PMTCT A Decision Model for Evaluating Strategies to Prevent Mother-to-Child Transmission of HIV

Publication date: 2002

By Lori Bollinger John Stover Katharine Cooper-Arnold The Futures Group International A Decision Model for Evaluating Strategies to Prevent Mother-to-Child Transmission of HIV Spectrum System of Policy Models The POLICY Project Spectrum PMTCT PMTCT Version 1 A Decision Model for Evaluating Strategies to Prevent Mother-to- Child Transmission of HIV January, 2002 _____________ Spectrum System of Policy Models _____________________________ Lori Bollinger John Stover Katharine Cooper-Arnold The Futures Group International PASCA is funded by the U.S. Agency for International Development under Contract No. 596-0179-A-00-5127-00, implemented by the Academy for Educational Development in collaboration with The Futures Group International. POLICYII is a five-year project funded by the U.S. Agency for International Development under Contract No. HRN-C-00-00-00006-00, beginning July 7, 2000. It is implemented by The Futures Group International in collaboration with Research Triangle Institute (RTI) and the Centre for Development and Population Activities (CEDPA). ii Table of Contents Section Page I. INTRODUCTION. 1 A. Description of the Spectrum System . 1 1. Components. 1 2. Software Description . 2 B. Uses of Spectrum Policy Models. 2 C. Organization of the Model Manuals . 3 D. Information about the POLICY Project . 4 E. What is the PMTCT Model? . 5 F. Why Use the PMTCT Model?. 8 II. STEPS IN MAKING PROJECTIONS. 11 III. PROJECTION INPUTS. 13 A. Context . 13 1. Births (Public). 13 % Public Sector Births. 13 % Antenatal Visits. 15 % Planned C-Section . 16 % Breastfed for 6 Months . 18 % Mixed Feeding . 19 Increase in CMR. 20 MMR (per 100,000). 22 B. Policy/Programs. 23 1. Testing/Counseling . 23 % Offered HIV Test . 23 % Accepting Test . 23 % Receiving Results . 24 # Women Seen/Counselor . 25 % Breastfeeding Spillover . 26 Infections Averted/Woman . 26 2. Treatment Option. 27 3. Treatment . 28 % Eligible for Treatment . 28 iii % Offered Treatment. 30 % Accepting Treatment . 30 % Adhering to Treatment . 30 % Elective C-Section . 31 % HIV+ Breastfeed for 6 Months . 31 % HIV+ Mixed Feeding . 31 MMR Increase: C-Section. 32 4. Costs . 34 HIV Test . 34 Pre-Test VCT, Post-Test VCT: HIV-, Post-Test VCT: HIV+, Post- Test VCT: Feeding. 35 User Fee: VCT. 35 Cost: C-Section Birth . 35 Monthly Formula Costs . 36 # Months Paid by Government . 36 User Fee: Tx . 37 Total Child Treatment Costs. 37 Total Adult Treatment Costs. 37 C. Epidemiology . 38 1. Transmission. 38 2. Reduction. 38 3. ARV Costs . 39 4. Transmission Probabilities - Results . 40 IV. PROJECTION OUTPUTS. 43 V. PROGRAM TUTORIAL . 45 A. Before You Get Started. 45 B. Installing the Spectrum Program . 46 C. Creating a New Projection. 47 1. Starting the Spectrum Program. 47 2. Opening Demographic and AIM Projections . 47 3. Adding the PMTCT Module to the Projections . 48 D. Entering the Projection Inputs . 50 1. About the Editors. 50 2. Specifying the PMTCT Projection Inputs . 51 3. Entering the PMTCT Context Inputs. 53 Births (Public) . 53 4. Entering the PMTCT Policy/Programs Inputs . 54 Testing/Counseling. 55 Treatment Option. 56 Treatment . 57 Costs . 58 iv 5. Entering the PMTCT Epidemiology Inputs. 59 Transmission . 59 Reduction . 60 ARV Costs. 61 Transmission Probabilities - Results. 62 6. Leaving the PMTCT Data Editors . 63 7 Saving the Input Data . 63 E. Making the Projection . 63 F. Examining the Output. 64 1. Graphs and Bar Charts . 67 2. Tables . 68 G. Saving the Projection. 68 H. Opening an Existing Projection . 68 I. Closing a Projection . 69 VI. SAMPLE APPLICATION. 71 A. PMTCT Inputs for Sample Application . 71 B. Changes in Child Deaths and HIV Infections Averted . 73 C. Changes in Treatment Cost Savings, Intervention Costs, and Net Costs. 74 1. Treatment Cost Savings . 75 2. Total Costs Intervention. 75 3. Net Cost of Intervention. 75 4. Benefit-Cost Ratio . 76 5. Net Cost per Child Death and Infections Averted . 76 D. Changes in Policy and Programs Outputs. 77 1. Additional Maternal Deaths. 77 2. The Number of HIV Test Kits Needed . 77 3. The Amount of Extra Formula Required . 77 4. The Number of Extra Cesarean Sections Required. 77 5. The Number of Counselors Needed . 77 6. Distribution of Costs . 78 E. Exploring Alternate Program Configurations. 79 1. Treatment Option. 79 2. Number of Infections Averted . 81 3. Parameters in DemProj and AIM . 82 4. User Fees . 82 v VII. METHODOLOGY . 83 VIII. REFERENCES. 89 IX. GLOSSARY OF TERMS . 103 X. ACRONYMS AND ABBREVIATIONS . 105 APPENDIX A: CALCULATION OF VERTICAL TRANSMISSION PROBABILITIES AND CHILD MORTALITY RATE WHILE BREASTFEEDING . 107 Registration . 111 vi List of Figures Figure 1. Eligibility .6 Figure 2. PMTCT Model Structure.7 Figure 3. Distribution of Costs .78 List of Tables Table 1. Percentage of Births Taking Place in a Health Facility .14 Table 2. Percentage of Women with Antenatal Care Visits.15 Table 3. Percentage of Births by Cesarean Section .17 Table 4. Percentage of Children Breastfed.18 Table 5. Percentage of Children that Use Mixed Feeding .19 Table 6. Effect on Child Morbidity and Mortality of Formula Feeding.20 Table 7. Under-5 Child Mortality Rates .21 Table 8. Maternal Mortality Rates .22 Table 9. Percentage of Pregnant Women Accepting the HIV Test .23 Table 10. Percentage of Pregnant Women Receiving HIV Test Results .25 Table 11. Number of Infections Averted per Person Counseled and Tested by Woman's HIV Status and Country .26 Table 12. Effect of Testing/Counseling for Heterosexual Couples.27 Table 13. Percentage of Women with Antenatal Visits by Pregnancy Stage.29 Table 14. Treatment Compliance Rates .31 Table 15. Effect of Cesarean Section on Maternal Mortality Rates (MMR) .33 Table 16. HIV Testing Costs (US$) .34 Table 17. HIV Counseling Costs (US$) .35 Table 18. Additional Costs of a Cesarean Section Birth.36 Table 19. Formula Feeding: Cost and Duration.36 Table 20. Discounted Lifetime Paediatric AIDS Treatment Costs (US$) .37 Table 21. Reductions in MTCT Transmission Probabilities by Treatment Option .38 Table 22. Summary of Treatment Costs (US$).40 Table 23. Vertical Transmission Probabilities .41 Table 24. Context Parameters for Sample Application.72 Table 25. Testing/Counseling Parameters for Sample Application.72 Table 26. Treatment Parameters for Sample Application.72 Table 27. Cost and Duration Parameters for Sample Application (US$).73 Table 28. ARV Treatment Costs for Sample Application (US$) .73 Table 29. Changes in Child Deaths and Infections Averted for Sample Application .74 Table 30. Changes in Costs for Sample Application (US$) .76 Table 31. Changes in Policy/Programs Outputs .78 Table 32. Transmission Probability - Results for Sample Application.80 Table 33. Changes in Deaths/Infections Averted for Nevirapine vs. Short-Course ZDV.80 Table 34. Costs for Nevirapine vs. Short-Course ZDV (US$) .81 Table 35. Vertical Transmission Probabilities .85 vii viii I. Introduction I. A. Description of the Spectrum System 1. Components The POLICY Project and its predecessor projects have developed computer models1 that analyze existing information to determine the future consequences of today’s reproductive health programs and policies. The new Spectrum Policy Modeling System consolidates previous models into an integrated package containing the following components: • Demography (DemProj) – A program to make population projections based on (1) the current population, and (2) fertility, mortality, and migration rates for a country or region. Spectrum consolidates DemProj, FamPlan, Benefit-Cost, AIM, RAPID, and PMTCT models into an integrated package. • Family Planning (FamPlan) – A program to project family planning requirements in order for consumers and/or nations to reach their goals of contraceptive practice or desired fertility. • Benefit-Cost – A program for comparing the costs of implementing family planning programs, along with the benefits generated by those programs. • AIDS (AIDS Impact Model – AIM) – A program to project the consequences of the AIDS epidemic. • Socioeconomic Impacts of High Fertility and Population Growth (RAPID) – A program to project the social and economic consequences of high fertility and rapid population growth for sectors such as labor force, education, health, urbanization and agriculture. • Prevention of Mother-to-Child Transmission (PMTCT) – A program to evaluate the costs and benefits of programs to reduce mother-to-child transmission of HIV. 1 1 The terms “model” and “module” are used interchangeably in the Spectrum manuals to refer to the computer programs within the system. 2. Software Description Spectrum is a Windows-based system of integrated policy models. The integration is based on DemProj, which is used to create the population projections that support many of the calculations in the other components—FamPlan, Benefit-Cost, AIM, RAPID, and PMTCT. Each component has a similarly functioning interface, which is easy to learn and to use. With little guidance, anyone who has a basic familiarity with Windows software will readily be able to naviga the models to create population projections and to estimate resource and infrastructure requirements. The accompanying manuals contain both instructions for users, and equations for persons who wa te nt to know exactly how the underlying calculations are computed. B. Uses of Spectrum Policy Models to factors that can be hanged or influenced by public policy. th ng those variables and present e results in an accessible way. s commonly addressed by the Spectrum set of models include: • ecisions s can accumulate during a period of policy stasis. • ore single set of actions is acceptable for the benefits gained. Policy models are design answer a number of “what if” questions. The “what if” refer to factors that can be changed or influenced b ed to s y public policy. Policy models are designed to answer a number of “what if” questions relevant to entities as small as local providers of primary health care services and as large as international development assistance agencies. The “what if” refers c Models are commonly computerized when analysts need to see the likely result of two or more forces that might be brought to bear on an outcome, such as a population’s illness level or its degree of urbanization. Whenever at least three variables are involved (such as two forces and one outcome), a computerized model can bo reduce the burden of manipulati th Some of the policy issue the utility of taking actions earlier rather than later. Modeling shows that little in a country stands still while policy d are stalled and that many negative outcome the evaluation of the costs vs. the benefits of a course of actions. Modeling can show the economic efficiency of a set of actions (i.e., whether certain outcomes are achieved m effectively than under a different set of actions), or simply whether the cost of a 2 • the recognition of interrelatedness. Modeling can show how making a change in one area of population dynamics (such as migration rates) may necessitate changes in a number of other areas (such as marriage rates, timing of childbearing, etc.). • the need to discard monolithic explanations and policy initiatives. Modeling can demonstrate that simplistic explanations may bear little relationship to how the “real world” operates. A set of policies under consideration may not be acceptable to all stakeholders. • the utility of “door openers.” A set of policies under consideration may not be acceptable to all stakeholders. Modeling can concentrate on favored goals and objectives and demonstrate how they are assisted by the proposed policies. • that few things in life operate in a linear fashion. A straight line rarely describes social or physical behavior. Most particularly, population growth, being exponential, is so far from linear that its results are startling. Modeling shows that all social sectors based on the size of population groups are heavily influenced by the exponential nature of growth over time. • that a population’s composition greatly influences its needs and its well being. How a population is composed—in terms of its age and sex distribution—has broad-ranging consequences for social welfare, crime rates, disease transmission, political stability, etc. Modeling demonstrates the degree to which a change in age and sex distribution can affect a range of social indicators. • the effort required to “swim against the current.” A number of factors can make the success of a particular program harder to achieve; for example, the waning of breastfeeding in a population increases the need for contraceptive coverage. Modeling can illustrate the need for extra effort—even if simply to keep running in place. C. Organization of the Model Manuals Each manual begins with a discussion of what the model does and why someone would want to use it. The manual also explains the data decisions and assumptions needed before the model can be run, and possible sources for the data. It defines the data inputs and outputs. The manual contains a tutorial, information on the methodology behind the model, a glossary, and a bibliography. 3 D. Information about the POLICY Project The POLICY Project is a USAID-funded activity designed to create a supportive environment for family planning and reproductive health programs through the promotion of a participatory process and population policies that respond to client needs. To achieve its purpose, the Project addresses the full range of policies that support the expansion of family planning and other reproductive health services, including: • national policies as expressed in laws and in official statements and documents; • operational policies that govern the provision of services; • policies affecting gender roles and the status of women; and • policies in related sectors, such as health, education and the environment that affect populations. The POLICY Project is implemented by The Futures Group International in collaboration with Research Triangle Institute (RTI) and the Centre for Development and Population Activities (CEDPA). More information about the Spectrum System of Policy Models and the POLICY Project is available from: Director, The POLICY Project The Futures Group International 1050 17th Street NW, Suite 1000 Washington, DC 20036 U.S.A. Telephone: (202) 775-9680 Fax: (202) 775-9694 E-mail: policyinfo@tfgi.com http://www.tfgi.com or The POLICY Project U.S. Agency for International Development Center for Population, Health, and Nutrition 1300 Pennsylvania Avenue Washington, DC 20523 U.S.A. Telephone: (202) 712-5787 or -5839 4 E. What is the PMTCT Model? In November 1996, it was estimated that more than 90 percent of the 2.6 million paediatric AIDS cases existing worldwide were contracted due to vertical transmission from mother to child (UNAIDS, 1996). The AIDS Clinical Trials Group (ACTG) study 076 developed a treatment protocol to address this issue, where a long course of zidovudine (ZDV) was found to be effective against mother-to-child transmission of HIV (Connor et al., 1994). This protocol was further modified in studies in Thailand, which developed a short course of ZDV treatment (CDC, 1998), the UNAIDS-sponsored PETRA studies that tested short-course ZDV treatment in breastfeeding populations, and the HIVNET 012 trials in Uganda that demonstrated the effectiveness of a single dose of nevirapine for mother and child. In addition, new information about the risk of transmission through breastfeeding has recently become available from several studies. Thus, viable treatment options now exist which reduce the rate of vertical transmission of HIV. The prevention of mother-to-child transmission (PMTCT) model is a computer program that evaluates the costs and benefits of intervention programs to reduce vertical transmission of HIV. As part of the intervention, the PMTCT model contains seven possible treatment choices: long course ZDV; short course ZDV following the Thailand, PETRA Arm A, and PETRA Arm B regimens; intrapartum and neonatal ZDV only; the HIVNET 012 regimen of single dose nevirapine to the mother and child; and universal nevirapine (nevirapine provided to all women and children at the time of delivery without checking for HIV status). In addition to these seven treatment options, the model also allows the percentage of women undergoing a Cesarean section, as well as the percentage of women that breastfeed their infants exclusively and with mixed breastfeeding and food supplements, to be specified. For each of the treatment options, PMTCT requires various data, including the costs of these interventions, and possible user fees to offset these costs. The vertical transmission rate of HIV is provided as a program default and also varies by treatment option, mode of delivery, and method of feeding. In order to be eligible for one of the treatment options, a pregnant woman must attend a public antenatal clinic, where she may be offered a test to identify her HIV status (Figure 1). If she agrees to take the test, she is counseled both before and after she receives her test result. Once she receives the result and is identified as HIV positive, she may be offered one of several treatment options. The 5 end result of the PMTCT intervention on the vertical transmission rate depends on several factors: 1. whether the woman has attended the clinic early enough to be eligible for the testing, counseling, and treatment or had her birth attended by a trained health care provider who could administer the selected treatment; 2. whether she accepts the treatment; 3. whether she complies fully with the treatment requirements; 4. the type of treatment; 5. whether delivery is vaginal or Cesarean; 6. whether or not the baby is breastfed. Figure 1. Eligibility 0 100 200 300 400 500 600 Thousands per Year Infections averted-50% Adhering to treatment -90% Accepting treatment-75% Offered treatment-100% Receiving result-90% Accepting test-90% Offered VCT-80% Ante-natal care-80% Potential HIV+ births The complete PMTCT model structure is shown below in Figure 2. PMTCT compares the intervention costs with the benefits achieved from reducing the number of children who are HIV positive. The model also calculates additional outputs including child and total infections averted, treatment cost savings, net cost of the intervention, and net cost per death or infection averted. The number of deaths and infections averted are calculated based on a comparison with a no-treatment scenario. Sensitivity analyses may also be performed using a second set of assumptions. 6 Figure 2. PMTCT Model Structure Number of births in public sector who attend an antenatal clinic Mother not offered testing/ counseling HIV- Mother HIV- Child HIV+ Mother Vaginal Birth Cesarean Birth Breast- feed Bottle- feed Breast- feed Bottle- feed HIV- Child HIV+ Child HIV- Child HIV+ Child HIV- Child HIV+ Child HIV- Child HIV+ Child Mother offered testing and counseling Mother refuses test/results not returned HIV- Mother HIV- Child HIV+ Mother Mother accepts test & receives results Mother tests HIV- True HIV- Mother HIV- Child False HIV- Mother Mother tests HIV+ Mother eligible for & accepts & complies with treatment Mother not given/doesn’t accept ZDV or other treatment ✫ = flow chart continues in the upper right-hand box labeled "✫" The purpose of the PMTCT model in Spectrum is to help planners evaluate the costs and benefits of policies and programs to reduce mother-to-child transmission of HIV. These programs may include VCT services as well as interventions that deliver antiretroviral drugs. Since governments are the primary decision-makers about these policies, the model focuses on the public sector. When used in conjunction with the DemProj and AIM models, PMTCT encourages policy makers and planners to think about strategies to reduce vertical transmission and the needs of pregnant women and their children. 7 By projecting estimates of future program needs and costs, PMTCT can help planners respond to changes in the HIV epidemic. It should be noted that the provision of high-quality PMTCT intervention treatments, including counseling and testing as well as the logistics of delivering antiretroviral drugs, is a complicated process requiring skilled and dedicated personnel; appropriate technologies, legislation, and infrastructure; and adequate funds. Although the PMTCT model can be used to forecast future program needs and costs to achieve a particular goal, it does not imply that the goals can be achieved merely by allocating the necessary funds. F. Why Use the PMTCT Model? Prevention of mother-to-child transmission projections may be made for several reasons. One of the most common reasons is to estimate the service and resource requirements necessary to meet a goal when introducing an intervention such as VCT and antiretroviral treatment to reduce vertical transmission. The goal may be expressed in terms of demand (i.e., meeting needs for antenatal counselors), resources available (how much antiretroviral therapy can be provided with a given level of funding), continuance or improvement of past trends (in terms of annual decreases in HIV incidence and prevalence rates), or health indicators (reducing the number or percentage of HIV-positive births). Setting up an intervention to reduce vertical transmission of HIV may have important secondary benefits, including opportunities to reduce further heterosexual transmission by motivating women to practice safer sex, preventing acquisition of HIV infection in those identified as negative, and protecting partners in those who are positive. Knowledge of the risk of transmission to the babies of HIV- infected women also allows for vigilant follow-up care of the infants and prophylactic medication to prevent opportunistic infections. These benefits have not been quantified in developing countries and, thus, are not included in the model. Other issues to consider in the introduction of PMTCT programs into health systems include affordability, availability of human resources, infrastructure, equity, and acceptability. Since prevention of vertical transmission should also be considered as part of the wider management of prenatal, delivery, and postnatal care, PMTCT projections can help in setting maternal and child health goals. By determining the resources required to achieve any proposed goal, these projections can help planners choose goals that are attainable and useful. 