Malawi - Aids Assessment Study (Vol. 2 of 2) : The Annexes

CONFIDENTIAL 17740-MAI VOL. 2 MALAWI AIDS ASSESSMENT STUDY ANNEXES Volume II MALAWI AIDS ASSESSMENT STUDY ANNEXES Volume II August 1998 This study was prepared jointly by the Government ofMalawi and the World Bank. ACRONYMS AND ABBREVIATIONS ADRA Adventist Development and Relief Agency MTEF Mid-Term Expenditure Framework AIDS Acquired Immune Deficiency Syndrome MTP Mid-Term Plan BLM Banja La Mtsogolo MOWYCS Ministry of Women, Youth CAPAR Canadian Physicians Agency for Refugees Affairs and Community Services CBD Community Based Distributors MOLGS Ministry Local Government and Sports CCA Community Change Agents MYP Malawi Young Pioneer CEP Community Empowerment Process NAC National AIDS Committee CBC Community-Based Care NACP National AIDS Control Program CHAM Community Health Association of Malawi NAS National AIDS Secretariat CHSU Community Health Services Unit NFWC National Family Welfare Council CNA Community Nexus Approach NGO Non-Governmental Organization CIDA Canadian International Development Assistance PSI Population Services International DAAC District AIDS Coordinating Committee PLWA Person Living with AIDS DAC District AIDS Coordinator RAC Regional AIDS Coordinator DFID Department for International Development RHO Regional Health Offices DHO District Health Officer RHU Reproductive Health Unit EU European Union SADC Southern African Development FLEP Family Life Education Program Community GDP Gross Domestic Product SCF Save the Children Fund GMS Government Medicine Stores STAFH Support to AIDS and Family Planning GNP Gross National Product STDs Sexually Transmitted Diseases GOM Government of Malawi SWCD Social Welfare and Community HBC Home Based Care Development HRMD Human Resources Management Development TB Tuberculosis HSA Health Surveillance Assistants UNAIDS United Nations Program on AIDS IEC Information, Education, Communication UNESCO United Nations Education, Scientific IEF International Eye Foundation & Cultural Organization LCH Lilongwe Central Hospital UNFPA United Nations Population Fund MACRO Malawi AIDS Counseling Resource UNICEF United Nations Children's Fund Organization USAID United States Agency for International MASAF Malawi Social Action Fund Development MBC Malawi Broadcasting Company VCT Voluntary Counseling and Testing MCH Maternal Child Welfare WHO World Health Organization MDHS Malawi Demographic Health Survey MK Malawi Kwacha MKAPHS Malawi Knowledge, Attitudes, Practices Health Survey MOE Ministry of Education MOR Management, Overhead and Recurrent Costs MOHP Ministry of Health and Population ANNEXES Page ANNEX A: Terms of Reference: Malawi AIDS Assessment Study..................................................... ANNEX B: The Impact of AIDS on Demographic Projections....................................... .....................6 ANNEX C: HIV and AIDS Prevalence in Malawi...............................................................................l ANNEX D: Impact of AIDS Compared to the Other Leading Causes of Illness and D eath in M alaw i......................... .. ................ ..... ........................ ..... ..... .14 ANNEX E: Total Expenditure (1996) On AIDS Control by Donor, Government and A ctivity C ategory.......................................................................................................17 ANNEX F: Which Interventions are Cost Effective?.......................................................................... 8 ANNEX G: Attitudes and Practices Among Youth in Consecutive Surveys.......................................33 ANNEX H: Reaching School Children with AIDS Prevention.................. ................. ......... ......34 ANNEX I: Testing the Efficacy of Voluntary HIV Counseling and Testing Services... ...................36 ANNEX J: The Cost-Effectiveness Of HIV Counseling And Testing................................................40 ANNEX K: Distribution Of Extension Staff In The Districts............................................. ................47 ANNEX L: Comparison of Neighboring Epidemics............................................................................48 ANNEX M: Evaluation In-Service Training In Syndromic Management Of ST s.............................70 ANNEX N: The Increased Health Costs from the AIDS Epidemic.................................................. ..74 A N N EX 0 : B lood Safety in M alaw i....................................................................................................76 ANNEX P: Health Care Conditions in Malawi........................................................................... .......79 ANNEX Q: Community-based Care in Malawi..................................................................................80 ANNEX R: Expenditures (1996) For Eight HBC Programs By Input Category................................86 ANNEX S: Description Of Six Home Based Care Programmes........................................................87 B IB L IO G R A PH8Y : ............................................................................................................................. 8 ANNEX A MALAWI AIDS ASSESSMENT STUDY TERMS OF REFERENCE OBJECTIVE: To undertake an assessment of the AIDS epidemic and the strategy to stop its spread in Malawi, with three outcomes: A. Consensus on 1) the factors influencing the spread of HIV; 2) the impact AIDS is having and will have on individuals, communities, and selected sectors and; 3) the interventions that are working best to prevent HIV and mitigate the adverse consequences of AIDS. B. Active participation of representatives from the public and private sector at the local and national levels in the assessment. C. Design of a proactive AIDS plan of action for subsequent government, World Bank and donor support. Four themes: prevention, quality of care, cost-effectiveness and greatest impact will guide the study's consensus building efforts in determining what interventions do and do not work. Thus, emphasis will be accorded to what interventions prevent the greatest number of HIV infections, provide the highest quality of care at the lowest cost and will have the greatest impact on slowing the epidemic. The study will assess geographic coverage and access, targeting of individuals most at risk, and opportunities gained in service delivery for the main interventions. The study will also consider what new interventions should be piloted and build upon the reviews and recommendations already made to date. APPROACH: The assessment team consultants will collect and analyze the relevant information and work closely with the appropriate parties to summarize best practices and lessons learned. A variety of informal meetings among NACP and UNAIDS staff, the consultants and a wide spectrum of representatives from the public and private community will assess the country's experience to date in preventing and mitigating the epidemic. OUTCOME: The AIDS Assessment Report will be written by the consultants for the national and international parties involved in AIDS prevention efforts in Malawi. The report will be directed towards an audience well informed about AIDS. Most likely the report will be organized into four main sections: 1) Understanding the True Magnitude of the Problem--an epidemiological description of the epidemic; 2) Ensuring that the Interventions Address the Risk Factors-- a summary of the known risk behaviors and assessment of the interventions' effectiveness to preventing the risk behaviors; 3) Understanding the Impact AIDS Has on Development and Ensuring the Interventions Mitigate the Adverse Effects--a description of the impact AIDS is having and will have on individuals, communities and selected sectors and assessment of the interventions' effectiveness in mitigating the impact; and 4) the plan of action. AREAS OF ANALYSIS: A. Understanding the True Magnitude of the Problem (epidemiological description) Al. Review the STI (HIV/AIDS and other STDs) epidemiological situation. Summarize STI trends in prevalence and if possible, incidence, and distribution in terms of sex, age, occupation, socio-economic status and geographic location. Identify the risk factors for STI transmission. Briefly review the modes of STI transmission. Briefly summarize the overall reproductive health status of the population. Briefly compare AIDS with the other main causes of mortality and morbidity. Answer the following questions: What is the estimated national prevalence among sexually active adults and all ages? What appears to be the cause for the plateau/decrease in HIV prevalence? What groups are most at risk? A2. Develop a monitoring and evaluation plan of action for the NACP. Assess the STI surveillance and reporting systems, take note of the previous recommendations made for their improvement and incorporate these recommendations into the plan of action. In collaboration with the other consultants, review and summarize in matrix format, the baseline- data available for measuring impact in STI levels and behavior change (use WHO Program Priority Indicators (PPI) as reference) and identify further evaluation activities in the action plan. Answer the following questions: What improvements in the STI surveillance and reporting systems can be made during the study? What is the sensitivity and specificity of the guidelines for diagnosing AIDS. Consultant: Jim Chin B. vinsuring That The Interventions Address The Risk Factors Bl. Summarize what is known about the sexual knowledge. attitudes and practices of the Malawi population with particular attention to youth and gender differences. Seek the answers to the following questions: What sexual networks exist in Malawi and what role does formal and informal commercial sex play in the epidemic? What role do cultural, socio-economic and gender factors play? Why are condems not used? Summarize what is documented about STI knowledge and myths, practice of hygiene and prevention, duration of STD symptoms before treatment. Ascertain what research on sexual behavior has been performed and identify areas requiring additional research. Work closely with the AIDS Sec., Center for Social Research, National Statistical Office, JSI-STAFH Research Unit, UNFPA and UNICEF. Note works by Helizter-Allen, Komfield, Namate, McAuliffe, abstracts from National Conference on AIDS, proceedings from First Annual STAFH Project Coordination Meeting, preliminary analysis from MKAPHS-97, and summary in the UNFPA Programme Review and Strategy Development Report. B2. Assess how well prevention programs use behavior change theory to guide interventio efforts. Answer these questions: What behavior change theories and approaches are used by the main prevention programs, including that of the National IEC Strategy, NACP, Project Hope's Training for Transformation Approach, and the Social Research Center's AIDS Risk Reduction Model? What monitoring and evaluation efforts are used to help gauge the model's effectiveness? B3. Describe the interventions to prevent sexual transmission of HIV supported by the main programs Answer the following questions: What interventions exist and are the interventions targeting the individuals most at risk and most involved in HIV transmission to others? What biological and socio-economic factors have the programs identified, influence behavior change? Are the messages and communication channels used in the interventions addressing the factors that influence behavior? What progress has been made in using -2- existing initiation rites to promote healthy and responsible sexual behavior? What condom promotion and distribution programs exist and do they meet all public and private needs? What rates of condom use among what type of individuals are required to significantly slow HIV spread? Are condom promotion strategies addressing the current resistance to condoms? What data suggest the interventions are or are not working? What interventions appear to be the most successful in changing behavior? What are their costs? Consultants: Mary Bassett. Anne Martin and/or Clement Chela on costing. B4. Summarize the data in Malawi that demonstrates the Impact STDs Have on HIV Transmission. What data exist to show STD facilitation of transmission? How are or can these data be used for decision-making? (Note study on viral load and presence of STD). B5. Describe the Interventions to Treat STDs. Assess the NACP's STD control program's status of implementation and effectiveness, with particular attention to availability of drugs and counseling. Include reference to STAFH's quality of STC care studies and others on sensitivity and specificity of syndromic approach. Answer these questions: What operations research has been performed to determine the most cost- effective approaches? What health-seeking behaviors have been documented? How can counseling and condom promotion efforts in STD facilities be more effective in prevention recurrent infections and subsequent HIV infection? Consultant: Sam Okware. Anne Martin and/or Clement Chela on costing. Brief discussion of the following or complete omission?: B6. What is known about the risk of non-circumcision among males and genital mutilation among females and HIV transmission? Additional questions to answer: Aside from the research performed by Moses, Plummer, Caldwell et al. on circumcision, what work has been performed in Malawi? Any data on practice and risk of genital mutilation among females and HIV transmission? What public health policy and prevention intervention strategies are needed in this regard? B7. What is the Impact of HIV-contaminated blood mode of transmission on the epidemic in Malawi? Additional questions to answer: Assess the effectiveness of the safe blood prevention strategy. What role do other modes play: contaminated needles/syringes, scarification, circumcision rites? B8. What role do other traditional practices play in HIV transmission. Summarize the traditional practices such as ritual cleansing, spouse inheritance heirship for chieftancy that exist and ? Interventions for prevention? Public policy and law required to change? B9. Summarize What Socio-economic Conditions Place Persons Vulnerable to HIV. Answer the following questions: What strong data exist demonstrating link between socio-economic status and STI levels? (See TOR above on risk factors.) Include reference to gender inequality, migration, urbanization, military mobilization, spouse separation, lack of economic and education opportunities in your assessment. What development strategies place persons more vulnerable to HIV in Malawi? BIO. Describe the Multi-sectoral interventions used to mitigate HIV vulnerability. Answer the following questions: What interventions provide employment alternatives for prostitutes, shorten the distances for long distance truck drivers, change harmful traditional practices? What World Bank/UNDP social funds/safety net programs should be linked to AIDS prevention activities (income generation for prostitutes)? -3- Consultants: The importance of covering these areas will be determined in Malawi in consultation with NACP staff. (The whole team would work together on the above 5 areas?) C. Understanding the Impact AIDS Has on Development and Ensuring the Interventions Mitigate the Adverse Effects Cl. Impact on individual and community. Answer the following: What data exist on life span of person with AIDS (PWAs) (i.e. number of episodes of illness, expenditures, savings, medical costs, burial costs, affect on family, labor, resources, moving to village, numbers of orphans now and projected?). How relevant is the UNICEF-Malawi Family and Community Care situation analysis of March 1994 today? Should another be performed? How many of the UNICEF recommendations were carried out? C2. Summarize the Impact of AIDS on Sectors (especially Health). Industry. Government. Answer the following: What sectors are most affected by AIDS (health, education, defense, and agriculture) and ascertain in what areas (recruitment, productivity, benefits, costs, training) impact is and will be felt. Regarding the health sector: Estimate number and direct costs of AIDS patients seeking care at all facilities and requiring assisted care at homes annually for next five years. Regarding agriculture: Since Malawi is an agrarian based economy, what data exist on AIDS impact on farming systems? Can Tabia's (year?) AIDS cost estimates used in FHI's study be updated ? C3. Describe the Interventions to Help Individual. Family. Communities Cope. Answer the following questions: What community-and facility-based care strategies exist? Make an inventory of HBC projects and summarize variations in approach.. How effective are they in quality of care (physical and psycho-social), coverage, access (gender specific) , costs? What is the availability of drugs? What operations research has heen performed to ascertain most cost-effective approaches? Is there a gold standard (model) of practice? What criteria are used to target beneficiaries for program assistance? What non-health related assistance ("program assistance") is provided (food, clothing, school fees, agricultural tools) What World Bank and UNDP social funds/safety net programs should include AIDS as criteria for assistance? What is being done with agricultural extension workers as to reproductive health messages that are gender sensitive? How is the District AIDS Coordinating Committee Initiative doing to facilitate coping mechanisms? What is status of response to External Review recommendations and proposed UNDP workshop?) Consultants: Sam Okaware, Anne Martin, Clement Chela C4. What Impact Will AIDS Have On Demographics and Macroeconomy. Describe briefly and project the change seen or projected in adult and child mortality, life expectancy, dependency ratio, age makeup, fertility. Refer to Lodh, FHI, Stover and Cunnington studies. Briefly discuss what effect AIDS will have on per capita GDP, GDP, via savings, productivity and labor loss, population growth. Consultants: Jim Chin and Anne Martin D. Assessing the Capacity of the NACP to Oversee the Nation's Response DI. Examine the NACP structure in relation to its function and goals. How well does the current structure of the NACP facilitate its ability to coordinate the multisectoral response to the epidemic? What are the weaknesses and strengthens in its current setup? What changes are recommended? Who is responsible for carrying out these changes? -4- D2. Asses the NACP's management of the nation's response. Ascertain the effectiveness of the NACP's procedures to plan, implement program activities and mobilize resources. Examine the NACP's coordination of donor and NGO contributions. What constraints exist? What actions should be taken to strengthen communication and management within the NACP and among all major players? D3. Examine the availability and utilization of human and financial resources in Malawi to respond to the epidemic's magnitude. What skills are abundant or deficient in Malawi in both the public and private sector response to the epidemic? What financial constraints exist? To what degree is availability of resources the problem as opposed to the effective utilization of existing resources? What actions are required to build capacity? Consultant: Amusaa Inambao WLR. June I1, 1997 -5- ANNEX B The Impact of AIDS on Demographic Projections In countries with relatively high HIV prevalence rates (i.e., >2 percent to 3 percent of the total adult population). AIDS will have a significant and measurable impact on mortality rates. In these countries, AIDS deaths will be the single largest determinant of life expectancy. In the absence of AIDS, the method for calculation of life expectancy was the use of model life tables that are region specific. However, because of the large impact of AIDS deaths in countries with high HIV prevalence, the recommended approach is to first develop a population projection that ignores the effect of AIDS, and then to make assumptions about the future level of adult HIV prevalence to calculate the effects of AIDS on population projections. Population projections for Malawi were made using the computer program, DemProj developed by the Futures Group International. DemProj requires information on the number of persons by age and gender in the base year, as well as current year data and future assumptions about the total fertility rate, the age distribution of fertility, life expectancy at birth by gender, the most appropriate model life table, and the magnitude and pattern of international migration. To model the impact of AIDS, DemProj also requires estimates and projections of the annual HIV prevalence rate in the population ? 15 years of age. Dr. John Stover of The Futures Group provided the baseline demographic file for Malawi in the absence of AIDS and a demographic projection file for Malawi that used annual HIV prevalence estimates and projections that were derived by the United States Bureau of the Census (USBOC) using the Interagency Working Group AIDS model (iwgAIDS). HIV Prevalence Estimates To use DemProj to assess the i, ipact of AIDS on demographic projections requires development of annual HIV prevalence estimates and projections. Figure BI presents three independent HIV prevalence estimates for Malawi that were used to project the impact of AIDS on demographic projections. The shapes of the estimated HIV prevalence curves are distinctly different. Dr. Bimal Lodh is an economist who used HIV serosurvey data from Malawi to develop district specific HIV prevalence estimates in 1995. He used Epimodel to calcula'- the annual HIV incidence from the start of the HIV epidemic in Malawi. After 1996 he departed from Epimodel's default HIV prevalence curve and decreased HIV prevalence until it reached zero after the year 2010. Dr. Lodh then inputted the annual HIV prevalence estimates he crafted into an earlier version of DemProj to derive demographic projections for his economic calculations. -6- Figure B1. Estimatecl/Projected Annual Prevalence of Adult HIV Infections - Malawi Percent positive Percent positive 36 36 30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .30 0 * 84 86 88 90 92 94 96 98 00 02 04 06 08 10 JX16-22-97 Year The USBOC used the iwgAIDS model (a complex deterministic HIWAIDS model) to estimate and project the annual adult HIV seroprevalence rate in Malawi. The iwgAIDS model projected that adult HIV prevalence rates would rise from about 20 percent in 1996 to about 35 percent in the year 2010. The conservative HIV scenario used by Epimodel used the 1996 seroprevalence estimate derived from the available HIV serosurvey data (11.4 percent for the population 15 years of age) and projected that the adult HIV prevalence in 2010 would be less than 10 percent. The demographic projections using the Epimodel/DemProj (Epi/Dem) programs compared with the USBOC are, as expected, very different. The demographic impact of AIDS projected by the USBOC are in general about 2 to 3 times greater than that obtained by use of Epi/Dem - primarily because the HIV seroprevalence levels used for the USBOC projections were 2 to 3 times higher than the conservative HIV scenario developed using Epimodel. The impact of AIDS on several major demographic parameters as modeled by the USBOC and by Epi/Dem are described below. 1. Rate of Natural Increase - The rate of natural increase (expressed as a percent) can be used to analyze the effect of AIDS on population size. The rate of natural increase for Malawi is projected to decrease more rapidly in the presence of AIDS. As a result of AIDS according to Epi/Dem, the growth rate is projected to decline from 3.3 in 1985 to 2.5 in 2010; without AIDS the growth rate would decline more slowly from 3.3 in 1985 to about 3.1 in 2010. According to the USBOC's projection, the RNI would fall sharply to 0.55 in 2010, as shown in Figure B2. -7- Figure B2. Rate of Natural Increase (Percent per year) - Malawi RNI RNI 3.6 3.6 NoAID S 3 2 9 . . . . . . . ... .. . .. . .. . .. . .. . 2.6 4*1 Epi/DemProj 1.6 . . . . . . . . . . . . . . . . . . . .44 .. . . . . . . . . . . . 1.6 USBOC 0. . .0.........- .- .- .- . ..-.-................-- .- - .- ................ 0.5 0 0 84 86 88 90 92 94 96 98 00 02 04 06 08 10 Malettom7 . Year 2. Age-Specific Mortality Rates - Infant and under five mortality rates are of particular interest in assessing the pandemic's impact on children; while for adults, mortality rates for those aged 15-49 need to be examined. The crude death rate for Malawi was projected to fall from 21.3 per 1000 population in 1985 to 10.6 per 1000 population in 2010 in the absence of AIDS. However with AIDS, the crude death rate is projected by Epi/Dem to fall to only 16.6 by 2010--about half the decrease without AIDS. The USBOC's projection has the crude death rate rising from 21.3 in 1985 to 34.4 in 2010. Infant mortality rates in the absence of AIDS would ha*e been expected to fall from 150.6 in 1985 to 82.2 in 2010, but according to Epi/Dem will fall to only 91.1 per 1000 live births by 2010. The USBOC's projection has the infant mortality rate falling only to 119.4 by the year 2010. The probability of dying by age five in the absence of AIDS was expected to fall from 226.4 per 1000 live births in 1985 to 117.7; with AIDS according to the Epi/Dem projection it is expected to decrease from 226.4 in 1985 to only 141 7 in 2010. The USBOC's projection has the child mortality rate falling only slightly to 206.9 by the year 2010. 