Organisation mondiale de la santé (OMS) · Publications

Meeting report on framework for metrics to support effective treatment as prevention, 2-3 April 2012 Geneva, Switzerland

17 septembre 2012 Organisation mondiale de la santé

Le texte intégral est hébergé par l’organisation qui le publie. lawenc.com indexe les métadonnées et renvoie vers la source officielle.

Texte intégral

hiv/aids Programme

meeting rePort on

Framework For metrics to support eFFective treatment as prevention

2–3 aPril 2012 GEnEVA, SWItzErlAnD

hiv/aids Programme

meeting report on

Framework for Metrics to Support Effective Treatment as Prevention

2–3 april 2012 Geneva, switzerland

hiv/aids Programme

meeting report on

Framework for Metrics to Support Effective Treatment as Prevention

2–3 april 2012 Geneva, switzerland

WHO Library Cataloguing-in-Publication Data Meeting report on framework for metrics to support effective treatment as prevention, 2-3 April 2012 Geneva, Switzerland. 1.HIV infections – drug therapy. 2.HIV infections – prevention and control. 3.HIV infections – transmission. 4.Antiretroviral agents – therapeutic use. 4. I.World Health Organization. ISBN 978 92 4 150433 1 (NLM classification: WC 503.2)

© World Health Organization 2012 All rights reserved. Publications of the World Health Organization are available on the WHO web site (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: bookorders@who.int). Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press through the WHO web site (http://www.who.int/about/licensing/ copyright_form/en/index.html). 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 the World Health Organization 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 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 the World Health Organization 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 the World Health Organization 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 the World Health Organization be liable for damages arising from its use.

Layout by L’IV Com Sàrl, Villars-sous-Yens, Switzerland. Printed by the WHO Document Production Services, Geneva, Switzerland..

Framework for Metrics to Support Effective Treatment as Prevention

contents 2

Acronyms and abbreviations.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I. Background of Treatment as Prevention (TasP) and the development of relevant metrics .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . What is TasP? Implementing TasP for HIV infections at scale. . . . . . . . . . . . . . . . . . . . . . . .. A need for guidance in implementation and monitoring and evaluation (M&E) .. . . . . . . . . . . . . . . . . . . . . . . . . Development of a TasP metrics framework

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 3 4 5 5

II. Programmatic context and implications for the development and application of TasP metrics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. A monitoring and evaluation framework for TasP ..

. . . . . . . . . . . . .

7 11 11 12 13

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A. Logical framework for TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Summary status of the feasibility of measuring logical framework indicators. C. General design considerations for impact evaluation of ART programmes. . .

IV. Monitoring adverse events and unintended consequences to guide programmatic action. . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. E.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

HIV drug resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pharmacovigilance for adverse drug reactions to ART. . . . . . . . . . . . . . . . . . . . Risk behaviour compensation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring the allocation of ART resources against national programme priorities. Ethical concerns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 15 17 18 18 19

V. Use of clinical/community trials and mathematical models for measuring the impact of TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. E.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Current status of clinical and community trials on TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of mathematical models for TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Current status of mathematical modelling for TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Essential partnership between modelling/clinical trials, epidemiological and programmatic expertise. Moving toward more routine use of modelling for crude estimates of TasP impact. . . . . . . . . . . . . .

20 20 . 21 . 21 22 22 24 24 25 26 26 28 30 32

VI. Assessing country readiness for adopting essential TasP metrics. .. A. Checklist to estimate the readiness for measuring the effectiveness of TasP B. Investing in stronger M&E systems for TasP.. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Annexes.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Annex A. List of participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex B. Indicator options for the TasP logical framework. . . . . . . . . . . . . . . . Annex C. Checklist of routine monitoring indicators for ART cascade of services.

References.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Acronyms and abbreviations

ARI ART ARV EWI FSW GARP HIV-DR HPTN HTC LQAS M&E MMC MSM PEP PITC PLHIV PMTCT PrEP PWID STI TasP TB UTT WHO

annual risk of infection antiretroviral therapy antiretroviral early warning indicator female sex worker Global AIDS Response and Progress HIV drug resistance HIV Prevention Trials Network HIV testing and counseling lot–quality assurance sampling monitoring and evaluation medical male circumcision men who have sex with men postexposure prophylaxis provider-initiated testing and counselling people living with HIV prevention of mother-to-child transmission pre-exposure prophylaxis people who inject drugs sexually transmitted infection treatment as prevention tuberculosis universal testing and treating World Health Organization

Framework for Metrics to Support Effective Treatment as Prevention

I. Background of Treatment as Prevention (TasP) and the development of relevant metrics

A. What is TasP? In this document, the term TasP refers to the use of antiretrovirals (ARVs) for treating people living with HIV (PLHIV). When ARVs are effective in reducing viral load, they also reduce a person’s likelihood of transmitting HIV to others, independent of CD4 cell count.1 TasP should not be perceived as being separate from the use of antiretroviral therapy (ART) for therapeutic benefits. It includes the use of ARVs for the prevention of HIV and tuberculosis (TB) (1) regardless of CD4 count but it does not include the use of ARVs for postexposure prophylaxis (PEP), pre-exposure prophylaxis (PrEP) and the use of ARV-based microbicides. The factors that are critical for ART to reduce AIDS-related morbidity and mortality are consistent with those necessary for effective TasP, i.e. high coverage and quality across the cascade of services – from testing, linkage to care, initiation on ART, adherence to the regimen, monitoring viral suppression and early detection of drug resistance. Assuming constant levels of risk behaviour and cofactors, such as the prevalence of sexually transmitted infections (STIs) and coverage of male circumcision, the potential number of HIV infections averted through the use of ARVs by PLHIV depends upon the number of individuals and unprotected contacts a person on ARVs is likely to have over a period of time, and the duration of time the infected person maintains viral suppression while on ARVs. Similarly, for vertical transmission, the number of infections averted among HIV-exposed infants is related to the proportion of infected pregnant women using ARVs during pregnancy and the period of breastfeeding. Optimizing the use of TasP at a programmatic level requires evidence-based decisions about which groups of PLHIV are prioritized for receiving ART, how early ART is initiated among PLHIV, and service delivery models to achieve and maintain viral load suppression through good adherence and clinical/laboratory monitoring of different types of patients on ART.

1 The effect of ART on reducing HIV transmission is related to reduced viral load and not to immunological status (i.e. CD4 count).

Meeting Report – Geneva, Switzerland, 2–3 April 2012

The term “TasP” encompasses a wide range of programmatic approaches. Countries developing TasP strategies must define (a) which types of outcomes and among which groups primary prevention is being prioritized; (b) what immunological, virological and clinical criteria to use for initiating ART among PLHIV; and (c) what service delivery bottlenecks are to be addressed. These options include the following: • Main prevention outcomes: prevention of new infections is one of the main impacts/outcomes of TasP. This would include, for example, preventing HIV infection among sexual or injecting partners and newborns, incidence or recurrence of TB disease among those on ART, and reduced morbidity and mortality among those coinfected with viral hepatitis. • Eligibility criteria for initiating ART: these include, for example, any person infected with HIV regardless of CD4 count; PLHIV with a higher likelihood of transmitting infection (e.g. persons with a high viral load, persons in serodiscordant sexual partnerships, those belonging to key populations at higher risk, pregnant women); PLHIV with co-morbidities (e.g. TB or living in areas where TB is endemic or epidemic, HIV-associated nephropathy, liver disease including hepatitis B, cardiovascular risk factors); or PLHIV with a CD4 count <500 cells/mm3;or 2010 treatment guideline recommendations for PLHIV with a CD4 count <350 cells/mm3. • Enhancement of the cascade of ART service delivery: this can be done through by more efficient case-finding of PLHIV through targeted HIV testing and counselling (HTC) services and/or more effective routine testing (i.e. provider-initiated testing and counselling [PITC]); coverage of priority populations in high-burden areas; stronger linkages and follow up between HIV testing and care and treatment services for PLHIV; improved patient support to facilitate better adherence; improved patient monitoring to ensure adherence and achieve/maintain higher rates of viral load suppression among those on ART.

B. Implementing TasP for HIV infections at scale In mid-2011, the effectiveness of early ART in preventing sexual transmission of HIV was confirmed by the HIV Prevention Trials Network (HPTN) 052 trial conducted among serodiscordant couples, which showed a 96% reduction in HIV transmission. (2) This trial provided definitive evidence, adding to more than ten years of data from observational and basic science studies, on the effectiveness of ART in preventing transmission through control of viral replication. (3) A large number of planned and ongoing clinical trials will provide more information about the efficacy of different TasP approaches in different settings. (4) TasP is now considered an increasingly important component of combined prevention strategies for HIV, which has implications for optimizing and measuring the potential synergies of a combination of components. As countries begin to implement TasP in different populations and settings, data from

Framework for Metrics to Support Effective Treatment as Prevention

country programmes, demonstration projects and implementation research sites will provide important information on how to optimize TasP outside clinical trial settings.