8 Another important use of the PMTCT model is the examination of alternative program configurations. Rather than simply project the current configuration into the future, the analyst can ask, “What would happen if different treatment options were used? How much would it help if the commercial sector could be stimulated to contribute more for counseling and testing? Would the addition of greater user fees for public sector services have a significant impact on the ability of the program to achieve its goals?” 9 10 II. Steps in Making Projections II. There are seven key steps in making projections within PMTCT. The amount of time spent on each step may vary, depending on the application, but most projection activities will include at least these seven steps. 1. Prepare a demographic projection. PMTCT requires the annual number of births prepared with DemProj. This projection should be prepared first or at the same time as the PMTCT projection. The first year and final year of the DemProj projection will determine the span of the PMTCT projection; the DemProj manual contains instructions on the steps associated with this module. 2. Prepare an HIV/AIDS projection. PMTCT requires HIV prevalence and incidence in the population. The number of pregnant women who have a positive HIV test and the number of women who become infected with HIV after birth as well as treatment cost estimates are linked to the AIM portion of the Spectrum model. This projection should be prepared when creating the DemProj projection or at the same time as the PMTCT projection. The AIM manual contains instructions on the steps associated with this module. 3. Collect data. At a minimum, two types of data are needed: (1) information on the base year demographic and health conditions including the percentage of annual births that will take place in the public sector, the percentage of those births where the mother has attended an antenatal clinic, the percentage of births in the public sector that are delivered by planned Cesarean section, and the maternal mortality rate; and (2) policy and program costs for the years of the PMTCT program, including counseling and testing costs, possible user fees, lifetime discounted treatment costs, and ARV drug prices. Some additional information (such as the marginal cost of a Cesarean section) is required for the projection. Many of these data are difficult to obtain; however, default values are available in the tables in the inputs section. Since the projection will only be as good as the data on which it is based, it is worth the effort to collect and prepare 11 appropriate and high-quality data before starting the projection. 4. Make assumptions. PMTCT projections require assumptions about the future resources available for prevention interventions as well as context, program, and cost characteristics. These assumptions should be carefully considered and based on reasonable selection guidelines. 5. Enter data. Once the base year data are collected and decisions made about projection assumptions, the PMTCT module can be used to enter the data and make a benefit- cost projection. 6. Examine projections. Once the projection is made, it is important to examine it carefully. This includes consideration of the various PMTCT indicators produced. Careful examination of these indicators can act as a check to ensure that the base data and assumptions were understood and entered correctly into the computer program. This careful examination is also required to ensure that the consequences of the assumptions are fully understood. 7. Make alternative projections. Many applications require alternative projections. Once the base projection has been made, the program can be used to quickly generate alternative projections as the result of varying one or several of the model assumptions. 12 III. Projection Inputs III. PMTCT requires (1) data describing the context and HIV counseling, testing, and intervention conditions in the country or region being studied, and (2) data about the use and costs of the different delivery, breastfeeding, and treatment options. Most of these data (e.g., HIV prevalence, breastfeeding practices, and rates of planned Cesarean sections) must be specific for the area being studied, although some of the data (e.g., percentage of women who accept the HIV test and receive the results) can be based either on local data or on international averages when local data are unavailable. PMTCT contains a number of different options that can be used to tailor the model to the needs of different situations. The model contains seven possible treatment choices as well as a no- treatment option: long course ZDV; short course ZDV following the Thailand regimen, PETRA Arm A, and PETRA Arm B regimens; intrapartum and neonatal ZDV only; the HIVNET 012 regimen of single-dose nevirapine to the mother and child; and universal nevirapine (nevirapine provided to all women and children at the time of delivery without checking for HIV status). Additionally, you may specify the mode of delivery and method of infant feeding when you input the data. The purpose of this chapter is to discuss sources of data that can be used to specify the required inputs and to suggest default values that may be used when local data are unavailable. Each of the required variables is discussed below. A. Context 1. Births (Public) % Public Sector Births In the PMTCT module, HIV counseling, testing, and ARV treatment for pregnant women occur at antenatal clinics, as part of antenatal care. Since the model will be used primarily by government policymakers, only antenatal clinics in the public sector are considered. Therefore, the percentage of births that will take place in the public sector is required. The annual number of births can be calculated using the Spectrum model, or taken from the demographic projection provided by the United Nations (United Nations [1998]. World Population Prospects: The 1998 13 Revision). The percentage of births in the public sector is avai from Demographic and Health Surveys (DHS), or it may be a policy variable that can be supplied. Representative percentages of lable births occurring in public health ble 1. Percentage of Births Taking Place in a Health Facility facilities from various DHS reports are shown in Table 1. Ta Country Year of Fieldwork Percent of Births in Health Facility only if ava Burkina Faso 1992/93 42.9 62.2 (59.6) ca Rep 1 43.8 (42.0) scar 55.3 (41.3) 1 e R EAST/ NORTH AFRICA Cameroon 1991 Central Afri 994/95 49.8 Cote d’Ivoire 1994 44.8 Ghana 1993 42.1 Kenya 1993 Madaga 1992 44.7 Malawi 1992 Namibia 1992 66.9 Niger 1992 15.4 Nigeria 1990 32.4 Rwanda 1992 25.5 Senegal 1997 46.9 Tanzania 991/92 53.3 Uganda 1995 35.4 Zambia 1992 50.6 Zimbabw 1994 69.1 ASIA/NEA Bangladesh 1993/94 3.5 32.5 (17.9) sia tan 1 s MERICA/CARIBBEAN Egypt 1995/96 India 1992/93 25.5 (14.6) Indone 1994 17.5 (7.9) Jordan 1990 78.5 (57.1) Kazakhs 1995 98.4 Morocco 1992 28.3 Pakistan 990/91 13.3 Philippine 1993 28.3 Turkey 1993 59.8 LATIN A Bolivia 1993/94 42.2 ia Republic 92.0 (70.8) 1994/95 ay 1 Brazil 1996 91.5 Colomb 1995 76.8 Dominican 1991 Guatemala 1995 34.3 (28.3) Haiti 16.4 (12.7) Paragu 1990 53.3 (43.3) Peru 991/92 49.6 various D ealth (Public: ilable) SUB-SAHARAN AFRICA Source: Stewart et al., 1991, p. 13, plus emographic and H Surveys. 14 % Antenatal Visits The next parameter is the percentage of public sector births where the mother has attended an antenatal clinic. It should be noted, however, that some women might use the public sector for antenatal care, and then deliver either at home or in a private facility. These women may undergo the testing and treatment process offered at the antenatal clinic, yet will not be represented in the DHS data as public sector births. Since the HIV testing occurs at antenatal clinics, the percentage of women who make an antenatal visit is needed. This percentage is generally available in the most recent national family planning survey. Representative percentages from various DHS surveys are shown in Table 2 below. Note that, in general, antenatal care attendance is higher than public sector births. Specific statistics for each country can be calculated by examining the relevant DHS survey directly. Table 2. Percentage of Women with Antenatal Care Visits Country Year of Fieldwork Percentage of Women with One or More ANC Visits Median ANC Visits SUB-SAHARAN AFRICA % DK/Missing Burkina Faso 1992/93 56.9 3.1 2.3 Cameroon 1991 75.5 3.5 4.0 Central Afr Rep 1994/95 75.3 2.6 3.5 Cote d’Ivoire 1994 83.8 0.7 3.0 Ghana 1993 85.6 1.7 4.7 Kenya 1993 92.8 3.4 4.7 Madagascar 1992 83.6 2.8 3.7 Malawi 1992 90.5 2.3 4.6 Namibia 1992 80.4 7.8 4.6 Niger 1992 30.0 0.5 0.7 Nigeria 1990 63.4 1.9 4.3 Rwanda 1992 94.8 0.1 2.9 Senegal 1992/93 75.1 3.8 3.1 Tanzania 1991/92 93.9 2.4 4.9 Uganda 1995 90.7 1.8 3.9 Zambia 1992 90.9 2.7 5.1 Zimbabwe 1994 91.9 2.5 5.6 15 Country Year of Fieldwork Percentage of Women with One or More ANC Visits Median ANC Visits ASIA/NEAR EAST/NORTH AFRICA Bangladesh 1993/94 26.7 0.3 0.7 Egypt 1995/96 38.9 0.4 0.8 India 1992/93 49.9 0.0 1.0 Indonesia 1994 86.5 0.8 5.4 Jordan 1990 80.0 0.8 6.3 Kazakhstan 1995 89.5 3.2 10.7 Morocco 1992 32.3 0.1 0.7 Pakistan 1990/91 27.4 2.9 0.7 Philippines 1993 91.2 1.1 4.2 Turkey 1993 62.5 0.6 2.5 LATIN AMERICA/CARIBBEAN Bolivia 1993/94 53.1 0.4 1.5 Brazil 1996 84.0 2.8 6.9 Colombia 1995 82.8 0.4 6.2 Dominican Rep 1991 96.5 0.6 7.5 Guatemala 1995 86.0 0.3 5.2 Haiti 1994/95 70.0 1.0 3.1 Paraguay 1990 90.1 1.2 5.8 Peru 1991/92 67.2 1.0 3.7 Source: Stewart et al., 1997, p. 6. % Planned C-Section The vertical rates of HIV transmission vary with the mode of delivery as well as with the type of ARV treatment. The probability of transmission decreases when delivery is performed with a planned C-section compared with a vaginal birth. Therefore, PMTCT requires an estimate of the percentage of public sector births that are planned Cesarean sections. The best source for this figure will usually be a national census or national survey such as the DHS. Note that the Cesarean section rate must be for a planned Cesarean section, not an emergency Cesarean section, since the transmission rate for an emergency Cesarean section is about the same as that for a vaginal birth. 16 Table 3 presents estimates available on the percentage of births undertaken by elective Cesarean sections for a number of countries. For most of these countries, the information comes from the most recent DHS. Table 3. Percentage of Births by Cesarean Section Country Year of Fieldwork % of Births by C-section SUB-SAHARAN AFRICA Burkina Faso 1992/93 1.3 Cameroon 1991 2.3 Central Afr Rep 1994/95 1.9 Cote d’Ivoire 1994 1.8 Ghana 1993 4.5 Kenya 1993 5.2 Madagascar 1992 1.0 Malawi 1992 3.4 Namibia 1992 6.6 Niger 1992 0.9 Nigeria 1990 2.5 Rwanda 1992 1.7 Senegal 1997 2.2 Tanzania 1991/92 2.5 Uganda 1995 2.6 Zambia 1992 2.6 Zimbabwe 1994 6.0 ASIA/NEAREAST/NORTH AFRICA Bangladesh 1993/94 6.6 Egypt 1995/96 2.5 India 1992/93 2.5 Indonesia 1994 5.6 Jordan 1990 4.6 Kazakhstan 1995 2.0 Morocco 1992 2.7 Pakistan 1990/91 5.9 Philippines 1993 8.1 Turkey 1993 LATIN AMERICA/CARIBBEAN Bolivia 1993/94 10.6 Brazil 1996 36.4 Colombia 1995 16.9 Dominican Rep 1991 22.3 Guatemala 1995 8.2 Haiti 1994/95 1.6 Paraguay 1990 13.1 Peru 1991/92 9.7 Source: Stewart et al., 1997, p. 15; various Demographic and Health Surveys; UNICEF, 1998. 17 % Breastfed for 6 Months The percentage of public sector births that are breastfed exclusively for six months must be specified in the model. Representative percentages of children that are exclusively breastfed from various DHS surveys and other sources are shown in Table 4. Specific statistics for each country can be calculated by examining the relevant DHS survey directly. Table 4. Percentage of Children Breastfed Country Year of Fieldwork % Exclusively Breastfed (0-3 months) SUB-SAHARAN AFRICA Burkina Faso 1992/93 12 Cameroon 1991 7 Central Afr Rep 1994/95 23 Cote d’Ivoire 1994 62 Ghana 1993 19 Kenya 1993 17 Madagascar 1992 47 Malawi 1992 11 Namibia 1992 22 Niger 1992 - Nigeria 1990 2 Rwanda 1992 90 Senegal 1997 9 Tanzania 1991/92 40 Uganda 1995 70 Zambia 1992 13 Zimbabwe 1994 16 ASIA/NEAREAST/ NORTH AFRICA Bangladesh 1993/94 51 Egypt 1995/96 53 India 1992/93 51 Indonesia 1994 47 Jordan 1990 32 Kazakhstan 1995 12 Morocco 1992 31 Pakistan 1990/91 16 Philippines 1993 33 Turkey 1993 14 LATIN AMERICA/CARIBBEAN Bolivia 1993/94 53 Brazil 1996 42 Colombia 1995 16 Dominican Rep 1991 25 Guatemala 1995 50 Haiti 1994/95 3 Paraguay 1990 7 Peru 1991/92 63 Source: Various DHS reports; UNICEF, 1998. 18 % Mixed Feeding The percentage of public sector births that use mixed breastfeeding and formula feedings is required by the model. Table 5 shows data on the percentage of children that use mixed feeding from various Demographic and Health Surveys. Table 5. Percentage of Children that Use Mixed Feeding Country Year of Fieldwork % Breastfed w/ Complementary Food (6- 9 mth) SUB-SAHARAN AFRICA Burkina Faso 1992/93 - Cameroon 1991 77 Central Afr Rep 1994/95 - Cote d’Ivoire 1994 54 Ghana 1993 63 Kenya 1993 90 Madagascar 1992 80 Malawi 1992 78 Namibia 1992 65 Niger 1992 67 Nigeria 1990 52 Rwanda 1992 68 Senegal 1997 87 Tanzania 1991/92 - Uganda 1995 64 Zambia 1992 88 Zimbabwe 1994 93 ASIA/NEAREAST/ NORTH AFRICA Bangladesh 1993/94 - Egypt 1995/96 37 India 1992/93 31 Indonesia 1994 85 Jordan 1990 48 Kazakhstan 1995 61 Morocco 1992 33 Pakistan 1990/91 31 Philippines 1993 52 Turkey 1993 17 LATIN AMERICA/CARIBBEAN Bolivia 1993/94 78 Brazil 1996 30 Colombia 1995 61 Dominican Rep 1991 47 Guatemala 1995 56 Haiti 1994/95 83 Paraguay 1990 61 Peru 1991/92 83 Source: Various DHS reports; UNICEF, 1998. 19 Increase in CMR The proportional increase in the child mortality rate (CMR) among infants not breastfed is the proportional increase in mortality from formula feeding. The child mortality rate is the number of children per 1,000 live births who die before the age of five. Breastfeeding has a positive impact on the initial health status of an infant and subsequent health status, as the infant grows older. Various studies have shown the increase in the child mortality rate from not breastfeeding to range widely. A proportional increase of two implies that the effect of formula feeding doubles the child mortality rate. A recent DHS report is a good source for this information. Table 6 presents the effects of formula feeding on child morbidity and mortality from several studies. Table 7 presents under-five years of age child mortality from various Demographic and Health Surveys and other sources for a selection of countries. Table 6. Effect on Child Morbidity and Mortality of Formula Feeding Country Findings Source Model assumption Used 2.5 proportionate increase in maternal mortality from no breastfeeding; 1.5 increase from early cessation of breastfeeding Kuhn and Stein, 1997 Model assumption Used 0.3 proportionate increase from breastfeeding; based on 0.2 increase observed from European Collaborative Study Group Hu et al., 1992 39 developing countries (late 1970s) Increased child mortality by 50% Hobcraft et al., 1985 Brazil (1984-85) Relative risk of death from gastrointestinal and respiratory infection are 14.2 and 3.6 Victora et al., 1987 Meta-analysis of nine studies from five countries Increased relative risk of death from gastrointestinal illness by 25 (based on pre-1950 literature) Feachem and Koblinsky, 1984 Malaysia (1976-77) 1.8 to 2.6 proportionate increase (controlled for sanitation, income, maternal education and age, etc.) Habicht et al., 1986 20 Table 7. Under-5 Child Mortality Rates Country Under-5 Mortality Rate (per 1,000 live births) 19961,2 SUB-SAHARAN AFRICA Burkina Faso 158 Cameroon 102 Central African Rep 164 Cote d’Ivoire 150 Ghana 110 Kenya 90 Madagascar 135 Malawi 217 Namibia 92 Nigeria 130 Rwanda 205 Senegal 88 Tanzania 144 Uganda 141 Zambia 202 Zimbabwe 86 ASIA/NEAR EAST/NORTH AFRICA Bangladesh 112 Egypt 66 India 85 Indonesia 60 Jordan 35 Kazakhstan 30 Morocco 67 Pakistan 123 Philippines 44 Turkey 47 LATIN AMERICA/CARIBBEAN Bolivia 102 Brazil 42 Colombia 31 Dominican Republic 47 Guatemala 56 Haiti 130 Paraguay 45 Peru 58 Sources: 1 Various Demographic and Health Surveys, World Development Report 1998/99, pp. 192-3; 2 UNICEF, 1998, pp. 118-21. 21 MMR (per 100,000) The maternal mortality rate (MMR) is the number of deaths from pregnancy-related causes per 100,000 women of reproductive age (15-49 years). A recent DHS report is a good source for this information. Representative maternal mortality rates are presented in the table below from various Demographic and Health Surveys and other sources. Table 8. Maternal Mortality Rates Country Maternal Mortality Rate (per 100,000 women of reproductive age) 19901,2 SUB-SAHARAN AFRICA Burkina Faso 930 Cameroon 550 Central African Rep 700 Cote d’Ivoire 810 Ghana 740 Kenya 650 Madagascar 490 Malawi 560 Namibia 370 Nigeria 1000 Rwanda 1300 Senegal 1200 Tanzania 770 Uganda 1200 Zambia 940 Zimbabwe 570 ASIA/NEAR EAST/NORTH AFRICA Bangladesh 850 Egypt 170 India 570 Indonesia 650 Jordan 150 Kazakhstan 80 Morocco 610 Pakistan 340 Philippines 280 Turkey 180 LATIN AMERICA/CARIBBEAN Bolivia 650 Brazil 220 Colombia 100 Dominican Republic 110 Guatemala 200 Haiti 1000 Paraguay 160 Peru 280 Sources: 1 World Development Report 1998/99, pp. 192-3; 2 UNICEF, 1998, pp. 118- 21, various Demographic and Health Surveys. 22 B. Policy/Programs 1. Testing/Counseling The testing and counseling inputs include several policy decisions. Note that the values for the testing and counseling projection inputs are based on births in the public sector. Women who are offered voluntary testing and counseling services are pregnant women who attend programs to prevent mother-to-child HIV transmission and who attend public antenatal clinics. See Births (Public) for a more complete discussion. There are a number of policy inputs required by the model. % Offered HIV Test A policy decision must be made as to what percentage of pregnant women will be offered the HIV test. The percentage of women offered the HIV test is the percentage of pregnant women in a PMTCT program at public antenatal clinics. This decision reflects the coverage the program is trying to achieve. Program managers should base this decision on the costs of the HIV tests and the counseling, as well as the availability of service facilities throughout the country. % Accepting Test The next parameter required by the model is the percentage of pregnant women offered the HIV test who accept it. Studies conducted in several sub-Saharan African countries and Thailand suggest that this percentage varies between 75 and 100 percent. The results of these studies are summarized in Table 9. Table 9. Percentage of Pregnant Women Accepting the HIV Test Country % Accepting Test (Duration of pre-/post-test counseling in minutes: GC=group counseling) Source Cote d'Ivoire 68-76 Seyler, 2001 Cote d'Ivoire 90 Beguet et al., 2001 Senegal 56 Sall et al., 2001 Botswana 46 Mazhani et al., 2000 Burkina Faso Cote d’Ivoire: I Cote d’Ivoire: II Kenya: I Kenya: II Tanzania Malawi: I Malawi: II 92 (15/25) 78 (10/20) 77 (GC/30) 99 (GC/30) 95 (GC/30) 88 (30/30) 53 (GC/30) 90 (15/25) Cartoux et al., 1998 (Note: 100% receiving rate in Zambia based on using rapid tests; also, these interventions took place in Fall 1997, before the short course results in Thailand were reported; 23 Country % Accepting Test (Duration of pre-/post-test counseling in minutes: GC=group counseling) Source Zambia Zimbabwe South Africa: I South Africa: II Thailand Median 81 (15/15) 100 (10/15) 97 (10/30) 98 (GC/25) 100 (GC/40) 92 (15/30) the lack of efficacious treatment programs may have affected women’s decisions to undergo testing and/or return for results.) Cote d’Ivoire 94.5 Ramon et al., 1998 (gynecology clinics) Kenya 99 Kiarie, 1996 Cote d'Ivoire Less than one-half of 13,000 women both accepted & returned for results (HIV+ less likely) www.unaids.org/unaids/p ress/wadep98.html, 1998 Soweto, South Africa - Hospitals only >95 McIntryre et al., 1998 United States 70-80 Carusi et al., 1998 Model assumption- personal communication 83 (counseling/testing participation rates) Mansergh et al., 1996 Model assumption 75-100 Mauskopf et al., 1996 % Receiving Results The PMTCT module also requires the percentage of women tested who received the HIV test results. Although a certain percentage of pregnant women may agree to undergo testing, not all of these women may receive the test results. There are two possible reasons for this. First, the results may not be returned to the clinics due to various logistical difficulties. Second, women themselves may not return for the results; some evidence suggests that women who are HIV-positive are less likely to return for their results. For example, in Soweto, South Africa, almost all women agreed to be tested (McIntyre et al., 1998). Of those who were tested, 80 percent returned for the test results; yet only half of those who were HIV- positive returned. As more clinics adopt rapid HIV testing, the percentage of women receiving the test results will approach 100. Table 10 below summarizes these studies. 24 Table 10. Percentage of Pregnant Women Receiving HIV Test Results Country % Receiving Result (Note: overall rates =% accepting * % receiving) Source Cote d'Ivoire 46-93 Seyler, 2001 (higher acceptance for rapid test) Cote d'Ivoire 69 Beguet et al., 2001 Botswana 87 Mazhani et al., 2000 Burkina Faso Cote d’Ivoire: I Cote d’Ivoire: II Kenya: I Kenya: II Tanzania Malawi: I Malawi: II Zambia Zimbabwe South Africa: I South Africa: II Thailand Median 82 (81 if HIV+) (e.g.: overall rate=.92*.82=.75) 58 (39 if HIV+) 63 (59 if HIV+) 70 (54 if HIV+) 68 (62 if HIV+) 86 (72 if HIV+) 83 (75 if HIV+) 68 (69 if HIV+) 100 (100 if HIV+) 33 (36 if HIV+) 97 (83 if HIV+) 98 (98 if HIV+) 100 (95 if HIV+) 92 (82 if HIV+) Cartoux et al., 1998 (Note: 100% receiving rate in Zambia based on using rapid tests; also, these interventions took place in Fall 1997, before the short course results in Thailand were reported; the lack of efficacious treatment programs may have affected women’s decisions to undergo testing and/or return for results.) Cote d’Ivoire 90 Ramon et al., 1998 (gynecology clinics) Rwanda 69 (64 if HIV+) Ladner et al., 1992-93 Kenya 70 (HIV+ less likely to return) Kiarie, 1996 Cote d'Ivoire <50 (HIV+ less likely) www.unaids.org/unaids/press /wadep98.html, 1998 Natal, South Africa 17 when tests sent away; 96 when testing performed on-site Wilkinson, 1998 Model assumption based on early UNAIDS trials 60 (remain through post-test counseling) Marseille et al., 1998 # Women Seen/Counselor The number of women seen by a VCT counselor is the number of women each counselor will see annually. This value is used in calculating how many counselors will be necessary to meet the needs of those being tested. Very little country-specific data are available; field tests suggest that a reasonable figure is 1,000 women seen by a VCT counselor annually. 