3. Life Expectancy at Birth - The effect of AIDS on average length of life can be determined by comparing life expectancy at birth in situations for which AIDS either is or is not considered. The difference between the two averages represents life expectancy lost due to AIDS. DemProj estimated that a Malawian was expected to live 45.2 years in 1985- close to 30 years less than a person in a developed country--but in the absence of AIDS, the life expectancy was projected a rise to 57.4 years by 2010. With AIDS, according to the Epi/Dem projection, life expectancy is expected to decrease slightly to 44.1 in 2010, or by more than 13 years compared to projected life expectancy in the absence of AIDS. The USBOC's projection has the life -8- expectancy falling by 20 years in 1997 (Epi/Dem projects a nine year decrease in 1997) and over 40 years by 2010, as shown in Figure B3. Figure B3. Projections of Life Expectancy - Malawi Life expectancy Life expectancy so ---o 1-9 40. ... . . .4 f -20 EpilDemProj 20 usoc 0'0 88 88 90 92 94 96 98 00 02 04 06 08 10 t Year O 4. Dependency Ratio - The dependency ratio is used to analyze the economic implications of changes in age distribution: a higher dependency ratio indicates a heavier burden on persons aged 15-64. According to DemProj's projections for Malawi, AIDS deaths are expected to reinforce the declining trend in the dependency ratio which was calculated to be 1.0 in 1985. By the year 2010 this ratio without AIDS is projected to be 0 91 and with AIDS the projected ratio (by both Epi/Dem and the USBOC) is projected to be 0.89. However, this result must be interpreted cautiously since thousands of deaths in adults at productive ages will show little effect on the dependency ratio when the size of the reference population is several million or more. An analysis of the expected number of orphans, for example, is needed to examine more fully the increasing social and economic burden on certain segments of the population as a consequence of the age-specific effects of AIDS. 5. Total Population - The total population size with or without AIDS is a demographic parameter that is easy to understand. In the absence of AIDS, the DemProj program projected that the current (1997) population size of about 10 million would increase to 16.7 million in the year 2010. According to the conservative Epi/Dem projection, there will be close to 2 million less persons in the year 2010, and about 5 million less according to the USBOC's projection, as shown in Figure B4. -9- Figure B4. Population Projections - Malawi Millions Millions NOAIDS 16.7 1 Ep. /Dem USBOC 11. 10. 100 1986 1990 1996 2000 2006 2010 Ia Year - 10- ANNEX C HIV and AIDS Prevalence Data in Malawi' HIV Prevalence Surveillance on HIV prevalence among antenatal care attendees has been conducted annually in 19 sites in Malawi since 1992. In a few sentinel sites annual testing began earlier than 1992; Blantyre began testing in 1985. Figure C.1 below, shows that in the few sentinel sites with available data across years, prevalence grew quickly in the 1980s and early 1990s. (See the main text, Chapter 1 for data on the 1996 survey of sentinel sites.) Figure C1 Fig 1. Trends in HIV seroprevalence among women attending antenatal care 35 30- ,~.4 '-- QUEC 30 - - - - - - - - - - - - - - - - - - - - - - - L C H 20 ------------------- ----- ----- -- StJohn 20 --- -------M c h in j 15 - mulanje Ksoba 10 - ---- - - -------- -------- - - - - - 4- Kam boni 5 ------ ------------------------- -- -- - - - M dlepa 0 , I T honje LO (0 r- C O a 0 - Ln') ~ U CO CO 0 0 00 CO 0) 0 ) 0 ) 0 Year Among antenatal care attendees and women giving birth in QECH in Blantyre the HIV prevalence increased between 1990 and 1993, but has slowed between 1993 and 1996 (Figure C2, next page). Information and survey data in this annex were summarizea oy a variety of reports and correspondence located in the NAS files. -11- Figure C2 SERIAL SEROPREVALENCE OF HIV-1 %o AMONG PREGNANT WOMEN, QECH, MALAWI 40- 35- 32.8 30.1 30.2 30- 25.9 27.2 25- 23 - 18.6 20. 10 8.2 5 2%- 1985 1987 1988 1990 1991 1992 1993 1994 1996 N = 200 85 247 6603 404 291 2471 6929 829 The highest age-specific HIV prevalence has shifted over time from younger to older age groups, (highest HIV prevalence between 20 and 24 years of age in 1990, between 25 and 29 years of age in 1993, and between 30 and 34 years of age in 1996). A decline of HIV prevalence among first year college students has been documented from 5.2 percent in 1993 to 2.2 percent in 1994. (Liomba, Taha et al. June 1996) The highest rates of HIV prevalence in Malawi are found in the southern region, where urbanization, transportation and migration are greatest. Studies suggest education status, particularly among males, is positively associated with HIV status. Professionals, skilled workers and the armed forces have higher rates of HIV than farmers, villagers and housewives. Yet most of the persons infected with HIV in Malawi are poor, since most Malawians are poor. (USAID, 1993; Dallabetta, and Miotti et al. 1994;) A'though most developing countries use IV prevalence data from sentinel sites among antenatal facilities, there are some limitations to these convenient data: The limitations are the following: * the sentinel sites are not randomly selected; * the sample size is inadequate to check for trends at all sites; * the antenatal attendees tested may not be representative of all antenatal attendees in Malawi; * data on HIV prevalence on men are absent; * data on age groups below 15 and above 40 are absent; -12- * the blood samples were tested for HIV using Elisa without other confirmatory tests and the false positive rate is not known. However, recent data from Uganda and Zambia suggest these rates approximate those for the general adult population (Wawer, et al., 1998). AIDS Prevalence AIDS cases are reported by 61 health institutions involved in the AIDS reporting system Between 1985, when the first AIDS case was diagnosed in Malawi, and June 1996, a total of 43,067 cases have been reported. The Malawi MOHP follows the WHO criteria for the diagnosis of AIDS but requires a positive serology result for reporting AIDS cases. As mentioned in the main text of this report, several reasons exist for the low level of reporting of AIDS cases. The lack of reagents and reluctance by providers to perform pre-test counseling act as disincentives to reporting Other reasons include the small number of persons with AIDS seeking care at the health units, the irregularity with which the 61 health units included in the surveillance system send their returns, the incomplete information on the returns, and lack of interest by the medical personnel in filling the forms. Even when patients die from AIDS, it is common for providers to record other causes of death because of the stigma attached to AIDS. Among those cases reported, AIDS is concentrated in the under five years of age and sexually active group The age distribution is characterized by a peak below 5 years of age, a decline between 5 and 14 years of age, an increase between 15 and 49 years of age and a decline afterwards. The ratio of AIDS cases between males and females is nearly 1, but more females are affected in the age group 15-29 and more males are affected after the age of 30. One possible explanation for this sex difference is a transmission of HIV from older males to younger gir,s. According to the NAS, and based on interviews with AIDS patients, 90 percent of cases are estimated to stem from heterosexual transmission; about 8 percent form mother to child transmission; and the remaining 2 percent from blood transfusion, unclean needles and unsterilized medical equipment. -13- ANNEX D Impact of AIDS Compared to the Other Leading Causes of Illness and Death in Malawi2 Morbidity Estimating the impact of AIDS on both morbidity and mortality in any country in Africa is difficult because of unreliable reporting and diagnosis, and paucity of studies. In Malawi, data on morbidity can be found from two sources: 1) reported illnesses from public and private health facilities; and 2) surveys from small samples of select populations and large samples of household populations. Comparisons among numbers of illness reported from health facilities are not recommended because of the incompleteness and unreliability of the data. And comparisons can not be made among household population responses on frequency of illnesses in surveys, because survey questions on AIDS have not posed. Data on illnesses reported from health facilities are incomplete and out of date. Malaria, upper respiratory infections and diarrhea are the most common illnesses reported in the country. The table below provides the most recent (1994) data available on causes of out-patient attendances in Malawi. Table DI Top Five Causes of Out-patient Attendances, 1994 Illness Number (000s) % Malaria 6,977 33 Upper Respiratory 2,857 14 Infections Diarrhea 1,465 7 Abdominal complaints 1,146 5 Other respiratory infections 858 4 All others 7,769 37 Source: CHSU (1994) Basic Health Statistics Report, MOHP Reports on AIDS cases are even more incomplete because of the requirement for an HIV positive serology and pre-and post test counseling. Thus, most persons with AIDS are diagnosed and reported as the 2 A portion of the discussion above was taken from two papers presented at the World Bank, UNAIDS and IUSSP Committee on AIDS Workshop on the demographic impact of HIV/AIDS, Washington, DC 12-14 January, 1998.: "The impact of HIV/AIDS on mortality To sub-Saharan Africa: evidence from national surveys and censuses." Ian M Timaeus, Centre for Population Studies, London School of Hygiene & Tropical Medicine. "Projecting the Impact of AIDS on Mortality." John Stover, The Futures Group International, and Peter Way, International Programs Center, US Census Bureau. -14- "AIDS opportunistic infection" causing the overt symptoms at the time of examination--TB, pneumonia and diarrhea. A recent sero-prevalence study in Malawi showed two thirds of TB cases were HIV-infected. Because of this lack of reliable data for AIDS as well as for the other diseases, comparing these reported cases to one another is not recommended. As to data on morbidity from surveys, forty-five percent of children under five were said by caretakers in households, to have had symptoms (fever) of malaria in the two weeks preceding the 1996 MKAPH Survey. Prevalence of fever peaks at 54 percent in the 6-23 month age group. Sixteen percent of children 0-59 months of age were said to have diarrhoea. And 12 percent of children age 0-59 months were said by caretakers to have symptoms (fast or difficult breathing due to chest problems) of ARI during the two weeks preceding the survey . It should be borne in mind that morbidity data collected in surveys depend on the subjective judgment of informants, in this case caretakers of children under five, and are not validated by medical personnel. No questions in the 1996 MKAPH Survey were asked about AIDS-related illnesses for either children or adults, thus no comparison can be made. Mortality The main sources of information on mortality in Malawi are national censuses and household surveys, for no vital registration system exists in the country. Data from these sources show AIDS to be the most important determinant of adult mortality in SSA.. Under-five mortality n Malawi has declined slowly in the 80's compared to rates in Kenya and Zimbabwe. The subsequent slowdown or reversal of the decline in infant and child mortality in the early 90's in Kenya, Malawi, Rwanda, Uganda, Zambia, and Zimbabwe is most likely a result of the HIV epidemic.3 HIV may have become sufficiently prevalent by the mid 1980's that by the early 1990s, pediatric AIDS mortality in these five countries could have risen to 25-30 per 1000 births. Recent DHS surveys in Zimbabwe and Kenya show this increase. Collecting data on adult mortality retrospectively in censuses and surveys has proved more challenging than collecting data on under-five mortality. In Malawi, the mortality of both men and women rose in the 1980s, although the data from this period are insufficiently detailed to indicate exactly when this trend developed. A more recent source of data now available is the sibling histories collected as part of the maternal morality module of phase 2 and phase 3 DHS surveys (1992-1997). These surveys asked women about the birth and death of each of their siblings. The results from the sibling histories are striking. According to these, huge increases in adult mortality occurred in Uganda, Zambia and Zimbabwe between late 1980's and mid- 1990s. Although adult mortality in Malawi remained fairly high in the 1980's, this has not stopped mortality from rising further. Over a six year period, men's death rates rose by about 55 percent and women's deaths by 15 percent (figure DI below). 3 If about one-third of infected women giving birth transmit the HIV virus to their child one would expect each I percent rise in seroprevalence among women attending antenatal clinic to raise the under five mortality rate by about 3 per 1000. 6.4 percent: rate could rise by about 20 per 1000. 20 percent, increase in rate of nearly 60 per 1000. -15- Figure D1 Percentage change in the adult age-specific death rates by sex, DHS sibling histories 300% - 250% - 200% - 150% - aMe 100% -: 0% 0O Women 50%- 0% a - -50% - -100%- -150% j -I C NNOZ Wa Z Cn WN CLD Source: Timaeus, 1998 In contrast, the sibling histories data for four Western African countries (Mali, Niger, Senegal and Benin) where HIV prevalence is low, do not suggest that mortality in these countries is rising. Thus, mortality has risen more in the countries where HIV seroprevalence is known or thought to be high, and the differential trends in adult mortality revealed by these surveys are very consistent with what is known about the spread of HIV in Africa. Furthermore, the evidence on mortality trends from these DHS sibling histories agrees closely with the evidence of earlier inquiries and recent data on orphanhood. In summary, in Malawi (and most of Africa), the severity of the AIDS epidemic is now the dominant determinant of adult mortality. AIDS mortality is overwhelming the progress made up to the 80's in controlling infectious disease mortality. -16- ANNEX E 1996 TOTAL EXPENDITURE (in US$) ON AIDS CONTROL BY DONOR, GOVERNMENT AND ACTIVITY CATEGORY FUNDING CATEGORIES SOURCE YOUTH CBC STD/AID NACP/RHO COM.SEX IEC CONDOM OTHER TOTAL S SUPPORT WORK UNICEF 217,694 238,495 24,026 41.505 521,720 UNDP 186,750 100,000 115,000 98,250 500,000 USAID/STAFH 341,852 392,097 1,465,653 15,000 127,498 300,00 1,025.000 3,332,900 7,000.000 ODA 78,720 53,300 73,800 37,720 106,600 46,740 396,880 EU/EEC 334,149 168,339 81,121 34.003 745,389 1.363.001 UNFPA* __868 868 ACTION AID 10,496 62,092 72,588 GOM 57,458 190,911 35,080 283,449 OTHER 149,722 4,000 27,717 181,439 Project Hope 225.359 225,359 TOTAL 972,415 1,030,8 1,869,250 491,495 161,501 300,00 1,131,600 4.588.183 10,545,304 60 1 0 *This figure is projected to increase to at least $5 Million in 1997 -17- ANNEX F CONFRONTING AIDS WHICH INTERVENTIONS ARE COST-EFFECTIVE?' Which Interventions Are Cost-Effective? Only a few HIV/AIDS interventions have been rigorously evaluated with respect to their impact on the incidence and prevalence of HIV; among those evaluated, interventions targered to those who practice high-risk behavior rend to be more effective (Aral and Peterman 1996; Choi and Coates 1994; National Research Council 1996; Oakley, Fullerton, and Holland 1995). Appendix A of this report presents the results of 22 of the more rigorous evaluations of HIV/AIDS interven- tions that have taken place in developing countries. Unfortunately, in- formation on the costs of these interventions is usually unavailable; their cost-effectiveness has rarely been evaluated. The impact of alternative prevention strategies in four epi- demics. The effectiveness of alternative interventions will be strongly influenced by the nature of the intervention itself and by the herero- geneity of the behavior that is fueling the epidemic. To illustrate this point, Van Vlier and others (1997) have simulated the impact of increased condom use and increased treatment of curable STDs (chlamydia, gonorrhea, and syphilis), on a heterosexual HIV epidemic in the four hypothetical populations described in chapter 2 using the STDSIM simulation package.6 The simulations show the impact of increased condom use and STD treatment in various groups in each of the four populations fifteen years after the start of the epidemic. (As seen in chapter 2, in each population earlier intervention among those most likely to contract and spread HIV would be more effective than the later behavior change discussed here.) The simulations show the impact of increased condom use in three groups of people with different rates of partner change-female sex The following is an excerpt from Confronting AIDS- Public Priorities an a Global Epidemic' A World Bank Policy Research Report. Oxford University Press 1997. pp. 139 - 153 -18- workers, men with casual or commercial partners, and women in stable relationships. These groups are commonly the focus, respectively, of out- reach programs to sex workers, socially marketed condom programs, and reproductive health services. In these simulations, the term " sex worker" refers to women with the highest rate of partner change- 10 new part- ners per week, or more than 500 partners per year. In the real world, of course, some women who have very high rates of partner change do not regard themselves as "sex workers" and they may contact their male partners in a variety of settings. There are other men and women in these imaginary populations with large numbers of partners, but fewer than 500 per year. The impact of different interventions in all of these groups can be simulated, individually and simultaneously. However, for expositional purposes, we present simulated interventions with only three groups. In the baseline scenario, before any intervention, we assume that only 20 percent of sex workers and 5 percent of men having sex with ca- sual or commercial partners use condoms consistently, that is, in every act of intercourse. We also assume that none of the women in steady re- lationships are using condoms.17 The simulations show the impact of instantaneously raising consistent condom use among sex workers to 90 percent, and among the other two groups to 20 percent.18 These levels were selected because the authors believed they are realistically achiev- able in some countries. In other countries, it may be possible to ex- ceed the levels of condom use simulated here. Among those who use condoms, the failure rate through breakage or misuse is assumed to be 5 percent. The simulations also show the impact of increased STD treatment on HIV prevalence. The baseline scenario assumes that 25 percent of all STD cases that produce symptoms are effectively created and that there is no specific screening or treatment program for sex workers. The sim- ulations show the impact of increasing the share of symptomatic STDs created to 75 percent in the general population and, in a separate sce- nario, the impact of implementing a monthly screening and treatment program that covers 90 percent of sex workers. In the latter intervention, it is assumed that 5 percent of sex workers are not cured. The assump- tions behind the baseline scenario and the five condom and STD inter- ventions are summarized in table 3.4. The simulated impact of increased condom use and STD treatment on the HIV prevalence of adults in the four hypothetical populations is -19- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIVIAIDS Table 3.4 Sununary of the Assumptions before and after Intentions, STDSIM Modeling (percent) Anwmp:ion Baseine internsvion Using cndoms corirarv Se workers 20 90 Men with casual or commercial partners 5 20 Women ages 15-50 in stable relationships 0 20 Sympsomatic ST trated 25 75 Sex workers idih monthly screening and tyramentfor STDs 0 90 Source- Bacrundpap, Van Vher and others 1997. shown in figure 3.3. Despite the different underlying patterns of sexual behavior, the impact of specific interventions shows some striking con- sistencies across populations: w Achieving 90 percent condom use among sex workers results in a dramatic drop in HIV prevalence in all three populations where there is commercial sex (a, b, d), even though sex workers comprise only a very small share of each population (0.25 percent or fewer women). STD screening and treatment for sex workers is far less powerful. n Increased STD treatment among the general population is less effective than raising condom use among those with many part- ners. This is not surprising, since people with high-risk behavior generate a disproportionate number of STD cases, and condoms prevent transmission of both HIV and STDs. STD treatment among the general population and condom use by women in sta- ble relationships have the largest impact in the populations with concurrent casual sex (b, c). a The impact of greater condom use among women in stable rela- tionships is very small, and in the serial monogamy population (d) it has almost no impact on the epidemic. In the population where the epidemic is driven by commercial sex (a), condom use by monogamous women slightly accelerates a decline in HIV preva- lence, while in the other two populations it merely slows the growth of a still-expanding epidemic. -20- CUNFRONTING AIDS PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 3.3 The Impact of Changes in Condom Use and STD Treatment in Four Populations with Different Pattens of Sexual Behavior (a) Commercial sex only (b) Commercial and casual sex Adult HIV Prevalence (%) Adult HIV prevalence (%) - Baseline 30 - 90% STD treatment, as sex workers i ........... 75% symptomatic STD treatment, general 20 population **** 90% condom use. o is sex workers .. *... 20% condom use, men 'N .******... In commercial and *, casual sex 20% condom use, women In stable relationships o 0V,a..........t.l.....i..i.l.o 0 5 10 15 20 25 30 0 s 10 i5 20 25 30 Years since the beginning of the epidemic Years since the beginning of the epidemic (c) Casual sex only (d) Serial monogamy Adult HIV prevalence (%) Adult HIV prevalence (%) 30 2 25 15 20 Raising condom use among sex 1 '0 workers to 90 percent was the most effective way to reduce 10 HIV prevalence in three of the four simulated populations; in s the remaining simulated popula- tion there is no commercial sex. ........" .. ...".o...0. .."........ 