 years FSW, female PWID • Age (<25 years, 25+ years) This indicator measures the percentage of female participants who gave birth in the last 5 years who were tested for HIV in an ANC during their most recent pregnancy. This indicator reflects one component of PMTCT, which is to increase the number of pregnant women who know their HIV status.a Women living with HIV are at greater risk of adverse birth outcomes and have higher maternal mortality rates overall.b Identification of HIV-positive women when they are pregnant is a key entry point into care and treatment services. A low percentage or decreasing trend in this indicator may suggest that women in the target population are not accessing antenatal services or are not being tested for HIV at ANC. In such cases, countries may consider training for health providers, media and community campaigns promoting ANC, and outreach to FSW and female PWID who may be pregnant.
Corresponding This indicator is constructed from responses to the following questions from Module 13: questionnaire Reproductive health. item(s) (if applicable) Numerator: • RHANCUSE = 1 (yes) [Think about the last time you were pregnant. Did you go to a health facility for antenatal care?] AND • RHHIVTST = 1 (yes) (one time) or (two or more times) [During any of your last visits to the antenatal clinic the last time you were pregnant, were you tested for HIV?] Denominator: • RHEVBRTH = 1 (yes) [Have you given birth in the last 5 years?] Related existing indicators • PEPFAR MER PMTCT_STAT: Percentage of pregnant women with known HIV status (includes those who already knew their HIV status prior to ANC) • GF PMTCT-1: Percentage of pregnant women who know their HIV status • GF HSS O-1: Percentage of women attending antenatal care
ANC, antenatal clinic; PMTCT, prevention of mother-to-child transmission
a b
Next generation indicators reference guide addendum: recommended indicators reference sheets. President’s Emergency Plan For AIDS Relief (PEPFAR); 2009. Horton R. Maternal mortality: surprise, hope, and urgent action. Lancet. 2010;375(9726):1581–1582.
SECTION 3: INDICATORS
493
Seeking care for sexually transmitted infection symptoms Percentage who visited a health-care provider in the last 12 months to seek care for STI symptoms Numerator Number of participants who visited a health-care provider in the last 12 months to seek care for STI symptoms Number of participants who experienced STI symptoms SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who experienced any STI symptoms in the last 12 months and sought care from a health-care provider. STIs increase the risk of getting and passing on HIV.a,b A person who seeks care for STI symptoms should receive risk-reduction counselling, and those who are not already HIV-infected should be tested for HIV.c However, KP may not seek the health care they need due to perceived shame and feeling disrespected by doctors or other healthcare workers.d This has implications for timely STI treatment and is a missed opportunity to test for HIV. A low percentage or decreasing trend in this indicator may suggest a lack of access to services, lack of sensitivity to KP among health-care providers, or decreased STI prevalence among the target population. In such cases, countries may work with STI service providers to help KP gain access to treatment services. This indicator is constructed from responses to the following questions from Module 25: Sexually transmitted infections. Numerator: • STSTDX = 1 (yes) [Did you see a health-care provider because of these problems?] Denominator: For males • STABNMX = 1 (yes) [In the last 12 months, have you had an abnormal discharge from your penis?] OR • STULCM = 1 (yes) [In the last 12 months, have you had an ulcer or sore on or near your penis?] For females • STABNF = 1 (yes) [During the last 12 months, have had an abnormal discharge from your vagina? This may include an unusual smell, colour or texture.] OR • STULCF = 1 (yes) [In the last 12 months, have you had an ulcer or sore on or near your vagina?] Related existing indicators Not applicable
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
OVP, other vulnerable populations
a b
McClelland RS, Sangare L, Hassan WM, Lavreys L, Mandaliya K, Kiarie J et al. Infection with Trichomonas vaginalis increases the risk of HIV-1 acquisition. J Infect Dis. 2007;195(5):698–702.
Solomon MM, Mayer KH, Glidden DV, Liu AY, McMahan VM, Guanira JV et al. Syphilis predicts HIV incidence among men and transgender women who have sex with men in a preexposure prophylaxis trial. Clin Infect Dis. 2014;59(7):1020–1026. Miranda AE, Nasser SO, Simon M, Guerra PB, Sant’ Anna PM. Risk factors and prevalence of HIV infection in people seeking health care in an STI clinic in Brazil. AIDS Care. 2007;19(1):75–78.
c d
Veldhuijzen NJ, van Steijn M, Nyinawabega J, Kestelyn E, Uwineza M, Vyankandondera J et al. Prevalence of sexually transmitted infections, genital symptoms and health-care seeking behaviour among HIV-negative female sex workers in Kigali, Rwanda. Int J of STD AIDS. 2013;24(2):139–143
494
INDICATORS
Sexually transmitted infection screening Percentage who were screened for STIs in the last 12 months Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who were screened for STIs in the last 12 months Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who were screened for STIs in the last 12 months. Globally, STIs are a major cause of acute illness, infertility, long-term disability and death.a Several STIs increase the risk of both getting and passing on HIV.b,c Because of their sexual risk behaviours, STI screening is recommended for all KP.a A low percentage or decreasing trend in this indicator may suggest lack of knowledge of STI screening guidelines for KP. In such cases, countries may consider improving the quality of STI services through partnerships with KP advocacy groups in order to make services friendlier to these populations and screening more frequent. This indicator is constructed from responses to the following questions from Module 25: Sexually transmitted infections. Numerator: • STISCREEN: In the last 12 months did a health-care provider check or test you for STI? Yes = 1 Denominator: • STISCREEN: In the last 12 months did a health-care provider check or test you for STI? Yes = 1 or No = 2 Related existing indicators Not applicable
Corresponding questionnaire item(s)
OVP, other vulnerable populations; STI, sexually transmitted infection
a
Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016). Solomon MM, Mayer KH, Glidden DV, Liu AY, McMahan VM, Guanira JV et al. Syphilis predicts HIV incidence among men and transgender women who have sex with men in a preexposure prophylaxis trial. Clin Infect Dis. 2014;59(7):1020–1026. McClelland RS, Sangare L, Hassan WM, Lavreys L, Mandaliya K, Kiarie J et al. Infection with Trichomonas vaginalis increases the risk of HIV-1 acquisition. J Infect Dis. 2007;195(5):698–702.
b
c
SECTION 3: INDICATORS
495
Received sexually transmitted infection treatment Percentage who screened positive for an STI and received treatment in the last 12 months Numerator Number of participants who screened positive for an STI and received treatment in the last 12 months Number of participants who screened positive for an STI in the last 12 months SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who screened positive for an STI and received treatment for the STI in the last 12 months. Globally, STIs are a major cause of acute illness, infertility, long-term disability and death.a Several STIs increase the risk of getting and passing on HIV.b,c STI screening and treatment is recommended for all KP.a,d A low percentage or decreasing trend in this indicator may suggest lack of knowledge of STI screening and treatment guidelines for KP. In such cases, countries may consider improving the quality of STI services through partnerships with KP advocacy groups in order to make services friendlier to these populations and screening more frequent. This indicator is constructed from responses to the following questions from Module 25: Sexually transmitted infections. Numerator: • STPHARM = 1 (yes) [Did you get treatment for these problems?] Denominator: • STSTDX = 1 (yes) [Did the health-care provider tell you that you had a sexually transmitted infection?] Related existing indicators Not applicable
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
OVP, other vulnerable populations; STI, sexually transmitted infection
a
Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016). McClelland RS, Sangare L, Hassan WM, Lavreys L, Mandaliya K, Kiarie J et al. Infection with Trichomonas vaginalis increases the risk of HIV-1 acquisition. J Infect Dis. 2007;195(5):698–702. PEPFAR. Next generation indicators reference guide addendum: recommended indicators reference sheets. President’s Emergency Plan For AIDS Relief (PEPFAR); 2009.
b c
d PEPFAR. Technical guidance on combination HIV prevention. President’s Emergency Plan For AIDS Relief (PEPFAR); 2011.