25 % Breastfeeding Spillover This parameter is the percentage of HIV-negative women who cease breastfeeding due to the impacts of the spillover effect (these impacts could be from a mass media campaign, an intervention, or counseling). There are no quantitative studies on which to base this figure. The default value for the model is zero. This can be changed if new data become available. Infections Averted/Woman An additional benefit from the counseling and testing that women receive as part of any PMTCT intervention is that some women will change their sexual behavior to avoid becoming infected in the future or to avoid passing on an existing infection to a partner. The number of infections averted will depend on many factors including the quality of the counseling, the infection level in the population, and the social and political environment. Many studies have found that HIV-positive clients are more likely to change their behavior than HIV-negative clients (Weinhardt et al., 1999). A study of voluntary counseling and testing in Kenya and Tanzania estimated that condom use increased significantly among clients (Sweat et al., 2000). The behavior change recorded by that study implies that about 2.8 infections are averted for every 100 VCT clients. The figure would be somewhat less in low-prevalence settings. Programs that provide very little counseling as part of the PMTCT intervention would have fewer infections averted than shown in this table, while those that focus on high-quality counseling might avert somewhat more infections per client. The default value may be set to 0.028. Table 11. Number of Infections Averted per Person Counseled and Tested by Woman's HIV Status and Country Country HIV Status Infections Averted per Person Counseled and Tested Kenya HIV+ 0.034 Kenya HIV- 0.030 Tanzania HIV+ 0.027 Tanzania HIV- 0.022 26 Further background information from the literature regarding VCT is shown in Table 12 below. Table 12. Effect of Testing/Counseling for Heterosexual Couples Country Result Source Rwanda After 1 year, condom use increased from 7% to 16% of HIV-, 35% of HIV+, and 57% of discordant couples Allen et al., 1992 Allen, Tice et al., 1992 Zaire Discordant couples increase condom use from <5% to 71% and abstinence from 0% to 28%; HIV+ women were almost uniformly unwilling to disclose status to partners Kamenga, Ryder et al., 1991 Ryder et al., 1991 Uganda Significant increase in condom use, especially among HIV+ UNAIDS, 1999 Zaire Increase in condom use, though less than stated plans (20% of HIV+ actual vs. 71% planned) Heyward et al., 1993 Gambia Counseling had no effect on condom use by sex workers Pickering et al., 1993 Kenya Single-session counseling had no effect on condom use Temmerman et al., 1990 Rwanda HIV+ more likely to use condoms following HIV CT, though less likely to reveal HIV status Van der Straten et al., 1995 Rwanda 8.8% of HIV+ used condoms 4 months after HIV CT vs. 3.9% of HIV- Ladner et al., 1996 India Level of consistent condom use by sex workers increased with follow-up time proportionately Bentley et al., 1998 Source: Wolitski et al., 1997, for a complete review of recent findings, and Higgins et al., 1991 for an older review of results, including those for MSM, IV drug users, etc. 2. Treatment Option The group of pregnant women who have tested HIV positive is then eligible for one of the ARV treatment options. There are seven different treatment choices as well as a no-treatment (None) option in the model. • Long-course ZDV. The long-course treatment is identical with the ACTG 076 treatment protocol. Women begin zidovudine (ZDV) treatment sometime between 14 and 34 weeks gestation; costs will vary, depending on when treatment starts. The medicine is continued intravenously through labor and birth. In addition, ZDV is given to the newborn for the first six weeks of life. 27 • Short-course ZDV (Thailand regimen). The first short-course treatment consists of the protocol followed in the North Thailand Perinatal HIV Prevention Trial (NTPHPT). The HIV-infected women are offered ZDV treatment beginning at the 34th week of pregnancy, and continue through labor and birth. • Short-course ZDV (PETRA Arm A). PETRA Arm A includes 300mg of ZDV and 150mg of 3TC provided twice daily to the mother starting at 36 weeks gestation, 300 mg of ZDV and 150 mg of 3TC orally every 12 hours during labor, 300 mg of ZDV and 150 mg of 3TC provided to the mother twice daily for one week postpartum, and 4 mg/kg body weight of ZDV and 2 mg/kg body weight of 3TC provided orally twice daily to the newborn child for one week. • Short-course ZDV (PETRA Arm B). PETRA Arm B is the same as Arm A but without the antepartum treatment. • Neonatal only. The neonatal treatment consists of oral ZDV administered to infants within the first 48 hours of life, and continued for the first six weeks of life. If the results of the rapid test given to the woman in labor are received quickly, ZDV is also offered intrapartum. • Nevirapine (HIVNET 012). The HIVNET012 treatment consists of one dose of 200mg of nevirapine for the mother at the onset of labor and one dose of 2 mg/kg of body weight of nevirapine administered to the child within 72 hours of birth. • Universal Nevirapine. Nevirapine provided to all women and children at the time of delivery without checking for HIV status. • None. No intervention. 3. Treatment % Eligible for Treatment Eligibility for a treatment option depends on when a woman begins her antenatal care. If the intervention option is changed, the group of women eligible is changed. For example, with long-course AZT, antenatal visits early in the pregnancy (24 weeks) are required to be eligible. With nevirapine it can be much later. The table below shows the distribution of antenatal care provided to women for various countries, by timing of initial visit. Updated data for each of the variables can be found in the most recent DHS survey for each country. 28 Table 13. Percentage of Women with Antenatal Visits by Pregnancy Stage Country Year Time of First Visit No ANC <7 months 7+ months Median Months Pregnant at First Visit SUB-SAHARAN AFRICA % % % Burkina Faso 1992/93 40.0 48.6 8.1 4.8 Cameroon 1991 21.0 73.6 5.0 4.3 Central African Republic 1994/95 22.1 70.3 6.7 4.8 Cote d’Ivoire 1994 15.5 67.7 16.0 5.4 Ghana 1993 12.7 78.2 7.5 4.5 Kenya 1993 3.8 78.9 15.7 5.6 Madagascar 1992 13.6 72.3 12.3 5.4 Malawi 1992 7.2 74.0 17.8 5.9 Namibia 1992 11.8 76.8 9.3 4.8 Niger 1992 69.5 26.5 3.3 4.6 Nigeria 1990 34.7 52.6 10.2 5.3 Rwanda 1992 5.1 51.1 43.5 6.9 Senegal 1992/93 21.1 68.8 6.1 3.9 Tanzania 1991/92 3.7 83.2 12.6 5.6 Uganda 1995 7.5 71.3 20.8 5.9 Zambia 1992 6.4 80.8 12.3 5.6 Zimbabwe 1994 5.6 81.0 12.0 5.1 ASIA/NEAR EAST/NORTH AFRICA Bangladesh 1993/94 72.4 21.9 5.5 5.0 Egypt 1995/96 60.7 37.1 2.0 2.8 India 1992/93 50.1 39.7 10.2 5.0 Indonesia 1994 12.7 80.7 6.1 3.5 Jordan 1990 19.2 76.0 3.9 3.0 Kazakstan 1995 7.3 89.6 2.3 3.5 Morocco 1992 67.6 28.5 3.8 3.6 Pakistan 1990/91 69.7 21.7 6.0 4.0 Philippines 1993 7.7 83.4 8.3 4.3 Turkey 1993 36.9 58.6 3.9 3.1 LATIN AMERICA/CARIBBEAN Bolivia 1993/94 46.5 47.0 6.2 3.6 Brazil 1996 13.2 83.9 1.5 2.9 Colombia 1995 16.8 80.2 2.9 3.0 Dominican Republic 1991 2.9 93.9 2.9 3.0 Guatemala 1995 13.7 82.5 3.4 2.7 Haiti 1994/95 29.0 65.1 5.2 3.8 Paraguay 1990 8.7 84.3 6.6 3.5 Peru 1991/92 31.8 61.3 6.2 3.4 Source: Stewart et al., 1997, p. 6. 29 % Offered Treatment After the percentage of eligible women is determined, a policy decision must be made about the percentage of eligible women that are offered treatment. Factors affecting this decision include financial constraints, the available facilities in a country, and other logistical requirements. These factors may vary by treatment option. % Accepting Treatment The PMTCT module requires the percentage of pregnant HIV- positive women who are offered the intervention and who then choose to accept it. Women may not accept treatment after it is offered for various reasons, including a belief that the treatment will not work, a lack of desire to add complexity to their life, and a fear of side effects that may result from taking the medication. These concerns may vary by treatment option; for example, a woman may be more likely to accept the neonatal treatment than the long-course treatment, as it entails less effort on her part. Various estimates for treatment acceptance have been made; the UNAIDS ZDV clinical trials experienced an acceptance rate of 75 percent (Marseilles et al., 1998). Several recent studies have examined acceptance rates for short-course ZDV protocols: in Senegal, the acceptance rate was 50 percent (Sall et al., 2001); in Botswana, 56 percent of women and 92 percent of newborns received short- course ZDV (Mazhani et al., 2000); in Thailand, initially 58 percent accepted treatment, increasing to 69 percent by the end of the intervention (Kanshana et al., 2000). % Adhering to Treatment The model requires the percentage of HIV-positive women who adhere to the full course of the intervention once they accept it. Compliance may vary by treatment option; for example, in the case of nevirapine, women not attending a health facility may forget to take the nevirapine supplied to them. The model is designed so that the efficacy of the intervention is based on the number of women complying fully with their treatment. That is, if someone does not comply fully with her treatment, she drops out of the group of treated women, and is subject to the vertical transmission rates experienced by untreated women. Although some studies suggest that incomplete compliance may still result in some reduction of the untreated transmission rate, it is not clear what this reduction should be. Thus, no reduction is used here. 30 The compliance rates experienced for the different treatment options are listed in Table 14 below. Table 14. Treatment Compliance Rates Protocol Rates Source Pilot program (Thailand) 89% of women took at least 90% of ZDV doses Kanshana et al., 2000 NTPHPT (Thailand) 99% of the women took at least 90% of the antenatal ZDV doses and 99% took at least 1 dose during labor Weekly Epidemiological Record, 1998 UNAIDS trials 7.7% of women did not take any of the prescribed medication Marseille et al., 1998 % Elective C-Section Another policy decision to be made is the percentage of HIV- positive women who undergo an elective Cesarean section. Note that the Cesarean section must be scheduled and not an emergency section; emergency Cesarean sections have been shown to have approximately the same vertical transmission rate as a vaginal birth. Table 3 (above) describes the percentage of C- sections performed for all women regardless of HIV status and, thus, may be used as a benchmark when making this policy decision. A country's capacity to deliver Cesarean sections will have an impact on this policy decision. % HIV+ Breastfeed for 6 Months A policy decision must be made on the percentage of pregnant HIV-positive women who will breastfeed their baby exclusively for six months. Breastfeeding has a positive impact on the initial health status of an infant, and subsequent health status as the infant grows older. Various studies have shown the increase in the child mortality rate from not breastfeeding to range widely. Studies find that breastfeeding decreases both the occurrence and severity of illnesses such as diarrhea, ear infections, and bacterial meningitis. Updated data for breastfeeding can be found in the most recent DHS survey or national survey for each country. Table 4 (above) shows the percentage of infants breastfed for up to three months, regardless of HIV status, and may be used as a benchmark when making this policy decision. % HIV+ Mixed Feeding The final policy decision to be made for the model is the proportion of HIV-positive women who feed their infants with formula in combination with breastfeeding. There are significant costs associated with this policy, including both financial costs and the 31 effects on child mortality and morbidity. If infant formula is not prepared properly, the severity of illnesses such as diarrhea, ear infections, and bacterial meningitis may increase. One source suggests that for rural areas in Uganda, 85 percent of formula-fed infants will die due to water contamination (Miro, 1998). However, a randomized clinical trial in Kenya found that, with access to clean water, formula-fed and breastfed infants had similar initial mortality and morbidity rates, although HIV-1-free survival at two years was significantly higher in the formula group (Mbori-Ngacha, et al., 2001). Table 5 (above) shows the results of various Demographic and Health Surveys and other data sources of the percentage of live births using mixed feeding. Although these data do not include consideration of HIV status, they may be used as a benchmark when making this policy decision. MMR Increase: C-Section HIV transmission during birth can be reduced if, instead of a vaginal birth, a scheduled Cesarean section is performed. The PMTCT model requires that for HIV-positive women accepting the intervention, the proportional increase in maternal mortality due to a scheduled Cesarean section must be specified. Evidence suggests that maternal mortality is somewhat higher with a Cesarean section than with a vaginal delivery; morbidity is certainly higher. These health effects can be exacerbated in some developing countries by poor sanitary conditions. Evidence on the increase in maternal mortality from having a Cesarean section is mixed; many times Cesarean sections are performed due to underlying medical conditions, which themselves may have an effect on maternal mortality. Furthermore, sanitary conditions vary by country; sometimes the principal effect of Cesarean sections on maternal mortality is due to postoperative infections. Many detailed studies have been conducted for a small number of countries on the effects of Cesarean sections on maternal mortality rates. The results of these studies are summarized in Table 15. 32 Table 15. Effect of Cesarean Section on Maternal Mortality Rates (MMR) Country Results Source Cape Town, South Africa Overall relative risk of mortality from C- section compared with vaginal delivery was 7 decreasing to 5 after the exclusion of complications due to pre-existing conditions Lilford et al., 1990 Senegal MMR associated with C-section was 4.7%, compared with overall MMR of 850/100,000 (29% of deaths during surgery, 71% were post-operative) Boullin et al., 1994 Indonesia Maternal mortality rates following C- sections were more related to maternal illness than the surgery. Between 1981- 83, MMR was 985.7/100,000 overall and 4.5/1,000 for C-sections; between 1988- 90, overall MMR was 757.5/100,000 and 2.6/1,000 for C-sections Wirakusuma, 1995 Kaziba, rural Zaire Between 1971-92, overall MMR decreased from 0.3% to 0.12%, and deaths associated with C-sections decreased from 3.2% to 0.7%, but risk of dying remained 13 times higher for C- section births vs. vaginal births Onsrud and Onsrud, 1996 Nigeria MMR associated with 1,973 C-sections between 1985-89 was 11/1,000 Ozumba and Mbadiwe, 1993 Turkey C-section delivery had a higher MMR than vaginal delivery (44.7 vs. 14.1); probably a result of pre-existing medical conditions rather than the C-section itself Tuncer et al., 1995 Nigeria/Developing countries The mortality rate in a Nigerian teaching hospital after C-section was 0.6%; serious maternal morbidity after C-section occurs in 9-15% of cases in developing countries Van Roosmalen and van Does, 1995 United States Maternal mortality is 2-4 times higher and morbidity is 5-10 times higher after a C-section compared to vaginal birth Shearer, 1993 Nigeria Sepsis was the most important cause of deaths after C-sections occurred (81.5%) Ojo et al., 1988 33 Table 15 (cont.) Country Results Source Tegulcigalpa: MMR Per 1,000 Number Cumana, VEN: MMR Per 1,000 Number Santiago, CH: MMR Per 1,000 Number Campinas, BR: MMR Per 1,000 Number Total Vaginal Cesarean 0.12 0.13 0 8,258 7,850 408 0.59 0.33 2.50 13,586 11,987 1,599 0.34 0.21 0.78 5,939 4,656 1,283 0 0 0 6,416 3,763 2,653 Janowitz et al., 1983 4. Costs Total testing and counseling costs will be a function of whether counseling is offered on an individual or group basis and the total number of sessions offered, as well as the overall HIV prevalence in the country. The PMTCT model allows for the input of costs associated with counseling, testing, C-sections, formula, and ARV treatment. Cost default values can be changed in the model or set to zero. HIV Test The cost of the HIV test depends on the type of HIV test used and the HIV prevalence in the country. Representative costs that have been experienced in various countries are presented in Table 16. Table 16. HIV Testing Costs (US$) Country HIV Testing Costs Source Based on Lusaka, Zambia clinic $1.00/initial kit + $3.00/confirmatory test = $4.00/total per woman Includes facility rent, counselors’ salaries and administrative costs Marseille et al., 1998 Developing countries $1.00/kit Nearly half are simple-and-rapid tests UNAIDS 1998 United States $4.80/ELISA test $63.40/3 ELISA + 1 Western blot Mauskopf et al., 1996 Based on Uganda clinic data $18.50 (includes 2 ELISA tests and staff time for pre- and posttest counseling up to 30 min) Mansergh et al., 1998 South Africa $2.70/Rapid HIV test $2.70/Double ELISA HIV test $1.40/Single ELISA HIV test Wilkinson et al., 1998 34 Pre-Test VCT, Post-Test VCT: HIV-, Post-Test VCT: HIV+, Post-Test VCT: Feeding The PMTCT model allows for input of voluntary counseling and testing (VCT) costs. The number of counseling sessions may be estimated and then the total counseling costs can be calculated accordingly. At present, the usual practice is to provide one session prior to the HIV test (pre-test VCT), perhaps in a group setting, and a second session after the test results are received (post-test VCt: HIV+ or HIV-). Those testing positive usually receive more intensive counseling than those whose outcomes are negative. Women testing HIV positive should also receive counseling on infant feeding options (post-test VCT feeding). Representative HIV counseling costs that have been reported in studies in several different countries are presented in Table 17. Table 17. HIV Counseling Costs (US$) Country Test/Counseling Costs Source Based on Lusaka, Zambia clinic $4.00/total per woman Includes facility rent, counselors’ salaries and administrative costs Marseille et al., 1998 United States $55.20/pre- and posttest counseling $99.34/pre- and posttest counseling) Mauskopf et al., 1996 Based on Uganda clinic data $18.50 (includes 2 ELISA tests and staff time for pre- and posttest counseling up to 30 min) Mansergh et al., 1998 Thailand $154.60/HIV testing and counseling Walker, 1997 Tanzania Kenya $29/client $27/client (general clinic, not ANC) Sweat et al., 1998 User Fee: VCT The user fee is the cost paid per woman participating in the intervention; in certain countries, it may be feasible to assess user fees for counseling and testing. If this is the case, the amount of the user fee per woman needs to be supplied for the model. In some cases, it will be possible to shift all of the testing and counseling costs to the women undergoing the process; usually, however, only a certain percentage will be feasible. The default value of this parameter is set to zero. Cost: C-Section Birth The additional cost of a Cesarean section birth, which is the difference between the cost of a vaginal birth and a Cesarean section birth, is required by the model. This includes the cost of the operating room, drugs, fluids, and extra days in the hospital. The 35 additional costs for Cesarean sections are shown in the table below. Table 18. Additional Costs of a Cesarean Section Birth Country Results Source Philippines Extra cost from a C-section was about $576 InterPress Service, 1998 Bolivia Cost with drugs: vaginal, $47.90; C- section, $101.80; cost without drugs: vaginal, $33.72, C-section, $46.33. Between 1985-90, C-section cost with drugs averaged about twice the cost of a vaginal delivery; without drugs, the C- section cost averaged about one-third more than vaginal delivery. Rosenthal and Percy, 1991; International Monetary Fund, 1998, exchange rate for 1990 period average Akpakwume, Nigeria If normal delivery took place at home or without a trained birth attendant, a penalty of $12 is paid to the women’s union Okafor, 1991; Turner, 1991 England $959.50 net C-section increase (approximately 125% more than normal cost of $765) Ratcliffe et al., 1998 Monthly Formula Costs The total monthly costs of formula per infant must be supplied for the model. In order to calculate these costs, both the cost per feeding and the duration of formula feeding must be specified. Estimates from the literature, both on costs of formula and on the length of time that a baby will use formula, are listed in Table 19. Table 19. Formula Feeding: Cost and Duration Country Findings Source England $410/year for formula costs McDaniel, 1997 Zimbabwe US$280/year for formula costs Inter Press Service, 1998 South Africa 6 months of formula feeding Kinghorn, 1998 # Months Paid by Government The model requires the number of months of formula paid for by the government. Formula may be provided to HIV-positive women to discourage breastfeeding. Government support for formula purchase should continue until infants no longer require formula in their diet. The recommended range for the number of months is based on field applications of the PMTCT model. Approximately six months may be considered a reasonable value. 