0 5 10 15 20 25 30 0 5 £0 Is 20 25 30 Years since the beginning of the epidemic Years since the beginning of the epidemic Source Background paper. V2n Vher and others 1997 Looking at the ranking of interventions within specific populations, we see that: M In the epidemic fueled by commercial sex alone (a), all of the in- terventions produce an absolute decline in HIV prevalence; in the epidemic fueled solely by casual sex (c), none of the simulated in- cerventions is sufficient to cause an absolute decline. -21- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTI\G HIV/AIDS " In the population with concurrent commercial and casual sex (b), prevalence declines in response to 90 percent condom use among sex workers; 20 percent condom use by men in casual sex keeps prevalence from rising. Other interventions in this population merely slow the expanding epidemic. " In the population practicing serial monogamy (d), increased con- dom use by sex workers is the only intervention that results in an absolute decline in HIV prevalence. In reality, no intervention is an "either/or" proposition. There is al- ways spillover between interventions intended for people with different rates of partner change, so any particular intervention is likely to change behavior in more than one group of people but to differing degrees. Fur- ther, combined interventions to multiple groups will have greater impact than single interventions; for example, working only with sex workers to raise condom use will be less effective than simultaneously working with both sex workers and their clients. However, these simulations show that the greatest share of the impact will be achieved through the interven- tions that succeed in changing the behavior of those with the highest rates of partner change. The broad conclusion that can be -drawn from these simulations is that, although the overall pattern of sexual behavior in the population does affect the impact of interventions, prevention of infection among those with the highest rates of partner change has a large effect irrespec- tive of the underlying patterns of sexual behavior in the population. As- suming that it is not a great deal more expensive to increase rates oFcon- dom use among those with high rates of partner change-sex workers or others-fG'using condom subsidies and promotion efforts on changing their behavior is likely to be highly cost-effective. Studies of the cost-effectiveness of HIV interventions in develop- ing countries are rare and not transferable. Fewer than half a dozen studies have documented the costs and effects of preventive interven- tions in developing countries (Beal, Bontinck, and Fransen 1992; Gilson and others 1996; Moses and others 1991). An overview of the results of several cost-effectiveness studies in developing countries is in appendix B of this report. Most evaluation studies measure impact by changes in in- termediate behaviors that are believed to affect risk-such as the increase in condom use, or knowledge about HIV prevention, or the number of people receiving sterile syringes. The number of HIV infections averted -22- CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC is then extrapolated on the basis of assumptions about the relationship between the behavior and HIV incidence. However, the lack of good data on sexual behavior and on the relation between sexual behavior and incidence makes these benefits very difficult to assess. Virtually no stud- ies, except those based on simulations, have measured the impact of in- terventions on secondary cases where they are thought to exist (Over and Piot 1996; Stover and Way 1995; background paper, Van Vliet and oth- ers 1997). None to our knowledge have taken into account the external benefits of interventions or the issue of complementarities between interventions. While cost-effectiveness studies can be very useful in deciding among alternative interventions in a given setting and stage of the epidemic, their conclusions are usually not easily transferable to other settings (background paper, Mills and Watts 1996). For example, an evaluation of the effectiveness of enhanced treatment of symptomatic STDs in reduc- ing HIV incidence in rural Mwanza Region of Tanzania found that the intervention lowered HIV incidence by 42 percent at a cost of roughly $10 per person treated, or $234 per primary HIV infection averted (Gilson and others 1996; Richard Hayes, personal communication).'9 However, treatment costs clearly could be much higher in a middle- income country, and effectiveness may have been quite different in an area with lower HIV prevalence than the Tanzanian study site (4 percent of adults were infected).20 Moreover, without estimates of the costs and impact of alternative interventions in the same area, we cannot say whether a particular intervention is more or less cost-effective than other interventions in reducing HIV transmission. Ideally, we would like to know the costs and effects of alternative interventions implemented in the same setti..g, but this has rarely been done (box 3.9). Interventions focused on those most likely to contract and spread HIV should be more cost-effective from the public perspective, because preventing infection in a person with risky behavior prevents many sec- ondary infections among individuals with whom they mix-some of whom practice high-risk behavior and some of whom practice lower-risk behavior. In fact, the extent to which ongoing programs affect those who practice high-risk behavior is often unknown. For example, condom so- cial marketing programs improve the access of the poor to condoms, but it is still not known to what extent these condoms are used by those in the highest-risk groups. Information on the level, distribution, and type of high-risk behavior, the number of people involved; and their charac- -23- EFFICIENT AND EQLITABLE STRATEGIES FOR PRE% E;TING HIV;AIDS Box 3.9 The Cost-Effectiveness of Prevention among Those with the Highest Risk HOW MANY HIV INFECTIONS COULD BE AVERTED attending STD clinics in California (low risk). Very by spending an additional $1 million per year on low risk includes most of the general population, prevention in groups with different levels of risk of including women of childbearing age in 41 of the HIV infection? The answers generated in a recent country's 50 states. study in the United States demonstrate the high The results of the exercise depend on the as- cost-effectiveness of focusing prevention expenditure sumptions with respect to the costs and impact of on those most likely to contract and spread HIV, as different interventions. The author initially assumed well as the additional benefits from early interven- that prevention costs for one individual in any given tion (Kahn 1996). group would be $200 per year, then examined the The study defined four risk groups in terms of sensitivity of the results to this assumption. This the level of steady-state HIV prevalence they would $200-per-year figure was based on a survey of the attain without intervention-high risk (50 percent annual cost of various prevention interventions for prevalence); medium risk (15 percent); low risk high- and low-risk groups in the United States I In (1 percent); and very low risk (0.1 percent). Steady- terms of the impact of interventions, the author state HIV prevalence is defined as the point at assumed that interventions lower risk by 10 percent, which the number of new infections exactly matches which he believes is a conservative estimate. the number of people exiting the group through Box table 3.9 shows the number of HIV infec- death or eliminating the risk factor (for example, tions prevented, given these assumptions, by spend- stopping injecting drug use). Examples of these ing $1 million on each of the several groups; either groups in the United States include young homosex- with late intervention (after steady-state prevalence ual men in San Francisco (high risk), injecting drug has been reached) or with early intervention (before users in San Francisco (medium risk), and women (Box continues on the folowng page.) Box Table 3.9 HW Infections Averted by $1 Million Annual Spending on Prevention, U.S. Estimates HIV infections averted Baseline HIV 5-year 20-year Risk group prevalence (%)horizon oorzon High-rfsk Steady state 50 164 681 Pre-steady stare 10 93 837 Medium risk Steady state 15 58 348 Pre-steady stare 3 14 112 Low rirk Steady state 1 4 262 Pre-steady stare 0.2 0.8 5.4 Very low risk Steady state 0.1 0.4 2.6 Source Kahn 1996. -24- CONFRONTING AIDS- PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 39 (c-Iangua steady-state prevalence has been reached). The averted for the low-risk steady state rises to only 18 impact of each intervention is shown for a five-year and 93, respectively. Even if prevention is substan- time horizon and for a 20-year time horizon. One tially more successful at changing behavior in the million dollars prevents the most cases if focused on low-risk groups, the higher effectiveness of preven- the highest-risk group in the steady state. However, tion in high-risk groups remains. the benefits of intervening early with this group only While interventions in high-risk groups are more become evident in the 20-year time horizon. effective, they are also potentially more costly. These figures understate the impact of preven- However, the study estimated that interventions in tion in high-risk groups, however, because sec- the low-risk groups (steady state) would have to ondary inficdons prevented in the partners or chil- be roughly one-fortieth to one two-hundredth dren of people in high-risk groups are not included. (1/40-1/200) the cost of an intervention in high- The number of infections averted in low-risk groups risk groups (steady state) to prevent the same num- will not be much affected by this omission, but ber of infections averted by intervening in high-risk among high-risk groups the total of averted infec- groups in the steady state. In other words, in the tions might be several times higher, depending on high-risk steady state, interventions in the low-risk the group and the degree of sexual mixing with population would have to cost $1 to $5 per person lower-risk groups. per year, compared with $200 per person per year in The result-that prevention in high-risk groups the high-risk population, to prevent an equivalent is most cost-effective--is robust to large changes in number of HIV infccuons for a $1 million outlay. the assumptions about the effectiveness of the inter- vention. If programs reduce risky behavior by 50 I7e programs and ca per person per year included- annual percent instead of 10 percent, the number of infec- testing and counseling ($40-110); bleach distribution and out- tions averted in the high-risk steady state rises to reach ($60); 3-session group counseling for IDUs ($75); needle 80for the five-year simulation and 3,750 for the xchange ($40-800); peer workshops for high-risk gay men 830 o($250); 5-session counseling for low-risk women ($269); 12- 20-year simulation, while the number of infections session counseling for medium-risk gay men ($470). teristics is a basic public good and will enhance efforts to improve cost- effectiveness by helping to improve the targeting of programs. Cost-effectiveness and the accessibility of target populations. Although it is highly desirable to focus public interventions on those who are most likely to contract and spread HIV, identifying and reach- ing these individuals can be difficult, especially where legal sanctions and social stigma cause these people-to want to avoid being discovered. The costs of reaching those most likely to contract and spread the virus can have a significant impact on the cost-effectiveness of interventions. Figure 3.4 shows a stylized classification of groups of people accord- ing to the extent to which they practice high-risk behavior and their pre- -25- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVE\TING HIVIAIDS Figure 3.4 Classilfication of Groups by Riskiness of Their Behavior and Their Accessibility Higher-risk behavior Sex workers in brothels Streetwalkers IDUs in treatment programs Out-of-treatment. IDUs Prisoners Homosexual/bilsexual men Military, police, sailors Street children STD patients Truck drivers, bar workers More difficult ___ Easier access access Government employees Elderly in rural areas Employees of large firms Women attending antenatal clinics Children in school Lower-risk behavior IDL Injecting drug user Soce Adapted from Adler and others 1996. figure 8. Used by permission sumed accessibility. Of course, the extent to which members of these Prevention programs should focus on identifiable groups practice risky behavior varies considerably across- people most likely to contract and spread HIV; some of these individuals settings and with the effectiveness of prior prevention efforts. Thus, such are easy to identify and reach, others a figure would need to be modified according to the situation in a spe- less so. cific country, on the basis of the results of HIV and behavioral monitor- ing systems. In the upper-right quadrant are groups with higher-risk behavior that are relatively easy for government agencies and collaborating prevention -26- CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC partners to reach. The benefits of behavior change in these individuals are relatively high, especially given the large number of secondary infec- tions generated, while the costs of locating them will be relatively low, enhancing cost-effectiveness. In the upper left quadrant are groups who also practice high-risk behavior but who are less easy to reach. In these instances the benefits of behavior change will again be great but the costs of locating and working with these individuals may be high, reducing the net benefits. In the lower-right quadrant are people who, on average, are presumed to practice lower-risk behavior but to whom access is easy. The benefits of intervening in these groups may not be great, but the costs of reaching them may be very low. Inexpensive interventions for these groups may still be cost-effective relative to some alternatives (box 3.10). The lower left quadrant includes people who are very unlikely to contract and spread HIV and are very difficult and costly to reach; of the four types of groups, this is the lowest priority for public-sector HIV pre- vention efforts. Note that the accessibility of most groups can be im- proved through government actions to reduce stigma, decriminalize be- havior, and educate the public on the nontransmissability of HIV by casual contact and the benefits of working with these groups. Of course, the "groups" identified in figure 3.4 are not homogeneous with respect to their risky behavior. Since individuals with high-risk be- havior cannot be easily identified, programs need to focus interventions on people with characteristics that are highly correlated with risky be- havior. However, some sex workers consistently use condoms, and some government employees who have many partners do not. Intervening to change the behavior of people with specific characteristics like age, sex, occupation, or geographic area is not a perfect way to reach those with high-risk behavior. Some members of these groups will be exposed to in- terventions even though they practice low-risk behavior. Moreover, oth- ers with high-risk behaviors who don't belong to any of these groups will be missed. The lack of perfect criteria for focusing interventions on those with the highest-risk behaviors is one source of leakage of program re- sources. This reduces the cost-effectiveness of interventions if resources go to people with lower-risk behavior. On the other hand, leakage may improve cost-effectiveness if resources-go to people with even riskier be- havior than those targeted. Surveys of sexual behavior, such as those con- ducted by WHO/GPA and DHS, can help to overcome these problems by establishing the characteristics and geographic location of those who have unprotected sex and high rates of partner change. Unless programs -27- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.10 Educating Adolescents on HW/AIDS: A Sound Investnent IN COUNTRIES WHERE SEXUAL ACTIVITY BEGINS Such programs are sometimes unpopular with at an early age and young people have high rates of parents who worry that information on reproductive partner change, promoting safer behavior among health, STDs, and contraception might cause their adolescents is clearly important to slowing the children to become sexually active at an earlier age. spread of HIV. There are many possible inter- Research has shown that this is not the case. Reviews ventions to address risky behavior among adoles- of school-based programs have found that partici- cents, both in and out of school. However, even in pating youth have not begun sexual activity earlier societies where sexual activity does not generally (Gluck and Rosenthal 1995, Kirby and others 1994, begin until after young people have completed UNAIDS 1997). Moreover, a review of school- their schooling, reproductive health education in based programs in the United States found that pro- the school system--which includes information on grams that included sexual health education and the benefits of postponing sexual activity as well AIDS prevention not only delayed the start of sexual as how to prevent pregnancy, STDs, and HIV activity, but reduced the number of partners and for those who do not abstain-is a potentially raised contraceptive use among those who became powerful intervention. Besides preventing HIV sexually active (Gluck and Rosenthal 1995). among students who might otherwise adopt risky Given the other broad social benefits and the rel- behavior, these programs have many other bene- atively low cost of adding HIV/AIDS education to fits. They prevent STDs and associated infertil- existing programs, HIV/AIDS education is likely to icy, and they prevent unwanted pregnancy, which be a good investment in preventing HIV. The over- may lead to abortion or to girls' dropping out of whelming majority of AIDS program managers school. More broadly, reproductive health educa- who responded to the AIDS in the World lI survey tion that includes education on preventing HIV felt that, as of 1993, reproductive health education may help to alter the social norms of the next needed expansion in their countries (Mann and generation of adults in ways that encourage safer Tarantola 1996). This was true for countries at all behavior. stages of the epidemic. are carefully designed and implemented, targeting prevention programs to groups with specific characteristics may stigmatize members and lead to discrimination, making future prevention efforts more difficult and less effective. Government Priorities, Resource Constraints, and the Stage of the Epidemic Considerations discussed in the preceding sections suggest a broad pre- vention strategy that assigns priority to activities based on the principles of epidemiology, public economics, and cost-effectiveness. Whatever the stage of the epidemic, this strategy calls for a strong emphasis on preven- -28- CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC tion activities starting with those at highest risk of contracting and spread- ing HIV and covering as many others as available resources allow. As the epidemic expands, containing it will increasingly require efforts to prevent infection among people with relatively lower levels of risk, which will raise the cost of prevention activities. If the epidemic is to be contained, this ex- pansion of activities must not weaken the fundamental commitment to work with those most likely to contract and spread the virus. This section presents a minimum set of activities to improve the efficiency and efficacy of national prevention programs, and suggests a framework for deciding the order in which to expand activities if additional resources are available. Provision of public goods or ensuring their provision by regulation is an important role for government at all stages of an HIV/AIDS epi- demic. Governments should invest in the information-collecting infra- structure they need to monitor the epidemic and to identify where high- risk behaviors are practiced and how to reach those at highest risk. In a nascent epidemic, understanding levels of HIV and STD infection in the subpopulations most likely to contract and spread HIV, the underly- ing patterns of sexual behavior, and the nature of the links to lower-risk subpopulations are critical pieces of information for assessing the proba- bility of a more extensive epidemic. As the epidemic progresses, govern- ments need to monitor, in addition, the spread of HIV to populations at lower risk and to support evaluation of the costs and effectiveness of alternative interventions. The need to reduce the negative externalities of high-risk behavior, as well as the insights of epidemiology, argue for heavily subsidizing safer behavior among those most likely to contract and spread HIV in coun- tries at all stages of the HIV/AIDS epidemic. It is not necessary to wait for HIV to take hold to make widespread knowledge, 100 percent con- dom use, and quick recognition and creatment of other STDs the norm among those more likely to contract and spread HIV, such as sex work- ers, bar maids, long-distance truck drivers, the military and police, min- ers or factory workers living away from home, and homosexual and bisexual men with multiple partners. The case for rapid action is even stronger in the case of injecting drug users because of the potential for very rapid increases in HIV prevalence in this subpopulation, and the subsequent spread of HIV to others, including their sex partners (some of whom may be sex workers) and their children. In a nascent HIV epi- demic, such highly focused actions may be sufficient to dramatically slow the spread of the virus. In countries with concentrated and general- -29- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING Hil/AIDS ized epidemics, preventing HIV among those with the highest chances of contracting and spreading the virus is still essential to slowing the epi- demic. However, behavior change among those with less-risky behavior who may nonetheless inadvertently spread the virus will be necessary to reverse the course of the epidemic. With respect to ensuring equiy, in countries with nascent epidemics, government can protect the poor best by acting early and decisively to prevent an epidemic. In countries with generalized epidemics, the risk of infection has risen for everyone and poverty should not restrict access to preventive services. Government can ensure that the poor have access to the knowledge, skills, and means to prevent HIV. The cost-effectiveness of interventions aimed at people with different levels of risky behavior will also change as HIV spreads from those who practice more risky behavior to those at relatively lower risk. Interven- tions for those most likely to contract and spread HIV are still likely to be highly cost-effective. In countries where HIV has spread widely in the general population, the cost-effectiveness of preventive interven- tions for lower-risk populations, such as subsidies for STD treatment, provision of safe blood, and reproductive health and AIDS education in schools, improves. However, these programs do not generate many ex- ternalities: their benefits accrue mainly to the person who uses them. While such interventions will prevent HIV infections and save lives, they will not do so as efficiently as interventions focused on those more likely to contract and spread HIV and they will often not be sufficient to reverse the course of the epidemic. Furthermore, the costs of provid- ing these services to the entire population at low risk of spreading HIV are potentially very large. In light of the substantial individual benefits of these services, people who are not poor will usually be willing and able to pay for them. In countries with scarce financial resources, the priority should be first on guaranteeing equity in access to these services by the poor. Not all developing countries face equal resource constraints in pursu- ing this strategy. In the lowest-income countries, prevention should start among those whose behavior generates or is likely to generate the high- est reproductive rate for HIV; subsidized prevention to others with lower H IV reproductive rates that are still greater than 1 can be extended as re- sources permit. Middle-income countries may have the resources, even at the earliest stages of the epidemic, to finance interventions for a much larger share of those for whom the HIV reproductive rates would be -30- CONFRONTING AIDS- PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC greater than 1. They may also have the resources to subsidize to a greact extent services for the poor and to extend prevention subsidies to suE populations that are unlikely to spread HIV to others. At an operational level it is impossible to determine the actual or pc tential reproductive rate of HIV for any group of individuals. Howeve using information about the average number of partners, condom us4 and injecting behavior from surveys and epidemiological surveillanc< subpopulations in a given country can be ranked from those with th highest-risk behavior (those most likely to contract and spread HIV) those with the lowest-risk behavior. Figure 3.5 shows a stylized view c the ranking of several subpopulations in a hypothetical population ac cording to the extent of risky behavior at a specific point in time, an, how the scope of prevention efforts would expand to include group with increasingly less-risky behavior depending on the availability of re sources. Once the highest-priority subpopulations have been effectivel reached, programs can be expanded to cover those with progressivel less-risky behavior, provided that sufficient resources are available. In deed, if sustained behavior change is achieved in the highest-priorir groups, the relative priority assigned to other groups will increase. O course, no ranking of this sort can apply to all countries, or even to a sin gle country over time. To overcome problems in locating the peopi most likely to contract and spread HIV at any given point in time, it i essential that policymakers and program managers finance collection o the necessary information for cost-effective use of the scarce resource available for HIV prevention. In concluding, we return to the important epidemiological point tha countries at the nascent stage of the epidemic have a unique opportunit to act early, to make a few key investments, and largely prevent an HIB epidemic. Not all countries with low levels of infection will necessaril, go on to experience an HIV epidemic, even without government action However, our inadequate understanding of the distribution of differen behaviors in the population and the links between different subpopula tions make it difficult to predict which among the countries will be sc lucky and which will not. Furthermore, even in countries where high risk behavior is relatively rare, patterns of sexual and injecting behavioi can change