496
INDICATORS
Received lubricants Percentage who received packets of lubricant through an outreach service, drop-in centre or sexual health clinic in the last 12 months Numerator Number of participants who received packets of lubricant through an outreach service, drop-in centre or sexual health clinic in the last 12 months Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who received packets of lubricant through an outreach service, drop-in centre or sexual health clinic in the last 12 months. Using condomcompatible lubricants with condoms helps to prevent condoms from breaking and slipping during vaginal and anal sex.a Condoms breaking or slipping can put users at higher risk of getting HIV and STIs.b A low percentage or decreasing trend in this indicator may suggest low access to services and lack of knowledge or acceptance of lubricant use. In such cases, countries may consider expanding access to and distribution of condom-compatible lubricants through outreach or other services.
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 19: Lubricant questionnaire use. item(s) (if applicable) Numerator: • LUFREE = 1 (yes) [In the last 12 months, have you been given “packets” of lubricant for free? For example, through an outreach service, drop-in centre or health clinic.] Denominator: • Total number of respondents, except OVP Related existing indicators Not applicable
OVP, other vulnerable populations; STI, sexually transmitted infection
a b
Gabbay M, Gibbs A. Does additional lubrication reduce condom failure? Contraception. 1996;53(3):155–158.
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016).
SECTION 3: INDICATORS
497
HIV knowledge and perceptions Percentage who both correctly identify ways of preventing the sexual transmission of HIV and who reject major misconceptions about HIV transmission Numerator Number of participants who correctly identify ways of preventing the sexual transmission of HIV and who reject major misconceptions about HIV transmission Total number of participants SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who correctly identify ways of preventing the sexual transmission of HIV and who reject major misconceptions about HIV transmission. Knowledge of HIV prevention may show the extent to which outreach or other informational campaigns and educational materials have reached KP. However, misconceptions may still persist, such as the belief that a healthy-looking person cannot be infected with HIV, which can result in unprotected sexual intercourse with infected partners.a Rejecting major misconceptions about HIV transmission is as important as correct knowledge.a A low percentage or decreasing trend in this indicator may suggest lack of exposure to comprehensive HIV education. In such cases, countries may consider educational campaigns and promotion of materials about prevention of the sexual transmission of HIV and common misconceptions. This indicator is constructed from responses to the following questions from Module 38: HIV knowledge and perceptions. Numerator: • HKONEPAR = 1 (yes) [Can the risk of HIV transmission be reduced by having sex with only one uninfected partner who has no other partners?] AND • HKCONUSE = 1 (yes) [Can a person reduce the risk of getting HIV by using a condom every time they have sex?] AND • HKHEALTH = 1 (yes) [Can a healthy-looking person have HIV?] AND • HKMOSBIT = 2 (no) [Can a person get HIV from mosquito bites?] AND • HKSRFOOD =2 (no) [Can a person get HIV by sharing food with someone who is infected?] Denominator: • Total number of participants OVP, other vulnerable populations
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding questionnaire item(s)
a
UNAIDS/UNICEF/WHO. Global AIDS response progress reporting 2014: construction of core indicators for monitoring the 2011 United Nations political declaration on HIV and AIDS. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS)/United Nations Children’s Fund (UNICEF)/World Health Organization (WHO); 2014 (http://files.unaids.org/en/media/unaids/contentassets/ documents/document/2014/GARPR_2014_guidelines_en.pdf, accessed 2 September 2016).
498
INDICATORS
Incarceration Percentage who were ever arrested because they are a member of a KP Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of participants who were ever arrested because they are a member of a KP Total number of participants SW, MSM, TG, PWID • Sex for SW, PWID (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who were ever arrested because they are a member of a KP. Social exclusion and legal marginalization are important factors that influence health outcomes for KP.a Imprisonment is related to increased risk of HIV infection and delayed stopping of injecting drug use.b A high percentage or increasing trend in incarceration may indicate increased discrimination among KP and low sensitization of police towards KP. In such cases, countries may consider expanding sensitization training for police or legalizing behaviours that define KP groups.
Corresponding This indicator is constructed from responses to the following question from Module 29: Shame, questionnaire stigma, harassment and discrimination. item(s) (if applicable) Numerator: • STGARST = 1 (yes) [Have you ever been arrested because you [insert behaviour]?] Denominator: • Total number of participants Related existing indicators Not applicable
a
Csete J, Cohen J. Health benefits of legal services for criminalized populations: the case of people who use drugs, sex workers and sexual and gender minorities. J Law, Med Ethics. 2010;38(4):816– 831. Nosyk B, Li L, Evans E, Huang D, Min J, Kerr T et al. Characterizing longitudinal health state transitions among heroin, cocaine, and methamphetamine users. Drug Alcohol Depen. 2014;140:69–77.
b
SECTION 3: INDICATORS
499
Health-care stigma Percentage who were ever treated unfairly or denied services by health-care providers because they are a member of a KP Numerator Number of participants who were ever treated unfairly or denied services by health-care providers because they are a member of a KP Total number of participants SW, MSM, TG, PWID • Sex for SW, PWID (male, female) • Age (<25 years, 25+ years) • HIV status (positive, negative, unknown) This indicator measures the percentage of participants who were ever treated unfairly or denied services by health-care providers because they are a member of a KP. Discrimination and stigma in health-care settings contributes to the exclusion of KP from needed services.a KP are consistently underserved, and low-service coverage remains an important factor for ongoing HIV transmission.a,b A low percentage or decreasing trend in this indicator may suggest more accepting environments where the target population feels comfortable seeking needed services without harassment, discrimination or abuse.a In situations where this is not the case, countries may consider active outreach outside clinical settings to reach KP, who are often stigmatized or hidden.c
Denominator Applicable KP Recommended disaggregation Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 29: Shame, questionnaire stigma, harassment and discrimination. item(s) (if applicable) Numerator: • STGHEAL = 1 (yes) [Have you ever been treated unfairly or denied health-care providers because you [insert behaviour]?] Denominator: • Total number of participants Disaggregation (HIV status): • CSCTRS = 1 (positive) [What was the result of your last (HIV) test?] • CSCTRS = 2 (negative) [What was the result of your last (HIV) test?] • CSCTRS = 3 (unclear/neither positive or negative) or 4 (did not receive result) [What was the result of your last (HIV) test?] Related existing indicators Not applicable
a
Beyrer C, Grady C, Bekker L-G, McIntyre J, Over M, des Jarlais D. A framework for ethical engagement with key populations in PEPFAR programs. President’s Emergency Plan For AIDS Relief (PEPFAR); 2013 (http://www.pepfar.gov/sab/210110.htm, accessed 14 September 2016). Wolfe D, Carrieri MP, Shepard D. Treatment and care for injecting drug users with HIV infection: a review of barriers and ways forward. Lancet. 2010;376(9738):355–366. Underhill K, Operario D, Mimiaga MJ, Skeer MR, Mayer KH. Implementation science of pre-exposure prophylaxis: preparing for public use. Curr HIV/AIDS Rep. 2010;7(4):210–219.
b c
500
INDICATORS
Discrimination and social exclusion Percentage who experienced discrimination or social exclusion in the last 12 months because they are a member of a KP Numerator Number of participants who experienced discrimination or social exclusion in the last 12 months because they are a member of a KP Total number of participants SW, MSM, TG, PWID • Sex for SW, PWID (male, female) • Age (<25 years, 25+ years) • HIV status (positive, negative, unknown) This indicator measures the percentage of participants who experienced discrimination or social exclusion in the last 12 months because they are a member of a KP. KP are often subjected to stigma, discrimination and negative attitudes related to their behaviour, especially if also living with HIV.a,b Stigma and discrimination reduce uptake of HIV services and disclosure of HIV status.c A low percentage or decreasing trend in this indicator may suggest more accepting environments for KP. In cases where discrimination and social exclusion remains high, countries may consider promoting antidiscrimination and protective policies for all KP.b
Denominator Applicable KP Recommended disaggregation
Purpose/ rationale
Corresponding This indicator is constructed from responses to the following questions from Module 29: Shame, questionnaire stigma, harassment and discrimination. item(s) (if applicable) Numerator: • STGFRND = 1 (yes for at least one of the following response options: partner/spouse, family, friends/acquaintances, health-care providers, other) [In the last 12 months, which of the following have treated you badly or excluded you because you [insert behaviour]?] Denominator: • Total number of participants Disaggregation (HIV status): • CSCTRS = 1 (positive) [What was the result of your last (HIV) test?] • CSCTRS = 2 (negative) [What was the result of your last (HIV) test?] • CSCTRS = 3 (unclear/neither positive or negative) or 4 (did not receive result) [What was the result of your last (HIV) test?] Related existing indicators Not applicable
a
Buller AM, Devries KM, Howard LM, Bacchus LJ. Associations between intimate partner violence and health among men who have sex with men: a systematic review and meta-analysis. PLoS Med. 2014;11(3):e1001609. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. (2014). Geneva: World Health Organization Karim QA, Meyer-Weitz A, Mboyi L, Carrara H, Mahlase G, Frohlich JA et al. The influence of AIDS stigma and discrimination and social cohesion on HIV testing and willingness to disclose HIV in rural KwaZulu-Natal, South Africa. Global Public Health. 2008;3(4):351–365.