36 User Fee: Tx In certain countries, it may be feasible to assess user fees for the treatment. If this is the case, the amount of the user fee for drug treatment per woman needs to be specified in the model. In some countries, the government will be able to charge some amount to help defray the costs of the antiretroviral portion of the intervention. Usually, however, only a certain percentage will be feasible. Note that this figure is also based on the number of women accepting treatment. The default value of this parameter is set to zero. Total Child Treatment Costs The lifetime discounted paediatric AIDS treatment costs are required by the PMTCT model. The model is designed so that the costs of the treatment are based on the number of women who are given the medication. The implicit assumption here is that, even though a woman may not comply with the regimen, she is given a full supply of the medicine. Total intervention costs vary by country and by type of ARV treatment and other medical costs provided to infants and children who are HIV positive. Discounted paediatric AIDS costs are approximately two-thirds of the adult's treatment costs (see below). Table 20 provides suggestions from the literature for total child intervention costs. Table 20. Discounted Lifetime Paediatric AIDS Treatment Costs (US$) Country Costs Source Model/ Tanzania $195 Marseille et al., 1998 Model/ U.S. $98,915 (25% early progression/2 years; 75% late progression/7 years) Mauskopf et al., 1996 Model/ based on Tanzania and Zaire (Over et al., 1988) $396 (medical costs only; calculated as 2/3 of adult cost) Mansergh et al., 1998 Thailand $5,077 Walker, 1997 Thailand $5,000 Prescott, 1997 Total Adult Treatment Costs The total cost of treating HIV-positive adults is determined by the number of adult infections averted and the lifetime discounted AIDS treatment costs per adult. Although costs may vary substantially between countries, a general rule is that the annual expenditure per AIDS patient is between 100 and 200 percent of per capita GNP. A World Development Report is a good reference for per capita GNP. In countries where AIDS care is not well 37 developed, the annual cost is a good approximation for lifetime- discounted costs. The costs, however, could be much higher with the use of antiretroviral therapy. For suggestions of adult medical care costs for certain countries, see Table 5 in the AIM manual. C. Epidemiology 1. Transmission The probability of transmission of HIV from mother to child is a function of where transmission occurs. There are three points of transmission: in the uterus during gestation (intra-uterine), during delivery (vaginal or Cesarean), and during feeding (breastfeeding exclusive, or mixed feeding). The model provides vertical transmission default values from published studies (Ratcliffe et al., 1998) for the following inputs: • Intra-uterine • Vaginal - delivery • Cesarean - delivery • Breastfeeding - exclusive • Breastfeeding – mixed Note that the breastfeeding exclusive option uses a zero probability of transmission because, in a recent study in South Africa, infants exclusively breastfed for three months or more had no excess risk of HIV infection at 6 months compared with those never breastfed (Coutsoudis et al., 2001). These defaults can be changed if new information becomes available. 2. Reduction The reduction in the probability of HIV vertical transmission is a function of which treatment option is chosen. Default values for reduction in transmission probabilities are shown in Table 21 below. Defaults should be altered only if further scientific data become available. Table 21. Reductions in MTCT Transmission Probabilities by Treatment Option Treatment Reduction in Transmission Probability Source Long-course ZDV 0.68 Connor et al., 1994 Short-course ZDV (Thailand Regimen) 0.50 Shaffer et al., 1999 Short-course ZDV (PETRA Arm A) 0.50 Saba, 1999 38 Short course ZDV 0.37 Saba, 1999 (PETRA Arm B) Neonatal only 0.40 Guay et al., 1999 VNET 012) Nevirapine (HI 0.47 None 0 Guay et al., 1999 3. ARV Costs ach of the ARV treatments has costs associated with it, as well as are The Johns Hopkins AIDS Service Web site: http://www.hopkins-AIDS.edu To access the information after reaching the web site, select • The UNICEF PMTCT program web site: http://www.unicef.org/programme/health/focus/hiv/mtct/proj. • The WHO report: Sources and prices of selected drugs and rces0501. ote that other, nonmonetary, costs may be associated with or , ZDV E reductions in the vertical transmission rates. Costs vary by country and by treatment option. Additionally, some countries may be getting substantially reduced or no-cost nevirapine or ZDV. Representative cost scenarios that are used in the literature listed in Table 22. Other sources of average wholesale prices for various medications are: • “Publications,” then “Medical Management of HIV Infection.” The web site will be updated annually. htm diagnostics for people living with HIV/AIDS, May 2001. http://www.unaids.org/acc_access/access_drugs/Sou doc N treatments, including various side effects of the medication. F example, if a woman is already anemic before starting treatment the WHO guidelines suggest treatment not be offered, as anemia is a possible side effect of ZDV treatment (Lorenzi et al., 1998; Sperling et al., 1998; Meda et al., 1998). To date, no adverse long-term effects have been noted for children whose mothers received treatment (Culnane et al., 1999). 39 Table 22. Summary of Treatment Costs (US$) Type of Treatment Cost Source Long course $1000 $895 (drugs only) $150 (additional costs) $857 0.37/100 mg ZDV $58.54/hospital delivery $4.50/8 antenatal care visits UNAIDS, 1998, p. 2 Mauskopf et al.1996 Wilkinson, 1998 Walker, 1997 Short course $1/300 mg tablet $50 (special Glaxo-Wellcome price) $3.10/ZDV 200-mg dose $3.73/3TC 150-mg dose $340 Mansergh et al., 1996 UNAIDS, 1998,p. 3 Marseille et al., 1998 Wilkinson , 1998 Neonatal - only US$0.15/10 mg of paediatric syrup (U.S.) Mauskopf et al., 1996 NVP (HIVNET 012) $4/woman Marseille et al., 1999 4. Transmission Probabilities - Results As discussed above, vertical transmission probabilities are a combination of rates at three different points of transmission: in the uterus (during gestation), during delivery (vaginal or Cesarean), and during feeding. The antenatal treatment for pregnant women reduces the transmission rate in utero. However, HIV-positive women who receive a specific ARV treatment can be subdivided by whether they give birth through a Cesarean section or vaginally, and by whether they use exclusive breastfeeding, bottle (formula) feeding, or mixed feeding. Therefore, the inputs for the transmission probabilities for each treatment option are either vaginal births that are bottle-fed, vaginal births that are breastfed, vaginal births that used mixed feeding; or Cesarean births that are bottle-fed, Cesarean births that are breastfed, Cesarean births that used mixed feeding. Table 23 presents model defaults for vertical transmission probabilities by mode of delivery and method of feeding. 40 Table 23. Vertical Transmission Probabilities Type of Delivery Vaginal Vaginal Vaginal Cesarean Cesarean Cesarean Type of Feeding : Bottle Breast Breast Bottle Breast Breast Exclusive Mixed Exclusive Mixed Type of Treatment Long-course ZDV treatment 0.082 0.082 0.132 0.052 0.052 0.103 Short-course ZDV treatment (Thailand regimen) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm A) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm B) 0.159 0.159 0.301 0.101 0.101 0.253 Neonatal-only ZDV treatment 0.151 0.151 0.295 0.096 0.096 0.249 Nevirapine (HIVNET 012) 0.134 0.134 0.280 0.085 0.085 0.240 Universal Nevirapine 0.134 0.134 0.280 0.085 0.085 0.240 None 0.246 0.246 0.374 0.158 0.158 0.300 41 42 IV. Projection Outputs IV. PMTCT will calculate and display prevention of maternal-to-child transmission indicators by year. Recall that the number of deaths and infections averted are calculated based on a comparison with a no-intervention scenario. A number of health indicators and costs can be displayed. A complete list of indicators available is given below. • Child deaths averted. The number of child deaths averted during the year due to the intervention option. • Child infections. The number of child HIV infections during the year. • Child infections averted. The number of child HIV infections averted during the year due to the intervention. • Adult infections averted. The number of adult HIV infections that are averted during the year due to HIV- women who change their behavior, due to counseling and testing, to avoid infection. • Total infections averted. The sum of the adult and the child infections averted during the year. • Child treatment cost savings. The amount of costs saved on HIV-positive children's treatment due to the intervention. • Adult treatment cost savings. The amount of costs saved on HIV-positive adults’ treatment due to infections averted. • Total treatment cost savings. The total (sum of the adult and the child treatment cost savings) amount of costs saved on HIV-positive treatment due to infections averted from the intervention. • Total costs intervention. The costs of implementing the intervention including testing, counseling, and antiretroviral drug costs. • Net cost intervention. The difference between the total costs of the intervention (minus a user fee, if any) and the total treatment costs savings. 43 • Benefit-cost ratio. The ratio of the total treatment cost savings and the total costs of the intervention (minus a user fee, if any). • Net cost per child death averted. The net cost of the intervention for each child death averted. This figure is equal to the net cost of the intervention divided by the total number of child deaths averted. • Net cost per child infection averted. The net cost of the intervention for each child infection that is averted. This figure is equal to the net cost of the intervention divided by the total number of child infections averted. • Net cost per total infections averted. The net cost of the intervention per total number of child and adult HIV-positive infections averted. This figure is equal to the net cost of the intervention divided by the total number of infections averted. • • sary to supply all of the women who are offered the HIV • n . • the men who have C-sections as part of the • the omen that can be seen annually by one • idual feeding counseling, treatment, C-section, and formula costs. Additional maternal deaths. The number of additional maternal deaths due to the percentage of women who undergo Cesarean sections as part of the intervention. Number of HIV test kits needed. The number of additional kits neces test. Amount of extra formula required. The amount of formula required (32-oz. cans) by the country. This figure is based o the number of infants who will be formula-fed due to the intervention as well as the duration of formula feeding Number of extra C-sections required. The number of additional Cesarean sections in a given year based on number of wo intervention. Number of counselors needed. The number of additional counselors necessary to counsel all of the women who are offered the HIV test in one year. This figure is based on number of w counselor. Distribution of costs. The percentage of each of the indiv costs associated with the intervention including testing, pre/post test counseling, infant 44 V. Program Tutorial V. This tutorial covers the key steps in installing and running Spectrum and PMTCT. It assumes that you have an IBM-compatible computer running Windows 3.1 or Windows 95 or later, and that you are familiar with the basic operation of Windows programs and terminology. A. Before You Get Started First, you will need to run DemProj and AIM, part of the Spectrum system of policy models; please refer to their manuals for more information. Then, you will need to collect data and make certain decisions before running the model. The data you will need include: • the percentage of births covered in the public sector and percentage of births with antenatal visits. • the maternal mortality rate (MMR). • costs for VCT programs including HIV test, pre/post-test counseling (for HIV+/HIV- women), and possible user fees. • costs within the context of a specific region or country including the monthly formula costs per infant, lifetime discounted paediatric AIDS treatment costs, and lifetime discounted adult AIDS treatment costs. • additional (marginal) costs of C-section. This is the cost of the Cesarean section that is in addition to the initial cost of a vaginal birth. • costs of each of the ARV intervention options. Although cost inputs have default values that are based on international drug prices, you may modify them. All costs should be in the same currency throughout when running the Spectrum modules. • the number of women seen annually per counselor and the number of months of infant formula that will be paid for by the government. • the proportional increase in the child mortality rate (CMR) due to not breastfeeding. 45 • MMR increase due to C-section and the percentage of HIV- negative women who cease breastfeeding due to spillover effects. You will need to make policy decisions about the percentage of women who are offered the HIV test, the percentage who test positive, the percentage eligible who are offered treatment, and the percentage who adhere to the full course. You will also need to make certain policy decisions about which treatment option to choose, the percentage of women undergoing Cesarean sections, and the percentage of HIV-positive women who breastfeed exclusively and those who choose mixed feedings. These decisions can be changed at any time. Note that the vertical transmission probability values in the model are based on scientific studies and should be used as program defaults. It is strongly recommended that they be used in PMTCT unless there is a reason to overwrite the values. These inputs are all described in Chapter III of this manual. B. Installing the Spectrum Program The Spectrum program is distributed on floppy diskettes; it is also available through the Internet at http://www.tfgi.com. However, it must be installed onto a hard disk before it can be used. Spectrum will run on any IBM-compatible computer running Windows 3.1 or Windows 95 or later. It requires about 3MB of hard disk space. To install the Spectrum program from floppy diskettes, start by inserting the “Install” diskette into your floppy disk drive.2 For Windows 3.1: Select “File” from the Program manager menu, then select “Run.” In the dialogue box that appears, type the file name “a:\setup.exe” and press “Ok.” (If the install disk is in floppy disk drive b, then use the file name “b:\setup.exe.”) Follow the instructions on the screen to complete the installation. For Windows 95 or later: Select “Start” from the task bar. Then select “Run” from the pop-up menu. In the dialogue box that appears, type the file name “a:\setup.exe” and press “Ok.” (If the install disk is in floppy disk drive b, then use the file name “b:\setup.exe.”) Follow the instructions on the screen to complete the installation. 2 To remove the Spectrum program from your hard disk, run the Unwise program located in the Spectrum directory. 46 Complete instructions for installing Spectrum from the Internet are available on the web site. C. Creating a New Projection 1. Starting the Spectrum Program To start Spectrum, use one of the following methods3: In Windows 3.1: 1. Double click on the Spectrum icon on the desktop, or 2. Use the File Manager to locate the directory “c:\spectrum\” and then double click on the file named “spectrum.exe.” In Windows 95 or later: 1. Click the “Start” button on the task bar. 2. Select “Programs” from the pop-up menu. 3. Select “Spectrum” from the program menu. Alternatively, you can use Windows Explorer to locate the directory “c:\spectrum” and double click on the file named “spectrum.exe.” 2. Opening Demographic and AIM Projections d IM Before using PMTCT, you shoul use DemProj and AIM to prepare demographic and AIDS projections. DemProj and A are part of the Spectrum system of policy models; for more informa consult their tion, anuals. ction ) ted at least a year or two before the start of the AIDS epidemic. m DemProj. PMTCT in Spectrum requires a demographic proje prepared with DemProj and an HIV prevalence projection prepared with AIM; the two modules work together in iterative fashion. AIM influences the demographic projection by adding a number of AIDS deaths, and, possibly, specifying a lower fertility rate because of the effects of HIV infection. All the population figures required by AIM and PMTCT (e.g., number of annual births are provided by DemProj. The HIV/AIDS projections will be more accurate if the projection is star 3 The computer screen prototypes in this manual correspond to Windows 95. 47 AIM. An AIM projection must be made at the same time as the DemProj projection in order to generate the number of HIV- positive pregnant women. Therefore, before using PMTCT, consult the DemProj manual, DemProj: A Computer Program for Making Population Projections and the AIM manual, AIM: A Computer Program for Making HIV/AIDS Projections and Examining the Social and Economic Impacts of AIDS. It is important to note that a change in either DemProj or AIM will affect PMTCT. For example, if the adult female HIV prevalence rate in AIM changes, a new projection would be recalculated in PMTCT, with the inputs and outputs reflecting the new prevalence rate. Additionally, if the first year of the projection is changed in DemProj, the first year will also be changed in AIM and in PMTCT. The first step in preparing the PMTCT projection is to open the demographic projection. 1. Select “File” from the menu bar. 2. From the pull-down menu that appears, select “Open projection.” 3. Select the projection file from the “Open” dialogue box and press “Ok.” All existing projections that can be loaded will be listed here. 3. Adding the PMTCT Module to the Projections Once the demographic projection is open, you need to change the configuration to indicate that the AIM and PMTCT modules will be used as well. To do this, select “Edit” from the menu bar and “Projection” from the pull-down menu. You will see the “Projection manager” dialogue box. It will look similar to the display shown below. 48 The following information is displayed. Projection title: The title will be printed at the top of all printed output and will be used to identify the projection if more than one projection is loaded at a time. You can change the title to reflect the projection you are about to prepare. If more than one projection is loaded at a time, certain parameters (such as first and last year of projection) must be the same or an error message will appear. If you want to change the projection file name, the years, or the demographic projection interval, you will need to do so in DemProj. The options in the Projection manager were set when the demographic projection was created with DemProj. Projection file name: This is the name that will be used to store all data files associated with this projection. You cannot change the file name here. You can change it if you select “File” and “Save projection as” to save the projection to a new name. First year: This is the first year of the projection, determined in the DemProj projection. Final year: This is the final year of the projection, determined in the DemProj projection. Demography. The radio button labeled “Standard demographic projection <= 50 years” will be deselected ("grayed out") by default because a family planning projection will not be used with the PMTCT module. 49 Once all the information is entered for this dialogue box, click on the “Ok” button. You can always return to this screen and change some of the information by selecting “Edit” from the menu bar and then “Projection” from the pull-down menu. Active modules. These radio buttons (or options) let you select other modules that will be used with the population projection. You should select the PMTCT module by clicking on the check box next to the name. This choice will allow you to include the maternal to child transmission module in the projection. EasyProj. This is one method of quickly producing a data file for a demographic projection and is fully explained in DemProj: A Computer Program for Making Population Projections. Once all the information is entered for this dialogue box, click on the “OK” button. You can always return to this screen and change some of the information by selecting “Edit” from the menu or bar and “Projection” from the pull-down menu. D. Entering the Projection Inputs 1. About the Editors Each editor in PMTCT is similar. At the very top of the screen, the variable name appears. At the bottom of the screen are the special edit keys. "Cancel" allows you to cancel the entire screen if you do not want to use the information you have entered; “Duplicate” allows you to copy information from one cell, column, or row to another; “Interpolate” to enter a beginning and ending number and have the computer calculate numbers for the intervening intervals; and “Source” to write notes indicating the source of the data for future reference. To use the “Duplicate” button, 1. Highlight (select) the range (column, row, or cells to be affected). The first cell in the range should be the value you want to copy. 2. Extend the range to the last year by using the mouse (hold down the left button and drag the range) or the keyboard (hold down the shift key and use the arrow keys). 3. Click on the “Duplicate” key to copy the value at the beginning of the range to all the other cells in the range. To use the “Interpolate” button, 1. Enter the beginning and ending values in the appropriate cells. 2. Highlight the entire range from beginning to end. 3. Click on the “Interpolate” key to have the values interpolated and entered into each of the empty cells. 50 To use the “Source” button, lick on the “Source” button to open a small1. C word processor s This p a record of the data sources and is source information ill be maintained with the data file and printed whenever you window. 2. Enter the source of the data and make any special comment about the assumptions. 3. Click on “Close” to return to the editor. feature allows you to kee assumptions as you make the projections. Th w print the projection summary. It is strongly recommended that you use this feature to avoid later confusion. When you have finished entering all the necessary data for the component into the editor, If you decide that you do not want to keep the changes you have just made, click the “Cancel” button in any editor. This will exit the editors and restore all inputs to their values before you entered the PMTCT editor. Any changes made during the editing session w be lost. 1. Click the “Ok” button to return to the initial dialogue box. 2. Click the “Close” button t The “Cancel” button allows you to exit the editor without makin an changes to the data. 2. Specifying the PMTCT Projection Inputs of t e terms found in this section, Chapter III and To enter the data and PMTCT parameters: 1 ill o complete the editing process. g y For readers who feel they need additional review or explanations h the glossary of this manual may be useful. . Choose “Edit” from the menu bar. nu. A dialogue box such as the one shown below will appear. 