with economic and social conditions. Interventions at th< nascent stage are the most effective and will likely involve far less tota cost than if implemented after HIV has saturated subpopulations witf high-risk behavior. Further, because the number of people in these sub- -31- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVE\T[NG HIV/AIDS Figure 3.5 Resome Availabiilty and Program Coverage Spectrum of risky behavior Resource availability for prevention (eg, rate of partner change) Subpopulat Low Medium High S MHig Sex workers In brothels Most Ilkely to contract STD clinic patients and spread HIV Sex workers in bars/nightclubs Truckers, sailors Young military recruits In setting priorities, governments Factory workers should first ensure that the subpopu- Government employees lations most likely to contract and spread HIV are effectively reached University students with preventive interventions. As addi- Least Ikely to contract Secondary school students .mml resources become.available, and spread HIV prevention efforts can be expanded to Women at reproductive health clinics cover groups that are progressively Low Married women In rural areas less likely to contract and spread HIV. Note This is a hypothetical example only and is not meant to reflect the situation in any par- ticular country Soune Authors populations is small relative to the entire population, the absolute costs of prevention will be relatively low. These recommendations are not meant to limit the scope of govern- ment involvement if there are ample resources and public will to under- take even more. Rather, our intention is to point out the minimum set of actiVIties that all governments should be engaged in to improve the efficiency and equity of prevention programs and a rational order in %hich co expand these activities if HIV spreads or more resources be- come available. -32- ANNEX G ATTITUDES AND PRACTICES AMONG YOUTH IN CONSECUTIVE SURVEYS (McAuliffe 1994, Bisika 1996, Phiri 1997) 1994 1996 1997 Source of Information on Sex Male Female Male Female Male Female Friends 64% 78% 79% 84% 95% 90% School 11% 24% 13% 11% 58% 58% Methods of Prevention Condoms 52% 30% Avoid Multiple 90% 67% Partners Condom Use 55% 43% 34% 35% (Sexually Active) Possible to have one Sexual Partner Yes 54% 69% 74% 89% Partner who refuses 21% 56% 14% 44% 53% sex cares about me Ease obtaining 87% 63% 82% 93% condoms Sexually Active Total 53.3% 58% 44% Secondary 66% 78% 65% 26% Primary 41% 42% 46% 29% Out of School 62% 58% 58% 25% Age at First Intercourse (Among Sexually Active) 10 16% 13% 16% 16% 10-15 60% 64% 58% 56% 16-20 24% 21% 26% 29% Number of Lifetime Partners > 1 80% 60% 1- 68% 42% 5 122% 4% - 116% 5% -33- ANNEX H Reaching School Children with AIDS Prevention AIDS school curriculum. In 1995/96 an AIDS curriculum, first developed in 1991, was introduced by the MOE in collaboration with UNICEF and USAID in all government primary and secondary schools. The topics cover the modes of transmission, prevention, signs and symptoms, care for persons with AIDS (PWAs) and social impact. However, sexual transmission is not mentioned until Standard Five. Each year the same topics are repeated with more sophisticated biological information. Transmission of HIV with blood contact (razors, toothbrushes)--the least important modes of transmission--receives unfortunately, the most attention, even after Standard Five. Studies suggest that students know about sexual transmission of HIV, but they are not taught this in school. Success in implementing the AIDS courses has been mixed. While 66 percent of students reported they had been taught about AIDS in 1996, the highest coverage was in secondary schools (after students have formulated their sexual attitudes and for some, their sexual practices) followed by upper and lower primary schools. Of the 30 percent of teachers who actually received guidance in training--more were likely not to teach the curriculum because of embarrassment or because of other commitments than those who did not receive training. Forty-two percent of students did not receive books. Results of exam questions in 1966 showed adequate biological knowledge but students performed poorly on the question of prevention. The curriculum's focus on AIDS facts and examination performance suggests that few students are taught or understand prevention methods. In addition, the curriculum is only beginning to include a broader coverage of life skills. A curriculum for population (AIDS included), is planned in biology, social sciences and health in Standards five-eight. A syllabus has been printed and teacher resource materials are in draft and under review. This activity is led by the Malawi Institute of Education and supported by UNFPA. The effort to promote teaching about AIDS and life skills to youth has elicited a multi-sectoral response. Involved ministries include the Ministry of Health and Population (MOHP), MOWYES, Ministry Youth Sport and Culture (MYSC), and Ministry of Education (MOE). NGOs involved number about 30, including the Scripture Union and The Youth Arm. Work with youth also occurs in the setting of other activities which target the general population and care of t4e affected community. This input is more difficult to capture. AIDS clubs as activities for after school, have been formed in about 45 percent of primary and 66 percent of secondary schools. Most are Edzi Toto clubs, some are Why Wait? and Choose Freedom. Since many Malawians receive only two or three years of school, reaching youth through this channel (school) is limited. At present, over half of students between 15-19 are not in school and substantial numbers (30 percent of girls, 20 percent of boys) never attend. This pattern is likely to change in the future. The new free primary school initiative has already led to a dramatic increase in enrollment, from pre-1994 enrollment figures of 645, 000 students in Standard one enrollment to 1.3 million students in 1994-95. Total primary enrollment is estimated to have grown from 1.9 million in 1994 to around three million in 1997. -34- However limited, the opportunity to access youth through the school system nationwide, can be capitalized much further. Targets for reaching all teachers with training on AIDS and reaching close to 100 percent of all students should be set and attained. The monitoring of this coverage will require program managers to frequently estimate the number of teachers to be trained (approximately 50,000 in 1997; with 42,000 teachers for primary level alone) and number of students to be reached (approximately 3,200,000 in 1997; with 3 million for primary level, 108,844 for distance education centers and the remaining from secondary level and private schools). -35- ANNEX I Testing the Efficacy of Voluntary HIV Counseling and Testing Servicess A randomized controlled trial of the efficacy of HIV counseling and testing was conducted in Kenya, Tanzania and Trinidad. The study is still being completed in Trinidad. Objective: The objective of the study was to assess the impact of HIV counseling and testing (C&T) on behavior change among persons seeking such services and determine the cost-effectiveness of the C&T intervention. Design and Implementation: This was a randomized study in which one group receiving the C&T intervention was compared to a group receiving standard health information (HI). The study was conducted in three countries (Kenya, Tanzania and Trinidad) from May 1995 to April 1997. In each study site various strategies (including radio advertisement, television advertisement, posters, pamphlets, community outreach, etc.) were used to advertise the study. Informed consent was obtained from those presenting to the study centers and a standardized questionnaire was administered to obtain data on sexual behavior, psychological status, knowledge and attitudes about HIV/AIDS and other STDs, and care-seeking behaviors for STDs. In addition, a urine sample was obtained for later ligase chain reaction (LCR) test. Subjects were then randomized to one of the two interventions (C&T or HI). Subjects in the C&T arm received pre-test counseling and had blood taken for HIV antibody testing. They were provided with condoms and asked to return two weeks later for HIV serostaus notification and post-test counseling. Those in the IH arm were shown an informational videotape containing culturally appropriate information on the prevention of HIV and other STDs, and provided with a stock of condoms. All subjects were invited to come back to the study center at 6 and 12 months and all were offered counseling and testing at this point. At 6 months they were all administered a follow-up questionnaire and they underwent physical examination and laboratory testing for STD (and treatment if found positive for any STD). For the estimation of incidence of gonorrhea and chlamydia infections, a second urine sample was obtained for LCR test, and an LCR test was performed on the baseline urine samples for subjects with a positive LCR at six months. Those with a positive test at six months but a negative one at baseline were considered incident (new) cases. Data on STD will be used to corroborate the self-reported behavior. Key Findings: A total of 4290 participants were recruited (1512 in Kenya, 1427 in Tanzania, and 1351 in Trinidad) with an overall retention rate of more than 75 percent at the six-month follow-up. Preliminary analysis suggest the following key findings: * It was possible to recruit men, women and couples and to randomize them to counseling and testing and a comparison condition. * Scientifically acceptable follow-up rates were achieved at 6 and 12 months. Taken from "The Voluntary HIV Counseling and Testing Efficacy Study." Final Report Prepared by The Center for AIDS Prevention Studies University of California at San Francisco. January 15, 1998. -36- * It was possible to recruit high risk persons to counseling and testing. * Men reported more risk behavior than women (Figures II - 14 below). * Individuals randomized to C&T reduced risk behaviors to a significantly greater degree than individuals randomized to HI. * Couples assigned to C&T did not change behavior more than couples assigned to HI. A more detailed analysis by serostatus is required. * Individuals receiving C&T did not experience significantly greater psychological or social harms than did individuals receiving health information. Figure I.1 Participants Who Enrolled Individually. Percent Reporting Any Non-Primary Partner During the Previous 2 Months. N 46% 42%4% o 40% 40% 32%-- ---30 30% 26% ""'*** 26% Baseline 6-Month - C&T Males HI Males - - C&T Females - - - -- HI Females -37- Figure I.2 Participants Who Enrolled Individually. Percent Reporting Intercourse With Any Non-Primary Partner During the Previous 2 Months 0 30% 26% C-4 18% . 12% T0 Baseline 6-Month - C&T Males HI Males - - C&T Females - HI Females Figure 1.3 Individuals: Effects of Serostatus. Percent Reporting Any Non-Primary Partner During the Previous 2 Months. 47% = = = = = -_ 47% o 40% 32%/ 31% 31% 0 28% N Baseline 6-Month HIV + Males - - HIV - Males - HIV + Females - - - - HIV - Females -38- Figure I.4 Individuals: Effects of Serostatus Percent Reporting Unprotected Intercourse With Any Non-Primary Partner during the Previous 2 Months 36% 29% 25% 23 _R 22%== = 12% 10% 0 Baseline 6-Month HIV+ Males - = HIV - Males - HIV + Females - - - - HIV - Females Recommendations: Based on this preliminary analysis we are able to recommend that voluntary C&T should be made available to individuals. It appears to attract high risk individuals and to result in reduced risk behaviors at 6 months with no adverse psychological consequences. Voluntary C&T may result in adverse outcomes (i.e. break-up of marriage, physical abuse by partner) thus some strategies for reducing this impact may need to be incorporated. Further analysis (to be conducted January to June 1998) will pinpoint the impact of C&T on couples and its cost-effectiveness. Once all of these data are assembled firm recommendations can be made. -39- ANNEX J THE COST-EFFECTIVENESS OF HIV COUNSELING AND TESTING' EINDNGS Behavior change The overall analysis of the data from the C&T study indicates that C&T intervention is effective in inducing behavior change among individuals receiving this intervention. More details on this data can be found in the study final report. Cost per client Tanzania: The cost of provision of C&T per client was estimated at $29 in Tanzania. Assuming that the C&T site would provide the service to 3000 persons per year, we estimated the total annual cost at $86,791 (Figure.11). The break up of the cost per client shows that 72.4% ($21) represent recurrent costs, 24.1% (7%) client commodities, and 3.4% ($1) the start up. It is also interesting to note that most of the recurrent costs are for labor with more than 75% of the cost. The cost of the HIV test represents more than half the cost of client commodities. However, it represents less than 15% of the total cost to provide the intervention per client. Figure J1 Cost of the C&T Service, Tanzania * $86,791 total annual cost * $29 per client - Dar es Salaam only - 3,000 clients per year - 50% male and 50% female clients - male HIV prevalence 13% - female HIV prevalence 29% - 31% of clients coming as couples 6 Figures from Sweat, Michael: "The Cost-Effectiveness of HIV Counseling and Testing: Annex to the Final Report of the Voluntary HIV Counseling and Testing Efficacy Study" (pp 14-19): January 1998 -40- Kenya: The cost to provide the service per client was $27, resulting in a total annual cost of $79,950 if the site provides the service to 3000 persons per year cost (Figure J2). As seen in Tanzania, recurrent costs represent a large proportion (72.5%) of the cost per client in Kenya with 73.6% of this portion being spent for labor. The cost of HIV test represents more that 80% of the cost of client commodities (Figure 13) but it is only 18.5% of the total cost to provide the C&T intervention per client. Figure J2 Cost of the C&T Service. Kenya * $79,950 total annual cost * $27 per client - Nairobi only - 3,000 clients per year - 50% male and 50% female clients - male HIV prevalence 11.6% - female HIV prevalence 34% - 34% of clients coming as couples Figure J3 Cost of Client Comimodities Per Client Kenya HV Test Leaflets Follovup Client Recos Phleb 15.5% Letter Cant Supplie 3.9% 1.8% 0.3% Total: US$ 7 -41- Cost per HIV infection averted and variation of cost by target communities Tanzania: The average cost per HIV infection averted by the C&T intervention in Tanzania was estimated at $243. The data shown here indicate that it was more cost-effective to intervene with HIV-infected persons that with HIV-uninfected persons (Figure 4). The cost per case averted was also estimated for participants enrolling as couples and as individuals, stratified by HIV serostatus. Intervening with HIV- positive individuals appears to be the most cost-effective and intervening with HIV-negative individuals the least cost-effective of the four target communities on this figure. Looking at the cost by gender and serostatus the data suggest that it was generally more cost effective to intervene with HIV-infected males and females. HIV-infected females were the most cost-effective group to intervene with. It is interesting to not that from the Tanzania data the cost per case averted by providing the service to HIV-negative males was extremely high for this limited resource setting ($3,112). Figure J4 Cost Per HIV Infection Averted by C&T In Tanzania by Target Groups* $USD 6 0 0 - ............................................................ . . .. . -4482 5 0 0 . ................................................................. .. ... . . 4 0 0 - ...................................................... ..... 300 -243 2 0 0 --.-- ..--..124 ............ ....... .... 100 HIV- Avg. For HIV- Positive AII Negative Note: Cost affiected by number of people per group and the group!s average behavior change -42- Kenya: The average cost per HIV infection averted by the C&T intervention was estimated at $241 in Kenya. As seen with the Tanzania data, it was more cost effective to provide the intervention to HIV-infected than to HIV-uninfected clients (Figure J5). Intervening with HIV-infected couples was the most cost-effective whereas HIV-uninfected individuals represented the least cost-effective of the four target community in this analysis. The analysis of the data by gender and serostatus indicates that it was more cost-effective to provide C&T intervention to HIV-infected males and females than to those who were HIV-uninfected. Although like in Tanzania the cost per case averted was higher among uninfected than infected clients in Kenya, it is interesting to note that this cost was not as high as seen in Tanzania. This may suggest that the C&T intervention achieved more behavior change among uninfected clients in Kenya than in Tanzania. Figure J5 Cost Per HIV Infection Averted by C&T In Kenya by Target Groups* 6 0 .............................................................. 4 64 .464..... 400 ................................241.241. ... 2 0 0 . .......... 75 ..................... HIV- Avg. For HIV- Positive All Negative Note: Cost affected by number of people per group and the group's average behavior change Cost to client and potential for cost recovery Tanzania: Clients reported incurring various costs to receive the intervention such as: lost wages (about $7), travel cost (37 cents), non travel cost (61 cents), two hours and a half to travel to the C&T site, and two hours and a half spent at the site. When asked how much they would be willing to pay for the service if they had to pay, they reported willing to pay $5.11 on the average. Kenya: The costs incurred to receive the C&T intervention were of the same types as in Tanzania. Clients reported the following costs: lost wages ($2.25), travel cost (22 cents), non travel costs (10 cents), about two hours to travel to the C&T site, and three hours and a half spent at the site. Participants at this site reported willingness to pay $1.64 on the average for the C&T service if they were asked to pay. -43- DISCUSSION AND CONCLUSIONS Cost of Provision of HIV C&T The cost to provide HIV C&T in both Kenya and Tanzania was calculated based on estimates of the actual cost to provide the service to 3000 persons with similar demographic and risk characteristics as persons who enrolled in the study. We assume that persons who came to the study sites would be similar,to those who would come to receive a service. The number of 3000 per year was based on an estimate of the number of persons who would be serviced at the sites in the initial year of service provision, with the understanding that the number of persons served could be dramatically expanded as future resources become available. Note that if more people than 3000 per year are served the cost per client would drop due to an economy of scale. That is, it would be cheaper to offer HIV C&T to more people due to a more efficient use of the staff and infrastructure. The cost per client in Tanzania is estimated to be $29, and in Kenya $27. This would translate into an annual cost of $89,791 in Tanzania, and $79,950 in Kenya. In both Kenya and Tanzania the majority of costs are from labor and infrastructure. Despite the common belief that HIV C&T is expensive due to the cost of the laboratory testing, we found that it was the other costs that were most significant. Labor was approximately 2/3 of the total cost for provision of the service in both sites, reflecting the amount of time required for individual patient contact, training, and supervision. In general we feel that these costs are reasonable, especially given the significant impact on risk reduction that was demonstrated, as well as the number of HIV+ infections averted as described below. Cost per HIV Infection Averted In Tanzania it cost $243 for each HIV infection averted from C&T. In Kenya the cost per HIV infection averted was remarkably similar at $241. These costs per HIV infection averted from C&T are similar to the estimated cost per infection averted from enhanced STD treatment found in the Mwanza trial ($253 per infection averted). When compared to US based C&T services there is a dramatic difference, with the African- based service being significantly less costly per HIV infection averted. For example, Holtgrave (1995) estimated that for each life year saved due to HIV C&T in the US it cost $2000. This amount is for each year, thus is a person lived an additional 10 years it would be at a cost of $20,000. Similarly, a study by Owens (1993) of the cost per HIV infection averted if HIV counseling and testing was implemented in an acute care setting with an HIV prevalence of at least 1% the cost would be $60,000 per infection averted in the US. This comparison of the cost-effectiveness of HIV counseling and testing across countries points to the value of targeting resources to locations with high HIV incidence, such as Africa. Targeting Results of our study show that targeting the intervention can result in significant reductions in the cost per HIV infection averted, and thus generating greater preventative impact of the intervention at a lower cost. In general, we found that it is always better to target the intervention to persons and groups who are likely to test HIV positive + in essence, high risk groups. For the sub-groups examined (males & females, couples and individuals) it was always more cost-effective to target persons who were infected with HIV. Females and couples also were good groups to target, however the differences between males and individuals were not extremely high. The one exception was for HIV negative males in Tanzania. Males who tested HIV negative in Tanzania tended to make few behavioral changes, and thus the intervention was not cost-effective for this group. This implies that the service should be limited to high risk males to use resources more efficiently. -44- Yalidit Sensitivity analysis of the models utilized showed that the results were most sensitive to the base rate of HIV infection of partners, as well as the infectivity of HIV. Thus, for these parameters we used a wide range of values with a probability-based function that reflected the likely range of values most likely to be present. The range of output for the number of HIV infections averted from HIV C&T, however, was not too large for policy purposes + giving a reasonable range of results when the unknown values were considered using the probability functions. In sum, we feel confident that the results presented are valid. Cost to Clients and Cost Recovery In general we found that there were significant cost to clients to receive the service in terms of travel and lost wages. These factors should be considered when planning programs, and those services that are quickly delivered in locations convenient to the client population will reduce the cost to the client. Study participants also showed a willingness to pay for the service when asked at follow up. However, in practice we have since found that fees do significantly dissuade persons from accessing HIV C&T in this very poor setting. Moreover, fees are most likely to discourage the very poor from seeking the service. Thus, a sliding scale with moderate fees may be best, and more research in actual practice is needed. It is important to note that the literature on cost-recovery has repeatedly shown that people will significantly over estimate what they would actually pay for a service that they value. It seems that many people are inclined to overestimate their willingness to pay as a sign of how much they appreciate what services they received. We feel that this happened in this study due to the high quality of services that we provided, and given that they had already been through a difficult process of learning of their HIV test result. RECOMMENDATIONS The results of this study have demonstrated that HIV C&T can be a highly effective intervention in changing risk behavior. The cost analysis further demonstrates that the intervention is "cost-effective" in that it results in measurably reduced rates of new infections at a reasonable cost. Further, using prevention money in high incidence settings such as East-Africa is dramatically more cost-effective than in US settings. Targeting further enhances the cost-effectiveness of HIV C&T. Clients value the service and are willing to pay a moderate amount for it in both direct costs and in lost wages in transport costs. Thus we have several recommendations: * HIV counseling and testing should be supported as a primary HIV prevention intervention when feasible due to its effectiveness weighted against its cost; * Targeting of HIV C&T to HIV infected (high risk) groups and women can significantly enhance the preventative impact of the intervention at a lower cost; * Targeting of men should especially focus on high risk groups, as HIV negative men (especially in Tanzania) did not have significant reductions in HIV risk behavior making the intervention not cost-effective for such men. * HIV C&T services should be located in convenient locations and the time spent by clients at the service site (wait time and the time necessary to receive the service) must be as short as -45- possible to reduce the cost to clients. * Experiences after the study was completed showed that fees would significantly dissuade people from seeking HIV C&T, despite an expressed willingness to pay a moderate amount for the service. Further research on cost recovery is needed to better understand the impact of user fees on the cost-effectiveness of the service. -46- ANNEX K DISTRIBUTION OF EXTENSION STAFF IN THE DISTRICTS MINISTRY NAME OF CADRE STATUS VILLAGE T.A.