C. A need for guidance in implementation and monitoring and evaluation (M&E) As more national programmes consider how to begin implementing different TasP approaches at scale, there has been a call for updated guidance on ART. In preparation for developing an updated, consolidated guideline on ART use, the World Health Organization (WHO) has initiated a review of the evidence on the benefits (and possible harms – resistance, adverse effects, poor adherence) of earlier initiation of ART, in terms of preventing HIV and other coinfections, as well as decreasing morbidity and mortality due to HIV. In addition, WHO has undertaken a review of methods to monitor, evaluate and assess the impact of TasP. In particular, WHO is interested in knowing how to strengthen existing routine monitoring systems for HTC, prevention of mother-to-child transmission (PMTCT), ART, TB/HIV, pharmacovigilance and HIV drug resistance (HIV-DR) to ensure the effective management of TasP. Metrics for evaluating the impact and determining the cost effectiveness of different approaches in different contexts can support decision-making by programme managers on how and when to reallocate resources (financial, human and technical) to support TasP implementation. The use of mathematical modelling, including models used to compare costs and benefits, will play a critical role in assessing the impact of TasP in different epidemic settings. The development and use of TasP metrics are consistent with the renewed focus of development partners on using implementation science and operational research to ensure that global investment in the HIV/AIDS epidemic results in the maximum benefit. (5–9)

D. Development of a TasP metrics framework WHO convened an initial meeting among experts in the field to develop a framework for TasP metrics (see Annex A for list of participants), including an assessment of the status of indicators and methods of measurement that will help to track the effectiveness of TasP at both the national and global levels. This paper proposes a framework for TasP metrics resulting from this consultation. The key issues in measurement explored in this paper will be further refined through discussion by technical experts and country-level stakeholders with field experience in applying these metrics in different epidemic and resource contexts. The TasP metrics framework addresses the following key areas: • The programmatic context in which TasP is implemented and the implications for developing metrics;

Meeting Report – Geneva, Switzerland, 2–3 April 2012

• • • •

A logical framework for guiding impact evaluation of TasP, including discussion of the different options proposed for indicators at different levels; Monitoring systems for early identification of problems to mitigate the adverse events and unintended consequences of TasP; The use and availability of mathematical models as tools for guiding policy-makers and evaluating the impact of TasP; and Tools to assess country readiness for employing TasP metrics and approaches for strengthening M&E systems.

Framework for Metrics to Support Effective Treatment as Prevention

II. Programmatic context and implications for the development and application of TasP metrics

Monitoring and evaluation is a fundamental tool for planning and managing large-scale programmes to ensure that they achieve their intended goals and objectives. Both the development of TasP approaches and its associated measures of effectiveness must be grounded in the programmatic realities in which they are implemented. This section highlights some of the key programmatic contexts that influence the proposed TasP metrics framework. 1. TasP is a multi-component intervention. To achieve a reduction in HIV transmission, ART programmes must ensure the effectiveness and quality of a cascade of services from testing, referral to care and treatment, determining ART eligibility/initiation, ensuring adherence and ongoing patient management on ART. Current data collected as part of universal access and Global AIDS Response and Progress (GARP) reporting provide important information on the ability of national programmes to achieve high coverage based on ART needs as defined by the 2010 WHO treatment guidelines and good follow up of patients across this cascade of services. Measuring the effectiveness of TasP cannot be separated from the effectiveness of these other programme components.

Figure 1. Cascade of services necessary for effective implementation of TasP Corresponding routine monitoring indicators* People with HIV (100%) Aware of status (Testing) Linked to care Retained in care On ART VLsuppressed #/%tested,%positivity%asymptomaticamongHIV+ % of diagnosed linked to C&T* % of not-ART-eligible in care #/%onART(bysymptoms/CD4count),%retainedon ART at 12, 24, 36, & 60 months % of ART patients with viral load suppression

100

Percentage * Bold font indicates current global reporting indicators. Source: Adapted from Gardner EM et al. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clinical Infectious Diseases , 2011, 52:793–800.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

For example, TasP requires therapy for life, and there are already significant challenges in retaining people on treatment and supporting adherence among those on ART for therapeutic purposes. Programme reports and observational studies in selected countries in sub-Saharan Africa indicate that between 78% and 90% of PLHIV are still on treatment at 12 months but only 53%–78% at 60 months after initiation of ART. (10) It is unclear what effect earlier ART initiation may have on adherence. Some suggest that adherence and retention may be even more difficult to effect among patients who initiate ART when they may not yet have serious illnesses or perceive a direct benefit of treatment. Others posit that adherence and retention may increase with the added motivation of preventing HIV transmission to partners, and potentially less adverse effects and drug interactions, as experienced with late presenters with co-morbidities. The HPTN 052 trial has shown that adherence counselling facilitated more than 95% adherence to ART in 79% of participants who initiated ART with a CD4 count >350 cells/mm3 compared with 74% in participants who initiated ART with a CD4 count <350 cells/mm3.

2. Most data quality and data availability issues that will affect the measurement of TasP effectiveness apply to existing global monitoring indicators, such that general strengthening of data systems will improve TasP measures and vice versa. Many countries have recognized the weakness of routine ART monitoring systems in both clinical and community-based settings. For example, participants at a recent WHO meeting held in September 2011 (11) identified as considerable problems a lack of consistency in the definitions of terms (e.g. retention, loss to follow up, period definitions, etc.) and, in general, poor programme reporting (e.g. on the linkages between testing and care, retention and adherence, etc.). The availability of adequate on-site supervision was also cited as a contributing factor to weak systems. As a follow up from this meeting, it was agreed that WHO would prioritize supporting the development of consensus on these areas to promote better monitoring and reporting. Efforts to develop TasP indicators of programme effectiveness must be consistent and well integrated into the larger system of routine monitoring indicators at both the country and global levels. Efforts to strengthen routine monitoring for TasP should also contribute to strengthening the general capacity for routine monitoring for both HIV and other aspects of the health system. Proposals for including core indicators for TasP which require new data collection systems or adoption of new technology must be considered carefully, particularly if these indicators are expected to be measured routinely. Guidance for adopting these measures as core indicators for TasP must provide countries with feasible options for how to do so in a cost-effective manner.

3. The term “TasP” itself encompasses a number of different possible approaches, either with universal scope (universal testing and treating [UTT]) or targeted at different subgroups of PLHIV. As such, each approach has a different potential for reducing HIV incidence. Different measures of effectiveness may be more suitable or feasible for different TasP approaches. This includes the use of different technology to measure changes in incidence directly, the use of proxies or surrogates for changes in incidence, and methods of defining

Framework for Metrics to Support Effective Treatment as Prevention

the group among which to measure the effect of TasP. For example, in generalized epidemics, efforts to measure HIV transmission among serodiscordant couples in a stable relationship may encompass a large proportion of new infections, while in concentrated epidemics, measuring the effectiveness of TasP in key populations at higher risk may capture a larger proportion of new infections. Similarly, impact evaluations of the effectiveness of TasP in preventing infections among seronegative partners within discordant stable partnerships may be able to directly monitor and measure seroconversion through special studies. However, this approach may not be feasible for measuring the impact of TasP on populations with multiple, often anonymous, sexual or injection partners. Instead, the most feasible measure of the effect of TasP may be limited to measuring viral load suppression among key populations who are infected with HIV. Finally, different approaches to and target populations for TasP may also require programmes to collect information about different confounders or potential adverse effects. 4. Estimating the impact of current ART programmes on HIV and TB incidence based on a national programme’s eligibility criteria for therapeutic benefit, e.g. those with a CD4 count <350 cells/mm3, may comprise a large proportion of the total benefit of TasP and should be estimated in all countries (see Figure 2). By the end of 2010, 6.6 million people were on ART for their own health, accounting for around 47% of those in need, as defined by the 2010 WHO treatment guidelines. (12) Despite this achievement, an estimated 7.5 million people with CD4 cell counts <350 cells/mm3 are still in need of treatment. In many countries, the average CD4 count at the time of diagnosis is also quite low (e.g. <100 cells/mm3). In the context of limited resources, many countries may continue to focus on scaling up ART among people meeting treatment eligibility criteria, rather

Figure 2. Overlaps in programming for TasP

HTC Condoms ART as prevention (including PMTCT) STI management Targeted interventions ART for therapeutic benefit

MMC

A combined prevention strategy

Meeting Report – Geneva, Switzerland, 2–3 April 2012

than implementing TasP-specific strategies aimed at earlier enrolment of PLHIV on ART. It will be valuable to quantify the impact of ART on both morbidity/mortality and incidence, even among countries that are not implementing TasP-specific approaches. Measuring impact will be challenging and require the use of a mix of methodologies, including strengthening of the available data on which to base impact evaluations, and epidemic modelling. Because people on ART, regardless of the criteria for initiation, will receive services at the same facilities, it will be important to ensure that the methods for monitoring of retention in and adherence to ART for prevention is the same as adherence monitoring of ART for therapeutic purposes to prevent overburdening the health services with multiple monitoring and reporting systems.

