AFFIDAVIT I, the undersigned NICOLI JEAN NATTRASS do hereby make oath and state as follows: 1. The facts deposed to in this affidavit are within my personal knowledge except where I indicate otherwise. To the extent that I rely on information supplied by others, I believe that such information is true and correct. 2. Between 1981 and 1991 I obtained the following degrees: 1981: B.A. (cum laude) Stellenbosch University; 1983: Honours Soc.Sci. (first class) University of Cape Town; 1984: M.A. (Social Science) University of Natal Durban; 1985: M.Sc. (Development Economics) University of Oxford; 1991: D.Phil. (Economics) University of Oxford. I was awarded the Rhodes Scholarship to Oxford (1984) and a Southern African Research Fellowship to Yale University (1993). (Annexure: NN1 – Curriculum Vitae). 3. I am currently a full professor in the School of Economics, Director of the Centre for Social Science Research and founder of the AIDS and Society Research Unit at the University of Cape Town. I have held academic positions at: the University of Natal Pietermaritzburg; the University Colleges of Galway and Dublin, and Stellenbosch University. I have done consulting work for the World Bank, the United Nations Development Programme, the International Labour Organisation and the Organisation for Economic Co-operation and Development. I served on the Technical Team of the Development Bank of Southern Africa Transformation Team, was a Commissioner on the South African Presidential Comprehensive Labour Market Commission, and have done work for the Taylor Committee into Comprehensive Welfare Reform in South Africa. 4. I have produced a substantial body of academic research in the areas of economic policy, development economics, labour economics and political economy. More recently, I have been working on the economics of mother to child transmission (MTCT) of HIV in South Africa. It is my considered opinion that a programme to reduce MTCT of HIV is not only cost-effective, but will save the government money (by reducing the number of HIV+ children who will need health care for opportunistic infections). There is, in other words, no basis for the argument that the government cannot afford a MTCT programme. The Key Conclusions: 5. My central conclusions are: ? HIV+ children require health care for opportunistic infections over their short lives. Reducing the number of HIV+ children via a MTCT reduction programme reduces these paediatric costs. These cost savings must be taken into account when analysing the net costs of a MTCT reduction programme. ? My analysis shows that the total cost to the health sector of MTCT programmes (i.e. the costs of voluntary counselling and testing, the costs of the anti-retroviral regimen and the costs of treating all children born HIV+ despite the MTCT programme) is less than the costs of treating all children born HIV+ in the absence of a MTCT programme. This is true for all four of the MTCT programmes discussed here. ? In other words, saving children from HIV infection by implementing a MTCT programme will save the state money because the costs of a MTCT programme are less than the costs associated with treating the additional children who would be born HIV+ if no MTCT programme was in place. It is therefore not tenable to argue that a MTCT reduction programme is too costly. ? I estimate that the savings to the health sector per pregnancy as a result of a MTCT programme are as follows: o R171 (AZT Thai regimen and breast-feeding) o R197 (Nevirapine and breast-feeding) o R315 (AZT Thai regimen and substitute feeding) o R341 (Nevirapine and substitute feeding) 6. Most research into the cost-effectiveness of MTCT programmes does not consider potential cost savings (as in my research). This literature looks at cost- effectiveness measured in terms of a standard measure called the disability adjusted life year (DALY). The implication of this literature is that irrespective of whether MTCT reduction saves the state money, it is nevertheless a cost- effective intervention. (Annexure: NN2 – Geffen, 2001 ). Research Into Cost-Effectiveness Of MTCT Reduction 7. There is a wealth of international scientific evidence that treating HIV+ pregnant women with antiretroviral drugs significantly reduces MTCT of HIV. In conducting my analysis and formulating my opinions as an economist, I have relied on this evidence, some of which I summarise here. 8. In situations where resources are constrained, and where adherence to long and complicated drug regimens cannot be managed effectively, short course interventions are recommended. These include: ? the short-course AZT 'Thai' regimen (300 mg of Zidovudine every 12 hours from 36 weeks into the pregnancy and 300 mg every 3 hours during labour); and ? the HIVNET012 Nevirapine regimen (200 mg of Nevirapine during labour, and 2mg per kg for the baby). In breastfeeding populations, research indicates that MTCT is reduced by 37% with the AZT short course, and by 35% by the Nevirapine regime. 