* DISTRICT MOWCA Home Craft Workers Paid 600 Adult Literacy Volunteers Honorarium 3,000 Community Development Paid 500 Assistants Community Health Officers Paid 100 WID Project Officers Paid 50 MOHP Community Health Volunteers Voluntary 4,000 Health Surveillance Assistants Paid 3,500 Health Assistants Paid 160 Health Inspector Paid 99 Enrolled Community Health Paid 151 Nurse Enrolled Nurse/Midwife Paid 643 Registered Nurse/Public Health Paid 312 Nurse Community Based Distributor Paid 1,084 Traditional Birth Attendant Voluntary 5,000 MOA Field Assistant and Farm Home Paid 1,900 Assistant Technical Officer and Paid 150 Development Officer Subject Matter Specialist Paid 440 Project Officers Paid 30 SMS Paid 100 Area Agricultural Officers District Agricultural Officers MOE Primary School Teachers 46,000 Secondary School Teachers 3,000 District Health Education Paid 15 Officers MOI District Information Officers Paid 24 MOYSC District Youth Officer Paid 24 TOTALS $15,584 $51,966 $3,442 *T.A: Traditional Authority -47- ANNEX L Comparison of Neighboring Epidemics7 1.1 A comparison of HIV prevalence trends in ANC populations/sentinel sites: Kampala, Uganda; Blantyre, Malawi; and Lusaka, Zambia. Data on sentinel ANC trends were obtained in collaboration with the Malawi National AIDS Control Program and the Medical College of Malawi/Johns Hopkins University-sponsored study of HIV and reproductive health in pregnant women in Blantyre, Malawi. Data on age-specific HIV prevalence were obtained on study participants in Blantyre from 1989 on. Similar data from Lusaka has been collected though routine ANC sentinel surveillance as well as through a 1995 population based HIV survey. Preliminary results in Zambia were shared; however, data that would have allowed a thorough comparative analysis of the 3 sites were not available. They may nevertheless become available through further collaborative research. HIV prevalence trends from the Nsambya antenatal sentinel surveillance site in urban Kampala (pop. 800,000) were compared to those for the antenatal sentinel surveillance site at Queen Elizabeth Central Hospital (QECH) in urban Blantyre, Malawi (pop. 600,000) for the period 1985-96 (Figure 1). 40 HIV prevalence (%) L 30 -- - - 20 -J 'Kampala 10 _ 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Figure 1: HIV prevalence trends in ANC sentinel surveillance sites in Kampala, Uganda and Blantyre, Malawi, 1985-96. The circles and squares represent 95% confidence intervals for Blantyre and Kampala, respectively. Excerpt from "An Assessment of Emerging Patterns of HIV Incidence in Uganda and Other East African Countries", Rand Stonebumer, Family Health International, 1997, AIDS Control and Prevention Project. -48- Age-specific prevalences in pregnant women were compared for Blantyre (QECH) and Mulago antenatal clinic, Kampala for 1989-90 and 1996 (Figure 2). QECH, Blantyre, Malawi HIV prevalence (%) 40 3s L 1151989-90E 11996 3o - - --0- - - - -- --- ----- - i --- - 20 - 1 5 - - - - - - . . . -.- - 10 - - -- 0 15-19 20-24 25-29 30-34 Mulago clinic, Kampala, Uganda 40 ----- - - 35 - - - - - - - -- - - - - -- - - - - - 30 - --- -- - - -- I- _ T 25 --- - 20 - 15 - ----- 10 ---- --- 5 I- - - - - - -- - - -- - - - 0 15-19 20-24 25-29 30-34 Age Figure 2: Age-specific HIV prevalence in pregnant women are compared at antenatal clinic populations in Mulago clinic, Kampala, Uganda and Queen Elizabeth Central Hospital, Blantyre Malawi for 1989-90 and 1996. The difference in prevalence in the 15-24 age group in Kampala is statistically different between 1989-90 and 1996. The HIV epidemic in Blantyre appears to be about 1.5 to 2 years later than that in Kampala, but it is noteworthy that while HIV prevalence in Kampala has declined dramatically since 1991/92, the situation in Blantyre has thus far remained stable or may indeed be gradually increasing. In the second figure the decline in HIV prevalence among 15-24 year olds is evident in Kampala but is absent in Blantyre. While the Kampala epidemic exhibits the "hallmark"(a decline in prevalence in the 15-24 age group) of HIV incidence and prevalence reduction, and perhaps successful intervention and behaviour modification, the Blantyre epidemic reveals no such evidence. The epidemic in Blantyre perhaps is indicative of a growing stabilisation of HIV prevalence, continued high HIV incidence rates among young people, and less effective intervention strategies. -49- Lusaka ANC trends in HIV prevalence, illustrated in Figures 3 and 4, suggest an epidemic which is intermediate in both timing and dynamics, between that of Kampala and Blantyre. HIV prevalence (%) Blantyre 30 ----------- - --- - ------ - Lusaka 25 - - - ---( e s t i----ed- (estimatedsince 1993) 20 --- ----- - - ---- - -- -- -- 15 -- - ---- - -- - ------ - --- - - - Kampala 10 --- - ------ ----- - - - - 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Figure 3: HIV prevalence trends in pregnant women attending antenatal clinics, Kampala, Blantyre, and Lusaka, 1985-95/96. There was reported stability beginning in 1993 and incipient declines in the 15-19 year old group from 1993-95. (The lack of any significant decline in the 20-24 year old group marks a difference from the situation in Kampala, and may be indicative of differential patterns of behaviour change according to age. The data from 1996 surveys will be released soon, and should provide further information.) -50- HIV prevalence (%) 35 . ........ W1993 W1994 m1996 30 - 25 20 ...o. 15 . 10 . 5. 0 15-19 20-24 25-29 3G-34 age Figure 4 [above]: HIV prevalence trends in pregnant women age 15-19 and 20-24 attending antenatal sentinel surveillance sites in Lusaka, 1993-95. Preliminary findings reported by Drs Sichone, Kasumba, Musonda, Ndhlovu, Fylkesnes. Chi-square for linear trend for 15-19 year olds, p= .055. Prevalence in 20-24 year olds reported as stable (prevalence presented in figure for 20-24 age group is estimated). From these comparisons, it could be hypothesized that HIV risk behaviors in Kampala in the later part of the 1980s was similar to those in Blantyre in the early to mid-1990s, and that behaviour change in the early 1990s in Kampala, and perhaps s few years later in Zambia, may be linked to declines in IHV prevalence in younger age groups. 1.2 A comparison of HIV risk behaviour data in urban populations of Uganda, Malawi and Zambia: Sources of data on HIV risk behaviors, condom use and distribution were analyzed. In addition to the previously mentioned National KABP surveys which were undertaken in 1989 and 1995, as well as the 1995 DHS survey in Uganda, the following data sources were identified in Malawi and Zambia. > 1996 Malawi DHS with AIDS module; > 1996 Zambia DHS with AIDS module (with Macro); > 1995 Malawi Contraceptive Consumer KAP survey, 1995, Population Services International.(PSI) (%50 urban); > Somarc condom distribution data: Uganda, Malawi, and Zambia. -51- A comparison of risk behaviors in Uganda in 1989 and Uganda and Malawi in 1995 is shown in Figure 5, below. 35 ------ -------- -- - - - - - - - - 225 0 a. I MUganda overall 20 - -- -------- --- OUganda Urban 15 ------- --2- 13--- W Uganda Rural 1 Malawi (50% urban) 10 0 1989 1995 Figure 5: Proportion of respondents reporting a casual sex partner in the last 12 months in Uganda (overall, urban and rural) in the 1989 Uganda National KABP 1989 compared to the 1995 Uganda Ministry of Health KABP and to a 1995 Population Services International Malawi condom social marketing survey (50% urban sample). The frequency of risk behaviors, as measured by self-reported casual sex in the past 12 months in Uganda for 1989 (30% overall, 32% urban, and 28% rural) are closer to the 25% reported in Malawi in 1995. By contrast, the frequency of people reporting this particular risk behavior in Uganda in 1995 is substantially lower. Although more work is called for to analyze the raw data from the Malawi (PSI) survey, which was 50% urban and included Blantyre, preliminary results are compelling. Risk behaviour among males and females in Uganda in 1989 (KABP), 1995 (MOH KABP), and 1995 (DHS survey) is compared to that in Malawi in 1995 (PSI survey) and 1996 (DHS survey) in Figure 6, next page. -52- % affirmative responses I 41.3 atotal QDurban grural 401 38.4 X6& -- - -- -- - - - - - o --- -Uganda males 20 -98 1. -----2------------ 10 1989 KABP 1995 MOH KABP 1995 DHS 50 r-------------- ----- ----------- ----- --- 40 L- ------------------ -- ------- ------ -- 30- Malawi males 20 L---------- ------- -11 10 1995 PSI KABP 1996 DHS % affirmative responses 25 ------- ------------------------ ---------- - 22 1 Ototal Murban grural 20 -JS8 15 is7 Uganda females 10 --- 79------ 63 o L Ilk 2s- 1989 KABP 1995 MOH KASP 1995 DHS 25 ------------- ------------- -- ------- ---- 20- 1 is - - Malawi females 10 - -- - 5 -- - - - - - - - - - 2 -24 15s 1995 PSI KABP 1996 DHS Figure 6: A comparison of reported casual sex partners in the last 12 months among males (top) and females (bottom) in 1989 and 1995 in Uganda, and in 1995 and 1996 in Malawi. The Uganda data are from the 1989 and 1995 Uganda KABP surveys (stratified by residence) and the 1995 National Demographic and Health Survey -53- (DHS). The Malawi data are from the 1995 Population Services International Malawi condom social marketing survey, the sampling frame of which was 50% urban, and from the 1996 Malawi National DHS survey. In Uganda the pattern of risk behaviors in 1989 was similar to that in Malawi in 1995. The Uganda DHS data for 1995 are largely in keeping with the 1995 MOH survey reporting sex with non-regular partners in the range of 13.4 % (males) and 6.3% (females) overall, and 19.8 % (males) and 7.9 (females) in urban areas. If the DHS data from Malawi are accurate and representative, they suggest a dramatic change in the proportion of respondents reporting casual sex between 1995 and 1996: from 33% in 1995 to 13% in 1996 among males; from 19% in 1995 to 2% in 1996. It is not unrealistic to believe that once behaviour begins to change on a population level, it does so precipitously. Widely shared perceptions and common motivating forces prompt population-wide changes in behaviour. Growing awareness of mortality among friends and family would be one of those widely-shared perceptions, but intensive saturation with education and information is surely likely to be another. Similarly, the diffusion of information through informal social networks must be taken into account as a major force in bringing about widespread behaviour change at given time intervals and in ways which are shared by large numbers of people from different backgrounds. Validation of these behavioral results, linked with sustained monitoring of HIV prevalence trends in Malawi, would allow linkage of HIV incidence patterns to reported behaviors. If the data are accurate, a decline in prevalence among 15-19 year old people should begin to be apparent in few years, and perhaps earlier. Condom use in last casual sexual partner in 1995 is compared in Uganda Malawi and Zambia in Figure 7. 58 _ Kampala Uganda 33 other urban rural males Malawi 50% urban Zambia urban males 0 10 20 30 40 50 60 % affirmative responses Figure 7: Levels of reported condom use in the last casual sex partner from 1995 surveys in Uganda, Malawi and Zambia. The data sources are from the Uganda 1995 Ministry of Health KABP, the 1995 Malawi condom social marketing KABP performed by Population Services International, and from a National KABP performed -54- by the Zambia Ministry of Health. The Uganda and Malawi data are the results of combined responses of males and females. Reported condom use with the last sexual partner was relatively high in Uganda, Malawi, and Zambia in 1995. It is not clear how long these patterns of condom use have been in place, but in Malawi it may not have a major impact because 33% of males said they were still having non- regular partner sex in 1995. Though 45% of them may be using condoms, there is still considerable scope for more unprotected sex in Malawi than in Ugandan urban areas. It is also interesting to note that cumulative socially marketed condom sales are highest in Zambia, and then followed by Uganda and Malawi. Rates per capita are highest in Zambia, and then similar in Uganda and Malawi (Figure 8). These preliminary findings call for further analysis if the relationship between condom sales and HIV epidemic patterns is to be understood. 20 Millions rates per capita 5.0 18 4.3 15 13.5 aQ Cumulative sales a Cumlative rate 3.0 10 6.8 2.0 1.5 - 1.4 5 1.0 0 --Iiil0.0 Uganda Zambia Malawi Figure 8: Cumulative condom sales in millions and rates per population age 15-49, in Uganda, Zambia and Malawi. Data provided by Somarc. 2.0 HIV incidence dynamics and demographic impacts: Blantyre, Malawi . Applying modeling methods similar to those described earlier for Kampala, the following figures describe baseline and incidence reduction scenarios for Blantyre. -55- 120 thousands 100 ------------------------ 0 --- ------------ -------------- .wincidence 60 - --- - -- - - --- -- -- - - -------- ipWrevarenW - - baseline reeuctio 40 -- -- -------- ------------- ------------- -mortality - - -- 20 --------- ---- -- -- -- ------ --- ----- - - --- - - -- 0 -t.P.. 1981 1985 1990 1995 2000 2005 120 -- - - - - - --- 100 ----------------------- --- - - - 80 F------ -- -------- ----- ------ -dac-------ane a -n- ---- incidence 60 --- ------------ ------------------r-------- reul ti n 40 -------------- ---------------------------- - --- 20 --------- ---- ---------- --------------- --- -- - -- 0 - 1981 1985 1990 1995 2000 2005 Figure 9: Differential demographic impacts: Blantyre baseline and incidence reduction scenarios, 1985-2005. Cumulative incidence refers to total HIV infections; prevalence in 2005 refers to cumulative HIV infections minus deaths; cumulative mortality refers to all deaths due to HIV. 270.097 baseline I . I asI crulative incidence .cIumulative mortality I *prevalence in 2005 212.609 reduction 59.153 0 50 100 150 200 250 300 350 Thousands -56- Figure 10: Simulations of HIV prevalence (%) in pregnant women in Blantyre, Malawi under baseline and incidence reduction scenarios from 1985-2005, compared to HIV prevalence at the Queen Elizabeth Central Hospital (Blantyre) antenatal clinic sentinel surveillance site, 1985-96. The data fit relatively well. Age-stratified analyses comparing observed age-specific prevalence in QECH to those modeled under the baseline simulation also fit fairly well (not shown). Demographic impacts are shown in Figure 11 to illustrate the potential of the model to measure deaths averted. The scenarios are not comparable in 2005 to those for Kampala because of the later epidemic onset in Blantyre. 35 0 :Baseline U 30 -Mmciderice reducts, 25 QECH I >20------ 25 1 -- - 10 ------ ~ 0 Figure 11: HIV incidence, prevalence and mortality: Blantyre simulation 1983-2005. The epidemic onset in Blantyre was three years later than that in Kampala, and the period of highest HIV incidence growth was extended over a slightly longer period. The HIV incidence reduction is highly speculative but assumes a reduction on the same scale as that in Kampala, but beginning 5 years later, i.e. in 1995. Figure 9 describes the familiar incidence, prevalence and mortality curves while Figure 10 shows the ANC clinic simulations compared to empirical data. 2.1 Population movement and implications for the dynamics of emerging infectious diseases in sub- Saharan Africa: the social and demographic context. It is impossible to consider the emergence or spread of any communicable disease in sub Saharan Africa outside the context of the massive social and demographic changes which have characterized most countries in the region during the latter half of the 201h century. Most of sub Saharan Africa has been exposed to population movements of a magnitude and pace rarely seen before in that region, and migration has now become one of the most important determinants of demographic structures and social behaviour. Most of the population upheaval -57- and movement has been linked to two main factors, economically motivated movement, and politically precipitated uprooting. The massive movement of people from rural to urban areas for economic reasons has followed much the same lines (but possibly at a faster pace) as in other developing regions of the world. In Africa it began to pick up speed in the 1960's but has accelerated in recent years with the emergence of better communication and transportation systems. Economically motivated migration from rural areas, and to a lesser extent, from one urban area to another has picked up momentum as a result of mix of factors. Increasing land pressures in some regions, together with limited investment in rural development throughout the continent have made many rural areas socially and politically uninviting. Despite earlier attempts to improve coverage by health care, social and educational services, the gradual population depletion in these same regions has often made them less attractive and "necessary" to policy makers. At the same time, the fairly rapid economic growth and boom of urban areas in the 1960's gave birth to a widespread myth that cities in Africa would continue to grow and be able to absorb increasing numbers of people. Improved road and transportation systems have served to facilitate this increasingly rapid movement of people, and have opened up market options involving within and between- country temporary movements. It was not until the 1980's that it became apparent that the emerging cities of sub-Saharan Africa had long become saturated and no longer able to appropriately absorb the hundreds of thousands of people migrating annually to them. Most urban infrastructures quickly lost their capacity to keep pace with population growth due to in-migration and natural increase. The result has been the evolution of new urban ecological and social environments that are often worse than those which migrants felt obliged to leave. Today most urban conglomerations are increasingly typified by chronic unemployment, serious overcrowding, poor potable water and sanitation systems, and seriously limited access to quality health care. Health indicators in these cities reflect the change, and many communicable diseases which had previously been controlled, have now begun to reemerge. The second and by far more dramatic population shift has come in the wake of the many political upheavals and military-civilian conflicts throughout sub-Saharan Africa. At the end of 1996, of the well-over 50 million people displaced by complex disasters throughout the world, over a half were located in sub-Saharan Africa. Complex emergencies have been common and widespread throughout the region. They have involved countries as far afield as Mozambique and Angola in the south, to Rwanda in central Africa, to Uganda and Somalia in the east. It is conservatively estimated that more than 30 million people have been displaced at one time or another by these conflicts in Africa, and that people have been caught in temporary refugee and displaced persons settlements for periods of over ten years. The social conditions in displaced persons and refugee camps have often proved to be risk producing from the perspective of additional violence (often sexual) and communicable diseases spread. The implications of migration for HIV/AIDS are many. Irrespective of whether it is politically or economically prompted, migration invariably involves serious disruption of family life and a break with traditional values and social control mechanisms. It also involves younger people more so than older ones, and where this is the case, the psychosocial break with what are usually stable and value-conscious environments often provokes the need to compensate by seeking casual and serial partnerships. In the case of refugees and internally displaced persons the process of uprooting is fraught with opportunities for sexual exploitation and abuse. Rape and sexual violence has become a frequent tool of ethnic -58- cleansing, and safety from this is not necessarily found in refugee camps. Indeed many such camps, as was the case in Mozambique, often provided the occasion for chronic sexual abuse against women. Displaced women (who tend to represent over half of all displaced people throughout the world) are highly vulnerable to sexual exploitation because they are often alone with their children and have to find alternative ways of securing food and "security", be this in camps or "on the road". From the perspective of HIV transmission and the increased vulnerability of populations on the move, however, it is the broad psychosocial context surrounding migration that is perhaps the most important determinant of risk. Migration everywhere has become increasingly sex and age selective. The demand for labor, whether it is real or anticipated often creates sex biases in the populations who move. In some countries such as South Africa it has been the demand for male labor in mining industries, which has created the imbalance. In other locations such as the west coast of Africa market vending involving two and three day treks within the country and between countries has tended to involve more women than men. Market vending often means traveling away from home and seeking places to stay overnight. While by now a highly institutionalized phenomenon, it nevertheless often involves heightened risk for women. Sex biases in the demography of movement invariably have far-reaching implications for trends in sexual behaviour. Where male migrants predominate, the role of female prostitutes and serial relations with otherwise casual and often married partners has become much more evident. In light of the relatively smaller number of women selling sex to larger numbers of men, the opportunities for infection of women has also increased. Their capacity to secondarily transmit HIV has no doubt also increased. Where the converse demographic situation has been true, the implications often involve female migrants being "held at ransom" for sexual favors if they are to find work and shelter in highly competitive but poverty stricken markets. This has historically been the case in Europe as well as in Africa, but the magnitude of the contemporary migration process in Africa has meant a much higher degree and regularity of exploitation. The chances for these situations to gradually evolve into quasi-prostitute relations are strong, and also explain the elevated number of sex workers and casual providing of sex-for-favors in market towns of Africa. Another major form of population movement in Africa, which has had a far-reaching impact on patterns of sexuality and sexually transmitted diseases, has been the many military actions that have occurred throughout the continent in the past twenty years. These military movements within home territory and often outside it, have involved thousands of young men at the peak of their sexual careers, moving with the "authority" of invading or protective forces, and often with considerable leeway for personal behaviour. Even in the case of troops serving under the auspices of the UN, there is now considerable evidence that they have been a major source of sexually transmitted diseases among displaced people and refugees. Refugees, internally displaced people and rural-urban migrants share another characteristic. It is the limited access they often have to regular health care services. For a variety of reasons, including their physical marginalisation from service centers, their temporary lack of economic power, their often linguistic separation from the dominant group, their possible differences with respect to utilization of services, as well as the stress which invariably accompanies migration, migrants of all kinds rarely have the access to care that they need. By the same token, they may be personally ill-equipped to benefit from some of the AIDS prevention services including the education and information campaigns which have been available to urban communities in many countries. Migration is not always an uni-directional process, however. The market vending operations which have now come to characterize the lives of hundreds of thousands of women in West Africa are basically circular -59- circular movements in which women - and to a lesser extent men - are constantly rotating between their homes and other market locations. Miners, for example, have usually come to South Africa on temporary contracts just as "guest workers" have done in other parts of the world. They move backward and forwards from their home communities and countries. Military movements are equally likely to involve even more temporary "migration", and troops inevitably have the same ability to pick-up and take diseases back with them, as they do to introduce them into the communities they pass through. It is also noteworthy that many rural-to-urban migrants also tend to practice a return process, especially when they have left spouses behind. Anecdotal evidence suggest there has also been considerable return movement of people when they have found themselves ill and in need of care from family. To what extent this has been a factor in the movement of HIV is not clear but it must be considered. The analysis of available data which is presented below was not always able to address the possible role of migration, in particular because data records on the movement of people and the health of those who move have typically been weak. Where the movement of people deserves to be factored in to the understanding of what is occurring to HIV patterns, this is mentioned or is raised in a hypothetical fashion. Given the magnitude of the migration process in sub-Saharan Africa it would clearly be inappropriate not to assume that the mass movements of people throughout the continent are now having an important impact on infectious disease profiles. A new human and social ecological structure is emerging in Africa around, and determined by, the massive population disruption, movement, and unsatisfactory resettlement process which has been taking place over the past twenty years. Many changes are taking place in patterns of social organization and personal behaviour as a result of migration to urban slums and shanty towns and to high density areas of chronic unemployment. The relatively rapid deployment and movement of hundreds of thousands of military personnel, and the displacement of even greater numbers of civilians cannot have occurred without significant implications for health and illness in the region. There is no indication that this process is slowing down. Mass population movement of the kind, which that characterized sub-Saharan Africa to date, is likely to continue for some time and to fundamentally alter the way in which societies and cultures are organized there. They are also likely to influence the emerging pattern of diseases including HIV/AIDS, STIs and other infectious diseases. 