5. As TasP is implemented, it will be part of a combined prevention approach, which may make the TasP-specific impact on incidence difficult to measure (see Figure 2). Combined prevention strategies cover a wide range of activities, including PMTCT, medical male circumcision (MMC), condom promotion and distribution, targeted interventions for female sex workers (FSWs), men who have sex with men (MSM), and people who inject drugs (PWID), including harm reduction for PWID; management of STIs and HTC, in addition to TasP. The effect of these prevention interventions, which focus on overlapping but not coincident beneficiary populations, will influence transmission dynamics in a complex way. This necessitates the measurement of population-level effects on incidence for a combined package rather than for a TasP-specific intervention. (13) The intensity, specific components and duration of implementation of a combined prevention strategy may vary considerably in different geographical regions within a country. This may be due to both differences in epidemic context and resource availability. At the same time, the ability to measure impact will also be difficult in mature epidemics where HIV incidence is already decreasing (as is the case in more than 30 countries). (14) Measuring the effectiveness of a combined prevention strategy will require thorough documentation on programme implementation and a centralized approach to collating and analysing data for evaluation purposes. The specific contribution of TasP to the reduction of new HIV infections in the context of a combined prevention strategy could best be informed by randomized controlled trials. (4,15)

6. Comparing or extrapolating the impact or outcomes of TasP should be done thoughtfully among countries/geographical areas with similar epidemic contexts and similar types of combined prevention programmes. For both large-scale programmes and implementation science research, it will be important to acknowledge the contextual differences between and among countries when reporting results. Resources and capacity for different forms of M&E will vary from country to country and across different areas within a country. Geographical areas that are not able to undertake more resource-intensive approaches to evaluate TasP may look to other areas with more evidence of impact to extrapolate conclusions about their own programme’s success. This requires a clear understanding, potentially through modelling studies, of the key background parameters of the HIV epidemic context that can affect the success of TasP.

Framework for Metrics to Support Effective Treatment as Prevention

III. A monitoring and evaluation framework for TasP

A. Logical framework for TasP Although there is great diversity in the approaches to implementing TasP, a general logical framework describing the intended effect of TasP can be applied. The flow of goals, objectives, activities and resources is as follows: • At the impact level, TasP is expected to reduce the number of new HIV infections occurring in a specified population.1 • At the outcome level, TasP is expected to result in viral load suppression among those who receive ART, effectively reducing the transmission probability per risk act between an infected and uninfected person. Viral load suppression will be determined by the level of adherence, even though this indicator is very difficult to measure. At the output level, TasP is expected to enrol and retain PLHIV on ART, and assist patients in maintaining good adherence. Implicit in the measurement of enrolment is the estimation of the denominator, i.e. the number of undiagnosed PLHIV plus the number of PLHIV diagnosed at different stages, and the proportions that are successfully linked to care and treatment services. At the input level, TasP requires sufficient funding, staffing, equipment, management and capacity development for implementing TasP programmes. This includes the cost of longterm ART services.

•

Figure 3. A logical framework for TasP Resources — Input level Level of funding, staffing, equipment, management, and capacity development

Routine monitoring systems

Enrollment* — Output level # and % of eligible PLHIV enrolled in ART Retention — Output level % of patients retained in ART at 12, 24, 36, 60 months

Viral load suppression — Outcome level % of people on ART achieving viral load suppression Periodic surveys/ epidemic models Incidence — Impact level # of new HIV infections averted * Enrollment encompasses coverage of testing, and successful referrals between testing and care/treament

1 Early initiation of ART in the context of TasP may also result in the additional benefits of reduced AIDS-related morbidity and mortality, and sustained quality of life, which should be included in the cost–benefit analysis of such a programme.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

This type of logical framework forms the basis of impact evaluation of TasP. It is consistent with a logical framework developed for ART for therapeutic benefit, with the exception of the intended impact being reduced mortality/prolonged survival and improved quality of life of PLHIV who are on ART for therapeutic benefit. As discussed earlier, the TasP logical framework must also be integrated into that developed for the larger combined prevention package being implemented when conducting impact evaluation of national prevention programmes.

B. Summary status of the feasibility of measuring logical framework indicators As for most programmes, measurement of impact indicators is resource intensive and it may not be possible to conduct these frequently at all sites. The design of the impact evaluation of TasP will have to include a sampling plan specifying where and among which groups to conduct special studies of HIV incidence or seroconversion. The use of modelling will also be critical to project and test the plausibility of attribution of measured changes in incidence to TasP. (For more in-depth discussion, see Section V.) Outcome-level indicators require measurement of viral load among PLHIV. However, in many resourceconstrained countries, viral load is not routinely measured, even among PLHIV on ART. Obtaining representative samples of viral load measurements among patients on ART can build on efforts to collect similar data for monitoring HIV-DR, i.e. through the system for early warning indicators (EWI) for HIV-DR. (16) The most reliable source of data to use for TasP programme management purposes will be the output indicators generated through routine monitoring systems. As described earlier, many of the core indicators are already part of the universal access global reporting system (17) that countries currently report against annually.1 Other critical output indicators (e.g. linkages between testing and care and treatment, retention, loss to follow up, (11) and EWI for HIV-DR and pharmacovigilance) have been proposed by WHO guidance (17–19) but are at various stages of implementation in each country. In addition to strengthening the quality and reliability of routine monitoring systems, national programmes must also consider what additional disaggregation of data is warranted, i.e. by different patient populations, to more specifically evaluate the different TasP approaches adopted. At the input level, many countries have systems for tracking financial and human resources, but the manner in which these data can be used for programme M&E is often limited. As countries move towards integrating HIV clinical services into the regular health system, the costing and expenditure of tracking specific activities will be more difficult to disentangle. One of the greatest challenges in using these data to measure the cost effectiveness or availability of resources for specific TasP approaches will be to clearly define TasP activities and apportion the resources used for TasP, over and above what may have been used for services provided to the same beneficiary population or through the same 1 Indicators already included in the universal access global monitoring: the number of people tested for HIV; the proportion of people (including pregnant women, men and women of reproductive age [15–49 years], and male and female youth [i.e. those aged 15–24 years], key populations at higher risk, newly diagnosed TB patients) who have been tested in the past 12 months and know their status; the percentage of HIV-positive pregnant women completing a full course of ARV prophylaxis; the number and percentage of eligible PLHIV on ART; the percentage of PLHIV retained on ART after 12, 24, 36 and 60 months.

Framework for Metrics to Support Effective Treatment as Prevention

health facility and service provider. If TasP approaches result in significantly greater enrolment in ART programmes generally, potential economies of scale may also have secondary benefits that extend beyond measuring the impact of TasP. Table 1. Summary of status of development of TasP metrics by level of indicator Indicator type Impact – HIV incidence Routine monitoring or special study Special study Key action or development needed to use Feasibility/frequency measure for TasP Every 3–5 years, in selected sites and populations Tools to measure incidence or modelling Annually in selected sites and populations Measurement of viral load Use of existing system in all service sites Strengthen tools for direct measures of incidence; employ mathematical models to test plausibility of attribution HIV case reporting and new infections

Outcome – viral load suppression

Special study or routine monitoringa

Strengthen tools for field collection of specimens suitable for viral load measurement to enable representative sampling Adopt recommended indicators for linkages between testing and care and treatment, and early warning indicators for HIV drug resistance. Develop methods for collecting and analysing data disaggregated by priority patient populations on ART

Output – enrolment and retention

Routine monitoring

Inputs – resources and capacity

Routine monitoring

Define TasP-specific activities to apportion Use of existing administrative resources in order to assess cost systems in all service effectiveness sites Availability of data

a Special studies are required for countries that do not routinely measure viral load among ART patients, including through EWI for HIV-DR. Countries with more resources may adopt approaches for using existing patient viral load measures on a more routine basis.

A more detailed discussion of the specific indicator options for each level of the logical framework are provided in Annex B. This section reviews the indicators being used in current programmes, demonstration projects, operational research and trials, and some of the considerations in selecting one or many indicators for a standardized set that can be used for M&E of TasP in national programmes.

C. General design considerations for impact evaluation of ART programmes The design of impact evaluation studies will include multiple steps in data collection and synthesis, which are tailored to specific programme activities and epidemic context. In general, the complexity of designing impact evaluation studies for HIV prevention results from many sources. Two key challenges include establishing attribution and a counter-factual scenario for a specific intervention, and the ability to measure direct changes in HIV incidence over time. Addressing these evaluation challenges requires investment in a combination of tools that can include large-scale

Meeting Report – Geneva, Switzerland, 2–3 April 2012

community trials with mathematical modelling studies to assess the impact of TasP implemented in selected settings. (More detailed discussion of these tools is given in Section V.) While it is beyond the scope of this document to provide detailed recommendations on the design of impact evaluation studies for TasP, a number of key considerations for future impact evaluation of TasP are listed here: • To define measures of epidemic impact for geographical areas with epidemiological homogeneity. This often means subnational analysis or modelling of data. • • • In concentrated epidemics, to estimate the impact of TasP among specific networks of different key populations at higher risk To compare the impact in geographical areas which have different service availability and quality for both combined prevention and care and treatment To document explicitly how measures of impact from selected areas in a country are generalized or extrapolated to obtain national-level measures of impact.

Framework for Metrics to Support Effective Treatment as Prevention

Iv. Monitoring adverse events and unintended consequences to guide programmatic action

One aspect of optimizing the potential benefit of TasP is to monitor and mitigate or prevent potential adverse events or unintended consequences of rapid scale up of ART services. These issues have already begun to be addressed in the context of rapid ART scale up for therapeutic benefits, but are reiterated here to emphasize the importance of continued efforts to strengthen routine monitoring systems to detect potential adverse events.