9. Where substitute feeding is used rather than breast-feeding, MTCT is reduced in the AZT regime by 50%. Indications are that substitute feeding combined with a Nevirapine regime reduces MTCT by 44%. However the relative advantages of substitute feeding over breast-feeding for reducing MTCT in developing countries have yet to be established conclusively. Indications are that an exclusive breast-feeding regime followed by abrupt weaning may be more effective than the mixed feeding regimes typically followed in breast- feeding populations. And, given that substitute feeding is associated with higher infant mortality, the life-saving properties of formula-feeding will be reduced accordingly. After reviewing the available evidence, the WHO Technical Consultation team recommended that where substitute feeding is feasible, affordable, sustainable and safe, then breast-feeding should be avoided altogether. Otherwise, exclusive breastfeeding is recommended followed by abrupt weaning. The decision should be based on counselling the woman so that she can make an informed choice. 10. In this affidavit, I will consider the argument that the state cannot afford a MTCT reduction programme. I conclude that not only does it cost very little to save babies from HIV infection, but that unless the state denies HIV+ children health care, it almost certainly costs the government more to care for HIV+ children over their short lives than it does to save them from HIV infection. Put simply, the state cannot afford not to introduce a MTCT reduction programme. The methodology used follows that in Skordis and Nattrass (2001). Table 1: Summary With breast feeding With substitute feeding Nevirapine (HIVNET012) AZT (Thai regimen) Nevirapine (HIVNET012) AZT (Thai regimen) Number of children saved as a result of a MTCT programme for 1000 pregnant women 21 23 39 41 Cost of the MTCT programme per child saved R3,824 R5,831 R4,723 R5,802 Cost savings for the government of a MTCT programme for 1000 pregnant women R197,388 R171,349 R340,986 R314,837 11. Table 1 summarises the results. It shows that a programme to reduce MTCT by a single dose of Nevirapine is cheaper than a short-course AZT programme, but saves marginally fewer lives. Using substitute feeding rather than breast feeding saves more lives than breast-feeding regimes. Although the programmes with substitute feeding cost more per life saved than those using breast milk, the government saves more money by implementing a programme with substitute feeding. This is as a result of the lower incidence of HIV infection – and hence lower associated paediatric costs of HIV+ children – under substitute feeding regimes. Given that the Nevirapine interventions save the most money and are cost-effective and easy to administer, this suggests that the government should opt for a Nevirapine-based intervention to reduce MTCT in South Africa. 12. The data used in the calculations are 'best estimates' from the available local research and international literature. In order to test for the robustness of the finding, I subjected the calculation to a sensitivity analysis that cut the paediatric costs of HIV+ children by 25%, and simultaneously inflated all programme costs by 25%. The results (indicated in the last column of each table) remained robust: the government would still save money by introducing a programme to reduce MTCT. THE AFFORDABILITY OF MTCT REDUCTION PROGRAMMES 13. The affordability analysis is presented in terms of the costs associated with 1,000 pregnant women attending ante-natal clinics in South Africa. The first 6 lines of Table 2 indicate that in the absence of a programme to reduce MTCT, 74 babies will be born HIV+. These children will require medical attention to treat the opportunistic infections that will beset them over their short lives. Line 5 provides an estimate of the paediatric costs associated with each HIV+ child. Line 6 provides an estimate of the total health costs associated with all HIV+ children born in the absence of a programme to reduce MTCT. Table 2: For every 1000 pregnant women visiting antenatal clinics: Best estimate Cost sensitivity analysis 1. Percentage who will be HIV+ (from SA Antenatal survey 2000) 0.245 0.245 2. Number of HIV+ women (line 1 x 1000) 245.00 245.00 3, Percentage who will transmit HIV to their babies 0.30 0.30 4. Number of HIV+ babies (line 3 x line 2) 73.50 73.50 5. Hospital costs per HIV+ child (2001 prices) 13,342.70 10,007.03 ? 