2.2 Population movement in sub-Saharan Africa and its relationship to HIV. In the three decades since the main period of African independence, migration has played a major role in reshaping the demography of sub-Saharan Africa. However the emphasis devoted to better defining its dimension have been generally secondary to other demographic issues related to mortality and fertility. With respect to HIV, the modern emerging infectious disease, movement of population and their behaviors, are fundamental to the geographic diffusion of infection. Without population movement, HIV would have remained localized never reaching pandemic proportions. As the HIV pandemic matures, population movement will continue to play an important role in defining the evolving epidemiologic dynamics-its better understanding will improve the control of HIV, and strengthen the capacity to respond to future emerging infectious disease threats in the region. For definitional purposes population movement or migration is divided into international (cross border) and internal (within borders). -60- International migration The number of international migrants in Africa is estimated at about 3.6% of the population, comparable to the proportion of foreigners in the EC. However, their movement is much more volatile, unpredictable and, hard to measure. The major international flows of migrants in the 1980s are depicted in figure 12. Medsteran seems Red Sea Sf-C o ut - r eou Figure 12: International migration flows in sub-Saharan Africa. Source: Ricca, S, International Migration in Africa: Legal and administrative aspects. Geneva: International Labour Office (1989); Adapted from Demographic Change in Sub-Saharan Africa, Foote, Hill, and Martin, Editors, National Academy Press, Washington DC. 1993. -61- urbanized cities of coastal West Africa to the Sahel. Overall, the movement of refugees has been dominant in East Africa. Elsewhere migration for employment has dominated particularly in West and Southern Africa. An example of the relationship of migration to the geographic diffusion of HIV from East, Central and Southern Africa to South Africa is further illustrated in Figure 13. a) b) _____ J KENYA1 ~i2~ j AIE NETHERLANDS --a IMALAWI0T BOSWANAM r BIOUE CANADA ZAMBIA HAITI ESOTHO 0 km S0 BRAZIL SOUTH AFRICA ' S'MW Mtne S SWAZILAND Iworkers S .' Number of mKyant workers Others SOUTH 1' 40000 Moe workers - k0 000 (percentage) Number of nfec-aons AFRICA 0 kmn l00 a00 Number of cases I m 0 Wo Figure 13: (A) Nationality of foreign AIDS cases in South Africa by November of 1988. Arrow widths indicate number of cases originating from each country. (B). Mid-1980s estimates by country of origin of the number of HIV infected migrant workers in South Africa. Source: Atlas of AIDS, Smallman-Raynor, Cliff, and Haggett. London International, (1992). As of 1988 AIDS cases among migrants in South Africa by nationality revealed a link with migration coming predominantly from Zambia and Malawi, and early seeding of HIV in South Africa (13A). Figure 13(B) illustrates the total number of migrant workers originating from 5 African countries registered in South African in the mid-I 980s. The totals are shown in the form of proportional circles and those engaged in mining are shaded. These totals were then scaled by estimates of country specific HIV prevalence in cohorts of miners and other migrant workers in South Africa. The estimates are presented as proportional arrows. The large arrows suggest that the dominant flow of HIV into South Africa may have come with miners and other migrants, predominantly from Malawi. Studies of HIV prevalence in South African miners further support this hypothesis, and the return of these migrants to Malawi most likely exacerbated the epidemic there. -62- Internal Migration The principal internal migration patterns are circulation (circular), which involves repetitive, non- permanent moves, and definitive migration. A description of these patterns with African examples is shown in table 3. Direction Circulation Periodic Seasonal Lon, term Definitive mi,,ration Movement of Pastoral Labour Agricultural land dealers in displacement migration to colonization, resettlement fl; .produce and due to agriculture economic nodes and land stock environment wage sector consolidation; overspill and hazards and other rural. into marginal of spontaneous migrants from 4 ~ population pressure areas. Movement of Movement of Spontaneous migrants in dealers in employed and slums, shantytowns and RO agriculture underemployed suburbs. produce persons V Movement of Return "Repatriation" Return migration of retired dealers in migration of of unemployed person and unsuccessful .... *.. urban urbanites persons; labor urban migrants. UT, manufactures during "peak" migration to (e.g.) soap, agricultural rural agro- foods, etc. seasons. industrial mining nodes Movement of Movement of Prospective migration of self-employed transferred second-or later- generation persons workers; self- migrants out of touch with employed ancestral home persons (traders and business people relocating elsewhere) Table 3: Typology of internal migration with African examples. These patterns exclude daily movements such as cultivating, vacationing; and commuting. Adapted from Demographic Change in Sub-Saharan Africa, Foote, Hill, and Martin, Editors, National Academy Press, Washington DC. 1993. Rural to rural and rural to urban migration are the most common movement patterns in sub-Saharan Africa. However, since one- third or more of the population will live in urban areas in all sub-regions of Africa by the year 2000, rural to urban permanent and circulation migration are the patterns which may most contribute to spread of infectious diseases such as HIV. An important pattern of movement in the context of infectious disease dynamics and not well defined in the preceding typology, is the urban to rural and urban to urban -63- movement of "temporary migrants or travelers " who are connected through the social networks of the more traditional migrants. The potential effects of different migration patterns are addressed in the following sections: * The first (3.1) describes an investigation that provides empirical evidence that patterns of internal migration are associated with different urban and rural HIV prevalence levels. * The second investigation (3.2) describes results of the application of an HIV mixing model to simulate urban and rural prevalence scenarios when applied to different migration patterns. It demonstrates that circulation patterns from rural-urban-rural could result in insidious replenishment of susceptible populations and increasing HIV infections in urban areas-also providing a vector of HIV back to the rural community. Since 70 % of the population in sub-Saharan Africa is rural and more mobile with improving transportation, these population dynamics will be increasingly important with respect to health. * The third section (3.3) provides support that the extraordinarily high HIV prevalence in parts of Rakai district in Uganda may have related to encampment sites of military forces during the war in 1979-80. This highlights the classical importance of disease amplification during civil conflicts. 3.1 An investigation of population movement and its association with differential urban-rural HIV prevalence: patterns of internal migration as potential geographic vectors for the HIV transmission. A preliminary report, (E.Perkins, Cambridge University), explores the spatial interaction of migrant populations as a vehicular surrogate for the transmission of HIV. The purpose of the study is to superimpose the geographic vectors of population movement and displacement with the well-recognized differences in spatial transmission patterns of HIV. The linkage between migratory patterns and HIV urban and rural prevalence differentials is studied by comparing data on HIV prevalence and migratory patterns in Malawi and Rwanda. The hypothesis posed is that lower urban rural differentials in HIV prevalence in a country are related to patterns of rural-urban to rural migration. A summary of the findings follows: * Differences in HIV prevalence between urban and rural HIV prevalence are greater in Rwanda than in Malawi. * In contrast to the traditional idea that most migratory movement in Malawi is rural to urban, it appears that two-thirds of rural districts are experiencing population increases from 1977-1988 or 1988- 99 (projected) at rates greater than the national average. Three rural districts are experiencing the highest rates of population increases, while several of the urban districts are experiencing the lowest. * Analyses of migration flows of districts in Malawi revealed that those areas with net out migration were largely those that were more urban. Migration out of the city appears to be more important than migration in. * The urban exodus involves a combination of urban to urban and urban to rural migration. It is not merely a reversal of the previous rural to urban flow. -64- * The high degree of urban-rural-urban interaction in Malawi could facilitate the acceleration of the spread of HIV from cites to more rural areas. * In Rwanda (1981 census), most areas experienced a net out migration largely to urban areas. The pattern was different than that in Malawi. * An index of population mobility was calculated for by comparing migratory movements per 1000 residents and their population density index for districts in-Malawi and Rwanda. The findings suggest that districts in Malawi have much greater rate of population mobility than in Rwanda. * These findings suggest that differential migratory patterns in Malawi and Rwanda may explain different urban rural prevalence differences. What role these patterns of movement have played and are playing in determining HIV prevalence patterns is not entirely understood; but the issue merits further analysis. The HIV prevalence data used in this analysis need to be updated and applied to broader regions to further test the hypotheses and better understand its broader implications. Preliminary results nevertheless suggest that the impact of population movement in this region may be affecting behavioral dynamics relevant to the sexual transmission of HIV. 3.2 Simulation of population HIV prevalence under three migration patterns: evidence for the association of rural-urban-rural circulation pattern and a more severe and sustained HIV epidemic. A simple HIV mixing model, developed by Daniel Low-Beer, was used to test hypotheses regarding differential migration patterns and HIV incidence dynamics. Given the same probability of infection with exposure, three population mixing patterns were simulated: a baseline with no urban or rural mixing; a rural to urban migration; and a circulation pattern (circular) migration, from rural to urban to rural, (similar to that in West and Southern Africa and discovered in Malawi) in figure 14, next page. The utility of such a mixing model is to test the hypothesis rather than project an accurate scale of the epidemic with respect to time or magnitude. The major finding from this modeling exercise is that the circulation pattern of rural -- urban -- rural migration results in a situation in which HIV prevalence levels are maintained at high and stable levels in urban areas, while in rural areas the prevalence increases but the prevalence differential decreases over time. Although the urban prevalence is lower than in the rural to urban pattern, the increasing prevalence in the much larger rural sector of the population make for a much more severe epidemic. These findings are consistent with the empirical findings in Malawi and Rwanda, presented in section 3.1, and strengthen the hypotheses that rural- urban- rural circulation migration patterns are playing a major role in the spread of HIV in the region. -65- % HIV rural-urban circular no migration migration migration urban rural. 10 20 300 10 20 30 0 10 20 30 time in years Figure 14: IV dynamics under 3 migration patterns: Baseline, or no migration; rural to urban migration; and circular migration from rural to urban to rural. The details of the model and their parameters are available at ICMH. 3.3 Military conflict as a magnifier of iV spread: the ecological association of severely HIV impacted parishes and sites of troop encampments in 1979-80, Rakai District, Uganda. As described in section 2.1, the military has historically been implicated in the spread of infectious diseases. One hypothesis for the appearance of such high levels of HIV in certain areas of Rakai district in Uganda was that troops present in the region during the war between Uganda and Tanzania in 1979-80 might have acted as a magnifier of HIV spread. The topic has been a subject of speculation with mostly anecdotal associations. An analysis of 1991 Uganda census revealed several parishes in Rakai district with negative population growth as a result of AIDS deaths (Low-Beer D, Stoneburner R1 Mukulu A, Nature Medicine 1997:3; 553-57). Of particular note is the geographic heterogeneity of the impacts with severely affected parishes along side those with little or no impacts. Sero-prevalence studies in the region also support this heterogeneous pattern. Similar demographic impacts were not noted in other districts of Uganda. Figure 15 shows the overlay of Rakai and Masaka parishes with moderate and severely impacts of HIV deaths with troop movements, encampment sites and the village of Kasensero with first reports of AIDS. -66- AA A Very severe impact PA L Severe Impact KASENSERO x Encampment site X TPDF Movement Figure 15: The geographical distribution of severely impacted population pyramids in southeastern parishes of Rakai and Masaka districts, Uganda*. Kasensero is a village on Lake Victoria, with first reported AIDS cases. The road from Kasensero to Rakai was the site of Tanzanian and Ugandan military activities in 1979-80. The first village west of Kasensero was a known encampment site of a TPDF brigade. * Source: Nature Medicine 3, 553-57 (1997). Of particular interest is that the severely effected pyramid at the end of the road near Kasensero (the village with first reports of AIDS cases), was the site of a several month encampment of a brigade (5000 men) of the Tanzania Peoples Defence Forces (TPDF) in 1979-80. Information from sources familiar with troop movements in the region then, suggest that the other severely affected parishes seem to correlate with TPDF encampment sites. -67- Although, such evidence remains partially anecdotal, it provides the basis for an investigation with a testable hypothesis. Such an investigation would be hardly academic given the numbers of military conflicts since 1980 in the continent, and the potential for these conflicts to be seeds for the emergence of new infectious diseases. One long-term outcome of such an investigation could be more precise infectious disease surveillance systems geared to identifying situations in which population movement creates conditions for infectious diseases outbreak. Surveillance systems could be set-up accordingly in a more geographically focused manner and perhaps contribute to the earlier identification of future emerging infectious diseases. Preventive action, through conflict resolution and better education and discipline of troops, is implicit. 4.1 Linkage of HIV incidence and prevalence patterns deduced from surveillance and other data sources and relationship to interventions. Summary findings in Uganda and Malawi: > Substantial declines in HIV prevalence ranging between 30% to 50% have occurred among pregnant women at sentinel surveillance sites in urban areas of Uganda between 1992 and 1996; in rural areas the prevalence appears to have remained more stable. > Declines in HIV prevalence have been largely confined to 15-19 and 20-24 year old age groups and suggest that recent decreases in HIV incidence in urban areas (corroborated by epidemiolgic models) in Uganda could be as high as 50% or more. > Data from rural sites were not available for age-stratified analysis, but rural areas in Uganda do not appear to have experienced a major incidence reduction among younger age groups. > Biases related to changes in demographic mix of the populations sampled, postulated changes in fertility, laboratory measurement error, or natural HIV infection dynamics are unlikely to account for the observed changes in HIV prevalence in Uganda. > There is compelling evidence of substantial reductions in high-risk sexual behaviors including increases in condom use, particularly in urban Uganda. A causal link between this behavior change and the declines in prevalence in urban pregnant women is likely. There is also evidence of behavior change in rural areas, but not to the same degree as in urban. Use of condoms in rural areas is much less common than in urban areas, and may relate to access or knowledge. > Declines in HIV prevalence have been largely restricted to urban and semi-urban areas, and this is of great significance given that 89% of the 16 million Ugandans are rural. The apparent lack of decline in rural areas may relate to less effective rural interventions. > Age-stratified analyses of ANC sentinel surveillance data indicate recent changes in HIV incidence patterns and may prove to be a feasible, economic and sustainable method of assessing the effectiveness of intervention strategies in sub-Saharan Africa. This method may be more a more sensitive indicator of recent HIV incidence than heretofore recognized, and when linked to behavioral data, could be a useful evaluation tool. -68- > The intervention strategies, which have focused on behavior modification in Uganda, appear to have had a significant dramatic effect in urban and semi-urban areas, and now need to be sustained and extended to rural areas and other countries in the region. > The lack of decline in prevalence levels overall and in the 15-24 age group in pregnant women in Blantyre provide evidence of continuing epidemic growth and much higher incidence rates in younger populations than in urban Kampala. Behavioral data in Malawi suggest that this trend may relate to less behaviour change as compared to urban Uganda. The reasons for this difference need to be determined and lessons learned extended to other countries. 4.2 Summary findings on HIV dynamics and population movement: > Population movement patterns have an important impact on the epidemiological dynamics of infectious diseases. However, the role in which migration patterns play in further dissemination of HIV or in the facilitation of the emergence of other new infectious agents is a neglected area of research. > The increased mobility of populations in Sub-Saharan Africa, particularly with respect to urbanization and rural-urban-rural migratory circulation patterns, will increasingly become an important vector for infectious diseases. A better understanding of the relationship of migration to disease will contribute to improved prevention and control as well as surveillance strategies. > Preliminary findings from research on migratory patterns as potential geographic vectors for the transmission of HIV suggest that lower urban/ rural prevalence differentials in Malawi as compared to Rwanda may relate to a greater mobility of the Malawi population. In Malawi there is a tendency for greater urban to rural and urban to urban migration patterns than in Rwanda. If such migratory patterns in Malawi are present elsewhere, it may have important implications for HIV spread to rural Africa, with 70% of the continent's population. > The hypothesis that military activity in south-eastern Uganda in 1979-80 may have contributed to magnification of HIV spread is worthy of further research. It has implications for not only understanding better disease diffusion dynamics, and but in further understanding the broad implications of civil conflict on public health. -69- ANNEX M EVALUATION IN-SERVICE TRAINING IN SYNDROMIC MANAGEMENT OF STDS8 Major Findins A July 1997 evaluation sponsored by the STAFH Project selected 26 health units (health centers and hospitals that included STAFH-supported pilot sites) to assess quality of STD case management. The evaluation methodology included 158 questionnaires completed by service providers, 129 observations of client- provider interactions, interviews with 9 trainers of syndromic management of STDs; interviews with 85 clients both before and after being seen by a service provider, and interviews with 25 non-clinical ground staff at the units and interviews with 21 managers. In summary, the evaluation found that in-service training on syndromic management of STDs seems to have been reasonably successful in preparing for an adequate management of the STD approach-but only in the pilot sites. The pilot sites performed far better than non-pilot sites in almost all instances. The only factor that might explain this difference in performance was the constant support and attention in terms of frequent visits and feedback on performance. Other factors (drugs, equipment, staff, reference materials etc.), previous experience or other training were not different between pilot and non-pilot sites. Although the pilot sites had overall better performance, a number of components require improvement (Figure MI, next page). On the diagnostic side: The following of diagnostic algorithms was accurate and thorough in only 56 percent of observations in pilot sites (25 percent in non-pilot). The accuracy of STD treatment was only 67 percent overall in pilot sites (100 percent for urethral discharge, 63 percent for genital ulcers and 50 percent for genito-urinary symptoms in women) compared to 17 percent in non-pilot sites. Observers believed genital examination was generally very poor in quality. Male genitals were examined in only 69 percent of pilot site observations of STD clients (50 percent non-pilot), although female genitals were examined in 89 percent of pilot site STD observations (82 percent non-pilot). A speculum was used for female examination in only I out of 9 instances when a speculum was available in the pilot sites (compared to 7 out of 8 instances in non-pilot sites). The above two sets of findings imply that the basic training itself is modulated considerably by other factors in the workplace, and that the more important variable in ensuring good syndromic management is the frequency and extent of support provided. Only 43 (29 percent) of the 156 staff interviewed said they had any supervision at all (pilot sites were equal to non-pilot sites in this regard, and of these 43 only 5 said they felt their supervision was adequate. It should be noted that the pilot site staff were referring in this instance to supervision from their unit managers or coordinators-and not to the support from the STAFH project. 