b c
SECTION 3: INDICATORS
501
Violence Percentage who experienced violence in the last 12 months Numerator Denominator Applicable KP Recommended disaggregation Number of participants who experienced violence in the last 12 months Total number of participants SW, MSM, TG, PWID • • • • Sex for SW, PWID (male, female) Age (<25 years, 25+ years) Type of violence (physical violence, sexual violence, sexual coercion) Perceived reason (because of KP behaviour, not because of KP behaviour)
Purpose/ rationale
This indicator measures the percentage of participants who experienced violence in the last 12 months. KP are often targets of violence; criminalization of their behaviours helps create an environment in which violence against KP is tolerated.a KP may not report violence because their claims are usually dismissed or they may distrust or fear police.b Fear of physical violence is a barrier to providing services for KP and limits their use of existing services.a Sexual violence and sexual coercion compromise their ability to negotiate safer sexual behaviours and thus put them at greater risk of getting HIV and STI.bc Because violence is often an expression of stigma and discrimination, a high percentage or increasing trend may indicate a greater intolerance towards KP. In such cases, countries may consider implementing and enforcing antidiscrimination and protective laws to eliminate stigma, discrimination and violence against KP. Other measures may include providing timely access to postrape care, including emergency contraception, PEP, HBV immunization and psychosocial support.a This indicator is constructed from responses to the following questions from Module 30: Physical violence and Module 31: Sexual violence. Numerator: Physical violence [Module 25]: • PVYRFRQ = 2 (once), 3 (2–5 times), 4 (6–10 times) or 5 (more than 10 times) [In the last 12 months, how many times has anyone physically hurt you such as hit or choked you, or threatened you with a knife or other weapon?] OR Sexual violence [Module 26]: • SVCYRFRQ = 2 (once), 3 (2–5 times), 4 (6–10 times) or 5 (more than 10 times) [How many times has this (i.e. being physically forced to have sex) happened in the last 12 months?] OR • SVCYRFRQ = 2 (once), 3 (2–5 times), 4 (6–10 times) or 5 (more than 10 times) [In the last 12 months, how many times has someone tricked you, lied to you, or threatened you in order to make you have sex when you didn’t want to?] Denominator: • Total number of participants
Corresponding questionnaire item(s)
502
INDICATORS
Disaggregation (type of violence): • Physical violence [Module 25]: PVYRFRQ = 2, 3, 4 or 5 [see numerator above for detail] • Sexual violence [Module 26]: SVCYRFRQ = 2, 3, 4 or 5 [see numerator above for detail] • Sexual coercion [Module 26]: SVCYRFRQ = 2, 3, 4 or 5 [see numerator above for detail] Disaggregation (perceived reason): • Because of KP behaviour • Physical violence [Module 25]: PVYRBEH = 1 (yes) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] • Sexual violence [Module 26]: SVCYRBEH = 1 (yes) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] • Sexual coercion [Module 26]: SVCYRFRQ = 1 (yes) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] • Not because of KP behaviour • Physical violence [Module 25]: PVYRBEH = 2 (no) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] • Sexual violence [Module 26]: SVCYRBEH = 2 (no) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] • Sexual coercion [Module 26]: SVCYRFRQ = 2 (no) [In the last 12 months when these happened, do you think it was because you [insert behaviour]?] Related existing indicators • GARPR 7.1: Proportion of ever-married or partnered women aged 15–49 who experienced physical or sexual violence from a male intimate partner in the past 12 months
HBV, hepatitis B virus; OVP, other vulnerable populations; PEP, post-exposure prophylaxis; PWID, people who inject drugs; STI, sexually transmitted infection
a
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016).
b
Betron M, Gonzalez-Figueroa E. Gender identity, violence and HIV among MSM and TG: a literature review and a call for screening. Washington, DC: Futures Group International, US Agency for International Development Health Policy Initiative; 2009. Decker MR, Pearson E, Illangasekare SL, Clark E, Sherman SG. Violence against women in sex work and HIV risk implications differ qualitatively by perpetrator. BMC Public Health. 2013;13(1):876.
c
SECTION 3: INDICATORS
503
Condom use at last sex (sex workers only) Percentage of SW who used a condom at last sex Numerator Denominator Applicable KP Recommended disaggregation Purpose/ rationale Number of SW who used a condom at last sex Total number of SW SW • Age (<25 years, 25+ years) • Client type (regular, one-time) This indicator measures the percentage of SW who used a condom at last sex, and captures any differences in condom use between regular and one-time clients. Correct and consistent use of condoms can reduce sexual transmission of HIV and other STI in both vaginal and anal sex by up to 94%.a However, the inability of SW to refuse sex without a condom or to successfully negotiate condom use with unwilling clients can be a barrier to condom use.b A low or decreasing trend in this indicator may indicate a lack of knowledge about the modes of HIV transmission, low perceptions of personal risk, lack of access to condoms, low capacity for condom use and negotiation, low perceptions of condom effectiveness or perceived reduced sexual sensation.c In such cases, countries may consider increasing the availability, accessibility, affordability and use of male and female condoms and condom-compatible lubricants among SW through targeted distribution services. This indicator is constructed from responses to the following questions from Module 9: Sex work characteristics. Numerator with disaggregation: • SWCLNRC = 1 (yes) [The last time you had sex with a regular client, did you use a condom?] • SWCLNOC = 1 (yes) [The last time you had sex with a one-time client, did you use a condom?] Denominator: • Total number of SW Related existing indicators • GARPR 1.8: Percentage of sex workers reporting the use of a condom with their most recent client • GF HIV O-5: Percentage of sex workers reporting the use of a condom with their most recent client
Corresponding questionnaire item(s)
STI, sexually transmitted infection
a
WHO. Consolidated guidelines on HIV prevention, diagnosis, treatment, and care for key populations. Geneva: World Health Organization (WHO); 2014 (http://www.who.int/hiv/pub/guidelines/ keypopulations/en/, accessed 8 September 2016). Bharat S, Mahapatra B, Roy S, Saggurti N. Are female sex workers able to negotiate condom use with male clients? The case of mobile FSWs in four high HIV prevalence states of India. PLoS One. 2013;8(6):e68043. Sarkar NN. Barriers to condom use. Eur J Contracep Repr. 2008;13(2):114–122.