25% 6. Total inpatient costs for HIV+ children (line 5 x line 4) 980,688.45 735,516.34 7. Cost of pre-test counselling per woman 18.20 22.75 ? 25% 8. Pre-test counselling for 1000 women (line 7 x 1000) 18,200.00 22,750.00 9. 91.5% of the women will agree to a test = 915 915.00 915.00 10. Cost of the Rapid test 16.80 21.00 ? 25% 11. Cost of testing all those who accept the test (line 9 x line 10) 15,372.00 19,215.00 12. The number of HIV+ cases that will result (line 9 x line 1) 224.18 224.18 13. Cost of the confirmatory testing procedure 7.70 9.63 ? 25% 14. Total cost of all confirmatory tests (line 13 x line 12) 1,726.15 2,157.68 15. Cost of post-test counselling for each HIV- woman 3.70 4.63 ? 25% 16. Post-test counselling costs for all HIV- women (line 15 x 690.82) 2,556.05 3,195.07 17. Cost of post-test counselling for HIV+ (as in line 7) 18.20 22.75 ? 25% 18. Post-test counselling costs for all HIV+ women (line 17 x line 12) 4,079.99 5,099.98 19. Site costs (management, phones, transport etc) per pregnancy 33.00 41.25 ? 25% 20. Total site costs (line 19 x 1000) 33,000.00 41,250.00 21. Total voluntary counselling and testing (VTC) costs (line 8 + line 11 + line 14 + line 16 + line 18 + line 20) 74,934.19 93,667.73 22. Percentage of women who will accept ARV therapy 0.925 0.925 23. Number of participants in the programme (line 1 x line 12) 207.36 207.36 14. Lines 7 to 21 provide a costing exercise for a voluntary counseling and HIV testing programme for 1,000 women. This includes provision for management and administration, and of the costs of tests (Rapid tests and confirmatory tests) and of pre- and post-test counseling. It is assumed that after pre-test counseling, 91.5% will agree to an HIV test (line 9). Of those who test positive, it is assumed that 92,5% (i.e. 207 women) will agree to a short course of antiretroviral therapy (ARV) in order to prevent MTCT. 15. In this affidavit, I will present a costing exercise for four regimens: short- course AZT (Thai regimen) and short-course Nevirapine (HIVNET012) under a breastfeeding regimen; and short-course AZT and short-course Nevirapine under a substitute feeding regimen. Net Costs of a Short-Course AZT MTCT Reduction Programme 16. As can be seen from Table 3, 23 children (out of 1,000 born) would be saved from HIV infection by a MTCT reduction programme using a short course of AZT (the Thai regimen) with breastfeeding. According to the best estimate, this would cost the government R5,831 (in drugs, counselling, testing etc) per child saved (and R7,288) per child saved if all costs were 25% higher than expected. Notice that the government actually saves money by saving these children. This is because the total health costs under a MTCT programme (i.e. cost of the MTCT reduction programme and the costs of caring for all HIV+ children born) are about R171,000 less than would be the case in the absence of a MTCT reduction programme (i.e. caring for all the HIV+ children born if no programme to reduce MTCT was in place). Notice that the government still saves money if we assume that the costs of caring for HIV+ children are reduced by 25%, and if the costs associated with the MTCT programme are simultaneously 25 percent higher than expected. Table 3: AZT (Thai regime) with breastfeeding Best Estimate Cost Sensitivity Analysis 24. Cost of the AZT ARV Regimen for each woman 280.00 350.00 ? 