A. HIV drug resistance As more countries scale up their ART programmes, a common concern is the ability to ensure highquality follow up of patients and continuous drug supply, given the increased patient volume and more decentralized service provision. The consequence of poor-quality patient follow up is an increasing number of patients who develop HIV-DR due to poor adherence to their ARV regimen or lack of a reliable supply of ARVs. Drug resistance is associated with failure to suppress the viral load (“virological failure”) and should be followed up with a switch to second-line regimens. From a prevention standpoint, patients on ART with uncontrolled drug-resistant strains are more likely to transmit these strains to others than those without drug-resistant strains. The risk of HIV-DR is anticipated to increase because of a larger proportion of PLHIV who are on ART, for both prevention and treatment purposes. Largescale ART programmes must be informed by implementation research and demonstration projects whether lower adherence is observed among patients who may be asymptomatic for long periods of time, or who initiated treatment before becoming symptomatic, and perceive a less significant personal benefit of being on treatment. Several approaches to monitoring HIV-DR by national programmes have already been proposed by WHO and its partners. This includes the previously mentioned EWI system, using data that can be collected routinely by clinical sites providing ART and aggregated for management action at the regional or national level; and a methodology for conducting surveillance for both acquired and transmitted HIV-DR. The key benefit of the EWI system is that these data (with the exception of monitoring viral load suppression) should be readily available at any clinic providing ART, without the need for additional equipment or a high level of expertise. With this approach, local-level clinic managers can analyse and use their data to take management action on a regular basis, without waiting for information or feedback from higher levels. At the same time, the use of standardized and clearly defined indicators help more centralized managers to assess where in the system the greatest problems may be and allocate technical resources to support these areas. One of the limitations of the EWI system is that it does not explicitly include patients who are lost to follow up or who have died, which may be correlated with an increase in HIV-DR at a specific ART facility. This could result in an underestimation of the actual levels of HIV-DR experienced among all those who ever enrolled in ART.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Figure 4. Early warning indicators (EWI) for HIV drug resistance (HIV-DR) Indicators Description EWI 1 On time drug pick-up Target <85% 85-94% >95% EWI 2 Retention on ART at 12 months <75% 75-84% >85% EWI 3 ARV drug supply continuity <100% 100% EWI 4 Suboptimal prescribing practices >0% 0% EWI 5 Virological suppression at 12 months (age>2 yrs) <75% 75-84% Source: Phillips A, April 2012 (presentation)

>85%

For acquired DR, the methods include cross-sectional surveys to both measure raised viral load among patients on ART between 12–15 months and 24–26 months, as well as resistance to specific drugs among those with raised viral load. For transmitted HIV-DR, two survey sampling approaches are used: (1) truncated sequential sampling of ~50 patients1 with more recent infection (i.e. CD4 count >500 cells/mm3 or laboratory confirmation of recent infection) and who are treatment naive; (2) crosssectional sampling of patients initiating ART. Thresholds used for classifying survey areas as having high, medium and low levels of transmitted DR are set at >15%, 5–15% and <5%, respectively. (20) Because the factors associated with acquired HIV-DR are closely related to the quality of ART services (i.e. ability to support patient adherence or maintain a consistent drug supply), results from DR surveys are not easily generalizable beyond the survey area. Instead, the use of EWI and other similar quality assessments of ART clinics may help to identify areas where acquired HIV-DR is more likely to occur. When high levels of transmitted HIV-DR are detected in specific areas, national programmes may decide whether to switch first-line regimens (e.g. to integrase inhibitors or CCR5 antagonists) and/ or conduct individual-level resistance testing at ART initiation and/or institute more routine viral load monitoring for ART patients. It is currently unclear at what levels of transmitted HIV-DR such actions are indicated and this requires study by modelling exercises. 1 This method of sampling is also referred to as lot–quality assurance sampling (LQAS). It may be a potential way to assess viral load suppression in settings where measuring the percentage of ART patients who have achieved viral load suppression is costly or logistically difficult. This method is used to rate whether a unit (e.g. clinic facility or lot of goods) meets a standard of quality (e.g. <15% increase in viral load) by testing a minimal sample.

Framework for Metrics to Support Effective Treatment as Prevention

figure 5. Recommended programmatic actions for various levels of transmitted HIV drug resistance (HIV-DR) Low level transmitted HIVDR (<5%) (ANC, VCT/STI Clinics, high risk populations) • Repeat survey in 2 years in same geographic area. • Consider expansion to additional areas based on ART coverage and roll out. • No changes in ARV guidelines (PMTCT, ART, PEP, PrEP) based on survey data. Moderate level transmitted HIVDR (5–15%) (ANC) 1 Perform careful quality assurance of laboratory and epidemiological data. 2 Immediatelyrepeatsurveytoconfirmmoderateclassificationinsamearea;performTDRsurveillanceinadditionalareasofthecountry. 3 Critically review possible sources of HIVDR transmission: (1) Assess HIVDR EWI data and data from surveys of acquired HIVDR from clinics in the same geographic region; (2) Review performance of HIV prevention programmes for individuals aware of their HIV infection in the area of the survey; (3)Assess coverage of HIV testing services in the area of the survey to estimate the risk of unintended HIVDR transmission. 4 Toinformdecision-makingaround:intensifiedviralloadmonitoring,individualdrugresistancetestingpriortoARTinitiationandfirst-lineregimen selection, conduct HIVDR surveillance in populations initiating ART. 5 IfrepeatTDRsurveyinthesamegeographicareaconfirmsmoderateprevalenceclassification,considerperformingfullscalenationalsurveillancein allHIV-infectionpregnantwomentoestimatenationalpointprevalenceofTDRandtriggerpublichealthactions(switchtoP1-basedPMTCTorHIVDR testing of all HIV infected pregnant women) based on national cost-effectiveness thresholds. High level transmitted HIVDR (>15%) (ANC) • Take actions 1–4 listed in “moderate level of HIVDR (5–15%)”. • ImmediatelyperformfullscalenationalsurveillanceofallHIVinfectedpregnantwomentoestimatenationalpointofprevalenceofTDRandtrigger publichealthactions(switchtoP1-basedPMTCTorHIVDRtestingofallHIVinfectedpregnantwomen)basedonnationalcost-effectivenessthreshold. ANC = antenatal care; VCT = voluntary counselling and testing; STI = sexually trasmitted infection clinic; TDR = transmitted drug resistance; PMTCT = prevention of mother to child transmission; ART = antiretroviral therapy; PEP = post-exposure prophylaxis; PrEP = pre-exposure prophylaxis. Source: WHO. Pharmacovigilance for antiretrovirals in resource-poor countries . Geneva, WHO, 2007. (19)

B. Pharmacovigilance for adverse drug reactions to ART Another type of adverse event that national programmes must monitor is the toxicity experienced by patients on ART. These adverse events may be due to product quality, medication errors or adverse drug reactions among individual patients. Pharmacovigilance systems for ARVs in resource-constrained settings have been described in WHO guidance. (18) The lower-resource model for monitoring is spontaneous reporting of toxicity events by health facilities providing ART services. Such reporting is likely to be incomplete in a highly decentralized system and national programmes may opt to strengthen reporting at selected sentinel facilities, e.g. district- and provincial-level hospitals. Key variables for this type of reporting system include: patient demographics, type of treatment (e.g. for therapeutic benefit, PMTCT, PEP, PrEP, etc.), type of regimen, use of other concomitant drugs, duration and type of adverse event, treatment response or action to adverse event, and outcome/severity of event. The routine analysis of pharmacovigilance data can result in the following programmatic actions: 1. Identification and additional supervision of facilities with poor prescription practices or frequent medication errors

Meeting Report – Geneva, Switzerland, 2–3 April 2012

2. Early identification and removal from the drug supply of product lots that may be of poor or substandard quality 3. Special studies of high occurrences of specific adverse drug reactions to understand the characteristics of populations that are more prone to such events 4. Consideration of revising treatment regimens to those with lower levels of toxicity events for all patients, or for patient groups with higher occurrence of adverse drug reactions.

C. Risk behaviour compensation A key concern for the use of ART generally, and in the context of TasP, is that reduction in viral load among PLHIV on ART will be offset by increased risk behaviour, e.g. lower condom use, lower use of sterile injection equipment, and more frequent sexual or needle-sharing contact. Changing risk behaviours may be either expected among those on ART or more generally among those not infected, if the community perceives that increased ART coverage may reduce the overall risk of transmission. Risk behaviour is already a complex and constantly changing variable to measure; however, scaling up ART has stimulated greater interest in measuring patterns of behaviour change that may be related to being on ART for short or long periods of time, among different patient groups, as well as correlating general patterns of behaviour change among those still susceptible to infection. The initiation of patients on ART at higher CD4 counts, i.e. those who are asymptomatic, is likely to add another dimension to patterns of behaviour change that must be taken into account. A key area of development in TasP metrics will be to develop standardized approaches to assessing behaviour change attributable to the use of ART for prolonged periods, or earlier initiation of ART. A systematic review and meta-analysis of the association between ART use and high-risk behaviour or increases in risk behaviour have been commissioned by WHO. These findings will be reflected in the updated treatment guidelines in terms of how to strengthen counselling for patients on ART and monitor these trends in behaviour change. These findings may also inform epidemic models estimating the prevention benefit of ART in terms of the likely magnitude of effect that changes in risk behaviour may contribute.