2. Choose “PMTCT” from the pull-down me 51 Ch th to de sim 52 oose one of the input editors from the dialogue box. For each of e input editors required for the projection, there is a tab near the p of the screen. When you click on each of the inputs, a short scription line appears above the input box. The editors look ilar to the one shown below: 1. To enter data for any of the parameters, click on the appropriate tab to display the editor for that variable. 2. Then click anywhere inside the editor to make it active. 3. For most of the data editors in PMTCT, it will be sufficient to enter a base year value and a final year value for the PMTCT program and then interpolate between these two years. Be careful to only interpolate between the years of the actual PMTCT program and not between the first and final years of the projection. 4. You may specify which year the PMTCT program starts when you input the data. Note that some inputs may be zero for some years before the start of the PMTCT program. 3. Entering the PMTCT Context Inputs To enter the PMTCT Context inputs, select “Edit” from the menu bar and “PMTCT” from the pull-down menu and then select "Context" from the dialogue box. This step will display an editor like the one shown below. There is only one type of information required for “Context.” Bi s (Public) 1. 2. 3. rth Click somewhere inside the editor to make the scroll bar appear. Scroll to the right or left to see all the years. Enter the percentage of annual births that will take place in the public sector, the percentage of those births where the mother has attended a public antenatal clinic, and the 53 percentage of births in the public sector that are delivered by planned Cesarean section (C-section). 4. Enter the parameters relative to breastfeeding, which are the percentage of births breastfed exclusively for 6 months and percentage of births using mixed feeding. 5. Enter the proportional increase in the child mortality rate (CMR) among infants not breastfed. This is the proportional increase in child mortality from formula feeding. A proportional effect of two implies that the effect of formula feeding more than doubles the child mortality rate. Enter the maternal mortality rate (MMR) per 100,000 live births. When you have completed all the information in the Context dialogue box, click the "Ok" button to close. 4. Entering the PMTCT Policy/Programs Inputs To enter the PMTCT Policy/Programs inputs, select “Edit” from the menu bar and “PMTCT” from the pull-down menu or, from the Context dialogue box, simply select "Ok" and then select "Policy/Programs" from the dialogue box. This step will display an editor like the one shown below. There are four types of information required for "Policy/Programs." 54 Testing/Counseling 1. Click somewhere inside the editor to make the scroll bar appear. 2. Scroll to the right or left to see all the years. 3. For each year of the projection, enter the percentage of pregnant women offered the HIV test (usually based on costs of the test, counseling services and availability of service facilities), the percentage of pregnant women that accept the HIV test, and the percentage receiving test results. 4. Enter the number of women each counselor will see annually. Enter the percentage of HIV-negative women who cease breastfeeding due to the impacts of the spillover effect (these impacts could be from a mass media campaign, an intervention, or counseling). Enter the number of infections averted per woman with changed behavior. The default value is 0.028. When you have entered the information on Testing/Counseling, click the “Treatment Option” tab to move to the next editor. 55 Treatment Option From this dialogue box, you can choose one of the seven ARV treatment options or you can choose None (no treatment). The treatment options are described fully in the Inputs section of the manual. There are eight choices: 1. Long-course ZDV. 2. Short-course ZDV (Thailand regimen). 3. Short-course ZDV - PETRA Arm A. 4. Short-course ZDV - PETRA Arm B. 5. Neonatal only. 6. Nevirapine HIVNET 012 protocol. 7. Universal Nevirapine. 8. None - no ARV treatment. When you have entered the information on Treatment Option, click the “Treatment” tab to move to the next editor. 56 Treatment 1. Click somewhere inside the editor to make the scroll bar appear. 2. Scroll to the right or left to see all the years. 3. For each year of the projection, enter the percentage of pregnant women that are eligible for the treatment, the percentage eligible who are offered treatment, the percentage offered treatment who accept it, and the percentage who adhere to the full treatment course after accepting. Note that since treatment options differ by when in the antenatal/perinatal period they are given, the percentage of women who are eligible, offered, accept, and adhere to a specific treatment may differ as well. When you change treatment options, you may need to change these inputs. 4. Enter the percentage of pregnant HIV-positive women who undergo elective C-sections, the percentage of HIV-positive women who breastfeed exclusively for six months, and the percentage of HIV-positive women who practice mixed feeding (breastfeeding and formula). 5. Enter the proportional increase in the maternal mortality rate (MMR) due to a Cesarean section (C-section). A proportional increase of four implies that a C-section quadruples the maternal mortality rate. W en you have entered the information on Treatment, click the “C sts” tab to move to the next editor. h o 57 Costs 1. Click somewhere inside the editor to make the scroll bar appear. 2. Scroll to the right or left to see all the years. 3. Enter the costs of HIV testing and counseling. Individual costs include the HIV test, pre-test counseling, post-testing counseling for HIV-positive and HIV-negative women, counseling costs for infant feeding when women test positive, and the user fee, if any, for VCT. 4. Enter the additional costs associated with a Cesarean section birth (additional cost is the difference between a vaginal and a Cesarean-section birth), the monthly formula cost per infant, the number of months of formula (for HIV-positive women) paid for by the government, and the user fee, if any, an HIV- positive woman would be charged for the ARV treatment. 5. d , or a discussion of lifetime discounted AIDS treatment costs). s dialogue box, click the "Ok" button to close this di logue box. Enter lifetime discounted paediatric AIDS treatment costs an lifetime discounted adult AIDS treatment costs. For adults, a general rule is that the annual expenditure per adult AIDS patient is between 100 and 200 percent of per capita GNP while the cost for children is two-thirds the adult cost (see Inputs f When you have completed all the information in the Po cy/Program 58 a li 5. Entering the PMTCT Epidemiology Inputs To enter the PMTCT "Epidemiology" inputs, select “Edit” from the menu bar and “PMTCT” from the pull-down menu or, from the Policy/Programs dialogue box, select "Ok" and then select "Epidemiology" from the dialogue box. This step will display an editor like the one shown below. There are four types of information required for "Epidemiology." Transmission 1. Click somewhere inside the editor to make the scroll bar appear. 2. Although you can scroll to the right or left to see all the years, data are entered only for the first year. 3. The transmission probability values for the intra-uterine (gestation), delivery (vaginal or Cesarean), and feeding (breastfeeding exclusive or mixed) parameters are already entered in the editor and should be used as program defaults. It is strongly recommended that they be used in PMTCT unless there is a good reason to overwrite the values. A discussion of these values can be found in the Inputs section. When you have entered the information on Transmission, click the “Reduction” tab to move to the next editor. 59 Reduction 1. Click somewhere inside the editor to make the scroll bar appear. 2. Although you can scroll to the right or left to see all the years, data are entered only for the first year. 3. The values for the reduction in the probability of HIV transmission due to treatment options are already entered in the editor and should be used as program defaults. It is strongly recommended that they be used unless there is a good reason to overwrite the values. Equations for these values are discussed in the Methodology section. When you have entered the information on Reduction, click the “ARV Costs” tab to move to the next editor. 60 ARV Costs 1. Click somewhere inside the editor to make the scroll bar appear. 2. Scroll to the right or left to see all the years. 3. Enter the costs associated with different treatment options. Cost inputs have default values that are based on international drug prices and that are recommended for most projections; if these are not appropriate for this particular projection, make any changes necessary. All costs should be in the same currency throughout when running the Spectrum modules. When you have entered the information on ARV Costs, click the “Trans. Results” tab to move to the next editor. 61 Transmission Probabilities - Results 1. Click somewhere inside the editor to make the scroll bar appear. 2. Highlight a delivery method at the bottom of the screen (either Vaginal or Cesarean). The screen then shows the transmission probability results for that delivery method, treatment option, and for either the bottle-feeding, breastfeeding exclusive, or breast mixed feeding option. In this editor, the results come from the Transmission and Reduction probabilities (for a discussion of the calculations, see the Methodology section). Note that this editor is a "view only" screen; these results cannot be changed. If you do try to overwrite the values, a message that says, "Locked cells cannot be modified" will be shown. When you have finished viewing one delivery method, highlight the next method by selecting a tab near the bottom of the screen. 62 6. Leaving the PMTCT Data Editors Once you have entered all the necessary information, If you decide that you do not want to keep the changes you have just made, click the “Cancel” button in any editor. This will exit the editors and restore all inputs to their values before you entered the PMTCT editor. Any changes made during the editing session will be lost. 1. Leave the PMTCT editor by clicking the “Ok” button in any of the editors. When you click the “Ok” button, the program will record your changes and return to the “PMTCT” dialogue box. 2. Click the “Close” button to keep your work, and you will return to the main program. 7 Saving the Input Data Once you have entered the projection assumptions, it is a good idea to save the data onto your hard disk. To do this, select "File" from the menu bar and “Save projection” from the pull-down menu. The data will be saved using the file name you specified earlier. The earlier version of the file will be over-written. E. Making the Projection Whenever you enter data for a new projection or edit the assumptions, PMTCT will note that the data have been changed. The next time you try to display an indicator it will inform you that the data may have changed and ask if you want to recalculate the projection. Normally, you should answer “Yes” to this question. PMTCT will then make the projection. Once the projection is made, you will not be asked if you want to do the projection again unless you edit the assumptions. 63 F. Examining the Output To see the results of the projection, select “Display” from the menu bar. From the pull-down menu select “PMTCT.” You will then see another menu showing the indicators available: • Child Deaths Averted • Child Infections • Child Infections Averted • Adult Infections Averted • Total Infections Averted • Child Treatment Cost Savings • Adult Treatment Cost Savings • Total Treatment Cost Savings • Total Costs Intervention • Net Cost Intervention • Benefit Cost Ratio • Net Cost per Child Death Averted • Net Cost per Child Infection Averted • Net Cost per Total Infections Averted • Additional Maternal Deaths • Number HIV Test Kits Needed • Amount Extra Formula Required • Number Extra C-sections Required • Number Counselors Needed • Distribution of Costs 64 Choose one of these indicators. Then you will see the “Display” dialogue box. It will look similar to the one shown below. Th se in e exact choices available will depend on the indicator you have lected. You may select the chart type, final year, and the display terval. 65 When the “Chart Type” is set to “Table,” you may also see a check box labeled “Scale Table Values” below the final year box. This means that you can check that box to tell the program to scale the numbers in the table by a certain value such as thousands or millions. The scaler keeps the program from having to display very long numbers in the table while still leaving enough digits for adequate precision. However, the module will make the choice for you. The display will normally be in single years, but you can change it to display every five or ten years. The chart type is also set through this dialogue box. Click on the button next to the type of display you want. Normally the display will show all the years in the projection. However, if you want to see only part of the projection, you can change the final year by selecting a new final display year from the “Final Year” list box. Once you are satisfied with the type of display, click the “Ok” button, and the display will appear. It will look similar to the display shown below. 66 All the projections that are currently in use will be displayed on the same graph. You can change the configuration of the display by clicking the “Configure” button. You can also change the type of display by putting the mouse pointer anywhere inside the chart and clicking with the right mouse button. To close the display, click on the “Close” button. You do not have to close the display immediately. You can choose to display another indicator, and it will appear on top of the first display. The first display will be covered, but will still be there. You can return to any previous display that you have not closed by choosing “Window” from the menu bar and selecting the name of the display from the pull-down menu. From the “Window” selection you can also choose to tile or cascade all the existing display windows. 1. Graphs and Bar Charts PMTCT will display a variety of graphs and bar charts including: • Line charts • Two- and three-dimensional vertical bar charts (column charts) • Two- and three-dimensional horizontal bar charts • Two- and three-dimensional bar charts that appear in different windows (bars for multiple projections are shown on top of each other) 67 • Three-dimensional perspective bar charts. To print the current (selected) chart, select “File” from the menu bar and “Print” from the pull-down menu. 2. Tables PMTCT will also display data in the form of tables. In tables, each projection that is in use will be displayed in a separate column. You can scroll through the table to see all the years by using the PgUp and PgDn keys or by using the mouse. To print a table, select “File” from the menu bar and “Print” from the pull-down menu. G. Saving the Projection It is always a good idea to save the projection whenever you make a change to any of the assumptions. To save the projection without changing the name, choose “File” from the menu bar and “Save projection” from the pull-down menu. To save the projection with a different name, chose “File” from the menu bar and “Save projection as” from the pull-down menu. You will then have a chance to specify a new file name for the projection. Normally when you save the projection with a new name, you should also change the projection title. This step will avoid confusion if you have both projections loaded at the same time. H. Opening an Existing Projection If you have already created a PMTCT projection or are using a projection provided by someone else, you can immediately load that projection. 1. Select “File” from the menu bar. 2. Select “Open projection” from the pull-down menu. 3. Select the file you wish to use and click the “Ok” button to open the projection. You can open more than one projection at a time. Just repeat these steps to load a second or third projection. When you have more than one projection loaded, all projections will be displayed in the graphs and tables. The number of projections you can load at any one time is determined by the amount of available memory in your computer. 68 When you have more than one projection loaded, you will be asked to choose a projection when performing certain tasks, such as editing assumptions. The program will display a list of the e I. remove a projection that has already been opened, e menu bar, and ave more asked to select which Closi does You can open that projection gain at any time. projection names, and you choose the appropriate one from th list. Closing a Projection To 1. Choose “File” from th 2. Close projection” from the pull-down menu. If you h than one projection loaded, you will be projection should be closed. ng a projection just removes it from the computer’s memory; it not erase it from the hard disk. a 69 70 VI. Sample Application VI. This chapter describes a typical application of PMTCT. It uses a data file for a country with a medium HIV prevalence rate and a generalized spread of HIV in the population. We assume that a DHS is available for the sample country. We also assume that: • The source of health parameters such as child and maternal mortality rates are in the DHS report. • The projection of the annual number of births is taken from a DemProj projection. • The projection of female HIV prevalence and incidence rates is taken from an AIM projection. • The treatment option is nevirapine HIVNET 012, and the transmission probability values for nevirapine are the default values used in the model. • The PMTCT program starts in 1998. • HIV counseling and testing and antiretroviral treatment costs are derived from the average values presented in this manual. All costs are in US dollars (US$). • The number of infections averted per woman with changed behavior is set to 0.028. The basic model inputs for this example are shown in Tables 24-28. A. PMTCT Inputs for Sample Application The inputs include context and testing, counseling, and treatment parameters. Although inputs such as HIV counseling and testing costs and user fees may decline as PMTCT programs mature, in the sample application, costs are fixed throughout the projection. All user fees will be set to zero. In this example, the goal is to estimate the costs and benefits of the nevirapine treatment option from 1998, the first year of the PMTCT program, through the final year of the projection in 2010 compared with no treatment. Nevirapine in the peripartum period currently represents a valid option to reduce vertical transmission of HIV in 71 developing countries if appropriate prenatal, obstetrical, and postnatal care is provided and if alternatives to breast-feeding are considered according to the local situation and the mother's individual decision. Note that parameters specific to the PMTCT program should be entered only for the actual years of the PMTCT program and not for the years of the projection before the start of the program. Context parameters can be entered for all years of the projection. Table 24 represents the context parameters for the first year through the final year of the projection. Tables 25 through 28 present the policy and programs parameters for the first year of the PMTCT program through the final year of the projection. The base year of the projection (1985, set in DemProj) represents the year before the HIV/AIDS epidemic began. Table 24. Context Parameters for Sample Application Context Parameters 1985-2010 (%) % public sector births 42 % antenatal visits 92 % planned C-sections 3 % breastfed 6 months 7 % mixed feeding 90 Increase in CMR 2 MMR (per 100,000) 600 Table 25. Testing/Counseling Parameters for Sample Application Parameter 1998-2010 % offered HIV test 95 % accepting test 90 % receiving results 97 # women seen/counselor 1,000 % Breastfeeding spillover 0 Infections averted/woman .028 Table 26. Treatment Parameters for Sample Application Parameter 1998-2010 % eligible for treatment 92 % offered treatment 95 % accepting treatment 90 % adhering to treatment 90 % elective C-section 3 % HIV+ breastfed for 6 months 3 % HIV+ mixed feeding 47 MMR increase: C-section 4 72 Table 27. Cost and Duration Parameters for Sample Application (US$) Parameters 1998-2010 HIV test (US$) 4.90 Pre-test VCT (US$) 1.00 Post-test HIV- (US$) 2.18 Post-test VCT HIV+ (US$) 13.70 Post-test VCT: feeding (US$) 9.00 User fee: VCT (US$) 0 Cost: C-section birth (US$) 101 Monthly formula costs (US$) 25 # months paid by govt. 6 User fee (US$) 0 Total child treatment costs (US$) 4,214 Total adult treatment costs (US$) 6,320 Table 28 presents the treatment costs for all options. Note that in the sample application, the costs do not change from the first to the final year of the PMTCT program. Table 28. ARV Treatment Costs for Sample Application (US$) Treatment 1998-2010 (US$) Long-course ZDV 800 Short-course ZDV (Thailand regimen) 162 Short-course ZDV (PETRA A) 153 Short-course ZDV (PETRA B) 25 Neonatal only 80 Nevirapine (HIVNET 012) 4 Universal Nevirapine 4 None 0 Vertical transmission probabilities may vary by the treatment option, the mode of delivery (vaginal or Cesarean), and method of feeding (breastfed exclusively, bottle, or mixed). In this example, for nevirapine, the vertical transmission probabilities are model defaults and are set at 0.134 for vaginal delivery and bottle or breast exclusive, and at 0.280 for mixed feeding; for Cesarean delivery, the probabilities are set at 0.085 for bottle and breast exclusive, and 0.240 for mixed feeding. B. Changes in Child Deaths and HIV Infections Averted The number of child deaths and HIV infections averted are calculated based on a comparison of nevirapine with a no- treatment scenario. In the sample application, the number of child 73 deaths averted increased from zero in 1980-1998 (before the PMTCT program started) to 5,336 in the first year of the program and increased by 26 percent by 2010. The number of child infections increased from 26,282 in 1998 to a maximum of 31,377 in 2004 and then declined to 29,033, by the final year of the projection. The number of child infections averted was 7,393 in 1998 and increased to a maximum of 9,342 by 2004 before declining by almost 8.2 percent in 2010. Note that the number of child deaths averted is lower than the number of child infections averted. This is true because there will be some increase in the child mortality rate due to the decrease in breastfeeding in HIV-positive women. In the sample application, the number of adult infections averted per HIV-negative woman counseled and tested was set to 0.028. The number of adult infections averted was 9,219 in the first year of the program. The maximum number of infections averted was 9,226 in 1999 and then decreased to 7,989 by 2010. The number of total infections averted is determined by the sum of the number of child and adult infections averted. Total infections averted were 16,612 in the first year of the program and increased to a maximum of 18,320 by 2002. Total infections averted then decreased by 9 percent by 2010. Table 29. Changes in Child Deaths and Infections Averted for Sample Application Deaths/Infections Averted 1998 2010 Child deaths averted 5,335 6,717 Child infections 26,282 29,030 Child infections averted 7,393 8,638 Adult infections averted 9,219 7,989 Total infections averted 16,612 16,626 C. Changes in Treatment Cost Savings, Intervention Costs, and Net Costs Different antiretroviral options have different costs due to when in the antenatal/perinatal period treatment is initiated, how often it must be given, and the type of drug delivery system (oral or intravenous). Total intervention costs are also a function of testing and counseling fees, mode of delivery, and method of feeding. Therefore, different treatment options and their concomitant costs affect the total intervention costs, the total treatment cost savings, and, thus, the benefit-cost ratio. 