25% 25. Total cost of AZT therapy (line 24 x line 23) 58,061.33 72,576.66 26. Transmission under an AZT ARV regimen 0.19 0.19 27. Number of HIV+ children despite the ARV programme (line 26 x line 23) 39.40 39.40 28. Inpatient costs of children born HIV+ despite ARV therapy (line 5 x line 27) 525,685.79 394,264.34 29. Number of HIV+ children born to non-participants 11.29 11.29 30. Inpatient costs of the HIV+ children born to the non- participants (line 5 x line 29) 150,658.26 112,993.70 31. Total health costs under Thai regime (line 21 + line 25 + line 28 + line 30) 809,339.56 673,502.42 32. Number of children saved (line 4 – line 27 – line 29) 22.81 22.81 33. Total health cost savings (line 6 – line 31) 171,348.89 62,013.91 34. Cost of VCT + ARV per child saved (line 21 + line 24) / line 32 5,830.63 7,288.29 Net Costs of a Short-Course Nevirapine Regimen 17. Table 4 provides a costing exercise for a Nevirapine regime (with breast-feeding). The table shows that 21 children (out of 1,000 born) would be saved from HIV infection by a MTCT reduction programme using Nevirapine. According to the best estimate, this would cost the government about R3,824 (in drugs, counselling, testing etc) per child saved (and about R4,780) per child saved if all costs were 25% higher than expected. Notice that the government saves more money by introducing this MTCT reduction programme than it would with the AZT Thai regime. Total health costs under a MTCT programme (i.e. cost of the MTCT reduction programme and the costs of caring for all HIV+ children born) are about R197,000 less than would be the case in the absence of a MTCT reduction programme. The cost-savings result is highly robust even if we assume that the costs of caring for HIV+ children are reduced by 25%, and that the costs of the MTCT programme are 25% higher than expected. Table 4: Nevirapine (HIVNET) with Breastfeeding Best Estimate Cost Sensitivity Analysis 35. Cost of the Nevirapine ARV programme R21.00 26.25 ? 25% 36. Total cost of Nevirapine therapy (line 23 x line 35) 4,354.60 5,443.25 37. Transmision rate under a Nevirapine ARV regimen 0.20 0.20 38. Number of HIV+ children born under a Nevirapine ARV regimen (line 37 x line 23) 41.47 41.47 39. Inpatient costs of these HIV+ children (line 5 x line 38) 553,353.46 415,015.09 40. Inpatient costs of children born HIV+ to the non- participants (line 30) 150,658.26 112,993.70 41. Total health costs under a Nevirapine regimen (line 21 + line 36 + line 39 + line 40) 783,300.51 627,119.77 42. Number of children saved (line 4 – line 38 – line 29) 20.74 20.74 43. Total health cost savings (line 6 – line 41) 197,387.94 108,396.57 44. Cost of VCT + ARV per child saved (line 21 + line 36) / line 42 3,823.69 4,779.61 18. In the cost analysis performed here, it is assumed that the government purchases Nevirapine at the current state-tender price. However, Boehringer Ingleheim, the manufacturers of Nevirapine, have offered to donate Nevirapine to the government. Net Costs of a Short-Course AZT Regimen with Substitute Feeding 19. Table 5 provides a costing exercise for the Thai AZT short-course regimen with six months of substitute feeding for the infant (rather than breast-feeding). 20. One of the disadvantages of substitute feeding is that the risk of (non-AIDS related) infant mortality is higher. The calculation has been adjusted to include an estimate for the increased risk of infant mortality as a result of the substitute feeding regime. Table 5: AZT (Thai regime) with Substitute Feeding (SF) for 6 months Best Estimate Cost Sensitivity Analysis 45. Total cost of AZT therapy (line 25) 58,061.33 72,576.66 46. Cost of the substitute feeding for 6 months 499.20 624.00 ? 25% 47. Total cost of the substitute feeding (line 23 x line 46) 103,515.05 129,393.81 48. Transmission rate under an AZT and substitute feeding regimen 0.10 0.10 49. Number of HIV+ children born despite the treatment regimen (line 23 x line 48) 20.74 20.74 50. Inpatient costs of these HIV+ children (line 5 x line 49) 276,676.73 207,507.55 51. Inpatient costs of children born HIV+ to the non- participants (line 30) 150,658.26 112,993.70 52. Assume an increase in infant mortality of 4 in a 1000 due to substitute feeding 0.004 0.004 53. Number of children who die as a result of the substitute feeding (line 23 x line 52) 0.83 0.83 54. The number of additional deaths that will be amongst HIV+ children (line 53 x line 48) 0.08 0.08 55. The number of additional deaths that will be amongst HIV- children (line 53 x (1 – line 48)) 0.75 0.75 56. Assume these children die after 3 months. Savings on substitute feeding (line 53 x (line 46)/2 207.03 258.79 57. Assume medical costs of these children dying early is 30 percent of HIV+ children (line 5 x line 53 x 0.3) 3,320.12 2,490.09 58. Inpatient costs of HIV+ children adjusted for these early deaths (line 50 – (line 54 x line 5)) 275,570.02 206,677.52 59. Total health costs under Thai regime with substitute