8 Taken from the "Evaluation In-Service Training in Syndromic Management of STDs." Malawi, July 1997. STAFH Project. -70- Figure M 1 Malawi - Evaluation of Syndromic Management of STD's (July, 1992) Folloaing of chagnostic agort 2 Pilot Accumcy of STD Veatent 1 Non ilot Mae genitals examned 89% Femals genitals examined 29% Staff eceived superves2on Pat"nr notricaton 87% discussed 43% Avfabaity ol condoms 9% 89% 86% Condoms given to clints Education was provided on all the required aspects in over 83 percent of pilot site observations (compared to 35-50 percent in non-pilot sites), but this was often regarded as defective by observers. Providers' listening skills were seen to require some improvement: providers asked about the knowledge of the client in pilot sites in about 65 percent of observations, and in about 25 percent of non-pilot sites. Providers asked questions about a client's feelings in both pilot and non-pilot sites in approximately 60 percent of observations. Client interviews revealed a majority stating that they had not asked the questions they would like to. Attitudes to issues of gender, sexuality and STDs may be playing an important part in interactions between staff and other staff, as well as between staff and clients. One of the most obvious of these concerns beliefs about condoms. However, examination of genitals of the opposite sex also caused some problems and there were several stated beliefs about sexual practice that revealed considerable prejudice. There were many statements by service providers that the courses they attended were too short to allow of any practical experience or practice, or of adequate training in any of the varied communication skills required in the syndromic approach. In general, the key elements that require better preparation are: a better understanding of Genito-Urinary Symptoms in women and of Pelvic Inflammatory Disease; more accurate following of the flow charts and of prescribing practice; a better understanding of reproductive anatomy and physiology, improvement of listening and communication skills, better management of the impact of personal views and biases during the interactions with clients and other staff. Partner notification was seen as a difficulty by most staff, but the issue was discussed with clients and slips for partners given in 87 percent of observations in pilot sites (compared to 43 percent in non-pilot sites). -71- Attempts at follow-up were good, with all pilot sites asking the client to return (82 percent in non-pilot sites) and 92 percent of clients receiving a card or stamp to do so. However, only half had a method to check if clients returned. Counseling was done well in pilot and non-pilot sites only in certain respects and remarks by observers suggested a less than satisfactory performance. The availability of condoms was very high (96 percent of pilot sites, 89 percent of non-pilot sites). However, while they were given to clients in 86 percent of pilot site observations, they were only given in 44 percent of non-pilot site observations. The overall approach is very well received by the majority of managers of the units, who believe it is appreciated by both staff and clients, and that it is an easier system to mange than previously. This is despite the fact that some of the managers had had no orientation to the syndromic approach. Most of the staff involved had a good idea of why the approach was used, and the principles on which it was based. There is lack of consistency in the extent to which STD management services are integrated with those of the rest of the health service. In some units the staff and facilities are separated from the rest of the clinic work, and in other units people with STDs are seen in the same area and by the same provides as others. This at times led to problems in the separation of staff and responsibilities, and in the management of resources such as drugs. There was also a belief that separation led to stigmatization of clients. Many service providers commented to the observers that their training had been hampered because of the mixture of trainees with widely different experiences and abilities. There was a noticeable incompatibility of various management information systems. The old system of record-keeping had little relevance for the syndromic approach, and this made accurate assessments of numbers of patients seen or comparisons with other units difficult. Perhaps as a result of clinging to older systems, it was reported by 25 percent of service providers that they check their results for STDs with the laboratory. This is despite their training in a syndromic approach and occurs in both pilot sites (29 percent) and non-pilot sites (24 percent). This clearly wastes resource and time, as well as making pointless the syndromic approach. One major problem with the system, noted by all the observers is that despite the fact that nurses have been trained to prescribe drugs for the STDs, and are legally allowed to do so, in the majority of cases they continue to refer clients to Clinical Officers or Doctors for agreement with the diagnosis and prescribing. This practice takes away from the whole point of Syndromic Management, and considerably increases the amount of time clients have to wait. Interviews with several categories of non-clinical staff revealed their extensive involvement with clients and with their communities in terms of providing advice and education in relation to sexuality, AIDS and STDs. -72- RecomMendations 1. A series of orientation courses for locally-based trainers, supervisors and managers to the syndromic management of STDs so that they can understand the requirements of support and supervision for the syndromic approach. 2. Improvement of general management of staff. This would be as important as STD-specific supervision and would require improvement of the skills of managers/supervisors/coordinators/trainers so that they can provide better general and management support, consistently and frequently, to service providers as they come across difficulties. This measure should by itself improve current performance of those trained and maintenance of the system once established. 3. A series of refresher courses for all those currently trained in syndromic management. 4. Extension of initial in-service training which would be separated as follows: -five full days be allocated to the development of clinical skills. -five days of in-service training on providing adequate communication skills for counseling and education. 5. The training in counseling and education to concentrate on a few sharply focused areas of investigation into the client's knowledge, establishing whether the client needs to ask question(s), and providing education only that is relevant to the situation of the client. 6. Include in all training a set of values clarification exercises to review and understand feelings and beliefs for providers concerning sexuality, STDs and gender, and their potential impact on the work. 7. Mention in training that it is a function of trained staff to provide education to the non-clinical staff in their units. 8. Trainees to be of equal background and experience. 9. Standardisation of the Management Information System for the recording of data related to STDs, HIV and AIDS. -73- ANNEX N The Increased Health Care Costs From The AIDS Epidemic According to several assumptions9 outlined below, the AIDS epidemic is likely to be responsible for an increase in health expenditures, of about 81 percent from mid 1985 to 1997. If national seroprevalence increases beyond 13 percent, additional increases in health expenditures would also be expected. Assumptions: (a) The HIV prevalence rate of adults in Malawi has been stable at 13 percent for two years. HIV prevalence rose from zero in 1985 to 13 percent in 1996. (b) Crude death rate before the epidemic in the 15-50 year age group is equal to 5 per 1000. (c) The Future Group's "AIDSPROJ" model was used to project the number of adult AIDS cases, the consequent impact on adult mortality and the resulting impact on demand for health care. The following assumptions were also used: * Rate of progression to AIDS: Slow. * Year in which adult prevalence was around zero: 1985. * Year in which adult prevalence attained 13%: 1995. * Health care demand will increase by 1/4 the increase in adult mortality. * Baseline adult mortality is 5 per 1000. * Population crude birth and death rates from Zimbabwe. With these assumptions, AIDSPROJ predicts in Figure NI, next page, the following time path of the increase in adult deaths and of the increase in the demand for health care. The assumptions are from 'Confronting AIDS, Public Priorities in a Global Epidemic'. A World Bank Policy Research Report. Oxford University Press 1997. -74- Figure NI Estimated Percentage Increase In the Demand for Health Care In Malawi as a Result of the AIDS Epidemic 250% .0. Increased Adult Mortality, .- Increased Demand -.-Adult Prevalence 200% 150% 100% 50% - --h*-- - 1985 1990 1995 2000 2005 2010 Year (d) According to the above projections, in 1998 the increase in demand is expected to be around 50 percent. (e) The cost increase associated with an HIV epidemic uus not uepend on the size of the epidemic and it will be about 10 percent regardless of the size of the epidemic. There are three components of the increased cost per unit of equal quality care. These are (1) clean blood, (2) universal precautions and (3) improved training of and compensation for health care providers. All of these decisions are typically made system-wide on a once- and-for-all basis. One can argue that both the fixed costs and the political economy of implementing these innovations require that they be done system-wide once HIV-infection rises above some threshold, such as 0.1 percent. This view, which is the one adopted in Confronting AIDS leads to the conclusion that the cost increase associated with the epidemic does not depend on its size and it is assumed to remain around 10 percent. (f) Assuming that elasticity on demand is 0.8 percent, elasticity on supply is 0.5 percent and patients pay half of the health care impact on the Malawi health sector will be to increase all components of expenditure by 81 percent. Since we don't really know the subsidy rate in Malawi, the following table gives the calculations for three different subsidy rates, 0.25 (i.e. 3/4 of health expenditure is out-of-pocket), 0.5 (half is out of pocket) and 0.75 (only one quarter is out of pocket). Thus, the higher the subsidy rate, the higher will be the percentage impact of the same 13 percent prevalence rate. Subsidy is 0.25 Subsidy is 0.5 Subsidy is 0.75 Price increase (%) 44.3% 51.7% 64.9% Quantity increase (%) 14.0% 18.9% 28.0% Expenditure increase(%) 64.6% 80.5% 111.0% -75- ANNEX 0 Blood Safety in Malawi A. Current Status There is currently no centrally organized blood transfusion service in Malawi. Each hospital is responsible for organizing its own blood donor recruitment, blood collection and laboratory testing. For most of the district and CHAM hospitals the blood donors are relatives of the recipient patients. At Queen Elizabeth County Hospital (QECH), Lilongwe County Hospital and Zomba County Hospital there are volunteer networks called "friends" of each of these hospitals who staff mobile vans and collect blood from non-remunerated volunteer donors. At QECH about 40 percent of the blood is from these voluntary donationsi. The major recipients of blood transfusion are children with malaria and malnutrition related anemia, and women with complications of pregnancy. In 1993 there were about 30,000 blood donations, 25,000 of which were considered "suitable" for transfusion. Blood for transfusion in Malawi is tested for HIV, Hepatitis B Antigen (HBsAG) and syphilis. It is not tested for Hepatitis C Virus or malaria. About 22 percent of the donations are HIV positive, 9 percent are positive and 4 percent are positive for syphilis. Both for syphilis and HIV, seroprevalence levels are higher in relatives compared to volunteer donors (25.7 percent vs 17.2 percent for HIV). Reagents for HIV, HBsAG and syphilis testing are supplied by the NACP. Some CHAM hospitals also receive donations of test kits from volunteer organizations oversees. The NACP also provides the hospital with supplies such as gloves, syringes, needles, biological waste bags and needle disposal bags to supplement their need B. Lessons Learned Although nearly all blood is screened for HIV, the supply of test kits is intermittent and inadequate to meet national needs. This together with the high rates of HIV infection in blood donors result in high percentage of unmet transfusion needs. For example, at QECH in 1996, only 41 percent of the 7320 requests, for blood transfusion were met. Young children under 5 years of-age sometimes receive unscreened blood donated by their mother when reagent for HIV testing are not available. The shortage of blood for transfusion is also due to the reluctance of the majority of adult Malawians to donate blood voluntarily. Many rural Malawians have misconceptions about blood; "they do not have enough blood themselves, they cannot give it to a non-relative, bjood donation will weaken their sexual performance," 2 etc. Urban Malawians avoid donating blood, fearing they will be found to be HIV infected . The Malawi Red Cross Society has only been nominally involved in the mobilization and education of the community for blood donation. To address this growing gap between the national blood transfusion needs and the availability of safe blood, the Ministry of Health with the leadership of the NACP Secretariat has recommended the establishment of a centralized national blood transfusion service which will have four regional banks each responsible for an -76- assigned number of hospitals. This plan was endorsed by the World Health Organizations global blood safety initiative. The plan was submitted to the Netherlands Red Cross, the European Community and the Japanese Government between 1992 and 1996 for financial and technical support. There has been no favorable response to date. In addition, guidelines for safe blood transfusion practice have been developed. Among other things the guidelines limit the clinical situations under which blood transfusion is indicated. Orientation of clinicians in their use is yet to be done. C. Major Issues * Insufficient reagents are available for testing due to inadequate budgetary allocation and failure to estimate national needs. * High HIV seroprevalence is found among blood donors little or no donor selection and collection of a blood unit is carried out before laboratory testing is done. This results in wastage of a lot of resources including the collected blood. * There is a lack of interest among Malawian adults to donate blood due to held misconceptions or fear of being identified as HIV seropositive and lack of an education campaign to address these issues. * There is need for a centralized national blood transfusion service. However, national resources are lacking. D. Proposed Actions * An adequate budgetary allocation must be provided to ensure sufficient reagents and supplies for HIV, HBsAG and syphilis testing of blood. The laboratory coordinator must have accurate data of national test kit needs to provide a basis for financial allocation. * Improve donor selection and review current reliance on relative donors and the practice of collection of a unit blood before laboratory testing for HIV, HIVsAG and syphilis, in view of the large wastage of resources due to HIV infection. * Address through IEC campaign the reluctance of adult Malawians to voluntarily donate blood and to correct their misconceptions about blood and blood donation * Identify financial and human resources to establish the much needed national blood transfusion service. E. Estimated Annual Blood Safety (estimated) Budget 1. Blood Screening for HIV, Hbs Ag and syphilis U$200,000 2. Supplemental Safety Supplies U$120,000 3. IEC, promotion of blood donation US 10,000 4. Health care workers training U$ 25,000 Total US 355.000 -77- The proposal for the development of a national blood transfusion service was estimated to cost U$5 million over a five year period. References 1. Barnaba AK: HIV, Hepatitis B surface antigen and syphilis seroprevalence among blood donors at QECH Blood Bank (1995) 2. Chanika: Attitudes of Malawians to blood donation (1992) 3. Ministry of Health and Population: Development of a National Blood Transfusion service for Malawi 1992 4. Ministry of Health and Population: Recommended guidelines for the practice of safe blood transfusion in Malawi (1997) -78- ANNEX P Health Care Facility Conditions in Malawi Health personnel estimate that approximately 50 percent of patients admitted to health institutions throughout Malawi have HIV infection. In the urban areas, this percentage is higher. The central hospitals in Lilongwe and Blantyre have bed capacities of 1000 and 1056 respectively. During peak season, October to December, the hospitals use floor beds to accommodate a total of 500 extra patients each. The bed occupancy rate ranges from 80 percent to over 100 percent for all illnesses. Both hospitals are overwhelmed by HIV cases ranging from 50 perceht - 75 percent of admissions to the medical wards and 70 - 85 percent of admissions to the TB wards. As mentioned earlier, a national survey showed that the level of HIV in TB patients, both sputum positive and negative, is 66 percent. A study to determine what percentage of inpatients in a variety of health care settings, are HIV positive, is planned under the Mid-term Expenditure Framework (MTET) exercise. Fifty percent of all TB cases are diagnosed in the three urban centers of Lilongwe, Blantyre, and Muzuz. The average hospital length of stay ranges from 14 - 21 days for HIV disease and 60 days for sputum positive TB. Because of overcrowding in TB wards, this 60 day duration for TB is now often reduced to 30 days. In contrast, duration of hospital stay for malaria is four to six days and seven to ten days for an appendectomy. Health services at district level and several mission hospitals help meet many of the health care needs in Malawi's rural districts. The Mponela rural hospital sees approximately ten AIDS patients in the OPD a day and four are admitted every week. The admittance of terminally ill patients is felt as a strain by hospital staff. In addition, the effects of a cash budget handicap staff from providing adequate services. A home care program operating in the hospital's catchment area has helped ease the burden and HIV-related admissions appear to have fallen. The program is run by a community group, which has trained 120 volunteers as the main home care visitors. Their coverage ranges from five to ten kilometers. Findings are similar at St. Luke's hospital. About 60 percent of patients on medical wards are estimated to be HIV positive. The burden is relieved somewhat by an extensive home care program that caters to a total of 300 patients, 120 of whom are visited regularly. The success of the home care program is due to the participation of the community - over 100 volunteers are available to make home visits. -79- ANNEX Q COMMUNITY-BASED CARE IN MALAWI A. Current Status And Key Issues There are over 85 community-based care (CBC) support groups currently operating today in Malawi. Most CBC support groups are from small community-based NGOs or church organizations. This study examined eight of these programs (See annex R for a descriptive table). The annual cost of these programs are highly variable, from $1,883 for a small hospital based program to $184,232. This variation is due largely to differences in the size of the population served and the wide variety of services provided among the programs. The average annual cost for the programs is about $54,000. It is fair to estimate that a "typical" CBC support group providing a range of services would cost roughly $45,000 per year. A program limiting services to providing clinical and counseling care to PLWA's would cost roughly $24,000. Figure Q1 Percentage of Total Expenditure (1996) for Three HBC Program by Input Category Total Expenditure = $285,570 Salaries Direct Care 32% 18% Transport MOR Training 30% 12% Figure Ql shows how money is allocated among these programs. As would be expected, the largest proportion (62 percent) of money goes toward operating costs such as salaries. This is, a significant higher proportion of funds spent on operating costs than within prevention activities such as STD/AIDS (10 percent) or interventions directed toward youth (8 percent). About 18 percent is spent on direct patient care such as drugs, material supplies, Income Generating (IG) seed money or orphan support. -80- Figure Q2 Percentage of Total Expenditure (1996) for One HBC Program by Input Category Total Expenditure = $42,743 Direct care 24% Salaries 6% Transport Other 19% -4% Operating costs Training 4% 43% Figure Q3 Percentage of Total Expenditure (1996) for One HBC Program by Input Category Total Expenditure = $58,594 Salaries Direct care 5% 55% Other a& 9% Transport Training R 23%-8 -81- Figures Q2-Q3 show how resources from two programs are allocated. Their patterns of spending were consistent for operating, transport and salaries costs (about 4 percent, 20 percent and 5 percent of total expenditure, respectively). The programs showed a significant difference in philosophy about the need for training with a range of from 4 percent to 43 percent of their total costs going to this activity. Evaluation of programs in other countries have shown that careful planning for training agenda's is critical to CBC efficiency. It is not cost-effective to overly invest in training at the cost of direct patient care. In this example, the program spending a smaller proportion of its resources on training, spent more than twice the proportion of total costs on direct patient care (e.g. drugs, IG support) than the other program (56 percent compared to 24 percent). CBC Programs in Malawi provide a wide variation of services which was reflected in the differences in their costs. These included: * providing care and social support to PLWA's and elderly/orphans; * sensitizing the community on HIV/AIDS; * carrying out IEC activities; * creation of IGs; * counseling; * development of youth programs. The government has not developed a home-based care policy or program nor has it provided direct support to the programs. However, it is evident that RHOs and DHOs, with their AIDS coordinators and AIDS committee structures, offer support to community-based programs. Major achievement has been accomplished in the development of community-based care. These include: * Home-based care training manual used in many programs; * Many existing models of community-based care; * Some training of CBC providers; * Decreasing stigmatization of PWAs; * Strong community support. There are excellent models of CBC support programs in Malawi. The two major models are: * Hospital based programs usually found in CHAM hospitals; * "Comprehensive HBC models" which integrate preventive activities and other support activities along with care to its community members. The key elements of successful models include their ability to demonstrate: * Successful recruitment and sustainability of a core group of committed volunteers; * Strong integration with hospitals and other community resources; * Commitment to coordinate with the regional and district government infra-structures: * Ability to provide quality services with little resources; * Effective management/leadership and supervision infra-structures. -82- Three major issues which most negatively affect the success of CBC programs and require critical attention in Malawi include: * lack of comprehensive coverage; * inability to meet the medical and material needs of PLWA's, their caregivers and orphans; * sustainability. Services are not evenly distributed throughout the country, or even within program catchment areas. The key constraints to providing equitable CBC services are: * sporadic CBC program development; * inaccessibility to patients in,many outlying areas; * transport shortages among programs; * inability to identify many PWAs. IG activities have been successfully implemented in a number of CBC programs. There is evidence of profitability in some programs. Positive "ripple effects" from IG activities include: effective incentive for volunteers; a developed sense of comradeship/community by IG participants; an alternative activity for female youth who might otherwise turn to commercial sex; and, as avenue to employ PWA. IG activities developed by one CBC program were found to be sustainable, financially profitable and to offer