74 The following outputs are calculated based on the comparison of the nevirapine treatment and no-treatment results. 1. Treatment Cost Savings The total treatment cost savings is determined by the number of child and adult infections averted and their respective lifetime discounted AIDS treatment costs. Total treatment cost savings can be interpreted as the benefit of using nevirapine. Child treatment cost savings increased by almost 17 percent from the first to the last year of the PMTCT program; adult treatment cost savings decreased 13 percent from 58.26 million in the first year of the program to 50.49 million. Total treatment costs savings increased from 89.42 million in 1998 to a maximum of 96.37 million in 2001; total treatment cost savings then decreased by almost 10 percent by 2010. 2. Total Costs Intervention In addition to the ARV drug costs, total intervention costs are determined by the number of pregnant women that accept the HIV test, the costs of pre-and post-test counseling for both HIV- positive and HIV-negative women, additional counseling costs to discuss infant feeding options for women that test positive, formula costs, and the marginal costs of a Cesarean section for a specific region or country. In the sample application, the total cost of the intervention increased from $7.50 million in 1998 to a maximum of $8.77 million in 2003 before decreasing to $8.03 million in 2010. 3. Net Cost of Intervention The PMTCT model sets the total costs of testing, counseling, and treatment against the averted lifetime costs of paediatric and adult infections (treatment cost savings). The net costs of the intervention can then be determined by subtracting the total treatment cost savings from the total intervention costs (in this example, the user fees are zero). The conditions under which the nevirapine intervention may be considered cost saving can be seen by finding a break-even point between the total intervention costs and the total treatment cost savings. If the total intervention costs are less than the total treatment cost savings (including possible user fees), then the treatment is cost saving. In this example, the net cost of the intervention was -$81.92 million in 1998 and -$87.70 million in 2001 75 and was cost saving. The net cost of the intervention decreased b 10 percent from 2001 to 2010. y 4. Benefit-Cost Ratio enefits are measured in terms of child and adult infections ates sts. e benefit-cost ratio was 11.92 in 1998, the first year of the PMTCT ry dollar 5. Net Cost per Child Death and Infections Averted e net cost per child death averted and per total infections hild plus 010. ble 30 shows the changes in treatment cost savings, total erted ble 30. Changes in Costs for Sample Application (US$) B averted, which affect the total treatment cost savings. Estim are sensitive to the increased risk of mortality associated with vaginal delivery and mixed formula feeding and to formula co The benefit-cost ratio is determined by dividing the total treatment cost savings by the total cost of the intervention (minus the user fee, if any). When the benefit-cost ratio is greater than one, the treatment cost savings outweighs the intervention costs. Th program, and decreased to 10.83 by the final year of the projection. A benefit-cost ratio of 11.92 implies that for eve spent on the intervention, $11.92 is saved. Th averted is determined by the net cost of the nevirapine intervention and the number of child deaths and total (c adult) infections averted, respectively. The net cost per child death averted decreased from -$15,353 in 1998 to -$11,740 in 2 The net cost per total infections averted decreased from -$4,931 in 1998 to -$4,743 in 2010. Ta intervention costs, and net costs per death and infections av for the first and final year of the PMTCT program. Ta Parameter 1998 (US$) 2010 (US$) Child treatment cost savings 31,160,292 36,404,960 Adult treatment cost savings 58,262,576 50,487,800 Total treatment cost savings 89,422,864 86,892,760 Total cost of intervention 7,501,488 8,025,089 Net cost of intervention -8 -7 ed d 1,921,376 8,867,672 Net cost/child death avert -15,353 -11,740 Net cost/child infection averte -11,079 -9,129 Net cost/total infections averted -4,931 -4,743 76 D. Changes in Policy and Programs Outputs Additional outputs in the model are shown in Table 31 and include: 1. Additional Maternal Deaths There were zero additional maternal deaths due to C-sections in the sample application from 1998 through 2003. In 2004 through 2010, there was one additional maternal death. Since the number of maternal deaths is a function of the number of women who undergo C-sections as part of their intervention protocol, maternal mortality did not increase by much because, in the assumptions, the percent of elective C-sections was only set to 3 percent. 2. The Number of HIV Test Kits Needed This is the number of additional HIV test kits necessary to supply all of the women who are offered the HIV test. This number is determined by the number of women who accept the HIV test and the HIV prevalence rate. Test kits increased from 403,900 in 1998 to 417,611 in 2002 and decreased by more than 11 percent by 2010. 3. The Amount of Extra Formula Required sed on f 4,287 in 2003 and decreased by more than 8 percent in 2010. . The Number of Extra Cesarean Sections Required n. ed from 1,328 in 1998 to 2,591 in the final year of the projection. . The Number of Counselors Needed al counselors needed decreased from 377 in 1998 to 327 in 2010. The amount of extra formula required by the country is ba the number of infants who will be formula-fed due to the intervention, the duration of formula feeding, and the number o months of formula paid for by the government. Extra formula is measured in the number of 32-oz cans required. The number of cans increased from 3,984,817 in 1998 to 5,05 4 The number of extra Cesarean sections required is based on the number of women who have C-sections as part of the interventio Extra C-sections increas 5 The number of additional counselors necessary to counsel all of the women who are offered the HIV test is determined by the number of women that are offered the test and the number of women that can be seen annually by one counselor. The number of addition 77 Table 31 presents the changes in the policy and program outputs from the first to the final year of the PMTCT program. Table 31. Changes in Policy/Programs Outputs Parameter 1998 2010 Additional maternal deaths 0 1 Number HIV test kits needed 403,900 369,591 Amount extra formula required 3,984,817 4,624,554 Number of extra C-sections required 1,328 2,591 Number counselors needed 377 327 6. Distribution of Costs Figure 3. Distribution of Costs 18% 4% 18% 48% 1% 8% 3% Testing Pre-test Post-test Treatment Formula C-section Infant feeding counseling The distribution of the various components of the total cost can be seen in figure 3. Testing and counseling costs account for 40 percent of the total costs, with post-test counseling accounting for 82 percent of the counseling costs. The cost for the drug treatment is only 3 percent of total costs, while the cost of providing formula for six months is 48 percent of total costs. Note that, contrary to what might have been expected, drug treatment for nevirapine, at US$4, contributes a small amount to the overall cost of the intervention. The costs of counseling and testing are quite high in the sample application; post-test counseling is over 500 percent higher for HIV-positive women than for HIV-negative women. Additionally, providing formula is a considerable expense. ARV 78 treatment costs form a small portion of the overall costs of a PMTCT prevention program. E. Exploring Alternate Program Configurations Once the initial analysis has been prepared, it can be used to explore the consequences of different program configurations. If you would like to perform a sensitivity analysis, then only those values that the sensitivity analysis will be performed upon should be changed. Several of the issues that might be investigated using PMTCT are discussed briefly below. 1. Treatment Option Antiretroviral treatment options differ by when they are given during the antenatal/perinatal period and the length of treatment. Different treatments, modes of delivery, and methods of feeding affect the total treatment costs and number of infections averted, which in turn affect the total cost savings, the net cost of the intervention, the benefit-cost ratio, and the net cost per total infections averted. For example, if the base case assumes that nevirapine is offered, an alternative scenario might assume that short-course ZDV is used (Thailand regimen). The cost savings and effect on HIV transmission between the different scenarios can then be compared. Note that since treatment options differ by when in the antenatal/perinatal period they are given, the percentage of women who are eligible for a specific treatment differs as well. When you change treatment options, you must change the percentage of women who are eligible for and are offered treatment, the percentage of those who accept it, and the percentage of women accepting who adhere to the full course. In the example that compares short- course ZDV and nevirapine, none of the inputs has been changed from the base except the treatment option. The value of infections averted per HIV-negative woman is set to 0.028 for both the ZDV and nevirapine options. Table 32 presents the vertical transmission probabilities for the different antiretroviral treatment options, modes of delivery, and methods of feeding (with short-course ZDV and nevirapine in bold type). Table 33 presents the differences in infections and deaths averted between the two treatments. Table 34 presents the differences in treatment cost savings and net costs between the nevirapine and short-course ZDV options in the final year of the projection. 79 Table 32. Transmission Probability - Results for Sample Application Type of delivery: Vaginal Vaginal Vaginal Cesarean Cesarean Cesarean Type of feeding: Bottle Breast Exclusive Breast Mixed Bottle Breast Exclusive Breast Mixed Type of Treatment Long-course ZDV treatment 0.082 0.082 0.132 0.052 0.052 0.103 Short-course ZDV treatment (Thailand) regimen) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm A) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm B) 0.159 0.159 0.301 0.101 0.101 0.253 Neonatal-only ZDV treatment 0.151 0.151 0.295 0.096 0.096 0.249 Nevirapine (HIVNET 012) 0.134 0.134 0.280 0.085 0.085 0.240 Universal Nevirapine 0.134 0.134 0.280 0.085 0.085 0.240 None 0.246 0.246 0.374 0.158 0.158 0.300 Table 33. Changes in Deaths/Infections Averted for Nevirapine vs. Short- Course ZDV Deaths/Infections Averted 2010 - Nevirapine 2010 - ZDV Differences Child deaths averted 6,717 7,029 -312 Child infections 29,033 28,688 -345 Child infections averted 8,638 8,984 -346 Adult infections averted 7,989 7,989 0 Total infections averted 16,626 16,973 -347 Although slightly more child infections are averted by 2010, when the treatment option is changed from nevirapine to short-course ZDV, and thus more total infections are averted (adult infections remain the same), ZDV is considerably more expensive. As can be seen in Table 28, ZDV costs $162 versus nevirapine, which costs $4. As mentioned previously, treatment costs vary with treatment duration. Short- course ZDV is initiated in the 34th week of pregnancy and continued through labor and birth; nevirapine treatment consists of one dose for the mother at labor onset and one dose for the child within 72 hours of birth. Compared with nevirapine, ZDV has a higher total treatment cost savings (due to more child infections averted). However, the net cost of the intervention and the net cost per child death averted, per child infection averted, and per total infections averted is slightly better for nevirapine than for ZDV (Table 34). 80 Table 34. Costs for Nevirapine vs. Short-Course ZDV (US$) Parameter 2010- Nevirapine 2010-Short- Course ZDV Differences Child treatment cost savings 36,404,960 37,860,528 -1,455,568 Adult treatment cost savings 50,487,800 50,487,800 0 Total treatment cost savings 86,892,760 88,348,328 -1,455,568 Total cost of intervention 8,025,089 16,212,525 -8,187,436 Benefit-cost ratio 10.83 5.45 5.38 Net cost intervention -78,867,672 -72,135,800 -6,731,872 Net cost/child death averted -11,740 -10,263 -1,477 Net cost/child infection averted -9,129 -8,029 -1,100 Net cost/total infections averted -4,743 -4,250 -493 The benefit-cost ratio for ZDV is much smaller than for nevirapine by the final year of the projection (5.45 vs. 10.83). This implies that although the total treatment cost savings are comparable for the two ARV options, the total cost of the intervention for ZDV is approximately twice the total cost of the intervention for nevirapine. Therefore, when ZDV is used, the benefit-cost ratio decreases by almost 50 percent compared with nevirapine for the final year of the projection. 2. Number of Infections Averted In the sample application, the number of infections averted per woman counseled and tested is set to 0.028. The number of infections averted is secondary infections due to counseled women changing their behavior. If you do not want to include these secondary effects, you may change this parameter to zero. Note that the savings for adult treatment costs, which are included in the benefits in the calculations of the benefit-cost ratio, become zero when no adult infections are averted. These savings may be high, and thus their inclusion or exclusion may significantly affect the benefit-cost ratio. If the number of infections averted per woman counseled and tested is set to zero, the benefit-cost ratio becomes 4.15 in 1998 compared with 11.92, when the number of infections averted is set to 0.028. Additionally, the ratio increases to only 4.54 by 2010, which is a decrease of 58 percent, compared to the benefit-cost ratio when 0.028 infections averted is used in the model. 81 3. Parameters in DemProj and AIM Parameters could be changed in DemProj and AIM that would affect their outputs and subsequently affect the PMTCT model. For example, you may want to vary the parameters in AIM to generate high or low HIV population prevalence rates. The variation in the number of women with a positive HIV test (which affects the percentage of annual births treated) as well as other pre-treatment parameters may be compared as prevalence rates vary. If the HIV incidence rate from AIM is allowed to vary, then the number of women who may become infected with HIV after birth, and, therefore, transmit HIV during breastfeeding, will vary. Additionally, the number of child infections in PMTCT varies directly as the number of annual births created by DemProj changes. 4. User Fees As discussed in this sample application, PMTCT can also be used to investigate the effects of varying testing, counseling, and treatment costs as well as examining the differences in possible user fees for each treatment option. The changes in fees would impact the net cost of the intervention as well as the net cost per death or infection averted and the benefit-cost ratio. Depending on the country, some or all of the costs may be recovered by user fees. This change might occur through the introduction of new programs, such as a social marketing program, or through activities to encourage the commercial or NGO sector to expand services. The results might be used to plan for a scaling up of an existing counseling and testing program or to decide how to allocate scarce maternal and child health resources. 82 VII. Methodology VII. The model starts with the annual number of births occurring to women in the population of interest. It then follows those births to determine the number covered by the public sector, the number with antenatal visits, the number offered an HIV test, the number accepting an HIV test, etc. The result is to divide births into several different categories: • Those where the mother is known to be HIV+ • Those where the mother is known to be HIV- • Those where the mother’s HIV status is unknown In those cases where the mother’s HIV status is unknown we assume that a certain percentage, equal to the adult prevalence rate, are HIV+ and could give birth to an HIV+ baby. If the mother is HIV- then the baby will be HIV-. If the mother is HIV+ then she may receive treatment or not. The probability of an HIV+ birth depends on whether she receives treatment and the type of treatment. If the baby is born HIV-, there is the possibility of transmission through breast milk if the mother breastfeeds her baby. Costs are calculated at each stage of the process, including testing, counseling, treating, delivery and breastfeeding replacement. The final outcomes are compared with a scenario in which no PMTCT program is present to determine the number of infections and deaths averted by the intervention. The total costs of the intervention are compared to the number of infections and deaths averted to calculate the cost-effectiveness ratios. Costs to treat HIV+ babies may be avoided through the PMTCT program. These foregone costs are considered benefits of the program and used to calculate cost-benefit ratios. The details of these calculations are given below. Calculated variables are shown in standard font while input assumptions are shown in bold. 83 (1) Public Sector Births = Annual Number of Births * Proportion of Births Covered by the Public Sector (2) Births with Antenatal Clinic Visit = Public Sector Births * Proportion of Births with an Antenatal Visit (3) Women Offered HIV Test = Births with Antenatal Clinic Visit * Proportion of Births Offered HIV Test (4) Number Accepting HIV Test = Women Offered HIV Test * Proportion of Those Offered Test Who Accept It (5) Number Receiving Test Results = Number Accepting HIV Test * (Proportion Tested Who Receive the Test Result) (6) Number testing HIV+ = Number Receiving Test Results * HIV prevalence among pregnant women (7) Number With Unknown HIV Status Due to Lack of Testing = Number in Public Sector – Number Receiving Test Results After test results are received, the universe of women in the public sector can be split into three parts: women who are known to be HIV+, women who are known to be HIV- and all others. (A small proportion of women who receive a negative test result are assumed to be HIV+ but received a false negative result). For women who are known to be HIV+ they may be eligible for treatment or not. If they are eligible they may accept treatment or not. If they accept treatment they may adhere to the recommended regimen or they may not. All those who adhere to the regimen are assumed to receive treatment. All other women are assumed to receive no treatment. (8) Number Eligible for Treatment = Number Testing HIV+ * Proportion of Those Testing Positive Who Are Eligible For Treatment (9) Number Offered Treatment = Number Eligible for Treatment * Proportion of Those Eligible Who Are Offered Treatment (10) Number Accepting Treatment = Number Offered Treatment * Proportion of Those Offered Treatment Who Accept It (11) Number Complying With Treatment = Number Accepting Treatment * Proportion of Those Accepting Treatment Who Comply Fully 84 Both groups of women (those who receive treatment and those who do not) are further subdivided by whether they give birth through Cesarean section or vaginally and whether they use exclusive breastfeeding, mixed breastfeeding or substitute feeding. The transmission rates by type of treatment, delivery and feeding option are shown in the table below. The approach to calculating these rates is described in Appendix A. Table 35. Vertical Transmission Probabilities Type of delivery: Vaginal Vaginal Vaginal Cesarean Cesarean Cesarean Type of feeding: Bottle Breast Breast Bottle Breast Breast Exclusive Mixed Exclusive Mixed Type of Treatment Long-course ZDV treatment 0.082 0.082 0.132 0.052 0.052 0.103 Short-course ZDV treatment (Thailand regimen) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm A) 0.127 0.127 0.274 0.081 0.081 0.236 Short-course ZDV treatment (PETRA Arm B) 0.159 0.159 0.301 0.101 0.101 0.253 Neonatal-only ZDV treatment 0.151 0.151 0.295 0.096 0.096 0.249 Nevirapine (HIVNET 012) 0.134 0.134 0.280 0.085 0.085 0.240 Universal Nevirapine 0.134 0.134 0.280 0.085 0.085 0.240 None 0.246 0.246 0.374 0.158 0.158 0.300 The number of infants who will be HIV+ is calculated from the prevalence of HIV among pregnant women, the transmission rates shown in the table above and the distribution of women by type of treatment, delivery and feeding. (12) Number of HIV+ infants = Prevalence of HIV among pregnant women x Σd,f,t Proportion of birthsd,f,t x transmission probabilityd,f,t Where: d = type of delivery: Cesarean or vaginal f = type of feeding: exclusive breastfeeding, mixed breastfeeding or bottle feeding t = type of treatment All HIV+ children