feeding (line 21 + line 45 + line 47 + line 51 + line 57 + line 58 – line 56) 665,851.93 617,540.71 60. Number of children saved (line 4 – line 49 – line 29 – line 55) 40.73 40.73 61. Cost savings (line 6 – line 59) 314,836.52 117,975.62 62. Cost of VCT+ARV+SF per child saved (line 21 + line 45 + line 47 – line 56) / line 60 5,802.30 7,252.87 21. The results indicate that more children could be saved by providing women with substitute feeding so as to reduce the risks of MTCT through breast-milk (and mixed feeding). The costs per child saved are marginally higher than they are for the AZT regime with breastfeeding. However, as a result of the greater number of children saved, the government saves more money by introducing a MTCT programme with substitute feeding than it would by using a AZT or a Nevirapine regime with breastfeeding. Net Costs of a Single Dose Nevirapine Regimen with Substitute Feeding 22. Table 6 provides a costing exercise for the Nevirapine intervention with six months of substitute feeding (rather than breastfeeding). The calculation is based on research findings from Kenya that indicate that breastfeeding increases transmission by 44 percent. 23. Table 6 shows that a MTCT reduction programme using Nevirapine and substitute feeding has the greatest potential to save lives – and saves the government the most money. The amount saved by not having to treat as many HIV+ children as would have been the case in the absence of a MTCT reduction programme exceeds the costs of implementing a MTCT reduction programme by R341,000 (and by R164,000 if programme costs are 25% higher than expected, and if the costs of treating HIV+ children is reduced by 25%). This shows that there is no basis to the government's claim that it cannot afford a MTCT reduction programme. Unless the government is planning to deny HIV+ children health care, it costs the government more in terms of health costs to treat HIV+ children than it would to save many of them via a MTCT reduction programme. Table 6: Nevirapine Regimen with Substitute Feeding (SF) for 6 months Best Estimate Cost Sensitivity Analysis 63. Total cost of Nevirapine therapy (line 36) 4,354.60 5,443.25 64. Cost of the substitute feeding for 6 months (line 46) 499.20 624.00 ? 25% 65. Total cost of the substitute feeding (line 47) 103,515.05 129,393.81 66. Transmission rate under a Nevirapine and SF regimen 0.11 0.11 67. Number of HIV+ children born despite the treatment regimen (line 23 x line 66) 22.81 22.81 68. Inpatient costs of these HIV+ children (line 5 x line 67) 304,344.40 228,258.30 69. Inpatient costs of children born HIV+ to the non- participants (line 30) 150,658.26 112,993.70 70. Assume an increase in infant mortality of 4 in a 1000 due to substitute feeding (line 52) 0.004 0.004 71. Number of children who die as a result of the substitute feeding (line 23 x line 70) 0.83 0.83 72. The number of additional deaths that will be amongst HIV+ children (line 66 x line 71) 0.09 0.09 73. The number of additional deaths that will be amongst HIV- children (line 71 x (1 – line 66)) 0.74 0.74 74. Assume these children die after 3 months. Savings on substitute feeding (line 71 x (line 46)/2) 207.03 258.79 75. Assume medical costs of these children dying early is 30 percent of HIV+ children (line 5 x line 71 x 0.3) 3,320.12 2,490.09 76. Inpatient costs of HIV+ children adjusted for these early deaths (line 68 – (line 5 x line 72) 303,127.02 227,345.27 77. Total health costs under Nevirapine regime with substitute feeding (line 21 + line 63 + line 65 + line 68 – line 74 + line 76 + line 75) 639,702.21 571,075.06 78. Number of children saved (line 4 – line 67 – line 29 – line 73) 38.66 38.66 79. Cost savings (line 6 – line 77) 340,986.24 164,441.28 80. Cost of VCT+ARV+SF per child saved (line 21 + line 63 + line 65 – line 74) / line 78 4,723.08 5,903.85 The Cost of Orphans 24. There are, of course, various objections to this kind of costing exercise. One is that it does not take into account the costs of orphans. The problem of orphans is certainly going to increase significantly as deaths from the AIDS pandemic escalate. But this does not constitute an economic argument in favour 'orphan elimination', even when human rights arguments are put aside for the purpose of considering financial implications. 25. First, the vast majority of orphaned children are cared for by their extended families – and not by state institutions. Increasing the number of HIV+ children simply places extra burdens on these families. 