b
c
504
INDICATORS
BIOMARKER-BASED INDICATORS
Section
III
The following set of indicators is based on biomarkers. A subset of these indicators are constructed from both biomarker and self-reported data. Indicators derived from biomarkers complement data obtained from interviews, and provide a current objective description of disease status of the target population. They also contribute to an improved understanding of behavioural risk factors, determinants of disease, and the impact of prevention and service delivery interventions. Chapter A-6 provides additional information and considerations about biomarkers. Awareness of HIV-positive status Percentage of HIV-positive participants who are aware of their HIV-positive status Numerator Denominator Applicable KP Recommended disaggregation Description Number of HIV-positive participants who are aware of their HIV-positive status Number of participants who tested positive for HIV SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who are aware of their HIV-positive status. This means that they reported being HIV-positive during the interview and their HIV test was also positive. It can be used to assess the reach of HIV counselling and testing strategies and to estimate the percentage of KP with undiagnosed HIV infection. For PLHIV, knowledge of one’s HIV status is the entry point to HIV services, which prevents passing the infection to others and lowers HIVrelated illness and death. • The numerator is based on self-reported data from Module 27: Services uptake and includes only those who tested positive for HIV and reported an HIV-positive status. • CSCTRS = 1 (positive) [What was the result of your last test?] • The denominator is based on a biomarker to detect the presence of HIV infection. The denominator includes only those who test positive for HIV. • The indicator excludes participants who test negative for HIV, who were not tested for HIV, or who had an indeterminate test result. Related existing indicators Not applicable
Definitions/ considerations
OVP, other vulnerable populations; STI, sexually transmitted infection
SECTION 3: INDICATORS
505
Eligible but not on HIV treatment Percentage of treatment-eligible participants who are not on HIV treatment Numerator Number of HIV-positive participants who are eligible for treatment but report not being on HIV treatment or test negative for ARV metabolites Number of HIV-positive participants who are eligible for treatment (based on national or international guidelines) SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who meet treatment eligibility criteria per national guidelines based on CD4 count but who report not being on treatment currently. It can be used to estimate the unmet need for treatment. • Treatment eligibility should be based on national or international guidelines. • The numerator is based on self-reported treatment data from Module 27: Services uptake. • CSTRCURR = 2 (no) [Are you currently taking ARVs?] When feasible, the numerator should be based on biomarker results from ARV metabolite testing (see the HIV treatment coverage biomarker indicator). • The denominator is based on biomarkers to 1) detect the presence of HIV infection and 2) detect CD4 count to indicate ART eligibility. The denominator includes participants who test positive for HIV and meet national guidelines for treatment eligibility based on CD4 count. • The indicator excludes participants who were not tested for HIV or CD4 count or who had indeterminate results. Related existing indicators Not applicable
Denominator
Applicable KP Recommended disaggregation Description
Definitions/ considerations
ART, antiretroviral therapy; ARV, antiretroviral; OVP, other vulnerable populations
506
BIOMARKER-BASED INDICATORS
HIV treatment coverage Percentage of HIV-positive participants who are on HIV treatment Numerator Denominator Applicable KP Recommended disaggregation Description Number of HIV-positive participants who have detectable ARV metabolites Number of HIV-positive participants who were tested for ARV metabolites SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of HIV-positive participants currently on treatment. Results can be used to estimate treatment coverage among KP. They can also confirm treatment status through metabolite testing and be compared to responses obtained from the survey to calibrate selfreporting results. • • • • The indicator is based on ARV metabolite testing results. The numerator includes HIV-positive participants who have detectable ARV metabolites. The denominator includes all HIV-positive participants who were tested for ARV metabolites. The indicator excludes participants who were not tested for HIV or ARV metabolites or who had indeterminate results. • Note that in settings with PrEP programmes, it may be difficult to distinguish between individuals taking PrEP who are unaware of their seropositive status from those on treatment. Not applicable
Definitions/ considerations
Related existing indicators
ARV, antiretroviral; OVP, other vulnerable populations; PrEP, pre-exposure prophylaxis
SECTION 3: INDICATORS
507
Viral load suppression Percentage of HIV-positive participants with a suppressed VL Numerator Denominator Applicable KP Recommended disaggregation Description Number of HIV-positive participants with a suppressed VL Number of HIV-positive participants who received a VL test SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • Treatment status (currently on HIV treatment, not currently on HIV treatment) This indicator measures the percentage of HIV-positive participants with a suppressed VL. It can be used to assess the success of treatment. • Suppressed VL is based on WHO’s guidelines for monitoring the response to ART, which recommend a VL below 1000 copies of viral RNA per ml as the threshold for very low risk of HIV transmission.a • The numerator is based on a biomarker to detect the level of viremia (amount of virus in the blood). The numerator includes participants who have a VL of <1000 copies/ml. • The denominator is based on biomarkers to detect the presence of HIV infection. The denominator should include participants who test positive for HIV. • Disaggregation by treatment status is based on self-reported data from Module 27: Services uptake. • CSTRCURR = 1 (yes) [Are you currently taking ARVs?] • CSTRCURR = 2 (no) [Are you currently taking ARVs?] • When feasible, however, the disaggregation should be based on results from ARV metabolite testing (see HIV treatment coverage biomarker indicator). • The indicator excludes participants who were not tested for HIV or VL or who had indeterminate results. Related existing indicators Not applicable
Definitions/ considerations
ART, antiretroviral therapy; ARV, antiretroviral; OVP, other vulnerable populations; RNA, ribonucleic acid; VL, viral load
a
WHO. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – Section 7.3: Monitoring response to ART and the diagnosis of treatment failure. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/hiv/pub/guidelines/arv2013/art/artmonitoring/en/index3.html, accessed 8 September 2016).
508
BIOMARKER-BASED INDICATORS
Unsuppressed viral load Percentage of HIV-positive participants with an unsuppressed VL Numerator Denominator Applicable KP Recommended disaggregation Number of participants with an unsuppressed VL Number of HIV-positive participants who received a VL test SW, MSM, TG, PWID, OVP • • • • Sex for SW, PWID, OVP (male, female) Age (<25 years, 25+ years) Sexual risk behaviour (protected sex, unprotected sex) Drug use behaviour (shared injecting equipment, did not share injecting equipment)
Description
This indicator measures the percentage of HIV-positive participants with unsuppressed VL. It can be used to assess the potential for onward transmission and spread of the epidemic in the target population. • Unsuppressed VL is based on WHO’s guidelines for monitoring the response to ART, which indicate a VL above 1000 copies of viral RNA per ml as a sign of treatment failure.a • The numerator is based on a biomarker to detect the level of viremia (amount of virus in the blood). The numerator includes participants who have a VL of >1000 copies/ml. • The denominator is based on a biomarker to detect the presence of HIV infection. The denominator includes participants who test positive for HIV. • The disaggregation by sexual risk behaviour is based on self-reported data from Module 17: Last sex act: • LSACON = 1 (yes, for protected sex) [Did you use a condom the last time you had sex?] • LSACON = 2 (no, for unprotected sex) [Did you use a condom the last time you had sex?] • The disaggregation by drug use behaviour is based on self-reported data from Module 23: Injecting drug use: • IDLSTND = 1 (yes, for shared injecting equipment) [The last time you injected with another person, did you share a needle with that person (i.e. use it after they had used it)?] • IDLSTND = 2 (no, for did not share injecting equipment) [The last time you injected with another person, did you share a needle with that person (i.e. use it after they had used it)?] • This indicator excludes participants who were not tested for HIV or VL or who had indeterminate results.
Definitions/ considerations
Related existing indicators
Not applicable
ART, antiretroviral therapy; OVP, other vulnerable populations; RNA, ribonucleic acid; STI, sexually transmitted infection; VL, viral load
a
WHO. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – Section 7.3: Monitoring response to ART and the diagnosis of treatment failure. Geneva: World Health Organization (WHO); 2013 (http://www.who.int/hiv/pub/guidelines/arv2013/art/artmonitoring/en/index3.html, accessed 8 September 2016).
SECTION 3: INDICATORS
509
Population viral load Median VL of HIV-positive participants Numerator Denominator Applicable KP Recommended disaggregation Description Median VL of HIV-positive participants Not applicable SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) Population VL is an aggregate biological measure of VL for a particular geographical location or population. This indicator can be used to monitor the uptake and effectiveness of ART and to assess the potential for ongoing HIV transmission within KP. As HIV treatment coverage increases, the number of HIV-positive individuals who are virologically suppressed increases, leading to reduced population VL and, eventually, a reduction in HIV incidence. • This indicator captures the median VL of HIV-positive participants. It is based on biomarkers to detect the presence of HIV infection and measure VL. It is expressed as the median VL of all participants who test positive for HIV. • Median VL is more useful than mean VL, because VL distribution is often skewed and just a few very high values will skew the mean upward. • This indicator excludes participants who did not get tested for HIV or VL or who had an indeterminate result for either of these tests. • Note that if treatment coverage is high, and more than 50% of the target population has undetectable VL, then the median will be “undetectable” and survey investigators will need to use different methods to report VL, such as log transformation.a Not applicable
Definitions/ considerations
Related existing indicators
ART, antiretroviral therapy; OVP, other vulnerable populations; VL, viral load
a CDC. Guidance on community viral load: a family of measures, definitions, and method for calculation. Atlanta (GA), Atlanta, GA: Centers for Disease Control and Prevention (CDC); 2011 (https:// stacks.cdc.gov/view/cdc/28147, accessed 9 September 2016).