D. Monitoring the allocation of ART resources against national programme priorities At a programmatic level, new policies and guidelines may result in de facto shifts in deciding which groups of PLHIV are more likely to be enrolled in ART. National programmes should have clear guidelines and targets for allocating limited ART resources to populations that have a clear priority. To reinforce these guidelines, national programmes must monitor the numbers and proportion of different groups of patients prioritized for ART. Local (i.e. facility level) selection criteria for enrolling eligible patients should be assessed regularly to determine whether they are consistent with national priorities. Clinic-

Framework for Metrics to Support Effective Treatment as Prevention

level data can be aggregated by geographical region to determine whether ART allocation is uniform across geographical areas. Feedback should be given to specific ART sites where guidelines are not being followed. These data also have utility for impact evaluation and epidemic modelling, as the magnitude of infections averted should be related to the proportion of PLHIV on ART belonging to different subgroups.

E. Ethical concerns There are a number of ethical issues associated with administering ART programmes. Two of them are defining eligibility criteria and adopting methods for selecting patients from a pool of those who meet the defined eligibility criteria. These issues, which require ongoing tracking systems and periodic proactive assessment, are given below. 1. Confidential and voluntary testing practices, and enrolment in care and treatment services 2. Protection of PLHIV’s rights to voluntary self-disclosure of HIV status 3. Provision of strong linkages between HIV testing services and follow-up prevention, care and treatment services 4. Equity in selection and enrolment of PLHIV eligible for ART 5. Confirmation that patients enrolled in ART have HIV 6. Provision of clear and accurate information to patients about the benefits and risks of being on ART, and ensuring that there is patient choice on whether and when to enrol in ART.1 7. Provision of continuity of service for PLHIV who initiate ART, especially those who are incarcerated or institutionalized 8. Informed consent for patients included in implementation research or special studies to optimize ART programming. Monitoring these types of ethical issues requires the establishment of mechanisms for reporting, investigating and redressing ethical concerns or potential violations by specific health-care providers or facilities. The lack of policies to protect patients’ rights must also be addressed to enable the enforcement of ethical practices in ART programmes.

1 Formal assessment of the risk–benefit ratio for individuals has not been done among people with a CD4 count >350 cells/mm3 and is an area where there is continued need for study through future and ongoing randomized controlled trials of people with a CD4 count >350 cells/mm3 being randomized to immediate ART or deferred till the CD4 count is <350 cells/mm3. This may be particularly relevant for the subgroup of patients who initiate ART at a CD4 count >500 cells/mm3.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

V. Use of clinical/community trials and mathematical models for measuring the impact of TasP

Data that come from well-designed and implemented clinical and community trials play a critical role in efforts to measure the impact of TasP. Such data can provide key measures of new infections averted resulting from a thoroughly documented intervention scenario, as well as for comparing these results to an alternative situation. Using these types of results as inputs in mathematical models, it is possible to further project the effect of the intervention conducted at a larger scale or under other epidemic or programmatic contexts.

A. Current status of clinical and community trials on TasP A recent inventory of global research on ART as prevention was conducted by Granich et al. (4) Among those described are more than 20 randomized clinical or community trials planned or in progress. These trials represent a wide range of funding agencies and research teams from several countries. A majority of them are studying populations in sub-Saharan Africa or North America. These rigorous study designs aim to directly measure the impact of a range of TasP strategies, ranging from widerscale delivery of treatment based on current guidelines, through ART initiation based on viral load, to test-and-treat based on immediate initiation of ART for all those testing HIV positive in a range of epidemiological settings. There is a concerted effort to harmonize the conduct of these trials as far as is feasible to maximize the comparability of data and their value for subsequent joint analyses. These data will help to validate and calibrate the models so that they can then be applied more reliably for routine evaluation in a wide range of countries and settings. figure 6. Inventory of global efforts at research on ART as prevention (2011) Map representing countries planning or implementing antiretroviral therapy (ART) in prevention of HIV and/or tuberculosis (TB) research. Dark blue represents countries that are conducting ART in prevention of HIV and/ or TB research, light blue represents countrywide efforts (United States, Swaziland), and the red dots represent selected study sites within countries conducting research (some countries had too many sites to represent on this graphic).

Source: Granich R, Gupta S, Suthar AB, et al; ART in Prevention of HIV and TB Research Writing Group. Antiretroviral therapy in prevention of HIV and TB: update on current research efforts. Curr HIV Res. 2011 Sep;9(6):446-69. (4)

Framework for Metrics to Support Effective Treatment as Prevention

B. Types of mathematical models for TasP In the absence of empirical measures of HIV incidence, mathematical modelling plays a critical role in using what is known about HIV transmission to predict the impact of different TasP approaches on a wide range of epidemic contexts and populations. Modelling can be useful in testing hypotheses about the factors or epidemic context in which a TasP approach may be most successful. While mathematical models can only very rarely be used as a substitute for good data, they are invaluable for aiding the interpretation of limited or conflicting data, and in generalizing findings from one setting where data coverage is good to other similar settings where data coverage is more limited. They are also used for exploring hypothetical scenarios when considering new policy options. Some models choose simplified scenarios to demonstrate a theoretical effect of specific factors or corroborate results of clinical trials or operational research with modelled results. These types of models may be helpful in guiding national programmes in prioritizing different populations or approaches for implementing TasP. Other models are designed to estimate the impact of TasP in a specific real-world context. These types of models have direct application to impact evaluation and cost-effectiveness analyses, and the results can be applied for conducting advocacy, effecting policy change and adjusting resource allocation.

C. Current status of mathematical modelling for TasP The HIV Modelling Consortium is funded by the Bill and Melinda Gates Foundation to coordinate and guide teams engaged in epidemic modelling to address key questions that can inform HIV/AIDS programming. A meeting of the HIV Modelling Consortium was convened in early November 2011 to systematically review the development of mathematical models that could estimate the impact of treatment on HIV incidence. (21) For the exercise, eleven modelling teams were asked to project the effect of ART on HIV incidence in the context of the South African epidemic. Projections were developed for the period 2012–2020 for a number of different scenarios, varying ART coverage, CD4 count eligibility and levels of retention in ART programmes. The models participating in the exercise included both micro-simulation and deterministic models of various levels of complexity in model structure (such as altering transmission dynamics by gender, age, stage of HIV disease and in terms of heterogeneity in sexual mixing). Under a scenario of 80% ART coverage of persons with a CD4 count <350 cells/mm3 and 85% retention in the ART programme after 3 years, the model results ranged from between 35% and 52% reduction in incidence compared to a scenario where there was no ART coverage. The meeting participants identified a number of factors that may result in lower estimates of impact and called for additional efforts to compare the results of models to better understand the reasons for the underlying differences. Similar efforts to test epidemic models developed for ART coverage for different CD4 count thresholds, e.g. 350–500 cells/mm3 and low-level and concentrated epidemic settings, are also

Meeting Report – Geneva, Switzerland, 2–3 April 2012

needed to determine whether the potential impact of TasP is of similar magnitude or influenced by key factors to the same degree as modelling efforts in generalized epidemics in sub-Saharan Africa. This Consortium’s approach to peer review and developing standard scenarios with which to test the differences in models will be critical as national programmes make decisions about using models to support their planning and impact evaluation efforts.

D. Essential partnership between modelling/clinical trials, epidemiological and programmatic expertise There is already strong collaboration between modellers and those conducting clinical trials across a majority of the ongoing research efforts on TasP described earlier. By working together, the models can be validated, calibrated and improved on the basis of empirical population-based data collected through field trials. Similarly, modelling studies and clinical trials can be made more relevant and informative when study investigators collaborate with epidemiologists, biomedical researchers and programme experts. This type of collaboration is needed to identify the priority programmatic and planning questions that might be best answered through modelling, as well as to ensure the appropriate application of data used as inputs for the models, and to identify the underlying programmatic factors that might limit the ability to extrapolate modelled results from one country to another. The HIV Modelling Consortium and principal investigators of clinical trials are important resources for international development partners and country programmes for answering key questions on measuring the impact of TasP. This work by the Consortium includes testing the performance of ARVs and proposing appropriate measures for describing the effectiveness of ART in preventing new infections, including those proposed in this metrics framework, such as “mean or cumulative viral load among all PLHIV in a community,”1 annual risk of infection (ARI) or the reproductive number of HIV (R0).

E. Moving toward more routine use of modelling for crude estimates of TasP impact In the near future, it would be ideal to have a modelling tool that is feasible for use by national programmes to estimate the impact of ART on HIV incidence, using a standard limited set of data inputs characterizing their ART programmes. The current EPP/Spectrum model (22) used biennially by countries to estimate the burden of disease and project levels and trends in HIV incidence does include the effect of ART on reducing transmission, but because the model is designed to fit curves 1 “Community viral load” has been proposed as a summary measure for describing the potential for HIV transmission in a community. Changes in community viral load would ideally reflect the coverage and effectiveness of ART among PLHIV in a specific community. There are numerous ways to define community viral load, in terms of which population the measure pertains to and the mathematical formulation (e.g. arithmetic mean, geometric mean, log-transformed arithmetic mean, cumulative sum, etc.), each with different implications for the use of the measure to describe the performance of an ART programme. These types of measures, more specifically defined, may be used in various impact evaluation studies; however, given the current use and availability of viral load measures, for most national programmes, it is recommended that TasP outcome indicators be defined in terms of percentage of those who are virologically suppressed (see Annex B).

Framework for Metrics to Support Effective Treatment as Prevention

Defining programmatic questions to be answered with mathematical models of TasP: an example from PMTCT Background One of the key populations for measuring the impact of TasP is HIV-infected pregnant women and serodiscordant couples identified through couples counselling in PMTCT settings. A leading option for PMTCT regimens is Option B+, which involves treating all HIV-positive pregnant women with ART, regardless of CD4 count. Mathematical modelling of the number of infections averted through the scaled-up use of Option B+ could help: • to inform the WHO 2013 consolidated ART guidelines on the prevention effectiveness and cost effectiveness • • to advocate for the adoption of Option B+ by national programmes to identify where implementation of Option B+ would be best focused in the context of scarce resources.