26. But increasing the number of HIV+ children also increases the level of unproductive expenditure by the state. The most common form of state support for these children is the child support grant which, in the case of a child with a normal life expectancy, can be regarded as a form of investment in human capital. However, when spent on an HIV+ child who is likely to die before its fifth birthday, the child grant can only be regarded as an unrecoverable form of consumption spending. If we include an estimate of this 'unrecoverable' form of welfare expenditure into the estimate of the costs of HIV+ children, then the savings to the government of a Nevirapine MTCT programme (with substitute feeding) rise by over 50%. 27. Secondly, emerging anecdotal evidence suggests that orphans are more likely to be abandoned to state institutions by their extended families if they are HIV+ than if they are HIV–. Taken together, this suggests that the government could end up spending more on institutional child welfare than would be the case if a programme to reduce MTCT was in place. Other Research 28. Other researchers have also examined the cost-effectiveness of implementing an MTCT reduction programme. They have not given the same prominence to the cost-saving argument presented here. Instead, they have shown that MTCT reduction is cost-effective using a metric known as the Disability Adjusted Life Year (DALY). According to the World Bank, health interventions in developing countries that cost less than $100 per DALY saved, are worth considering. 29. There is a consensus in the South African academic research into MTCT reduction that such programmes are affordable and cost-effective. Wilkinson, Floyd and Gilks (1999 and 2000) , Marseille et al. (1998 and 1999) , Soderlund, Zwi, Kinghorn and Gray (1999) and Hensher (2000) have all examined the cost-effectiveness of MTCT reduction. Annexure NN2 includes an analysis of these studies. DEPONENT I CERTIFY THAT THE DEPONENT HAS ACKNOWLEDGED THAT SHE KNOWS AND UNDERSTANDS THE CONTENTS OF THIS AFFIDAVIT WHICH WAS SIGNED AND SWORN TO BEFORE ME AT CAPE TOWN ON THIS 15 DAY OF AUGUST 2001 AND THAT SHE HAS NO OBJECTION TO TAKING THE PRESCRIBED OATH AND CONSIDERS SAME BINDING ON HER CONSCIENCE. COMMISSIONER OF OATHS Geffen, N. 2001, Cost and Cost-Effectiveness of Mother-to-Child Transmission Prevention of HIV (TAC Briefing Paper), www.tac.org.za/mtctcost.rtf. See overview by Farley, T., Buyse, D., Gaillard, P. and J. Perriens. 2000. "Efficacy of Antiretroviral Regimens for Prevention of Mother to Child Transmission of HIV and Some Programmatic Issues", Background paper prepared for Technical Consultation on New Data on the Prevention of Mother to Child Transmission of HIV and their Policy Implications, Geneva, 11-13 October 2000. Available on http://www.who.int/reproductive-health. See Farley et al, 2000. See Farley et al, 2000, Table 4: Summary of different antiretroviral regimens, and Wood, R. 2001. Affidavit for the Treatment Action Campaign, p.11. See discussion in Wood (2001: 11-12). Marseille,E. and J. Kahn. 1999. Evaluating Antiretroviral Drug and Substitute Feeding Interventions to Prevent Mother-to-Child Transmission of HIV, Excel spreadsheet Version 1.0, Field Test Draft Version, December 1999. Available on www.unaids.org. See "New Data on the Prevention of Mother to Child Transmission of HIV and their Policy Implications", WHO Technical Consultation on Behalf of the UNFPA/UNICEF/WHO/UNAID Inter- agency Task Team on Mother-to-Child Transmission of HIV, Geneva., October 2000. Skordis, J. and N. Nattrass. 2001. What is Affordable: The Political Economy of Policy on the Transmission of HIV/AIDS from Mother to Child in South Africa. Paper presented to the AIDS in Context Conference, University of the Witwatersrand, April 2001. Hospital costs per HIV+ child assumes that the pediatric costs of a child with AIDS is equal to 10.8 days in hospital in a high-cost hospital bed. No additional costing is included for medicines. This proxy for the pediatric costs of HIV/AIDS is 15 percent lower than the average data for pediatric costs reported in Tanzania, Zaire and Thailand (reported in Marseille, E. and J. Kahn. 1999. Manual for Use of a Cost-Effectiveness Tool for Evaluating Antiretroviral Drug and Substitute Feeding Interventions to Prevent Mother-to-Child Transmission of HIV, Field Test Draft Version, December 