510
BIOMARKER-BASED INDICATORS
HIV/sexually transmitted infection coinfection Percentage coinfected with HIV and at least one STI Numerator Denominator Applicable KP Recommended disaggregation Description Number of participants coinfected with HIV and at least one STI Number of participants who were tested for HIV and at least one STI SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • STI (e.g. syphilis, gonorrhoea, chlamydia) This indicator measures the percentage of participants who are coinfected with HIV and at least one STI. HIV/STI coinfection increases the chance that HIV will be transmitted to a sexual partner.a It can be used to plan services and allocate resources for treating HIV-positive KP who are coinfected with STI.
Definitions/ considerations
• This indicator captures the prevalence of coinfection (i.e. the percentage of the target population living with HIV and at least one STI at a given time, regardless of the time of infection), whether the person has received a diagnosis (i.e. is aware of infection), or the stage of disease. • The indicator is based on biomarkers to detect the presence of HIV infection and selected STIs. STIs should be selected based on each country’s epidemiology. • The numerator includes participants who tested positive for HIV and at least one STI. • The denominator includes participants who were tested for HIV and at least one STI. • The indicator excludes participants who did not get tested for HIV and at least one STI, had an indeterminate HIV test result, or had indeterminate test results for all STI tests received. Not applicable
Related existing indicators
OVP, other vulnerable populations; STI, sexually transmitted infection
a
Kalichman SC, Pellowski J, Turner C. Prevalence of sexually transmitted co-infections in people living with HIV/AIDS: systematic review with implications for using HIV treatments for prevention. Sex Transm Infect. 2011;87(3):183–190.
SECTION 3: INDICATORS
511
HIV/hepatitis coinfection Percentage coinfected with HIV and HBV or HBC Numerator Denominator Applicable KP Recommended disaggregation Description Number of participants coinfected with HIV and HBV or HCV Number who were tested for HIV and HBV or HCV SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • Viral hepatitis type (HBV, HCV) This indicator measures the percentage of participants who are coinfected with HIV and HBV or HCV. HIV-positive individuals who become infected with HBV are at increased risk for developing chronic HBV infection and can have serious medical problems, including an increased risk for liver-related illness and death.a HCV infection progresses more rapidly to liver damage in HIV-positive persons and may affect the course and management of HIV infection.a This indicator can be used to plan services and allocate resources for treating HIV-positive KP who are coinfected with HBV or HCV. • This indicator captures the prevalence of coinfection (i.e. the percentage of the target population living with HIV and HBV or HCV at a given time, regardless of the time of infection), whether the person has received a diagnosis (i.e. is aware of infection), or the stage of disease. • The indicator is based on biomarkers to detect the presence of HIV, HBV and HCV infection. • The numerator includes respondents who tested positive for HIV and HBV or HCV. • The denominator includes all participants who were tested for HIV and HBV or HCV. • The indicator excludes participants with missing or indeterminate test results for HIV, HBV or HCV. Not applicable
Definitions/ considerations
Related existing indicators
HBV, hepatitis B virus; HCV, hepatitis C virus; OVP, other vulnerable populations
a
CDC. HIV and viral hepatitis. Atlanta, GA: Centers for Disease Control and Prevention (CDC); 2016 (http:// www.cdc.gov/hiv/pdf/library/factsheets/hiv-viral-hepatitis.pdf, accessed 8 December 2016).
512
BIOMARKER-BASED INDICATORS
HIV prevalence Percentage with HIV Numerator Denominator Applicable KP Recommended disaggregation Description Number of participants who tested positive for HIV Number of participants who were tested for HIV SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) This indicator measures the percentage of participants who tested positive for HIV. This indicator helps countries to assess progress in controlling HIV and is useful for planning and resource allocation as it reflects the number of people currently needing care for HIV infection. • This indicator captures HIV prevalence, the percentage of the target population living with HIV at a given time regardless of the time of infection, whether the person has received a diagnosis (i.e. is aware of infection), or the stage of HIV. • This indicator is based on a biomarker for HIV antibodies to detect the presence of HIV infection. • The numerator includes participants who tested positive for HIV. • The denominator includes participants who were tested for HIV. • The indicator excludes participants who did not get tested for HIV or who had an indeterminate test result. • • • • • • • • GARPR 1.10: HIV prevalence in sex workers GARPR 1.14: HIV prevalence in men who have sex with men GARPR 2.5: HIV prevalence in people who inject drugs GF HIV I-9a: Percentage of men who have sex with men who are living with HIV GF HIV I-9b: Percentage of transgender people who are living with HIV GF HIV I-10: Percentage of sex workers who are living with HIV GF HIV I-11: Percentage of people who inject drugs who are living with HIV GF HIV I-12: Percentage of other vulnerable populations who are living with HIV
Definitions/ considerations
Related existing indicators
OVP, other vulnerable populations
SECTION 3: INDICATORS
513
Sexually transmitted infection prevalence Percentage with any STI Numerator Denominator Applicable KP Recommended disaggregation Description Number of participants who tested positive for any STI Number of participants who were tested for at least one STI SW, MSM, TG, PWID, OVP • Sex for SW, PWID, OVP (male, female) • Age (<25 years, 25+ years) • STI (e.g. syphilis, gonorrhoea, chlamydia) This indicator measures the percentage of participants tested for at least one STI who had a positive STI test. Risk behaviours associated with STIs are similar to those for HIV, and some STIs may lead to increased risk of getting or passing on HIV. This indicator can help detect the potential emergence of HIV infection and highlights specific STIs that require public health attention. • This indicator captures STI prevalence, the percentage of the target population living with an STI at a given time regardless of the time of infection, or whether the person has received a diagnosis (i.e. is aware of infection). • This indicator is based on biomarkers to detect selected STIs of interest. STIs should be selected based on each country’s epidemiology. • The numerator includes all respondents who tested positive for an STI. • The denominator includes all participants who were tested for an STI. • The indicator excludes participants who were not tested for an STI or who had an indeterminate test result. • • • • GARPR 1.17.4: Percentage of sex workers with active syphilis GARPR 1.17.5: Percentage of men who have sex with men with active syphilis GF HIV I-3b: Percentage of men who have sex with men with active syphilis GF HIV I-3c: Percentage of sex workers with active syphilis
Definitions/ considerations
Related existing indicators
OVP, other vulnerable populations; STI, sexually transmitted infection
514
BIOMARKER-BASED INDICATORS
Supplemental Materials
Section
I
APPENDICES
Section
II
QUESTIONNAIRE MODULES
Section
III
INDICATORS
Section
IV
GLOSSARY
SECTION I: APPENDICES
515
Term Accuracy
Definition Describes how close an estimate is to the true value. Applicable to survey estimates or laboratory tests. A well-defined step-by-step procedure (e.g. describing the order of laboratory tests to be used). Describes something having no known name or identity. Achieved by removing all personal identifying information from an interview record or biological specimen. A substance to prevent blood from clotting (i.e. to prevent it from coagulating). Used as an additive in blood collection tubes. Drugs used to fight infections caused by retroviruses such as HIV. Treatment with antiretroviral drugs to inhibit the ability of HIV to multiply in the body; suppresses (reduces) the amount of virus (viral load) in the blood.
Algorithm
Anonymous
Anticoagulant
Antiretroviral drugs Antiretroviral therapy
Approach criteria
In time-location sampling, the criteria that staff use to approach particular venue attendees for survey participation.
Assent
An agreement by persons who cannot legally agree (consent) to survey participation; for example, minors (usually, consent by parents or caregivers is necessary in addition to assent by a minor potential survey participant).