are assumed to die before the age of five. Deaths to children who are HIV- depend on the child mortality rate. This rate is modified by breastfeeding, so that children who are not breastfed have a higher probability of 85 dying before age five than children who are breastfed. The calculation of these rates is described in Appendix A. (13) Child Mortality = Number of HIV+ infants + Σf Number of HIV- infantsf x mortality ratef Where f refers to the feeding option: exclusive breastfeeding, mixed breastfeeding or bottle-feeding. The costs for the intervention are composed of the costs of testing and counseling, treatment, additional Cesarean deliveries and formula. User fees for testing and treatment may offset some of these costs. (14) Cost of Testing = Number Accepting HIV Test * (HIV Test + Total Counseling Costs) (15) Testing Revenues = Number Accepting HIV Test * User Fee for Testing/Counseling (16) Cost of Treatment = Number Accepting Treatment * Costs per Person Treated (17) Treatment Revenues = Number Accepting Treatment * User Fee for Treatment (18) Cost of Additional C-sections = Number of additional Cesarean deliveries * Additional Cost of C-Section Birth (19) Cost of Infant formula = Number of Infants Bottle-Fed * Monthly Formula Costs * Number of Months Formula Feeding The net costs of the intervention are the total costs minus the revenues. (20) Net cost = Cost of Testing + Cost of Treatment + Cost of Additional C-sections + Cost of Infant Formula – Testing Revenues – Treatment Revenues The number of infant infections averted is the number of infant infections in the case of no intervention minus the number of infections with the intervention. (21) Infant infections averted = Infant infections with no intervention – Number of HIV+ infants 86 Women who undergo counseling and testing may change their behavior to avoid becoming infected or to avoid passing on an infection. The number of adult infections averted is equal to the number of women receiving counseling and testing multiplied by the number of adult infections avoided per women counseled. (22) Adult Infections Avoided = Number Receiving Test Results * Adult Infections Avoided per Person Counseled The total costs of the intervention can be compared to the number of infections averted to determine the cost per infection averted. (23) Cost per Infection Averted = Net cost / (Infant infections Averted + Adult Infections Averted) Each infection averted means that the government will not have to pay for treatment. These foregone treatment costs are the financial benefits of the program. (24) Benefits = Infant Infections Averted x Cost per Infant Treated + Adult Infections Averted x Cost per Adult Treated The benefit-cost ratio is the ratio of the total benefits of the program to the net costs. (25) Benefit-cost ratio = Benefits / Net cost Caveats Certain considerations have not been included in this model. For example, a woman could contract the virus upon receiving a transfusion, perhaps at delivery or as a treatment for malaria. The possible transmission of the virus to the infant in this case is not included in the model. Additionally, the infrastructure for the delivery of these services is assumed to be in place. Service availability could be a serious issue in some countries; additional counseling and testing facilities are necessary, along with increased numbers of trained counselors. Training of the pregnant women must also be provided, to ensure that the treatment protocol is followed. Cesarean section deliveries must take place at a hospital, with appropriate anesthesia and other inputs. Finally, infant formula must be made readily available, along with training of mothers to use safe water to mix the formula. No additional expenses are included in the model to include these factors. 87 Another consideration not included here is the effect on economic output of the intervention that takes place. Some models include as a benefit the productivity gains contributed by those children who are not HIV+, due to the intervention. These gains are not considered here. There are also some implications of the model that are not considered. For example, no treatment for the mother or the infant is included in the model after the initial postpartum period; thus these costs are not included in total treatment costs. Furthermore, if a woman is HIV+, she will probably die, leaving the child without a mother. If the father is also HIV+, and dies, the child is left an orphan. The costs of raising the orphaned children are not included in the model. These costs could be high; one source suggests that these costs are US$600 per year in Thailand, excluding staff salaries. Finally, the interventions evaluated here are not compared with other possible public health interventions, such as immunizations and oral rehydration solutions. The 1993 World Bank Development Report shows that the per capita cost for the EPI Plus immunization program in Tanzania and Uganda was US$0.50 in 1990, and the per- child cost was $14.60. The per capita cost for a health program to treat worm infections and micronutrient deficiencies was US$0.30, and the cost per child was US$3.60. These figures are provided as a benchmark to compare the results of this model. 88 VIII. References VIII. Allen, S, A Serufilira, J Bogaerts, P Van de Perre, F Nsengumuremyi, C Lindan, M Carael, W Wolf, T Coates, S Hulley. 1992. “Confidential HIV testing and condom promotion in Africa. Impact on HIV and gonorrhea rates.” Journal of the American Medical Association 268(23):3338-3343. Allen, S, J Tice, P Van de Perre, A Serufilira, E Hudes, F Nsengumuremyi, J Bogaerts, C Lindan, S Hulley. 1992. “Effect of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa.” British Medical Journal 304(6842):1605-1609. Bartlett, JG. Medical Management of HIV Infection. 1998. Baltimore, MD : Johns Hopkins University, Dept. of Infectious Diseases. Beguet L, Castelbon K, Ekouevi D, Leroy V. 2001. “Acceptibilité du dépistage prenatal utilisant les tests rapides pour la prevention de la transmission mère enfant du VIH (TME): DITRAME plus ANRS- 1201/1202.” Paper presented at the XIIth International Conference on AIDS and STDs in Africa, 9-13 December 2001, Ouagadougou, Burkina Faso (abstract no. 10BT2-5). Bentley, ME, Spratt K, Shepherd ME, Gangakhedkar RR, Thilikavathi S, Bollinger RC, Mehendale SM. 1998. “HIV testing and counseling among men attending sexually transmitted disease clinics in Pune, India: changes in condom use and sexual behavior over time.” AIDS 12(14):1869-1877. Bhatiasevi, A. 1998. 1998. “Better care for HIV-positive mothers and babies urged-Princess wants govt to allocate more funds.” The Bangkok Post, 25 November 1998. Bobat, R, Coovadia H, Coutsoudis A, Moodley D. 1996. “Determinants of mother-to-child transmission of human immunodeficiency virus type 1 infection in a cohort from Durban, South Africa.” Pediatric Infectious Disease Journal 15(7):604-610. Bobat, R, Moodley D, Coutsoudis A, Coovadia H. 1997. “Breastfeeding by HIV-1-infected women and outcome in their 89 infants: a cohort study from Durban, South Africa.” AIDS 11(13):1627-1633. “Bottle-feeding strategies.” 1999. http://family.go.com/Categories/Parenting/Features/family_0307_0 2/d…/dony77bafeeding2.html. Bouillin D, Fournier G, Gueye A, Diadhiou F, Cisse CT. 1994. “Epidemiological surveillance and obstetrical dystocias surgery in Senegal.” Sante 4(6):399-406. “Breast-Feeding and HIV: Weighing the Risks.” 1998. New York Times 19 August 1998. Bryson Y. 1996. “Perinatal HIV-1 transmission: recent advances and therapeutic interventions.” AIDS 10(suppl 3):S33-S42. Bulterys M, Chao A, Dushimimana A, Saah A. 1995. “HIV-1 seroconversion after 20 months of age in a cohort of breastfed children born to HIV-1-infected women in Rwanda.” AIDS 9(1):93- 94. Burger H, Kovacs A, Weiser B, Grimson R, Nachman S, Tropper P, van Bennekum AM, Elie MC, Blaner WS. 1997. “Maternal serum vitamin A levels are not associated with mother-to-child transmission of HIV-1 in the United States.” Journal of Acquired Immune Deficiency Syndrome and Human Retrovirology 14(4):321- 326. Cartoux, M, Meda N, Van de Perre P, Newell ML, de Vincenzi I, Dabis F, and the Ghent International Working Group on Mother-to- Child Transmission of HIV. 1998. “Acceptability of voluntary HIV testing by pregnant women in developing countries: an international survey.” AIDS 12(18):2489-2493. Cartoux M, Msellati P, Meda N, Welffens-Ekra C, Mandelbrot L, Leroy V, Van de Perre P, Dabis F, for the DITRAME Study Group (ANRS 049 Clinical Trial). 1998. “Attitude of pregnant women towards HIV testing in Abidjan, Cote d’Ivoire and Bobo-Dioulasso, Burkina Faso.” AIDS 12(17):2337-2344. Carusi D, Learman LA, Posner SF. 1998. “HIV Test Refusal in Pregnancy.” Obstetrics and Gynecology 91(4):540-545. Centers for Disease Control and Prevention. 1994. “Recommendations for the use of zidovudine to reduce perinatal transmission of human immunodeficiency virus.” Morbidity and Mortality Weekly Report 43(RR-11):1-20. 90 Centers for Disease Control and Prevention. 1998. “Administration of zidovudine during late pregnancy to prevent perinatal HIV transmission – Thailand 1996-98.” Morbidity and Mortality Weekly Report 47(8):151-153. Connor EM, Sperling RS, Gelber R, et al. 1994. “Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine therapy.” New England Journal of Medicine 331(18):1173-1180. Culnane M, Fowler M, Lee SS, McSherry G, Brady M, O’Donnell K, Mofenson L, Gortmaker SL, Shapiro DE, Scott G, Jimenez E, Moore EC, Diaz C, Flynn PM, Cunningham B, Oleske J. 1999. “Lack of long- term effects of in-utero exposure to zidovudine among uninfected children born to HIV-infected women.” Journal of the American Medical Association 281(2):151-157. Coutsoudis A, Kubendran P, Kuhn L, Spooner E, Tsai W-Y, Coovadia HM, for the South African vitamin A study group. 2001. “Method of feeding and transmission of HIV-1 from mothers to children by 15 months of age: prospective cohort study from Durban, South Africa.” AIDS 15(3):379-387. De Muylder X. 1993. “Cesarean section in developing countries: some considerations.” Health Policy and Planning 8(2):101-12. De Zoysa, I, Phillips KA, Kamenga MC, O’Reilly KR, Sweat MD, White RA, Grinstead OA, Coates TJ. 1995. “Role of HIV counseling and testing in changing risk behavior in developing countries.” AIDS 9(suppl A):S95-S101. “Early Data from Mother-to-Child HIV Transmission Study in Africa Finds Shortest Effective Regimen Ever.” 2 February 1999. http://www.hivnet.ch:8000/sea-aids/tdm. Fawzi WW, Hunter DJ. 1998. “Vitamins in HIV Disease Progression and Vertical Transmission.” Epidemiology 9(4):457. Fawzi WW, Msamanga GI, Spiegelman D, Urassa EJ, McGrath N, Mwakagile D, Antelman G, Mbise R, Herrera G, Kapiga S, Willett W, Hunter DJ. 1998. “Randomised trial of effects of vitamin supplements on pregnancy outcomes and T cell counts in HIV-1- infected women in Tanzania.” Lancet 351(9114):1477-1482. Feachem RG, Koblinsky MA. 1984. “Interventions for the control of diarrhoeal diseases among young children: promotion of breastfeeding.” Bulletin of the World Health Organization 62(2):271- 291. 91 Gottlieb S. 1998. “Elective cesarean and zidovudine cuts HIV transmission.” British Medical Journal 1998;317:11. Gray GE, McIntyre JA, Lyons SF. 1996. “The effect of breastfeeding on vertical transmission of HIV-1 in Soweto, South Africa.” Paper presented at the 1996 International AIDS Conference 1(2):237 (abstract no. Th.C.415). Guay LA, Musoke P, Fleming T, Bagenda D, Allen M, Nakbiito C, et al. 1999. “Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial.” Lancet 354(9181):795-802. Habicht J-P, DaVanzo J, Butz WP. 1986. “Does breastfeeding really save lives, or are apparent benefits due to biases?” American Journal of Epidemiology 123(2):279-290. d Harvard AIDS Institute. 1995. International Symposium on Paediatric AIDS in Thailand: A Public Health and Social Dilemma. Cambridge, MA: Harvard AIDS Institute. Heymann SJ. 1995. “HIV Transmission Through Breastfeeding: Perspectives on Prevention.” In Harvard AIDS Institute, International Symposium on Paediatric AIDS in Thailand: A Public Health and Social Dilemma. Cambridge, MA: Harvard AIDS Institute. Heyward WL, Batter VL, Malulu M, Mbuyi N, Mbu L, St Louis ME, Kamenga M, Ryder RW. 1993. “Impact of HIV counseling and testing among child-bearing women in Kinshasa, Zaire.” AIDS 7(12):1633-1637. Higgins DL, Galavotti C, O’Reilly KR, Schnell DJ, Moore M, Rugg DL, Johnson R. 1991. “Evidence for the effects of HIV antibody counseling and testing on risk behaviors.” Journal of the American Medical Association 266(17):2419-2429. Hobcraft JN, McDonald JW, Rutstein SO. 1985. “Demographic determinants of infant and early child mortality: a comparative analysis.” Population Studies 39(3):363-385. “Hope and Controversy Surround New AIDS Therapy for Babies.” Inter Press Service (June 30, 1998). “How much formula does your baby need?” 1999. http://www.carnationbaby.com/Y/Ya2g.html. 92 Hu DJ, Heyward WL, Byers Jr RH, et al. 1992. “HIV infection and breastfeeding: policy implications through a decision analysis model.” AIDS 6(12):1505-1513. International Perinatal HIV Group. 1999. “The mode of delivery and the risk of vertical transmission of human immunodeficiency virus type 1 – a meta-analysis of 15 prospective cohort studies.” New England Journal of Medicine 34(13):977-987. Janowitz B, Covington D, Higgins JE, Moreno LF, Nakamura MS, Nunez JA , Letelier MM. 1983. “Cesarean delivery in selected Latin American hospitals.” Pp. 43-55 in Potts M, Janowitz B, Fortney JA, ed. Childbirth in Developing Countries. Boston MA: MTP Press. Kamenga M, Ryder RW, Jingu M, Mbuyi N, Mbu L, Behets F, Brown C, Heyward WL. 1991. “Evidence of marked sexual behavior change associated with low HIV-1 seroconversion in 149 married couples with discordant HIV-1 serostatus: experience at an HIV counselling center in Zaire.” AIDS 5(1):61-67. Kanshana S, Thewanda D, Teeraratkul A, Limpakarnjanarat K, Amornwichet P, Kullerk N, et al. 2000. “Implementing short-course zidovudine to reduce mother-infant HIV transmission in a large pilot program in Thailand.” AIDS 14(11):1617-1623. Kennedy KI, Fortney JA, Bonhomme MG, Potts M, Lamptey P, Carswell W. 1990. “Do the benefits of breastfeeding outweigh the risk of postnatal transmission of HIV via breastmilk?” Tropical Doctor 20(1):25-29. Kiarie J, Nduati R, John G, Gakuru C, Musia J, Bwayo J, Rogo K, Kamau K, Kreiss J. 1996. “Acceptability of antenatal HIV-1 screening.” Paper presented at the 1996 International AIDS Conference 11(1):32(abstract no. Mo.C.211). Killewo JZ, Kwesigabo G, Comoro C, Lugalla J, Mhalu FS, Biberfeld G, Wall S, Sandstrom A. 1998. “Acceptability of voluntary HIV testing with counselling in a rural village in Kagera, Tanzania.” AIDS Care 10(4):431-439. Kinghorn A. 1998. “Interventions to reduce mother-to-child transmission in South Africa.” AIDS Analysis Africa 8(5):10-11. Kreiss J. 1997. “Breastfeeding and vertical transmission of HIV-1.” Acta Paediatric Supplement 1997 June;421:113-117. Kuhn L, Bobat R, Coutsoudis A, Moodley D, Coovadia HM, Tsai W-Y, Stein ZA. 1996. “Cesarean deliveries and maternal-infant HIV 93 transmission: results from a prospective study in South Africa.” Journal of Acquired Immune Deficiency Syndrome and Human Retrovirology 11(5):478-483. Kuhn L, Stein Z. 1997. “Infant survival, HIV infection, and feeding alternatives in less-developed countries.” American Journal of Public Health 87(6):926-931. Ladner J, Leroy V, Msellati P, Nyiraziraje M, De Clercq A, Van de Perre P, Dabis F, for the Pregnancy and HIV Study Group (EGE). 1996. “A cohort study of factors associated with failure to return for HIV post-test counselling in pregnant women: Kigali, Rwanda, 1992- 93.” AIDS 10(1):69-75. Leroy V, Newell ML, Dabis F, Peckham C, Van de Perre P, Bulterys M, Kind C, Simonds RJ, Wiktor S, Msellati P. 1998. “International multicentre pooled analysis of late postnatal mother-to-child transmission of HIV-1 infection.” Lancet 352(9128):597-600. Lilford RJ, van Coeverden de Groot HA, Moore PJ, Bingham P. 1990. “The relative risks of cesarean section (intrapartum and elective) and vaginal delivery: a detailed analysis to exclude the effects of medical disorders and other acute pre-existing physiological disturbances.” British Journal of Obstetrics and Gynaecology 97(10):883-892. Lorenzi P, Spicher VM, Laubereau B, Hirschel B, Kind C, Rudin C, Irion O, Kaiser L, and the Swiss HIV Cohort Study, the Swiss Collaborative HIV and Pregnancy Study, and the Swiss Neonatal HIV Study. 1998. “Antiretroviral therapies in pregnancy: maternal, fetal and neonatal effects.” AIDS 12(18):F241-F247. Lovvorn AE, Quinn SC, Jolly DH. 1997. “HIV testing of pregnant women: a policy analysis.” Journal of Public Health Policy 18(4):401-432. Luzuriaga K, Bryson Y, Krogstad P, Robinson J, Stechenberg B, Lamson M, Cort S, Sullivan JL. 1997. “Combination treatment with zidovudine, didanosine, and nevirapine in infants with human immunodeficiency virus type 1 infection.” New England Journal of Medicine 336(19):1343-1349. Lyons SF, Bredell WJ, McGillivray GM, Whistler T, Gray G, McIntyre JA. 1996. “Mother-to-infant transmission of HIV-1 in South Africa.” Paper presented at the 1996 International AIDS Conference 11(1):365(abstract no. Tu.C.2579). 94 Mansergh G, Haddix AC, Steketee RW, Nieburg PI, Hu DJ, Simonds RJ, Rogers M. 1996. “Cost-effectiveness of short-course zidovudine to prevent perinatal HIV type 1 infection in a sub-Saharan African developing country setting.” Journal of the American Medical Association 276(2):139-145. Mansergh G, Haddix AC, Steketee RW, Simonds RJ. 1998. “Cost- effectiveness of zidovudine to prevent mother-to-child transmission of HIV in sub-Saharan Africa.” Journal of the American Medical Assocation 280(1):30-31. Marlink R. 1995. “Preventing perinatal transmission of HIV in developing countries.” In International Symposium on Paediatric AIDS in Thailand: A Public Health and Social Dilemma. Cambridge, MA: Harvard AIDS Institute. Marseille E, Kahn JG, Saba J. 1998. “Cost effectiveness of antiviral drug therapy to reduce mother-to-child HIV transmission in sub- Saharan Africa.” AIDS 12(8):939-948. Marseille E, Kahn JG, Mmiro F, et al. 1999. “Cost effectiveness of single-dose nevirapine regimen for mothers and babies to decrease vertical HIV-1 transmission in sub-Saharan Africa.” Lancet 354(9181):803-809. Mauskopf JA, Paul JE, Wichman DS, White AD, Tilson HH. 1996. “Economic impact of treatment of HIV-positive pregnant women and their newborns with zidovidine.” Journal of the American Medical Association 276(2):132-138. Mazhani L, Phiri L, Keapoletswe K, Bowelo M, Mugabe M, Mouzin E. 2000. “Implementationof a population-based pilot program to reduce mother-to-child HIV transmission, Botswana.” Paper presented at the 2000 International AIDS Conference (abstract no. WeOrC550). Mbori-Ngacha D, Ndutai R, John G, Reilly M, Richardson B, Mwatha A, Ndinya-Achola J, Bwayo J, Kreiss J. 2001. “Morbidity and mortality in breastfed and formula-fed infants of HIV-1-infected women: a randomized clinical trial.” Journal of the American Medical Association 286(19):2413-2420. McDaniel, CG. 1997. “Mothers should breast feed for at least a year.” British Medical Journal 315(7121):1487. McIntyre JA, Gray GE. 1998. “The feasibility of short-course antiretroviral therapy for the prevention of mother to child 95 transmission of HIV in Soweto, South Africa.” Paper presented at the 1998 International AIDS Conference 12:406(abstract no. 23322). McIntyre, JA, Gray GE, Lyons SF. 1996. “Maternal and obstetrical factors in mother to child transmission of HIV in Soweto, South Africa.” Paper presented at the 1996 International AIDS Conference 11(1):246(abstract no. Tu.C.342). McKenna, SL, Muyinda GK, Roth D, Mwali M, Ng’andu N, Myrick A, Luo C, Priddy FH, Hall VM, von Lieven AA, Sabatino JR, Mark K, Allen SA. 1997. “Rapid HIV testing and counseling for voluntary testing centers in Africa.” AIDS 11(Suppl 1):S103-S110. Meda N, Dao B, Ouangre A. 1998. “HIV, maternal anemia and perinatal intervention using zidovudine.” International Journal of Gynecology and Obstetrics 61(1):65. “Medical Management of HIV Infection.” 1999. http://www.hopkins-AIDS.edu. Minkoff H, O’Sullivan MJ. 1998. “The Case for Rapid HIV Testing During Labor.” Journal of the American Medical Association 279(21):1743. Miro F. Makerere University, Uganda, quoted in New York Times, 19 August 1998, as cited in Global Health Council, Global AIDSLink #51, June/July/August 1998, p. 22. Nakchbandi IA, Longenecker JC, Ricksecker MA, et al. 1998. “A decision analysis of mandatory compared with voluntary HIV testing in pregnant women.” Annals of Internal Medicine 128(9):760-767. Newell, M-L. 1998. “Mechanisms and timing of mother-to-child transmission of HIV-1.” AIDS 12(8):831-837. Newell, M-L, Dabis F, Tolley K, Whynes D, for the Ghent International Working Group on Mother-to-Child Transmission of HIV. 1998. “Cost-effectiveness and cost-benefit in the prevention of mother- to-child transmission of HIV in developing countries.” AIDS 12(13):1571-1580. Nicoll A, Newell M-L, Van Praag E, Van de Perre P, Peckham C. 1995. “Infant feeding policy and practice in the presence of HIV-1 infection.” AIDS 9(2):107-119. Nieburg P, Hu DJ, Moses S, Nagelkerke N. 1995. “Contribution of breastfeeding to the reported variation in rates of mother-to-child HIV transmission.” AIDS 9(4):396-397. 96 Nishimoto MT, Gibbons AS, Segall-Correa AM, Teruya KM, Hayden RL. 1998. “Seroconversion in infants of positive mothers to HIV total breastfeeding impact.” Paper presen ted at the 1998 International AIDS Conference 12:46 (abstract no. 12175). gy and Obstetrics 279(2):171-176. l and d Obstetrics 34(4):331. 1):67-71. ORC Macro. Demographic and Health Survey, various years and Over M, Bertozzi S, Chin J, et al. 1988. “The direct and indirect cost M, FitzSimons DW, et al., eds. Global Impact of AIDS. New York, NY: Alan R Liss Inc. tors t no. PS-04-2). The Ramon R, La Ruche G, Sylla-Koko F, Boka-Yao A, Bonard D, nd practices of women of childbearing age in Abidjan, Cote d’Ivoire.” Journal of Acquired Immune Deficiency Syndrome and Human Retrovirology 17(5):470-476. Ojo VA, Adetoro OO, Okwerekwu FE. 1988. “Characteristics of maternal deaths following cesarean section in a developing country.” International Journal of Gynaecolo Okafor, CB. 1991. “Availability and use of services for materna child health care in rural Nigeria.” International Journal of Gynecology an Onsrud L, Onsrud M. 1996. “Increasing use of cesarean section, even in developing countries.” Tidsskr Nor Laegeforen 116( countries. Calverton, Maryland: ORC Macro. of HIV infection in developing countries: the cases of Zaire and Tanzania.” Pp. 123-135 in Fleming, AF, Carballo Ozumba BC, Mbadiwe IJ. 1993. “Preterm caesarean section in Nigerian obstetric practice.” Journal of the National Medical Association 85(11):857-860. Peckham CS. 1993. “Mother-to-child transmission of HIV: risk fac and timing.” Paper presented at the 1993 International AIDS Conference 9(1):10(abstrac Phillips KA, Bleecker T, Morrison KR, Sonnad SS. 1996. “HIV counseling and testing of pregnant women.” Journal of the American Medical Association 276(4):283-284. Pickering H, Quigley M, Pepin J, Todd J, Wilkins A. 1993. “The effects of post-test counselling on condom use among prostitutes in Gambia.” AIDS 7(2):271-273. “Problems of Unnecessary Medical Bills.” 