Specific modelling questions include: 1. Projecting the HIV prevention benefit attributable to providing ART to pregnant women currently eligible for ART (CD4 count <350 cells/mm3), i.e. the optimized delivery of current recommendations in terms of high coverage and retention rates, given an estimated HIV prevalence among pregnant women 2. Estimating the ART prevention benefit of providing couples counselling and ART for serodiscordant couples (regardless of CD4 count) in PMTCT settings, and of offering Option B+ to all HIV-positive pregnant women 3. Comparing the relative cost effectiveness of couples testing and ART for all serodiscordant couples in PMTCT settings compared to implementation of Option B+ for all HIV-positive pregnant women.

to empirical data, it is not well suited to estimate the change in incidence attributable to ART. Having a simple, standard modelling tool to give rough estimates of this impact will be helpful both for monitoring national programmes as well as generating regional and global estimates of effect. In the meantime, countries can concentrate on estimating the proportion of PLHIV who are not virologically suppressed as a core measure of the national response to the epidemic.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

VI. Assessing country readiness for adopting essential TasP metrics

A. Checklist to estimate the readiness for measuring the effectiveness of TasP As countries update their ART strategies, workplans and budgets to integrate TasP approaches, M&E systems must also be updated and strengthened. National programmes can review their existing information systems for the following: • What is the availability of data sources described in the TasP cascade of services? (see Annex C.) »» Is the indicator collected routinely? »» How complete or timely are the data for each indicator? »» Is the frequency of data reporting sufficient to take programmatic action? »» Is there a data bottleneck that prevents reporting from being used by different levels of management? • What is the programme’s ability to disaggregate each indicator in the TasP cascade of service by key variables, e.g. age group, gender, clinical stage or “CD4 status” by key population prioritized for TasP? What is the status of implementation of EWI for HIV-DR and routine pharmacovigilance monitoring/reporting systems? »» Which sites participate? »» How complete is reporting? »» Which agency/organization is responsible for reviewing the data and initiating programmatic action? • What is the availability of CD4 count and viral load measurements as part of routine patient monitoring? »» Is CD4 count available at the time of diagnosis; if yes, at which sites? »» Is CD4 count available at the time of care or ART enrolment; if yes, at which sites? »» If viral load is not routinely measured, where are viral load measurements possible, and what options for specimen transport are feasible? • For each TasP indicator from routine monitoring systems, at what level are data entered in an electronic data system and »» How frequently are the data transmitted centrally? »» Which service provider sites do direct data entry and for which indicators? What is the geographical scope of facilities able to do direct data entry? »» Which data entry personnel are available at district, provincial/state, or national level to support data entry? »» Is it feasible to do line-listed data entry, or only aggregated data entry for each indicator?

•

Framework for Metrics to Support Effective Treatment as Prevention

B. Investing in stronger M&E systems for TasP Based on the assessment of country readiness for TasP metrics, it should be possible to identify the most significant gaps in the current information system, by indicator and facility level or geographical region. The readiness checklist in the previous section is organized roughly by importance, and recommends that countries first address issues around the key indicators related to the TasP cascade, as needed. Strengthening these systems must be planned in conjunction with capacity building for M&E in other aspects of HIV/AIDS programming, as well as with efforts to improve the larger health information systems. In addition to stronger routine monitoring systems, more rigorous evaluation of the effectiveness of TasP approaches will require investment in a number of special surveys, facility assessment, more extensive laboratory testing, operational research and mathematical modelling.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Annex A. List of participants Dr Ann Kurth New York University 726 Broadway, 10 Floor Rm 1017, New York NY 10003 USA E-mail: akurth@nyu.edu

Dr David D. Celentano Charles Armstrong Chair and Professor Department of Epidemiology Johns Hopkins University Bloomberg School of Public Health 615 N. Wolfe St., Suite W6041 Baltimore, MD 21205 United States of America E-mail: dcelenta@jhsph.edu

Professor Richard Hayes London School of Hygiene and Tropical Medicine Room 359, Keppel Street London, WC1E 7HT United Kingdom E-mail: richard.hayes@lshtm.ac.uk

Dr William Miller Associate Professor of Medicine & Epidemiology Division of Infectious Diseases, CB#7030 UNC – Chapel Hill Chapel Hill, NC 27599 United States of America E-mail: bill_miller@unc.edu

Professor Christophe Fraser MRC Centre for Outbreak Analysis and Modelling Dept of Infectious Disease Epidemiology Imperial College London Norfolk Place London W2 1 PG United Kingdom E-mail:c.fraser@imperial.ac.uk

Dr Brian Williams 11B Chemin Jacques-Attenville CH-1218 Grand Saconnex Geneva Switzerland E-mail : williamsbg@mg.com

Dr CrispineMoyo National ART Coordinator Ministry of Health Ndeke House Haile Selassie Avenue 10101 Lusaka Zambia E-mail: crispin@moh.gov.zm

Dr Rui Wang Havard School of Public Health 400 Brookline Avenue APT 15D, Boston MA 02215 United States of America E-mail: rwang@hsph.havard.edu

Professor Andrew Phillips Research Department of Infection & Population Health UCL, London United Kingdom E-mail: andrew.phillips@ucl.ac.uk

Dr David Maman Epicentre/Médecins Sans Frontières 8 rue saint Sabin 75011 Paris France E-mail: david.maman@epicentre.msf.org

Dr Virginia Loo PEMA 98-644 Pualima Street Hawai, USA E-mail: ginialoo@gmail.com

Dr Moupali Das Director of Research HIV Prevention Section San Francisco Department of Public Health San Francisco United States of America E-mail: dasm@php.ucsf.edu

Dr Frank Tanser African Centre 28 Mimosa Drive PO Box 741, Mtunzini, 3867 South Africa E-mail l: ftanser@gmail.com

Dr Tim Hallet Dept of Infectious Disease Epidemiology Imperial College London Norfolk Place London W2 1 PG United Kingdom E-mail: t.hallet@imperial.ac.uk

Framework for Metrics to Support Effective Treatment as Prevention

UNAIDS Secretariat Tobias Alfven M&E Adviser, Data For Action Division UNAIDS Geneva Switzerland E-mail: alfvent@unaids.org

Nathan Shaffer Medical Officer, TAC Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: shaffern@who.int

Reuben Granich Medical Officer, TAC Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: granichr@who.int

WHO Secretariat Gottfried Hirnschall Director, Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: hirnschallg@who.int

Rachel Baggaley Medical Officer, KPP Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva E-mail: baggaleyr@who.int

Christopher Dye Director, Office of the Assistant Director-General, HTM Cluster World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: dyec@who.int

Sylvia Bertagnolio Medical Officer, TCO Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva E-mail: bertagnolios@who.int

Yves Souteyrand Coordinator, SIP Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: souteyrandy@who.int

AmitabhSuthar Consultant, TAC Department of HIV/AIDS World Health Organization E-mail: amitabh.suthar@gmail.com

Ying-Ru Lo Coordinator, KPP Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: loy@who.int

Jesus M Garcia Calleja Medical Officer, SIP Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva E-mail: callejaj@who.int

Joseph Perrïens Coordinator, TCO Department of HIV/AIDS World Health Organization 20 Avenue Appia, Geneva Switzerland E-mail: perriensj@who.int

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Annex B. Indicator options for the TasP logical framework

Indicator

Definition

Comments on adopting this indicator

Impact level: Adoption of a specific indicator depends on the impact evaluation methodology and type of TasP strategy being evaluated # and % of HIV infections averted, by age and gender Numerator: estimated # of HIV infections occurring over a period of time, with a specified ART coverage in a specified population Denominator: estimated # of HIV infections in the absence of any ART use in the same population The difference in # of new HIV infections per 100 person-years estimated in a population of specified ART coverage, compared to an appropriate control group or historical rate of incidence in the same population Requires epidemic models for the estimate and comparison to the counter-factual scenario. Also requires assumptions about the impact of other prevention interventions on incident infections over the same time period Appropriate for measuring impact of TasP in a defined, cohort/population. Requires reliable direct measures of incident HIV infections

Difference in incidence rate (per 100 person-years)

Outcome level: Four alternatives for measuring virological suppression are presented. Indicator selection is based on resource availability and intended target for realizing prevention potential % of PLHIV on ART who are virologically suppressed at 12months of treatment, by age and gender Numerator: # of PLHIV on ART who meet laboratory-based criteria of virological suppression at 12-month time point Denominator: the total # of PLHIV on ART with viral load measures at 12-month time point, during the reporting period Consistent with the definition of early warning indicator for HIV-DR measuring virological suppression Requires viral load measurement at 12 months to be routine practice, or to take a representative sample of viral load measurements among patients on ART at 12 months Measures performance of ART programme, but will underestimate the potential for averting new HIV infections if case-finding and linkage from diagnosis to care and treatment is poor