Autonomy, autonomous
Freedom from external control or influence; independence. Autonomous individuals act intentionally, with understanding, and without external controlling influences. The effort to secure the well-being of survey participants, both by avoiding or minimizing harm, and by maximizing possible benefits. Systematic error in a survey’s design, data collection, or analysis resulting in measurement errors. Causes of bias include over or under-sampling of subpopulations (e.g., proportionately more younger people in the survey than in the larger population), inaccurate tests (e.g., poor specificity), programming errors in electronic questionnaires (e.g., skip pattern incorrectly coded resulting in missing information), or untruthful or inaccurate answers provided by participants (e.g., denying ever having anal sex). The data derived from testing biological specimens.
Beneficence
Bias
Biological data or “biodata” Biological specimen
Biological material (e.g. blood, urine or oral fluids) collected from a survey participant.
516
GLOSSARY
Term Biomarker
Definition A biological characteristic measured through laboratory or other tests, generally using a biological specimen (e.g. blood) for testing. Examples of biomarkers are HIV antibody or viral load. The containment principles, technologies and practices to prevent unintentional exposure to pathogens and toxins, or their accidental release.
Biosafety
Biosecurity
Protection, control and accountability for valuable biological materials in laboratories to prevent their unauthorized access, loss, theft, misuse, diversion or intentional release. Quantifies the relationship between two variables. Typically, one of these is considered a “dependent” (outcome) variable and the other an “independent” (explanatory) variable (e.g. a characteristic, behaviour or exposure).
Bivariate analysis
Calibrate
Setting a device or instrument to a base level by matching it with a set value. Generally, laboratory testing equipment should be calibrated before use.
Capillary blood
Blood collected from the capillaries by finger prick; typically used for rapid tests or to collect dried blood spots. A blood test that measures the number of a subset of T-lymphocytes, the CD4+ cells. CD4+ T-cell counts can be used to indicate the strength of the immune system, with normal levels ranging from 500 to 1200 (values <500 indicate a weakened immune system). See also Snowball sampling. A sampling technique in which enrolled survey participants refer or recruit potential participants from among their acquaintances (peers). See Haemophilus ducreyi. A list of steps to follow when conducting a biobehavioural survey. A bacterial species that can be transmitted sexually and causes a wide range of diseases including urethritis, proctitis (rectal disease and bleeding), trachoma (conjunctivitis) and infertility. In men, it causes prostatitis and epididymitis; in women, it can cause cervicitis, pelvic inflammatory disease, ectopic pregnancy, and acute or chronic pelvic pain.
CD4+ T-cell count
Chain-referral sampling
Chancroid Checklist Chlamydia trachomatis, chlamydia infection
Cluster location
Defined venue where, in cluster sampling, an individual can be found for sampling purposes. The process of knowing, perceiving or understanding through the use of reasoning, intuition or perception. “Cognitive testing” relates to the testing of survey questions to investigate how well potential participants correctly understand survey questions and response categories.
Cognitive, cognitive testing
SECTION 4: GLOSSARY
517
Term Cohort study
Definition An epidemiological study that follows a group of people over time to observe which and how many individuals experience certain outcomes (e.g. HIV seroconversion). Also known as a longitudinal study.
Community liaison
A person acting as a representative of a population (e.g. sex workers) or acting as a link to a target population to facilitate survey planning, implementation, interpretation and dissemination of results. An organization largely composed of members from a given population (e.g. men who have sex with men), often working in advocacy or providing services.
Community-based organization
Compensation
Also known as reimbursement. The payment of money or item given for a participant’s effort to participate in the survey. Compensation may be made for travel costs, time spent, lost wages or other expenses incurred for participating in the survey. In respondent-driven sampling surveys, two types of compensation are known: primary (given at initial visit) and secondary (given at the follow-up visit, which includes compensation for the participant’s peer recruitment efforts). A range of values that is likely to include the true value. A 95% confidence interval includes a range (interval) wide enough to include the true point estimate in 19 of 20 surveys. The expectation or promise that information will be protected from disclosure to others. A factor associated with both the outcome of interest (disease) and exposure of interest (cause of disease), but not in the causal pathway of exposure to outcome. Permission given by an individual to proceed with a specific test or procedure, with an understanding of the risks, benefits, limitations and potential implications of the procedure itself and its results. The selection of individuals from a population based on accessibility and availability. Convenience samples are nonprobability samples and their findings are only applicable to those surveyed. A survey staff person who counts venue attendees, often with a tally counter.
Confidence interval
Confidentiality, confidential Confounder
Consent
Convenience sampling
Counter
Coupon manager
A study staff member who checks coupons for validity, enters coupon information into coupon manager software, trains participants on the recruitment process, and gives participants recruitment coupons and compensation.
518
GLOSSARY
Term Cross-recruitment
Definition In respondent-driven sampling surveys, a situation in which a participant who enrols at one site recruits a peer who enrols at another site. Observation of cross-recruitment may be used as a proxy for linked networks (i.e. networks of peers in different parts of a city that maintain social ties between them).
Data confidentiality Data instrument, data collection instrument
The protection of data against unintended or unauthorized access. Also known as a survey instrument or survey tool, the questionnaire used to collect data in a standardized manner, whether on paper or through computers.
Data quality Data redundancy
The completeness, validity, consistency and accuracy of data. Situation in which a data value can be determined from the values of other data items (variables). An activity through which the collected data are evaluated for correctness, documenting the type of error (e.g. missing values, logical errors or out-of-range values). In qualitative research, the point at which no new useful information emerges.
Data review
Data saturation
Data security
The requirements (technical and physical) that keep data secure and confidential. A chemical agent that absorbs moisture; a drying agent. Used to maintain integrity or quality of biological specimens during transport to a laboratory for testing or storage.
Desiccant
Design effect
A factor expressing how much larger a sample size for a complex survey design must be compared to simple random sampling. Blood collected on special filter paper; sample may be capillary blood (finger prick) or whole (venous) blood. Blood collected via dried blood spot is more stable at ambient temperatures than blood stored in tubes.
Dried blood spot
Drop-in centre
A safe physical venue that provides health, educational or counselling services to a particular population (e.g. people who inject drugs).
SECTION 4: GLOSSARY
519
Term Eligibility criteria
Definition The description of all the requirements a survey candidate must meet in order to participate. Most often conducted by asking potential participants a series of questions, or sometimes undertaking (laboratory) tests, to determine eligibility. A laboratory test that uses an enzyme-bound antibody to detect antigen or antibody. The enzyme catalyzes a colour reaction when exposed to substrate. For example, enzyme immunoassays are used as tests for the presence of HIV (antibody) in a blood specimen.
Eligibility screening
Enzyme immunoassay
Equal probability (of selection into a sample) Equilibrium
Also equal probability sampling. A method of sample selection in which all potential participants in the sampling frame have the same chance of selection.
In respondent driven sampling, the point at which the composition of the sample is stable and independent of the nonrandomly selected seeds.
Exposure
The condition of (potentially) being in contact with something, such as an infectious agent or an intervention. An evaluation of the testing process by an outside organization (e.g. a national reference laboratory).
External quality assessment External validity
Also generalizability. External validity involves the extent to which the results of a study can be generalized (applied) beyond the sample. Semistructured group conversations conducted with usually 6–8 members of the target population under the guidance of a facilitator, using an interview guide. Focus group participants are convened to confirm or refute observations made in key informant interviews and to gather additional information to facilitate survey preparation and implementation.
Focus groups
Formative assessment
The collection of initial information about a population to inform how best to prepare and conduct a survey of that population. Asking respondents to list as many items that fit in a particular domain (e.g. all the different drugs that are taken by local drug users). See Herpes simplex virus-2. A laboratory process that determines the DNA sequences of an organism’s genome (i.e. the entire genetic information of an organism).
Free listing
Genital herpes Genomic sequencing
520
GLOSSARY
Term Genotyping
Definition The process of determining differences in the genetic make-up (genotype) of an individual or infectious agent by examining its DNA sequence with biological assays and comparing it to a reference sequence. Genotyping is used for molecular epidemiology or to detect drug resistance.
Gonorrhoea Gram stain
See Neisseria gonorrhoea. A laboratory method used to classify bacteria into two large groups (negative or positive); frequently used to aid in identifying the cause of infection.
Haemophilus ducreyi, chancroid
The bacterial causative agent of the sexually transmitted disease known as chancroid.
Herpes simplex virus-2, genital herpes
The primary cause of genital and neonatal herpes and genital ulcer disease (genital herpes).