1998. Inter Press Service 19 October 1998. Coulibaly IM, Welffens-Ekra C, Dabis F. 1998. “HIV counseling and testing: behavior a 97 Ratcliffe J, Ades AE, Gibb D, Sculpher MJ, Briggs AH. 1998. “Prevention of mother-to-child transmission of HIV-1 infection: alternative strategies and their cost-effectiveness.” AIDS 12(11):1381-1388. Rosenthal G, Percy A. MotherCare/John Snow, Inc. 1991. “Maternity d AID Ryder RW, Batter VL, Nsuami M, Badi N, Mundele L, Matela B, Saba J. 1999. “Current status of PETRA study.” Paper presented at f HIV ontreal, Santos V, Cunha CS, Vasconcelos AL, Sudo E, Chequer PJ. 1998. Sall NM, Diouf A, Goumbala M, Gueye M, Signa H. 2001. “Bilan d’un r 2001, (abstract no. 10BT2-4). IDS: comparison of infected children born to screened and unscreened Seals BF, Hennessey M, Sowell R. 1996. “Factors influencing revent Seyler C. 2001. “Mortality of HIV-infected adults receiving RCI.” Services in Cochabamba, Bolivia: Costs, Cost-Recovery, an Changing Markets. Report of a Field Study: April 15-27, 1991.” Contract #DPE-5966-Z-00-8083-00. Utshudi M, Heyward WL. 1991. “Fertility rates in 238 HIV-1- seropositive women in Zaire followed for 3 years post-partum.” AIDS 5(12):1521-1527. the 2nd Conference on Global Strategies for the Prevention o transmission from Mother to Infants; September 2, 1999: M Quebec. “The challenge to reduce HIV vertical transmission in Brazil.” Paper presented at the 1998 International AIDS Conference 12:400(abstract no. 23295). an de prévention de la transmission du VIH de la mère à l’enfant (PTME) à Dakar.” Paper presented at the XIIth International Conference on AIDS and STDs in Africa, 9-13 Decembe Ouagadougou, Burkina Faso Scott GB, Tuomala R. 1998. “Combination antiretroviral therapy during pregnancy.” AIDS 12(18):2495-2497. Sculpher MJ, Gibb D, Ades AE, Ratcliffe J, Duong T. 1998. “Modelling the costs of paediatric HIV infection and A mothers.” AIDS 12(11):1371-1380. acceptance and adherence to zidovudine treatment to p vertical transmission of HIV.” Paper presented at the 1996 International AIDS Conference 11(1):368(abstract no. Tu.C.2594). antiretroviral therapy in the ANRS 1203 cohort study, Abidjan Paper presented at the XIIth International Conference on AIDS and 98 STDs in Africa, 9-13 December 2001, Ouagadougou, Burkin (abstract no. 10DT3-1). a Faso rean section: medical benefits and costs.” Social Science and Medicine 37(10):1223-31. r the prevention of postnatal transmission of HIV in an urban South African setting.” S of the ilized in the Pediatric AIDS Clinical Trial Group 076 Study.” AIDS 12(14):1805-1813. are. up. and Economic Impacts of AIDS. Glastonbury, CT: The Futures Group. willingness to pay for HIV counseling and testing in Kenya and ce 1998. “HIV counseling and testing in Tanzania and Kenya is cost-effective: results from the voluntary Shaffer N, Chuachoowong R, Mock PA, et al. 1999. “Short-course zidovudine for perinatal HIV-1 transmission in Bangkok, Thailand: a randomised controlled trial.” Lancet l353(9155):781-785. Shearer EL. 1993. “Cesa “Simple Drug Therapy Can Cut AIDS Transmission by Mother to Newborn.” 1999. New York Times 2 February 1999. Soderlund N, Gray GE, Zwi KJ. 1998. “Breast or bottle? A cost- effectiveness evaluation of formula feeding fo Paper presented at the 1998 International AIDS Conference 12:465(abstract no. 316/24124). Sperling RS, Shapiro DE, McSherry GD, Britto P, Cunningham BE, Culnane M, Coombs RW, Scott G, Van Dyke RB, Shearer WT, Jimenez E, Diaz C, Harrison DD, Delfraissy J-F, for the Pediatric AID Clinical Trials Group Protocol 076 Study Group. 1998. “Safety maternal-infant zidovudine regimen ut Stewart MK, Stanton CK, Ahmed O. 1997. Maternal Health C Macro International: Calverton, MD. Stover J. 1997. DemProj: A Computer Program for Making Population Projections. Glastonbury, CT: The Futures Gro Stover J. 1997. AIM: A Computer Program for Making HIV/AIDS Projections and Examining the Social Sweat M, Balmer D, Sangiwa G. 1998. “Cost to clients and Tanzania: Results from the voluntary HIV counseling and testing study.” Paper presented at the 1998 International AIDS Conferen 12:467-468 (abstract no. 24135). Sweat MD, Denison JA. 1995. “Reducing HIV incidence in developing countries with structural and environmental interventions.” AIDS 9(suppl A):S251-S257. Sweat M, Sangiwa G, Balmer D. 99 HIV counseling and testing study.” Paper presented at the International AIDS Conference 12:648 (abstract no. 580/3 1998 3277). Sweat M, Grigorich S, Sangiwa G, Furlonge C, Balmer D, Kamenga - Temmerman M, Moses S, Kiragu D, Fusallah S, Wamola IA, Piot P. Thaineua V, Sirinirand P, Tanbanjong A, Lallemant M, Soucat A, of ublic Health 29(3):424-442. a y ion get best maternity care.” International Family Planning Perspectives atus and Trends of the Global HIV/AIDS Pandemic. Symposium Final Report. Vancouver, 5-6 July 1996.” f UNAIDS. 1998. “Prevention of HIV Transmission from Mother-to- t l EF. The State of the World’s Children, various issues. Oxford University Press: New York. ision. C, Grinstead O, Coates T. 2000. “Cost-effectiveness of voluntary HIV-1 counseling and testing in reducing sexual transmission of HIV 1 in Kenya and Tanzania.” Lancet 356(9224):113-121. 1990. “Impact of single session post-partum counselling of HIV infected women on their subsequent reproductive behaviour.” AIDS Care 2(3):247-252. Lamboray J-L. 1998. “From efficacy to effectiveness: routine prevention of mother-to-child HIV transmission by a short course zidovudine in Northern Thailand.” South East Asian Journal of Tropical Medicine and P Tuncer RA, Erkaya S, Sipahi T, Kutlar I. 1995. “Maternal mortality in maternity hospital in Turkey.” Acta Obstetrics and Gynecolog 74(8):604-606. Turner R. 1991. “Rural Nigerians with highest educat 17(4):155-156. UNAIDS. 1996. “The St Geneva: AIDSCAP/Family Health International, Harvard School o Public Health, UNAIDS. Child: Meeting on Planning for Programme Implementation (Meeting Statement, Geneva, 23-24 March 1998).” http://www.unaids.org/highband/document/statement/mtcstat.h ml. UNAIDS. 1998. “Mother-to-child transmission of HIV: UNAIDS Technical Update (UNAIDS Best Practice Collection: Technica Update).” Geneva: UNAIDS, October 1998, revision. UNIC United Nations. 1999. World Population Prospects: The 1998 Rev United Nations: New York. 0 100 Urrio TF. 1991. “Maternal deaths at Songea Regional Hospital, southern Tanzania.” East African Medical Journal 68(2):81-87. termination.” Tropical Geo Medicine 47(1):19-22. 5. “Couple communication, sexual coercion and HIV risk reduction in mith PG, Vaughan JP, et al. 1987. “Evidence for protection by breast-feeding against infant diseases in Brazil.” harbonneau TT, French PT, Wang L, Baum JB, Tesoriero JM, Savicki R. 1998. “Abbreviated f 14. Weinhardt LS, Carey M, Johnson BT, Bickham NL. 1999. “Effects of -analytic l of Whynes DK. 1997. “Cost-effectiveness and cost-benefit issues in the d Wilkinson D, Floyd K, Gilks CF. 1998. “Antiretroviral drugs as a public Wirakusumah FF. 1995. “Maternal and perinatal mortality/morbidity wan RJ, Higgins DL, Jorgensen CM. 1997. “The effects of HIV counseling and testing on risk-related practices and Van Roosmalen J, van der Does CD. 1995. “Caesarean birth rates worldwide. A search for de Van der Straten A, King R, Grinstead O, Serufilira A, Allen S. 199 Kigali, Rwanda.” AIDS 9(8):935-944. Victora CG, S Lancet 2(8554):319-322. Wade NA, Birkhead GS, Warren BL, C regimens of zidovudine prophylaxis and perinatal transmission o the human immunodeficiency virus.” New England Journal of Medicine 339(20):1409-14 Walker D. 1997. “Reducing perinatal transmission of HIV in Thailand.” Paper presented at the Workshop of the Ghent International Working Group on Mother-to-Child Transmission. Ghent, November 1997. HIV counseling and testing on sexual risk behavior: a meta review of published research, 1985-1997.” American Journa Public Health 89(9):1397-1405. prevention of mother-to-child transmission of HIV.” Paper presente at the Workshop of the Ghent International Working Group on Mother-to-Child Transmission. Ghent, November 1997, 1-26. health intervention for pregnant HIV-infected women in rural South Africa: an issue of cost-effectiveness and capacity.” AIDS 12(13):1675-1682. associated with cesarean section in Indonesia.” Journal of Obstetrics and Gynaecology 21(5):475-481. Wolitski RJ, MacGo 101 help-seeking behavior.” AIDS Education and Prevention 9(suppl B):52-67. World Bank. World Development Report, various issues. New York: tive use of short-course ZDV for prevention of mother-to- child transmission of HIV.” Weekly Epidemiological Record Oxford University Press. World Health Organization. 1998. “Recommendations on the safe and effec 73(41):313-320. 1022 IX. Glossary of Terms IX. Some of the following terms are definitions employed by the United Nations World Wide Web site: http://www.unaids.org/. These terms are defined in the context of their use within PMTCT. Benefit-cost ratio. The ratio determined by dividing the total treatment cost savings by the total cost of the intervention (minus the user fee, if any). % Breastfeeding spillover. This is the percentage of HIV- women who cease breastfeeding due to the impacts of the spillover effect (these impacts could be from a mass media campaign, an intervention, or counseling). C-Section. A planned Cesarean section. Interpolation. Given two numbers that serve as boundary points, the estimation of values that lie at intervals between the two points. For example, if the total fertility rate for a country or region was actually measured only in 1980 and in 1995, by assigning a relationship between the values from year to year, it is possible to estimate a TFR for each intervening year. (Spectrum uses a linear form of interpolation so that the difference between each annual value is the same. Other nonlinear forms of interpolation also are possible, but are not used in Spectrum.) Intra-uterine. In the uterus, during gestation; refers to HIV transmission. Mixed feeding. Infant feeding that uses breastfeeding and formula feedings. Model. A computer system designed to demonstrate the probable effect of two or more variables that might be brought to bear on an outcome. Such models can reduce the effort required to manipulate these factors and present the results in an accessible format. Module. A synonym for “model.” 103 Net cost. The net public sector cost of PMTCT services. This figure is equal to the total intervention costs minus the total treatment costs savings. Normalization. The transformation of a series of data points into a percent distribution summing to 100 percent. Perinatal and perinatal transmission. Pertaining to or occurring during the periods before, during, or shortly after the time of birth; that is, before delivery from the 28th week of gestation through to the first seven days after delivery. The transmission of HIV from an infected woman to her fetus or newborn child is referred to as perinatal transmission. Prevalence. The proportion of a defined population with the infection, disease, or other health-related event of interest at a given point or period of time. Pop-up menu. A menu (shown on the computer screen) from which users can select items or actions. Pop-up menus can appear anywhere on the screen. Pull-down menu. A menu (shown on the computer screen) opened by clicking on key words at the top edge of the screen. Pull-down menus allow users to select operations. Radio button. These buttons (shown on the computer screen) emulate raised buttons on early radios, which were punched to select radio stations. The graphically portrayed raised “radio buttons” on interfaces permit users to select among at least three alternatives. Revenue. The total amount of revenue collected from PMTCT user fees. User fee. The portion of the testing and counseling or ARV costs that may be shifted to the woman undergoing the process. Usually, only a certain percentage of the total costs can be paid by the user. Vertical transmission. HIV transmission from an infected women to her fetus or nursing infant. 104 X. Acronyms and Abbreviations X. AIDS acquired immunodeficiency syndrome AIM AIDS Impact Model ANC antenatal clinics ARV antiretroviral therapy CDC U.S. Centers for Disease Control and Prevention CMR child mortality rate DHS Demographic and Health Survey GNP gross national product HIV human immunodeficiency virus MMR maternal mortality rate PMTCT prevention of mother-to-child HIV transmission NGO nongovernmental organization NVP nevirapine UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNICEF United Nations Children's Fund USAID United States Agency for International Development VCT voluntary HIV counseling and testing WHO World Health Organization ZDV zidovudine 105 106 Appendix A: Calculation of Vertical Transmission Probabilities and Child Mortality Rate While Breastfeeding The calculation of mother-to-child transmission probabilities is based on six parameters: u = probability of infection in utero Cs = probability of infection during elective Cesarean section v = probability of infection during vaginal birth be = probability of infection by exclusive breastfeeding bf = probability of infection by mixed breastfeeding z = reduction in transmission due to treatment The last parameter, z, will vary by treatment option. This will be explained further below. Thus the six parameters can be utilized to illustrate the transmission risk for the following probability tree (Ratcliffe et al., 1998, p. 1383). 107 Transmission Probabilities Zidovudine (P1) No Zidovudine (1-P1) Cesarean (P2) Vaginal (1-P2) Cesarean (P2) Vaginal (1-P2) Breast (1-P3) Bottle (P3) Breast (1-P3) v + b (1 - v) Breast (1-P3) Bottle (P3) u + v (1 – u) = V Breast (1-P3) Bottle (P3) Bottle (P3) u/z + Cs(1 – u/z) /z = Csz Csz + b (1 – Csz) / z u/z + v (1 – u/z) / z = Vz Vz + b (1 – Vz) / z u + c (1 – u) = Cs Cs + b (1 – Cs) Each of the three possible transmission points (in utero, delivery, and breastfeeding) have their own transmission risk probability, denoted as P1, P2, and P3. This probability will be reduced by a certain proportion, depending on whether treatment is received and what type of treatment is received. The relationships among the first four parameters, u, Cs, v, and b, are the same as the relationships in Ratcliffe et al., 1998. The derivation of these initial values is described in detail in the appendix to that article; basically, the results of many of the clinical trials undertaken, in Europe for the most part, are utilized to form a number of simultaneous equations. These equations are then solved for the values: u = 0.0577; Cs = 0.0590; v = 0.127; and b = 0.169. These parameters are the probability of transmission of HIV from mother to child by the specific mechanism. For example, breastfeeding children of HIV+ mothers would have a 16.9 percent chance of becoming infected while breastfeeding. These parameters yield a cumulative probability of transmission through vaginal delivery and no breastfeeding of 18.5 percent. Most of the literature suggests that a rate of 25 percent or more is common. Therefore, u, Cs and v were each increased by about 40 percent, to arrive at reasonable total transmission probabilities. The final probabilities for the first three parameters in the model are: u = 0.0815; Cs = 0.0833; and v = 0.1794. For example, the final 108 transmission probabilities that result are 0.246 percent for vaginal births and bottle-feeding, and 0.374 percent for vaginal births and breastfeeding. All of the other rates are calculated as shown in the tree above, and can be seen in Table 15 in the main documentation. Finally, the reduction in transmission probabilities due to ZDV treatment is derived from the clinical trial results, as well as the methodology described in the appendix to Ratcliffe et al., 1998. The ACTG 076 trial showed that the transmission rate for untreated mothers was 25.5%, while the transmission rate for the long course of treatment was 8.3%, a reduction of a factor of 3.072. Ratcliffe et al. (1998) assumes an elective Cesarean section rate of 10%, and modifies the overall reduction factor from 3.072 to 3.157 accordingly. The short course treatment in Thailand showed a reduction in transmission rates from 18.6% to 9.2%, or a factor of 2.02; the same adjustment for elective Cesarean sections is made here, to arrive at an overall reduction of a factor of 2.076. Finally, the reduction from intrapartum/neonatal treatment was from 17.2% to 10.8% in UNAIDS (1999). After the appropriate calculations, assuming a 10 percent rate for elective Cesarean sections, the overall reduction for neonatal treatments is 1.634. For example, the transmission rate for a woman who has received the long-course ZDV treatment, delivers by Cesarean section, and feeds with a bottle, is calculated using the first treatment probability listed: u/z + Cs * (1 – u/z) / z or: 0.0815 + 0.0833 * (1 – 0.0815/3.157) / 3.157 = 0.052 Inference of Child Mortality Rate While Breastfeeding Although the overall child mortality rate for the population is readily available, this rate reflects mortality rates for both breastfed and non-breastfed infants. Since the set of infants in this model are split into two subsets, breastfed and formula-fed, it is important to have the relevant child mortality rates for each subset. These two rates are calculated in the model based on two pieces of data, the child mortality rate and the proportion breastfed, and one assumption, the proportionate increase in child mortality if the child is not breastfed. These three variables are supplied by the user. 109 The following equation is used to calculate the child mortality rates: U5MR = [bf * U5MRbf] + [(1 – bf) * U5MRbf * z] where: U5MR = Child mortality rate for overall population (children under age 5) bf = Proportion of population that is breastfed z = Proportionate increase in child mortality if child is not breastfed U5MRbf = Child mortality rate for infants who are breastfed Thus the overall child mortality rate is the weighted sum of (1) the proportion of infants being breastfed multiplied by the (unknown) mortality rate for breastfed infants, and (2) the proportion of infants not breastfed multiplied by the (unknown) mortality rate for breastfed infants and the (assumed) factor increasing child mortality if the child is not breastfed. The equation must then be solved for the term U5MRbf. Re-arranging terms: U5MR = [bf + (1 – bf) * z] * U5MRbf U5MR = U5MRbf bf + (1 – bf) * z The child mortality rate for infants who are not breastfed then becomes: (1 – U5MRbf) Note that the solution to the equation depends on the assumption made by the user for z, the proportionate increase in child mortality if the child is not breastfed. As discussed in the inputs section, this number can vary by country, depending on sanitary and other conditions. For a complete description of the possible values for this variable, and a discussion of the underlying issues, please see the main documentation. 110 Registration If you have not already registered your copy of Spectrum, please take a moment to complete this form and return it to us. This will ensure that you receive information about future updates to Spectrum. Name: Title: Institution: Address: City: State or District: Postal Code: Country: Telephone number: Fax Number: E-mail address: Do you have access to the internet? Spectrum Version Number: What type of computer are you using with Spectrum? How large is your hard drive? What kind of printer are you using? What language are you using with DemProj? English Spanish French Other How do you plan to use Spectrum? What additions to Spectrum would you like to see? Additional comments: Please return this form to: Registration Department The POLICY Project The Futures Group International Suite 1000 1050 17th Street NW Washington, DC 20036 USA Fax: (202) 775-9694 For more information, please contact: Director, The POLICY Project The Futures Group International 1050 17th Street NW, Suite 1000 Washington, DC 20036 Telephone: (202) 775-9680 Fax: (202) 775-9694 E-mail: policyinfo@tfgi.com Spectrum Table of Contents SectionPageI.INTRODUCTION1A.Description of the Spectrum System11.Components12.Software Description2B.Uses of Spectrum Policy Models2C.Organization of the Model Manuals3D.Information about the POLICY Project4E.What is the PMTCT Model?5F. Why Use the PMTCT List of Figures List of Tables % Public Sector Births Year of Fieldwork % Antenatal Visits Percentage of Women with One or More ANC Visits LATIN AMERICA/CARIBBEAN % Planned C-Section C-section % Breastfed for 6 Months % Mixed Feeding Increase in CMR Under-5 Mortality Rate MMR (per 100,000) Maternal Mortality Rate SUB-SAHARAN AFRICA % Offered HIV Test % Accepting Test % Receiving Results # Women Seen/Counselor % Breastfeeding Spillover Infections Averted/Woman Country % Eligible for Treatment Time of First Visit SUB-SAHARAN AFRICA LATIN AMERICA/CARIBBEAN % Offered Treatment % Accepting Treatment % Adhering to Treatment % Elective C-Section % HIV+ Breastfeed for 6 Months % HIV+ Mixed Feeding MMR Increase: C-Section Country Results Source Country Results Source Total VaginalCesarean HIV Test Pre-Test VCT, Post-Test VCT: HIV-, Post-Test VCT: User Fee: VCT Cost: C-Section Birth Monthly Formula Costs # Months Paid by Government User Fee: Tx Total Child Treatment Costs Total Adult Treatment Costs Type of Delivery Births (Public) Testing/Counseling Treatment Option Treatment Costs Transmission Reduction ARV Costs Transmission Probabilities - Results 1998-2010 1998 (US$) 2010 - Nevirapine 2010-Short-Course ZDV Type of delivery: Inference of Child Mortality Rate While Breastfeeding PMTCT Title 9-19-02.pdf January, 2002 Spectrum System of Policy Models Katharine Cooper-Arnold The Futures Group International

View the publication