Where viral load measurements are % of PLHIV on ART who are virologically suppressed, by age routine: and gender Numerator: # of PLHIV on ART who meet laboratory-based criteria for virological suppression at last viral load measurement in the past year Denominator: # of PLHIV on ART at the end of the reporting period (Otherwise based on a representative sample of PLHIV on ART) % of PLHIV eligible for ART who are on treatment and virologically suppressed, by age and gender Numerator: # of PLHIV on ART who meet laboratory-based criteria for virological suppression Denominator: total estimated # of PLHIV eligible for ART using current national guidelines

Measures virological suppression achieved among those who are eligible for treatment (i.e. prevention potential realized according to national guidelines on ART eligibility). Indicator is comparable only among countries using the same treatment eligibility criteria

Framework for Metrics to Support Effective Treatment as Prevention

% of total estimated PLHIV who are virologically suppressed, by age and gender

Numerator: # of PLHIV on ART who meet laboratory-based criteria for virological suppression Denominator: total estimated # of PLHIVa

Measures virological suppression achieved among all PLHIV (i.e. total prevention potential realized, independent of national guidelines on ART eligibility)

Output level: Indicators include options for measures of case-finding (i.e. diagnosis of PLHIV), enrolment into ART and retention) # and % of PLHIV who know their status, i.e. have been diagnosed with HIV infection, by age and gender % of (specified population) who have been tested and know their status in the past 12 months Numerator: cumulative # of PLHIV who have been diagnosed with HIV and are still alive Denominator: total estimated # of PLHIV Based on surveys of specific populations, e.g. men and women aged 15–49 years, young people aged 15–24 years, sex workers, people who inject drugs, MSM, etc. Self-reported data from survey respondents Numerator: # of PLHIV receiving ART at the end of the reporting period Denominator: total estimated # of PLHIV Accurate numerator requires complete HTC reporting and HIV/AIDS death counts as well as sound estimates of current # of PLHIV Consistent with the Global AIDS Response and Progress reporting indicators. This indicator is more useful when the percentage of people who are HIV positive among those tested is known Measures ART coverage, independent of national guidelines on ART eligibility. Cannot distinguish between programmes with low levels of casefinding, poor linkage between diagnosis and enrolment in care and treatment, and programmes with inadequate ART resources or restricted treatment eligibility guidelines Measures ART coverage among those who have been diagnosed as HIV positive, independent of national guidelines on ART eligibility. Puts more emphasis on the effectiveness of linkages between diagnosis and care and treatment Measures ART coverage among the estimated # who meet national guidelines on ART eligibility. Puts more emphasis on the availability of adequate ART resources Measures the ability to follow up persons who initiate ART after a specific period of time. Does not distinguish between those with good and those with poor measures of regimen adherence (i.e. timeliness, completeness, etc.). Does not take into account virological failure or potential switch in regimen

# and % of PLHIV currently on ART, by age and gender

# and % of PLHIV diagnosed with HIV who are currently on ART, by age and gender

Numerator: # of PLHIV receiving ART at the end of the reporting period Denominator: the total # of PLHIV who have been diagnosed and who are still alive

# and % of PLHIV eligible for ART who are on treatment, by age and gender

Numerator: # of PLHIV receiving ART at the end of the reporting period Denominator: estimated # of PLHIV who are eligible for ART Numerator: # of PLHIV who initiated ART 12 months ago and are still receiving ART Denominator: number of those in the cohort of individuals who initiated ART 12 (24, 36 or 60) months ago

% of those initiated on ART who are retained in treatment after 12 , 24, 36 and 60 months

a This measure assumes that PLHIV who are not on ART are not virologically suppressed. This is supported by data presented in Williams BG. Determinants of sexual transmission of HV: implications for control . Cornell University Library, 2011.Available at: http://arxiv.org/abs/1108.4715 (accessed on 16 July 2012).

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Annex C. Checklist of routine monitoring indicators for ART cascade of services

People with HIV (100%) Aware of status (Testing) Linked to care Retained in care On ART VLsuppressed

100

Percentage Source: Adapted from Gardner EM et al. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clinical Infectious Diseases , 2011, 52:793–800.

Framework for Metrics to Support Effective Treatment as Prevention

Programme component People with HIV Aware of status

Data sources Estimates and projections (EPP/Spectrum) Surveillance data Probability surveys

Key measures Estimated number of PLHIV HIV prevalence estimates (for specific populations and geographical areas) % tested among different groups within past year (men and women aged 15–49 years, young men and women aged 15–24 years, sex workers, MSM and PWID) # of diagnosed HIV cases (new and cumulative) % positive among those tested from routine HIV testing and counselling data % of persons tested for HIV whose regular partner has also been tested Mean CD4 count at time of diagnosis % of pregnant women whose regular partner has been tested % of newly diagnosed cases who have been enrolled in HIV care and treatment services % of pregnant women receiving ARV prophylaxis (if not offered Option B/Option B+) Number currently receiving care (e.g. receiving cotrimoxazole prophylaxis, regular clinical monitoring/ assessment for ART eligibility) # who initiated ART # currently receiving ART % died, lost to follow up, stopped or switched regimens % retained after 12, 24, 36, 60 months % of patients with on-time drug pick-up % of clinics with drug supply continuity % of clinics with suboptimal prescribing practices # of patients on ART experiencing adverse drug events % virologically suppressed at 12 months % virologically suppressed at last viral load measurement in the reporting period % of transmitted DR detected among treatment-naive patients initiating ART % of patients with raised viral load at 12–15 months of ART and 24–26 months of ART. % detected with DR among those with raised viral load

HIV testing and counselling routine monitoring

HIV case reporting formats PMTCT routine monitoring Linked to care HIV testing and counselling routine monitoring / Care and treatment registry PMTCT routine monitoring Retained in care Care and treatment registry

On ART

ART register

EWI for HIV-DR

Pharmacovigilance monitoring Virological suppression EWI for HIV-DR Routine viral load measures HIV-DR surveillance survey

Meeting Report – Geneva, Switzerland, 2–3 April 2012

References 1. WHO. Antiretroviral treatment as prevention (TasP) of HIV and TB. Geneva, WHO, 2012. Available at: http://www.who.int/ hiv/pub/mtct/programmatic_update_tasp/en/index.html (accessed on 15 July 2012). 2. Cohen MS et al. Prevention of HIV-1 infection with early antiretroviral therapy. New England Journal of Medicine, 2011, 365:493–505. 3. Anglemyer A et al. Antiretroviral therapy for prevention of HIV transmission in HIV-discordant couples. Cochrane Database of Systematic Reviews , 2011, (5):CD009153. 4. Granich R et al. Antiretroviral therapy in prevention of HIV and TB: update on current research efforts. Current HIV Research, 2011, 9(6):446–469. 5. Padian NS et al. Implementation science for the US President’s Emergency Plan for AIDS Relief (PEPFAR). Journal of Acquired Immune Deficiency Syndromes, 2011, 56:199–203. 6. Herbst JH et al. Operational research to improve HIV prevention in the United States. Journal of Acquired Immune Deficiency Syndromes , 2012, 59 (5):530–536. 7. WHO, Alliance for Health Policy and Systems Research. Implementation research platform. Available at: http://www.who.int/ alliance-hpsr/projects/implementationresearch/en/ (accessed on 15 July 2012). 8. Schwartlander B et al. Towards an improved investment framework approach for an effective response to HIV/AIDS. Lancet , 2011, 377:2031–2041. 9. Global Fund to fight AIDS, TB and Malaria. The Global Fund strategy 2012–2016: investing for impact. 2011. Available at: http://www.theglobalfund.org/en/about/strategy/ (accessed on 15 July 2012). 10. Rosen S, Fox MP, Gill CJ. Patient retention in antiretroviral therapy programs in sub-Saharan Africa: a systematic review. PLoS Medicine, 2007, 4(10):e298. doi:10.1371/journal.pmed.0040298 11. WHO. Retention in HIV programmes: defining the challenges and identifying solutions towards decentralized and integrated services. Geneva, WHO, 2012. Available at: http://www.who.int/hiv/pub/meetingreports/retention_programmes/ en/ (accessed on 15 July 2012). 12. WHO. Global HIV/AIDS response: epidemic update and health sector progress towards universal access. Progress report 2011. Geneva, WHO, 2011. Available at: http://www.who.int/hiv/pub/progress_report2011/en/index.html (accessed on 15 July 2012). 13. Barnighausen T, Bloom D, Humair S. Health systems and HIV treatment in sub-Saharan Africa: matching intervention and programme evaluation strategies. Sexually Transmitted Infections , 2012, 88:e2; doi:10.1136/sextrans-2011-050303. 14. UNAIDS. Report on the global AIDS epidemic: chapter 2. Geneva, UNAIDS, 2010. Available at: http://www.unaids.org/ documents/20101123_GlobalReport_Chap2_em.pdf (accessed on 15 July 2012). 15. HIV Prevention Trials Network. HPTN 071: the PopART study. Available at: http://www.hptn.org/web%20documents/Index Docs/071StudyAnnouncement14Sep11.pdf (accessed on 15 July 2012). 16. WHO. HIV drug resistance early warning indicators: WHO indicators to monitor HIV drug resistance prevention at antiretroviral treatment sites. Geneva, WHO, June 2010 update. Available at: http://www.who.int/hiv/topics/drugresistance/ hiv_dr_early_warning_indicators.pdf (accessed on 15 July 2012). 17. WHO/UNICEF/UNAIDS. A guide on indicators for monitoring and reporting on the health sector response to HIV/AIDS . Geneva, WHO, February 2012. Available at: http://www.who.int/hiv/data/UA2012_indicator_guide_en.pdf(accessed on 15 July 2012). 18. Treatment as Prevention: issues relating to HIV drug resistance. Presented by Phillips A at the WHO Workshop on methods and tools for measuring and evaluating epidemiological impact of treatment as prevention. Geneva, Switzerland, WHO, 3–4 April 2012. 19. WHO. Pharmacovigilance for antiretrovirals in resource-poor countries. Geneva, WHO, 2007. 20. WHO. The World Health Organization HIV drug resistance prevention and assessment strategy: global, regional, and country progress. Clinical Infectious Diseases , 2012, 54 (suppl 4). 21. Vesga J; on behalf of the HIV Modelling Consortium. The potential impact of treatment on HIV incidence: meeting report . Stellenbosch, South Africa, 3–5 November 2011. Available at: http://www.hivmodelling.org/sites/default/files/uploads/ documents/meeting-reports/Meeting%20Report%203%20-%20Stellenbosch.pdf (accessed on 15 July 2012). 22. UNAIDS. Spectrum/EPP 2011. Available at: http://www.unaids.org/en/dataanalysis/tools/spectrumepp2011/(accessed on 15 July 2012).

notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Framework for Metrics to Support Effective Treatment as Prevention

contents Acronyms and abbreviations.. 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I. Background of Treatment as Prevention (TasP) and the development of relevant metrics.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. What is TasP?.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implementing TasP for HIV infections at scale. . . . . . . . . . . . . . . . . . . . . . . A need for guidance in implementation and monitoring and evaluation (M&E).. Development of a TasP metrics framework. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 3 4 5 5

II. Programmatic context and implications for the development and application of TasP metrics.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. A monitoring and evaluation framework for TasP. .

. . . . . . . . . . .

7 11 11 12 13

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A. Logical framework for TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Summary status of the feasibility of measuring logical framework indicators. . C. General design considerations for impact evaluation of ART programmes.. . .

IV. Monitoring adverse events and unintended consequences to guide programmatic action. . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. E.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

HIV drug resistance.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pharmacovigilance for adverse drug reactions to ART. . . . . . . . . . . . . . . . . . . . . Risk behaviour compensation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring the allocation of ART resources against national programme priorities. . Ethical concerns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 15 17 18 18 19

V. Use of clinical/community trials and mathematical models for measuring the impact of TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. D. E.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Current status of clinical and community trials on TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of mathematical models for TasP.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Current status of mathematical modelling for TasP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Essential partnership between modelling/clinical trials, epidemiological and programmatic expertise. Moving toward more routine use of modelling for crude estimates of TasP impact. . . . . . . . . . . . . .

20 20 . 21 . 21 22 22 24 24 25 26 26 28 30 32

VI. Assessing country readiness for adopting essential TasP metrics.. A. Checklist to estimate the readiness for measuring the effectiveness of TasP. B. Investing in stronger M&E systems for TasP . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Annexes. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

References. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Acronyms and abbreviations

ARI ART ARV EWI FSW GARP HIV-DR HPTN HTC LQAS M&E MMC MSM PEP PITC PLHIV PMTCT PrEP PWID STI TasP TB UTT WHO

Framework for Metrics to Support Effective Treatment as Prevention

I. Background of Treatment as Prevention (TasP) and the development of relevant metrics

1 The effect of ART on reducing HIV transmission is related to reduced viral load and not to immunological status (i.e. CD4 count).

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

country programmes, demonstration projects and implementation research sites will provide important information on how to optimize TasP outside clinical trial settings.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

• • • •

Framework for Metrics to Support Effective Treatment as Prevention

II. Programmatic context and implications for the development and application of TasP metrics

Figure 1. Cascade of services necessary for effective implementation of TasP Corresponding routine monitoring indicators* People with HIV (100%) Aware of status (Testing) Linked to care Retained in care On ART VL suppressed #/% tested, % positivity % asymptomatic among HIV+ % of diagnosed linked to C&T* % of not-ART-eligible in care #/% on ART (by symptoms/CD4 count), % retained on ART at 12, 24, 36, & 60 months % of ART patients with viral load suppression

40 Percentage

100

* Bold font indicates current global reporting indicators. Source: Adapted from Gardner EM et al. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clinical Infectious Diseases , 2011, 52:793–800.

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

Figure 2. Overlaps in programming for TasP

HTC Condoms ART for therapeutic benefit

ART as prevention (including PMTCT)

STI management Targeted interventions

MMC

A combined prevention strategy

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

III. A monitoring and evaluation framework for TasP

•

Figure 3. A logical framework for TasP Resources — Input level Level of funding, staffing, equipment, management, and capacity development

Routine monitoring systems

Enrollment* — Output level # and % of eligible PLHIV enrolled in ART Retention — Output level % of patients retained in ART at 12, 24, 36, 60 months

Periodic surveys/ epidemic models

Viral load suppression — Outcome level % of people on ART achieving viral load suppression Incidence — Impact level # of new HIV infections averted

* Enrollment encompasses coverage of testing, and successful referrals between testing and care/treament

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

Outcome – viral load suppression

Special study or routine monitoringa

Output – enrolment and retention

Routine monitoring

Inputs – resources and capacity

Routine monitoring

Define TasP-specific activities to apportion Use of existing administrative resources in order to assess cost systems in all service effectiveness sites Availability of data

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

Iv. Monitoring adverse events and unintended consequences to guide programmatic action

Meeting Report – Geneva, Switzerland, 2–3 April 2012

>85%

Framework for Metrics to Support Effective Treatment as Prevention

figure 5. Recommended programmatic actions for various levels of transmitted HIV drug resistance (HIV-DR)

Low level transmitted HIVDR (<5%) (ANC, VCT/STI Clinics, high risk populations) • Repeat survey in 2 years in same geographic area. • Consider expansion to additional areas based on ART coverage and roll out. • No changes in ARV guidelines (PMTCT, ART, PEP, PrEP) based on survey data. Moderate level transmitted HIVDR (5–15%) (ANC) 1 Perform careful quality assurance of laboratory and epidemiological data. 2 Immediately repeat survey to confirm moderate classification in same area; perform TDR surveillance in additional areas of the country. 3 Critically review possible sources of HIVDR transmission: (1) Assess HIVDR EWI data and data from surveys of acquired HIVDR from clinics in the same geographic region; (2) Review performance of HIV prevention programmes for individuals aware of their HIV infection in the area of the survey; (3) Assess coverage of HIV testing services in the area of the survey to estimate the risk of unintended HIVDR transmission. 4 To inform decision-making around: intensified viral load monitoring, individual drug resistance testing prior to ART initiation and first-line regimen selection, conduct HIVDR surveillance in populations initiating ART. 5 If repeat TDR survey in the same geographic area confirms moderate prevalence classification, consider performing full scale national surveillance in all HIV-infection pregnant women to estimate national point prevalence of TDR and trigger public health actions (switch to P1-based PMTCT or HIVDR testing of all HIV infected pregnant women) based on national cost-effectiveness thresholds. High level transmitted HIVDR (>15%) (ANC) • Take actions 1–4 listed in “moderate level of HIVDR (5–15%)”. • Immediately perform full scale national surveillance of all HIV infected pregnant women to estimate national point of prevalence of TDR and trigger public health actions (switch to P1-based PMTCT or HIVDR testing of all HIV infected pregnant women) based on national cost-effectiveness threshold. ANC = antenatal care; VCT = voluntary counselling and testing; STI = sexually trasmitted infection clinic; TDR = transmitted drug resistance; PMTCT = prevention of mother to child transmission; ART = antiretroviral therapy; PEP = post-exposure prophylaxis; PrEP = pre-exposure prophylaxis. Source: WHO. Pharmacovigilance for antiretrovirals in resource-poor countries . Geneva, WHO, 2007. (19)

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

Meeting Report – Geneva, Switzerland, 2–3 April 2012

V. Use of clinical/community trials and mathematical models for measuring the impact of TasP

Framework for Metrics to Support Effective Treatment as Prevention

Meeting Report – Geneva, Switzerland, 2–3 April 2012

Framework for Metrics to Support Effective Treatment as Prevention

Meeting Report – Geneva, Switzerland, 2–3 April 2012

VI. Assessing country readiness for adopting essential TasP metrics

What is the programme’s ability to disaggregate each indicator in the TasP cascade of service by key variables, e.g. age group, gender, clinical stage or “CD4 status” by key population prioritized for TasP? What is the status of implementation of EWI for HIV-DR and routine pharmacovigilance monitoring/reporting systems? »» »» »» Which sites participate? How complete is reporting? Which agency/organization is responsible for reviewing the data and initiating programmatic action?

•

What is the availability of CD4 count and viral load measurements as part of routine patient monitoring? »» »» »» Is CD4 count available at the time of diagnosis; if yes, at which sites? Is CD4 count available at the time of care or ART enrolment; if yes, at which sites? If viral load is not routinely measured, where are viral load measurements possible, and what options for specimen transport are feasible?

•

For each TasP indicator from routine monitoring systems, at what level are data entered in an electronic data system and »» »» »» »» How frequently are the data transmitted centrally? Which service provider sites do direct data entry and for which indicators? What is the geographical scope of facilities able to do direct data entry? Which data entry personnel are available at district, provincial/state, or national level to support data entry? Is it feasible to do line-listed data entry, or only aggregated data entry for each indicator?