High-performance liquid chromatography combined with mass spectrophotometry
A laboratory technique used to detect drugs or their metabolites (e.g. antiretroviral drugs).
High-risk behaviours
Behaviours that increase the risk of a person contracting a disease.
HIV drug resistance Homogeneity
A mutated form of HIV that reduces the effectiveness of one or more antiretroviral drugs. Often used in respondent driven sampling surveys; the degree to which people in a population have similar or identical characteristics.
Homophily
The tendency of individuals to associate (mix) with others who are similar or dissimilar to themselves; calculated by respondent driven sampling software packages.
Immunochromatographic Immunosuppression
A one-step test characteristic that enables rapid identification of various analytes (e.g. HIV antibodies) at the point of care. Reduces an individual’s ability to fight infection or cancer. HIV is one of the causes of immunosuppression.
SECTION 4: GLOSSARY
521
Term Indicator reporting Inference Informed consent
Definition The use of aggregate data for standardized reporting to national governments, United Nations agencies or donors. See Statistical inference. The provision of adequate information to potential survey participants to allow them to actively decide to consent (agree) to a survey or procedure (e.g. biological specimen). It protects potential participants’ freedom of choice and respects their autonomy with regard to decisions affecting their body and health. A condition that occurs when answers to different questions do not contradict each other. The extent to which a study has minimized systematic error or bias. The fair selection of research participants and the fair distribution of risks and benefits for the study participants. People who have first-hand knowledge of the target population or subject matter. In the context of HIV, populations distinguished by behaviours associated with higher levels of acquiring or transmitting HIV compared to the general population. Key populations commonly include female sex workers, men who have sex with men, transgender persons, and persons who inject drugs. A boundary placed on the difference between an estimate of something and the true value, reflecting the amount of sampling error in a survey’s result.
Internal consistency
Internal validity Justice
Key informants
Key populations
Margin of error
Molecular testing
Testing for nucleic acid sequences (e.g. DNA or RNA), as opposed to serologic or immunologic testing for antigens or antibodies. Causative bacterial agent for tuberculosis. The bacteria usually attack the lungs (referred to as pulmonary tuberculosis), but may attack any part of the body.
Mycobacterium tuberculos
Neisseria gonorrhoea, gonorrhoea Nonprobability sample
Gram-negative bacteria causing gonorrhoea, a sexually transmitted disease that affects the genital area and urinary tract, causing dysuria and penile discharge in men, and vaginal discharge or pelvic pain in women. Women often may be asymptomatic. A method of sampling in which the probability of sampling is unknown. Nonprobability samples generate estimates of unknown precision and reliability. Examples include convenience sampling, snowball sampling and quota sampling.
522
GLOSSARY
Term Nonresponse
Definition A participation category referring to individuals who are approached for interview in a study but choose not to participate. Tests that use nonspecific antigens to screen blood for syphilis. Examples include the rapid plasma reagin test or the venereal disease reference laboratory test. A laboratory test based on the amplification of specific genetic sequences through polymerase chain reaction or other methods. The amplification step allows these tests to be more sensitive than conventional methods; also, they may be used to detect sexually transmitted infections such as Neisseria gonorrhoea or Chlamydia trachomatis.
Nontreponemal tests
Nucleic acid amplification test
Odds ratio Optical density
A measure of association between an exposure and an outcome. Represents the intensity of colour (or absorbance) as measured by a spectrophotometer. In enzyme immunoassays, more intense colour (i.e. higher optical density) generally indicates more antibody (or antigen).
Outcome
The measure of interest or end-point in clinical trials or in data analysis, or the result or consequence of a public health intervention. Examples of outcomes include frequency of condom use or HIV status. Values that are extremely high or low compared to the mean or median values of a variable. Use of one population to recruit another related population (e.g. sex workers in a survey being asked to refer their clients for survey participation). Information that can be used to identify an individual (e.g. names, addresses, email addresses, phone numbers, identification numbers, photographs or date of birth). Asking the respondent to group items from the same domain, often using cards (e.g. sorting the locally used drugs [names] that are either stimulants or depressants). The fluid (liquid) portion of blood. Plasma can be obtained from the blood by removing all cellular components through centrifugation or gravity, and inhibiting its clotting capacity with anticoagulants.
Outliers
Partner survey
Personally identifiable information
Pile sorting
Plasma
Point-of-care technology
A technology that allows tests to be used at or near the point of specimen collection (i.e. without using a laboratory). Examples include HIV rapid tests and portable CD4 machines.
SECTION 4: GLOSSARY
523
Term Post-exposure prophylaxis Power Precision
Definition The use of antiretroviral medications to prevent HIV infection after an exposure has occurred (e.g. after a needle stick injury or a rape). The probability of not making a type II error (i.e. accepting a null hypothesis that is false). In surveys and estimates, refers to how close an estimate is to other estimates made using the same methodology (i.e. when a method [survey] is repeated, the degree to which a result is consistent). A precise estimate may not necessarily be an accurate estimate. A proportion (not a rate) describing the fraction in a population with a certain characteristic (e.g. HIV infection). See Compensation. The individual responsible and accountable for conducting a survey or study. The principal investigator assumes full responsibility for the ethical treatment and evaluation of human subjects, and for the integrity of the survey’s or study’s conduct and data. A sampling method used with surveys in which the probability of selecting a sampling unit (e.g. village, school, prison, brothel) is proportional to the size of the sampling unit. With PPS, larger sampling units have a higher probability of being sampled than smaller ones. PPS facilitates a probability (i.e. representative) sample. In sampling, methods in which the probability of a respondent being selected for the sample is known. The known sampling probability makes it possible to determine sampling weights and hence facilitates determination of adjusted (population) estimates. Request by the interviewer for more information or a method used during an interview to help a participant recall information. The testing of unknown samples sent to a laboratory by a reference laboratory. The test results are then reviewed to judge the laboratory’s testing quality.
Prevalence Primary compensation Principal investigator
Probability proportional to size (PPS)
Probability sample
Probe
Proficiency testing
Prostate-specific antigen
A protein produced by prostate cells. In surveys, prostate-specific antigen is sometimes measured in vaginal or rectal swabs, where its presence may indicate recent unprotected sex.
Proximal determinant Purposive sampling
A biological or behavioural variable that directly facilitates HIV acquisition or transmission (e.g. condom use). Also known as judgmental sampling. A method of sampling used to select individuals with specific characteristics (e.g. sex workers who take antiretroviral treatment).
524
GLOSSARY
Term P-value Quality assurance
Definition The probability that an observed result occurred by chance.
A system of procedures in place to prevent mistakes. The concept of quality assurance in biomarker measurements covers all steps from specimen collection to test result interpretation. A type of purposive sampling that specifies how many people with each characteristic will be sampled; usually done to ensure that diverse segments of the population are sampled. Biological assays that can be performed easily, even by lay people with minimal training, outside a laboratory, producing results in a short time.
Quota sampling
Rapid tests
Recall bias
Systematic error due to differences in accuracy or completeness of remembered past events. May lead to underestimates or overestimates of reported past events.
Recruitment area
In the context of time-location sampling, a well-defined space inside or outside a venue where potential participants who enter or cross the space are approached for recruitment. See Data redundancy. The degree to which a technique or approach provides consistent results if repeated multiple times.
Redundancy Reliability
Repository
A place where things are stored (e.g. a specimen repository or a data repository). Also called participant. Someone who responds to a question, survey or offer.
Respondent
Respondent-driven sampling
Combines “snowball sampling” (in which individuals refer those they know, who then in turn refer those they know, and so on) with a mathematical model that weights the sample to compensate for the fact that the sample was collected in a nonrandom way.
Risk
In ethical terms, the exposure to injury, loss or harm, expressed in terms of the probability and magnitude of that harm. Risks to subjects must be minimized and must be reasonable in relation to anticipated benefits to subjects and the importance of the expected knowledge.