a number of other positive results. The activities engaged one employee (home-based care provider), 23 volunteers (including 5 HIV+ and 3 orphans). The goal for the IG activities was to produce enough income to provide care for widows of PLWA's. Profitable activities included: recycled paper briquette making, sunflower oil and vegetable garden, piggery and chicken rearing projects. Other projects have not been successful either due to a lack of community cash flow or product quality (e.g. soap making projects). B. LESSONS LEARNED Currently, only 10 percent of PLWA's receive CBC and it is unlikely that universal, comprehensive coverage can be obtained. There is no strategic plan or national program for providing care for PLWA's in Malawi. Some provision of care is provided for many PLWA's living at home, however, care is not provided systematically or universally. Rather, coverage is sporadic and is not necessarily reaching the most needy patients. Programs are widely varied in the services they provide and do not provide equitable services to all patients. Finally, there is little integration of CBC with the rest of the health sector or with other CBC programs. There has not been any evaluation of the home-based care provided in Malawi. To adequately plan for optimal coverage of CBC, the level of demand/need for care and the cost-effectiveness of alternative levels of services provided need to be measured. Some CBC programs in Malawi have shown impressive ability to provide a wide scope of effective services with little resources. There are several low cost, successful programs that have significantly reduced volunteer turnover: These include: -83- * Routine meetings to discuss problems and concerns; * Monetary rewards for longevity such as bicycles, food allowance, IG training and seed money, and material support such as soap;, * Intrinsic rewards such as status titles, reward systems, simple uniforms such as t-shirts and recognition; * Effective management and supervision providing a structure for feed-back and communication. Counseling skills among community-care workers and providers requires strengthening. There is need for increased and sustained counseling training. Effective initial and refresher training of volunteers is a critical component to successful programs. C. MANAGEMENT ISSUES: Successful CBC programs require effective management infrastructures. Planning and management sKillS among CBC programs require strengthening. Program planning is required with emphasis on prioritizing services and coordinating with other CBC programs, hospitals and sectors. There is, for example, a lack of hospital discharge planning or referral mechanisms which compromises CBC coverage, equity, continuity of care and quality of care to PWAs. There is one model CBC pfogram reporting high referral patterns and consistent hospital reporting of discharged PWAs. There is a critical shortfall of basic, essential medicine and support supplies available to CBC programs. PLWA's living at home lack access to emergency essentials such as ointment to clean infected sores and soap to clean themselves. Stronger linkages with health centers who can access a consistent supply of low cost drugs and supplies is required. NACP must play a leadership role in defining the level of supplies that need to be available to CBCs and to offer the logistical systems and training to assure that the supplies are accessible to PWAs. D. SUGGESTIONS FOR THE WAY FORWARD: > Develop a cost-impact study of CBC resulting in a multi-sectoral strategy plan for developing CBC for PLWA's, in Malawi to include: * optimal level of coverage; * optimal range of services;. * assurance of access to essential, emergency medicines in CBC programs; * human resource strategies. > Develop two pilot CBC sites to demonstrate a model CBC program based on'the evidence from the above study. > Build capacity at tne national level through 2 planning workshops to plan a national CBC strategic plan of action. -84- > Develop a national strategic plan of action for providing low cost care to PLWA's. > Two-three days training of trainer workshops in two or three CBC sites, involving local stakeholders (district AIDS coordinators, DACC CBC technical sub-committee members, local representatives) to familiarize them with the need and "how to" for measuring and prioritizing CBC need. > Carry out a two person-weeks consultancy to develop a plan of action for providing two-three successful NGO/CBO's, engaged in CBC activities, skills required to effectively sustain their services through improved integration, efficiency, and resource mobilization. > Establish a social action fund or NGO Grant Facility to support CBC and to: * develop a strategy to build capacity in CBC programs; * improve NGO coordination, accountability and strategic planning/management capacity. > Develop a central reporting system for monitoring the total costs and impact of NGO/CBC activities. -85- ANNEX R 1996 EXPENDITURES (in USS) FOR EIGHT HBC PROGRAMS BY INPUT CATEGORY AGENCY TOTAL Salaries Mgmt./ Training Coordina- Transport Direct OTHER Overhead/ tion Costs Care Recurrent Meetings Costs Episc Conf. of MW 2,670 2,402 2,280 1,775 13,380 33,277 5,480 58,594 Mponela AIDS Info 426 2,379 2,742 5,121 & Council Centre Mulanje Miss. 1,496 58 329 1,883 Hospital Montfort HBC 164 3,388 3,552 Phalombe Comm 2,632 1,611 18,192 2,880 7,895 10,356 1,809 42,743 Care Project Blantyre Christian 13,145 2,128 395 342 37,688 40,553 Centre Save the Child. (US) 87,382 59,819 16,121 16,115 92,177 184,232 Save the Child. (MW) 33,000 65,684 98,684 TOTAL USS 106,255 98,960 41,027 7,797 24,663 59,748 203,167 435,362 -86- ANNEX S DESCRIPTION OF SIX HOME BASED CARE PROGRAMS NAME CATCH DIS #VOL # TRAINING SERVICES #REG # ISSUES SIZE (#KILO) STAFF VISIT (#VILLAGE S MANSO 6 6 25 3 Yes Counseling, 20 3 hrs. referrals, (24%) food, training hygiene PHALOMBE 7 40 135 3 Yes Support 270 1 x wk Excellent groups, PWAs, 1-2 hrs model for IEC, 166 mgmt counseling, orphan and moni- orphan , toring support, 157 H.E., guar. Training, household, drama, drugs. MULANJE 19 25 47 1 Yes Counseling NA Ix/wk co- MISSION gloves, I hr. ordination HOSP clothes, a problem condoms, drugs, IG EPISCOPAL NA NA 120 9 Yes IEC, NA I x per good CONF. OF HBCP Counseling wk. referral MALAWI nursing, 15- 16 and spiritual, clients mgmt. training, /wk. infra- material, Per structure IGAs, *HBC orphan care P) + Elderly drugs SAVE THE 6 (30,000 NA 50 25 Yes Youth, NA l x /wk transport CHILDREN pop) supplies, a problem MALAWI counseling, training MONTFORT NA NA NA 29 Yes Counseling, 400 NA Follow- HBC supplies, up by training health officer needed -87- BIBLIOGRAPHY Aggleton, Peter. 1996. "Global Priorities for HIV/AIDS Intervention Research." Thomas Coram Research Unit, Institute of Education, University of London, London, WC1HOAA, UK. International Journal of STD and AIDS; 7 (Suppl. 2): 13-16. Ainsworth, Martha. 1995. Deon Filmer and Innocent Semali. "The Impact of AIDS Mortality on Individual Fertility: Evidence from Tanzania." A paper prepared for the Workshop on "The Link between Infant and Child Mortality and Fertility". Committee on Population, National Research Council, Washington, D.C. November 6-7. Alliance Newsletter, Vol. 2, no. 3, 1996. "Malawi: Regional Policy Workshop on AIDS Prevention in a Civil-Military Environment." Mangochi, Malawi. Anderson, R.M. et al. 1991. "The Spread of HIV-1 in Africa: Sexual Contact Patterns and the Predicted Demographic Impact of AIDS." Nature. Vol. 352 no. 15. Aral, S.O. and T.A. Peterman. 1996. "Measuring Outcomes of Behavioral Interventions for STD/HIV Prevention." International Journal ofSTD and AIDS: 7 (Suppl. 2): 30-38. Baier, Erich G. 1997. "The Impact of HIV/AIDS on Rural Households/Communities and the Need for Multisectoral Prevention and Mitigation Strategies to Combat the Epidemic in Rural Areas (with Special Emphasis on Africa)." Food and Agriculture Organization of the United Nations (FAO), Rome. Behets, Lule G. and I. F. Hoffman et al.1994. "STD/HIV Control in Malawi and the Search for Affordable and Effective Urethritis Therapy: a First Field Evaluation." University of North Carolina at Chapel Hill, Department of Medicine, 547 Burnett-Womack, CB 7030, Chapel Hill, NC 27599. Behets, Freida M-T. and George Liomba et al.1995. "Sexually Transmitted Disease and Human Immunodeficiency Virus Control in Malawi: A Field Study of Genital Ulcer Disease." The Journal of Infectious Disease: 171:451-5. Bicego, George. 1997. "Using National Survey Data to Estimate Mortality Rates in the Context of the AIDS Pandemic in Sub-Saharan Africa." The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Bloom, David E. and Ajay S. Mahal. 1995. "Does the AIDS Epidemic Really Threaten Economic Growth?" Working Paper No: 5148. National Bureau of Economic Research, Cambridge, MA. Carpenter, Lucy M. and Jessica S. Nakiyingi et al.1997. "Estimates of the Impact of HIV-1 Infection on Fertility in a Rural Uganadan Population Cohort." Paper presented at The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Carr, Celeste and Jane Begala et al. 1992. "Technical Analysis of HIV/AIDS Situation in Malawi." Prepared for United States Agency for International Development (USAID)/Malawi in preparation for project paper for STAFH Project. (Processed). -88- Chilongozi, D.A. and N. G. Liomba et al. 1994. "Baseline Survey of Sexually Transmitted Disease Case Management." National AIDS Control Programme, Ministry of Health, Conducted in cooperation with Family Health International. Funding by United States Agency for International Development (USAID). Chilongozi, D. A., C. Costello Daly, L. Franco, N. G. Liomba, and G. Dallabetta. 1996. "Sexually Transmitted Diseases: A Survey of Case Management in Malawi." National AIDS Control Programme, Ministry of Health, Malawi. Chiwaya W.B. 1995. "Bar Girl study in Mangochi District." District AIDS Coordinator for Mangochi. (processed). Clotty, C. and G. Dallabeta. 1993. "Sexually Transmitted Diseases and Human Immunodeficiency Virus: Epidemiology Synergy ?" Infectious Disease Clinics of North America. 7: 753-770. Cuddington, John and John D. Hancock. 1992 "The Macroeconomic Impact of AIDS in Malawi: A Dualistic, Labor-Surplus Economy." (Processed). Dallabetta, Gina A., Paolo Miotti and others. 1994. "Traditional Vaginal Agents: Use and Association with HIV Infection in Malawian Women." Supported by the National Institute of Allergy and Infectious Diseases, Bethesda, Maryland, USA. Dallabetta, G.A. and P.G. Miotti et al. 1994. "High Socioeconomic Status Is a Risk Factor for Human Immunodeficiency Virus Type 1 in Women in Malawi But Not for Sexually Transmitted Diseases in Women in Malawi. Implications for HIV Control." Daly Costello, C. and others. "Validation of the WHO Diagnostic Algorithm and Development of an Alternative Scoring System for the Management of Women Presenting With Vaginal Discharge in Malawi." Lilongwe, Malawi. Daly Costello, C. and others. 1995. "Treatment of Women with Vaginal Discharge" Kudya Discovery Lodge, Liwonde, Background Report for Second National AIDS Control Programme Meeting on Sexually Transmitted Diseases. Department of Human Resources and Manpower Development (HRMD). 1997 "Events in the Civil Service." Processed Family Health International. 1995. Malawi AIDS Prevention Project. Final Report. Family Health International and The Center for AIDS Prevention Studies University of California. 1998. "The Voluntary HIV Counseling and Testing Efficacy Study." Final Report Prepared by The Center for AIDS Prevention Studies, University of California at San Francisco. Food and Agriculture Organization (FAO). 1995. "The Effects of HIV/AIDS on Farming Systems in Eastern Africa." Rome. -89- Forsythe, Steve. 1992. "The Economic Impact of HIV & AIDS in Malawi." Family Health International, Arlington, VA, USA. Foster, Susan. 1992. "Cost and Burden of AIDS on the Zambian health Care System. Policies to Mitieate the Impact on Health Services." (Processed). Glynn, J.R. and D.K. Warndorff et al. 1997. "Interactions Between HIV and Tuberculosis in a Rural Area of Malawi." The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Government of Zimbabwe. 1998. Ministry of Health and Social Welfare, AIDS Control Program. "Report of the Review of the Medium Term Plan 2 (1994-1998) for the Prevention, Control and Care of HIV/AIDS/STD." Harare. Gray, Ronald H. and David Serwadda et al. 1997. "Reduced Fertility in Women with HIV infection; A Population-Based Study in Uganda." Paper presented at The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Gregson, Simon and Basia Zaba et al. 1997. "Projecting the HIV/AIDS Epidemic in Southern Africa." Paper presented at The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Grosskurth, H. and F. Mosha et al. 1995. "The Impact of Improved Treatment of STDs on HIV Infection in Rural Tanzania: Randomized Control Trial." Lancet. 346: 530-536. Hayes, Richard. 1996. "Cost-Effectiveness of Improved STD Treatment Services as a Preventive Intervention Against HIV in Mwanza Region" Tanzania. Presentation at the International Conference on AIDS, Vancouver. Helitzer-Allen, Deborah and Hubert Allen. 1992. "Focused Ethnographic Study of Sexually Transmitted Illnesses in Thyolo, Malawi," prepared for Family Health International / AIDS Control and Prevention (AIDSCAP) Project. Helitzer-Allen, Deborah. 1993. "An Investigation of Community-Based Communication Networks of Adolescent Girls in Rural Malawi for HIV/STD/AIDS Prevention Messages." Final Report. Helitzer-Allen, Deborah. 1994. "An Investigation of Community-Based Communication Networks of Adolescent Girls in Rural Malawi for HIV/STD Prevention Messages." Women and AIDS Research Program, Research Report Series No.4. Hudson, Christopher P. 1994. "AIDS in Rural Africa: a Paradigm for HIV-1 Prevention." International Journal of STD & AIDS: 7:236-243. Joint United Nations Programme on HIV/AIDS (UNAIDS). 1996. "Fact Sheet - HIV/AIDS: The Global Epidemic." Geneva, Switzerland 1996a. "UNAIDS an overview. Geneva. -90- . 1998. "Expanding the Global Response to HIV/AIDS Through Focused Action". Geneva. Jones, Clair. 1996. "The Microeconomic Implications of HIV/AIDS." Dissertation. School of Development Studies, University of East Anglia, United Kingdom. Kaluwa, Owen. 1997. Personal communication. National AIDS Secretariat Kaluwa, 0 et al. 1996. "Malawi HIV/Syphilis Sentinel Surveillance Rates in Antenatal Clinics and HIV Seroprevalence Rate in STD Patients." (Processed). Kamali, A. and JAG Whitworth et al. 1997. "Impact of the HIV-1 Epidemic on Orphan Mortality in a Rural Ugandan Population Cohort." The Socio-Demographic Impact of AIDS in Africa Conference, 3- 6 February. Durban. Konde-Lule, Joseph K. 1997. "The Declining HIV Seroprevalence in Uganda: What Evidence ?" Health Transition Review, supplement to Volume 5,27-33. Konde-Lule, Joseph K. and Maria J. Wawer et al."HIV Infection in Rural Households" Rakai District, Uganda. Preliminary Draft. Kornfield, Ruth and Dorothy Namate. 1996. "Report on Preliminary Study of Cultural Practices Which May Be Associated With HIV/AIDS Transmission in Northern Mulanje." JSI-STAFH Research Unit. For the Salvation Army AIDS Education Project. Liomba, N.G., T.E. Taha et al. June 1996. "Trends of HIV and STD Among Pregnant Women in Urban Malawi." College of Medicine, University of Malawi and School of Hygiene and Public Health, Johns Hopkins University, Baltimore, MD, USA. Lodh, Bimal K. 1995. "The Demographic and Economic Impact of HIV-AIDS in Malawi," 1987-2022. November. (Processed). Maha D. and I. Hoffman. 1995. "Prevalence of Genital Infection in Medical In-patients in Blantyre, Malawi." Journal of Infectious Diseases: 1995:31:77-78 Manyasha, M. and D. Lindsey et al. 1997. "A Review of HIV/AIDS Data in Southern Africa, Questioning Collection, Reliability and Interpretation." The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Mason, Karen Oppenheim. 1993. "HIV Transmission and the Balance of Power Between Women and Men: A Global View." Paper presented at the seminar on "AIDS Impact and Prevention in the Developing World: The Contribution of Demography and Social Science." Annecy. McAuliffe, Eilish. 1994. "AIDS: The Barriers to Behavior Change." University of Malawi. Centre for Social Research. -91- McAuliffe, Eilish and Pierson Ntata. 1994. "Youth and AIDS Baseline Survey: Lilongwe and Blantyre Districts." United Nations Children's Fund - Malawi. Mercer, Mary Anne and Nicola Gates et al. 1996. "Rapid KABP Survey for Evaluation of NGO HIV/AIDS Prevention Projects." AIDS Education and Prevention: 8(2), 143-154. Ministry of Economic Planning and Development. 1992. National Statistical Office and Macro International Inc. "Malawi Demographic Health Survey (MDHS)." Ministry of Economic Planning and Development. 1995. National Statistical Office and Centre for Social Research. "Malawi Social Indicators Survey." Ministry of Economic Planning and Development. 1996. National Statistical Office. Macro International Inc. Demographic and Health Surveys. "Malawi Knowledge, Attitudes, and Practices in Health Survey." Ministry of Health and Population. 1997a. "Malawi Health Sector Strategic Plan, Essential Health Package Working Group." Background Paper. Draft. Ministry of Health and Population. 1997b. "Malawi Health Sector Strategic Plan, Health Finance Working Group." Background Paper. Draft. Ministry of Health and Population. 1997c. "Malawi Health Sector Strategic Plan, Hospital Autonomy Working Group." Background Paper. Draft. Ministry of Health and Population. 1997d. "Malawi Health Sector Strategic Plan, Rationalisation, Equity and Quality in Health." Background Paper. Draft. Ministry of Health and Population. 1996. "Planning the Health Sector" Background paper for the Workshop on Health Planning Issue in Malawi. Mangochi. Ministry of Health and Population. 1996. "Population, Health, Nutrition (PHN) Progress Report." Ministry of Health and Population. 1997. "Report on Human Resources Situational Analysis." Ministry of Health and Population. 1993. National AIDS Control Programme. "Annual Report 1993." Ministry of Health and Population. 1995. National AIDS Control Programme, "AIDS Cases Surveillance, Annual Report." Ministry of Health and Population. 1995. National AIDS Control Program, "AIDS Cases Update: 1995 Half Year Report." Data Processing Office, AIDS Secretariat. Ministry of Health and Population. 1995. National AIDS Control Programme, "1995 HIV Surveillance Study Protocol." -92- Ministry of Health and Population. 1995. National AIDS Control Programme, Workshop on "Proceedings of the Second National AIDS Control Programme Meeting on Sexually 1transmitted Diseases." Kudya Discovery Lodge, Liwonde, 25 May. Ministry of Health and Population. 1995. National AIDS Control Programme, "STD Management Guidelines." Ministry of Health and Popilation. 1996. National AIDS Control Programme, "National Conference on AIDS. HIV/AIDS in Malawi: Responses to the Epidemic." Zomba. Ministry of Health and Population. 1997. National AIDS Control Programme, "1996 IIV Surveillance Survey Results." Unprocessed. Ministry of Labour and Manpower Development. 1996. Malawi Congress of Trade Unions, Employers Consultative Assbeiation of Malawi, NACP in co-operation with Organisation of African Trade Union Unity Health, Safety and Environment Programme and CIDA. "Tripartite Workshop on Economic Policy, Employment and HIV/AIDS," Lilongwe. Moses, S., F. A. Plummer, E. N. Ngugi, N.J. Nagelkerke, A. 0. Anzala, and J.O. Ndinya-Achola. 1991. "Controlling HIV in Africa: Effectiveness and Cost of an Intervention in a High-Frequency STD Transmitter Core Group." AIDS 5:407-11 National Research Council. 1996. Preventing and Mitigating AIDS in Sub-Saharan Africa. Research and Data Priorities foi the Social and Behavioral Sciences. Washington, D.C.: National Academy Press. Over, Mead and M. Randall et al. 1992. "The Consequences of Adult Ill Health." In Richard A. Feachem, Tord Kjellstrom, Chistopher J. L. Murray, Mead Over, and Margaret A. Phillips, eds., The Health of Adults in the Developing World New York: Oxford University Press. Over, Mead and Peter Piot. 1993. "HIV Infection and Sexually Transmitted Diseases." In Dean T. Jamison, W. Henry Mosley, Anthony R. Measham and Jose Luis Bobadilla, eds., Disease Control Priorities in Developing Countries. New York: Oxford University Press. Over, Mead, Phare Mujinja, Daniel Dorsainvil, and Indrani Gupta. Forthcoming. "Impact of Adult Death on Household Expenditures in Kagera, Tanzania." Working Paper. World Bank, Policy Research Department, Washington, D.C. Project Hope. 1996. "Project Hope HIV/AIDS Prevention and Counseline Project 1994-1996 Final Evaluation." Draft. Quinn, T.C. "Population Migration and the Spread of Types 1 and 2 Human Immunodeficiency viruses. Proceeding of the National Academy of Sciences (USA) 91 (7): 2407-2414 STAFH. 1996. "First Annual STAFH Project Coordination Meeting. Papers and Proceedings." STAFH. 1997. "Evaluation In-service Training in Syndromic Management of STDs" Malawi. -93- Stoneburner, Rand and Manuel Carballo. 1997. "An Assessment of Emerging Patterns of HIV Incidence in Uganda and Other East African Countries." Final Report of Consultation for-Family Health International by International `Centre forMigration and Health. Geneva. Stover, John and Peter Way. 1998. "Projecting the Impact of AIDS on Mortality." Paper presented at the World Bank, UNAIDS and International Union for the Scientific Study of Population, (IUSSP) Committee on AIDS Workshop on the Demographic Impact of HIV/AIDS, Washington,,D.C. Stover, John. 1996. "The Future Demographic Impact of AIDS: What do,we know?",Prepared for AIDS in Development: The Role of Government-Conference. Chateau de Limelette, 17-19 June. Sweat, Michael. 1998. "The Cost-Effectiveness.of HIV Counseling and Testing." Annex to.the "Final Report of the Voluntary HIV Counseling and Testing Efficacy Study." Prepared for Family Health International. Taha, E.T. and Joesph K. Canier ei al. 1996. "Reported Condom Use Is Not Associated with Incidence of Sexually Transmitted Diseases in Malawi." AIDS. 10:207-212. Taha, E.T. and H.G. Liomba et al. 1996. "Trends of HIV and STD Among Pregnant Women in Urban Malawi." Department of Medicine, University of Malawi College of Medicine and School of Hygiene and Public Health, Johns Hopkins University, Blantyre. Timaeus, Ian M. 1998. "The Impact of HIV/AIDS on Mortality in Sub-Sahairan Africa: Evidence From National Surveys and Censuses." Paper presented at the World Bank, UNAIDS and IUSSP.- (International Union for the Scientific Study of Population) Committee on AIDS Workshop on the Demogragic Impact of IV/AIDS, Washington, DC 12-14 January. United Nations'Popihition Fund (UNFPA) - Malawi. 1996. "Programme Review and Strategy Develpment Report." 1997. "Support to the Population Programme of the Government of Malawi." United Nations Children's Fund - Malawi. 1994. "Family and Community Care Situation Analysis Report." 1996. "Programme Plan of Operation for Youth and Education for the Period 1997-2001." Draft. United States Agency for International Development (USAID). 1993. "Malaria Prevention in Pregnancy: The Mangochi Malaria Research Project." Africa Child Survival Initiative: Combating Childhood Communicable Diseases. Wawer, M. et al. 1998. Way, Peter O. and Karen A. Stanecki. 1994. "The Impact of HIV/AIDS on World Population" by U.S. Department of Commerce, Bureau of the Census. Washington, D.C.. -94- West, G.P. 1996. "The Integration of HIV/AIDS into National Development Planning." November. Eds. Alan Whiteside and Lieve Fransen. Whiteside, Alan. 1997. "Projecting the Epidemic: Policy Makers and Planners Needs." Prepared for The Socio-Demographic Impact of AIDS in Africa Conference 3-6 February. Durban. Whiteside, Alan and Greg Wood. 1994. "Socio-Economic Impact of HIV/AIDS in Swaziland." Capricorn Africa Economic Association. Windily B. and Makhumula P. Nkhoma. 1993. "A KABP Study to Establish Baseline Data in Selected Districts of Malawi for Future HIV Prevention Campaigns - A Survey Report." Infectious Diseases 167:36-42. Wilson, David. 1997. "Project Support Group Update Reports and Presentations." Department of Psychology, University of Zimbabwe. Harare. Processed. World Bank - Africa Sector Report. 1992. "Tanzania AIDS Assessment and Planning Study." Tanzania - Report No: 9825. Final. 1992-06. Washington, D.C. World Bank - Africa Sector Report. 1996. "AIDS Prevention and Mitigation in Sub-Saharan Africa - An Updated Strategy." Report No: 15569. Final. 1996-04-20. Washington, D.C. World Bank. 1997. "Accelerating Malawi's Growth: Long-term Prospects and Transitional Problems." Washington D.C. 1997. Confronting AIDS: Public Priorities in a Global Epidemic. Population, Health & Nutrition Department, Report No: 17285. New York: Oxford University Press. Zambia Ministry of Health and World Health Organization. 1994. "Cost and Impact of Home -Based Care for People Living with HIV/AIDS in Zambia." Lusaka. -95- CATALOGUERSIFOLE CONFIDENTIAL Report No.: 17740' MAI Type: SR