SECTION 4: GLOSSARY
525
Term Risk factor
Definition A behavioural, biological or environmental factor associated with a higher likelihood of acquiring or transmitting a disease. An exercise in which individuals (often other survey team members) pretend to be survey participants and interact with staff members who administer survey procedures. The degree with which a (participant’s) characteristic stands out or is relevant compared with other characteristics. Salient characteristics may include age and sex. The number of individuals (to be) enrolled in a survey. Some population members being systematically more or less likely to be sampled than others, leading to biased estimates. The error associated with nonobservation (i.e. with the variation in samples drawn from the same source population). The sampling error equals the square root of the variance. In cluster sampling, the time period where field staff are at a given venue recruiting and interviewing respondents. The target sample size divided by the number of people in the target population. If the sampling fraction is small enough, sampling with replacement is approximated. If the sampling fraction is too large, it becomes harder for recruits in respondent driven sampling surveys to find peers who have not yet participated. The source material or list from which a sample is drawn in order to represent the population. Describes how individuals will be selected and determines whether and how statistical inferences can be made from the sample to the source population.
Role play
Saliency, salient
Sample size Sampling bias
Sampling error
Sampling event
Sampling fraction
Sampling frame
Sampling strategy
Sampling stratification Secondary compensation Seed
See Stratified sampling.
See Compensation.
In respondent driven sampling surveys, the individuals who are chosen by the investigators and represent the start of the recruitment chains.
Selection bias
A systematic error in the process of selecting participants for a study or survey.
526
GLOSSARY
Term Semistructured interview
Definition Used in qualitative research. An interview where the interviewer uses an interview guide with themes or open-ended questions (grouped by topic areas) to conduct an in-depth, open-ended conversation with the respondent(s).
Serum
Fluid that rises to the top when coagulated (clotted) blood has been centrifuged or allowed to separate. The main difference between serum and plasma is that serum contains fewer proteins such as clotting factors. A sampling method whereby (eligible) individuals are chosen at random from the sampling frame, so that each individual has equal probability of being selected. Requires the use of a random number table or other method (e.g. computer-based) to generate random numbers that identify the persons to be included in the sample. See also Chain-referral sampling. A sample that is built by asking participants to provide the contact details of someone they know that the researcher can contact, or where survey participants refer others to the survey to enrol.
Simple random sampling
Snowball sampling
Social network
A social structure made up of a network of friends, colleagues and other personal contacts who know one another and who share certain defining characteristics that identify them as members of the target population.
Social structure
The pattern of social relationships in a society. Such structure regulates the interactions among members of the society or community.
Stakeholder
A party or group of people with an interest in, or who can contribute to or are affected by a given activity. A measure of the variability of a statistic; the standard deviation of the sampling distribution of a statistic. Although the standard error is used to estimate sampling error, it includes some nonsampling error.
Standard error
Standard operating procedures Statistical inference
A set of written instructions for distinct survey activities.
The process of using information from a sample to make inferences, or conclusions, about the population from which the sample was taken.
Statistical significance
Indicates that a certain observation (e.g. older sex workers are more likely to be HIV infected than younger sex workers) is unlikely to be the result of chance. Statistical significance is attained when a statistical test yields a p-value below a set (arbitrary) threshold. A p-value of 0.05 is a commonly used threshold to indicate statistical significance; other thresholds used include 0.1 or 0.01. Statistical significance does not necessarily equate to medical importance because p-values are dependent on the size of the sample being studied.
SECTION 4: GLOSSARY
527
Term Stratification
Definition See also Sampling stratification. In data analysis, stratification refers to the examination of a certain outcome (e.g. condom use) after the dataset (sample) has been grouped by certain characteristics (e.g. by gender).
Stratified sampling
The process of dividing members of the population into homogeneous subgroups before sampling, to reduce variability or random sampling error and thereby improve the precision of the estimates produced by the sample.
Strata
Partitions of the sampling frame; generally defined to create relatively homogeneous units within each strata.
Surveillance
The systematic, ongoing collection, analysis, interpretation and dissemination of (health) data to monitor the pattern of disease occurrence and potential in a community, in order to control and prevent disease in the community.
Survey
A method for collecting data, drawing a sample from a target population; conducted over a defined period of time (sampling duration).
Survey population Survey weights
Also called target population. Refers to all units (e.g. individuals or households) to which one desires to generalize survey results. Numbers associated with respondents that specify the weight (i.e. influence) their values should have in analysis, in order to ensure that the sample is representative of the population of interest. Two or more epidemics or endemics occurring simultaneously and acting synergistically. See Treponema pallidum. See Sampling error.
Syndemic
Syphilis Systematic sampling error Target population
In the context of surveys, the population from which the survey sample is drawn and to which the investigators will want to infer. In the case of HIV, disease-prevention interventions that are focused on specific populations often at higher risk for HIV exposure (e.g. a peer education strategy to deliver behaviour change communication and commodities such as condoms, lubricants and needles or syringes).
Targeted interventions
Technical report
A report based on survey data, which usually presents relatively simple descriptive statistics, including numerators, denominators, and the prevalence and distribution of specific characteristics.
528
GLOSSARY
Term Time-location cluster (time-location sampling)
Definition A sampling method that recruits individuals (target population members) from specific locations (venues) during specific time periods. Time-location sampling surveys represent only those who frequent venues.
Transcription errors
Mistakes made when a person enters data from one type of documentation to another. A spirochete bacterium with subspecies that cause treponemal diseases such as syphilis, bejel, pinta and yaws. Syphilis, a sexually transmitted bacterial illness, can cause heart and brain damage, paralysis, blindness and possibly death.
Treponema pallidum
Treponemal tests Trichomonas vaginalis
Tests using antigens from Treponema pallidum to screen for syphilis. A protozoan parasite that can lead to infection (trichomoniasis) and can be sexually transmitted. Trichomoniasis is a common cause of vaginitis and increases a woman’s susceptibility to HIV infection. See Trichomonas vaginalis.
Trichomoniasis (“trich”) Tuberculosis Two-sided test (two-tailed test)
See Mycobacterium tuberculosis. A test used when a deviation (change) of the outcome of interest in either direction (up or down) is considered possible. For example, in a survey conducted before (baseline) and after an intervention in the target population to increase HIV testing uptake, a two-tailed test considers the possibility that the intervention may have led either to an increase or decrease in HIV testing uptake. Investigators should always use a two-sided test unless certain that change is only possible in one direction.
Unique participant code
A code given to each participant that does not contain personally identifying information, but is used to verify participant identity. The degree to which an estimate is likely to be true and free of bias (i.e. systematic errors). The dispersion of a variable’s values around its average (mean). The formula for calculating variance is the sum of the difference between the observed values and the mean value divided by the sample size.
Validity
Variance
Venipuncture
The puncture of a vein with a needle to collect blood (also referred to as phlebotomy) or to administer intravenous therapy.
SECTION 4: GLOSSARY
529
Term Venous blood
Definition The blood that flows towards the heart. Venous blood is the most common specimen type for blood-based laboratory tests. In terms of biobehavioural surveys, a place where the target population congregates (e.g. bars or brothels). The amount of virus in the blood. The (relative) powerlessness to resist or survive the effects of an adverse environment. Compared to the general population, vulnerable populations may be at higher risk for contracting HIV or face greater challenges to access services or cope with the consequences of HIV/AIDS because of biological, cultural, social, gender, legal or religious practices and beliefs, including war and poverty. In contrast, key populations are defined by their HIV-related risk behaviours. Examples of vulnerable populations include people with disabilities, homeless people, refugees and displaced people, and youth. The set of RDS recruits who are at the same recruitment distance from a given seed. For example, seeds represent wave 0 of a respondent driven sampling survey, and their immediate recruits make up Wave 1. Recruits from Wave 1 make up Wave 2, and so on, until the sample size has been achieved.
Venue Viral load Vulnerability, vulnerable populations
Wave
Western blot test
Also known as immunoblot. A laboratory test that detects specific antibodies to components of a virus. Often used to confirm HIV antibodies in specimens found to be reactive using a screening test. Also known as standard score. Indicates the distance between a data point from the mean of a dataset, within a normal distribution. A z-score greater than 0 indicates that the observation is above the mean, a z-score below 0 indicates that the observation is below the mean, and a z-score of 0 indicates that the score is the same as the mean.
Z-score
530
GLOSSARY
ISBN: 978-92-4-151301-2
532
GLOSSARY