1999. Available on www.unaids.org.) These costs are over the life of a child. However, if we assume that the children live for a similar period and that the same number of children are born HIV+ each year, then the costs over the life of a child will be equal to the annual costs of HIV+ children. As reported in the West African pooled analysis, reported in Farley et al, 2000. Information obtained from the Western Cape HIV/AIDS Directorate. This is based on a salary of R2000 per month per counselor. Each counselor has 5 appointments a day for 22 days a month. According to the Directorate the Western Cape the national government has proposed a R500 per month stipend per lay- counselor. Using this number would substantially lower the cost of preventing MTCT. Information from the Western Cape HIV/AIDS Directorate and an adjustment for the cost of the nurses time taken to administer the test. This includes the cost of the Smartcheck confirmatory test and an additional Elisa test (and needles and tubes) for the 5% of indeterminate test results. Information from the Western Cape HIV/AIDS Directorate Based on counseling costs from the Western Cape HIV/AIDS Directorate (assuming 5 in a group). This includes the cost of stationary, phones, photocopies, transport, mentoring for nurses and the cost of a project manager for each site, assuming 5000 pregnancies per year per site (information from the Western Cape HIV/AIDS Directorate). Data has been obtained from the Western Cape HIV/AIDS Directorate and published sources (see footnotes to the items in the table). HIV Management Services. 1998. Projections of Costs of Anti-Retroviral Interventions to Reduce Mother to Child Transmission of HIV in the South African Public Sector, Technical Report to GlaxoWellcome April. HIV Management Services, 1998. WCDH (Western Cape Department of Health). 1999. Programme for the Prevention of Mother to Child Transmission of HIV in Khayelitsha, Western Cape, Presentation to the Provincial Health Restructuring Committee, 22 January. NB: This price has almost certainly fallen. Based on 37% effectiveness reported in the West African pooled analysis cited by Farley et al (2000). Information obtained from the Western Cape HIV/AIDS Directorate, 2001. Based on the 35% efficacy rate after 12 months breast-feeding (reported in Farley et al (2000)). This is based on a tin of formula milk (Pelargon costing R10.40, and a child needing 8 tins a month (information supplied by the Western Cape HIV/AIDS Directorate). The estimate uses data provided Marseille, E. and J. Kahn. 1999b. Evaluating Antiretroviral Drug and Substitute Feeding Interventions to Prevent Mother-to-Child Transmission of HIV, Excel spreadsheet Version 1.0, Field Test Draft Version, December 1999. Available on www.unaids.org.) Based on results surveyed in Farley et al (2000). Data in Marseille and Kahn, 1999b. Note that the hospital costs of HIV- negative children who die as a result of the substitute feeding are included as these lives are as wasted as they would have been if the child had died of AIDS. Reported in Wood (2001). Based on the evidence from a Kenyan study that breastfeeding increases transmission by 44 percent (reported in Wood (2001). Homedes, N., 2000, The Disability Adjusted Life Year (DALY) Definition Measurement and Potential Use., World Bank: Human Capital Development and Operations Policy (working paper), http://www.worldbank.org/html/extdr/hnp/hddflash/workp/wp_00068.html (see the addendum referred to on the paper as well). World Bank., 1993, World Development Report 1993 – Investing in Health., New York, Oxford University Press. See Nattrass and Skordis (2001) for a discussion of the limitations of the DALY measure, and of the limitations of the $100 per DALY rule of thumb. Wilkinson, D., Floyd, K., Gilks, C. F., 1999, A National Programme to Reduce Mother-To-Child HIV Transmission is Potentially Cost Saving: Evidence from South Africa., Medical Research Council and Wilkinson, D., Floyd, K., Gilks, C. F., 2000, National and Provincial Estimated Costs and Cost Effectiveness of a Programme to Reduce Mother-To-Child HIV Transmission in South Africa., SAMJ 90 (8), pp. 794-797. Marseille, E., Kahn, J. G, Saba, J., 1998., Cost-Effectiveness of Antiviral Drug Therapy to Reduce Mother-To-Child HIV Transmission in Sub-Saharan Africa. AIDS 1998, 12, pp. 939-948 and Marseille, E., Kahn, J. G., Mmiro, F., Guay, L., Musoke, P., Fowler, M. G., Jackson, J. B., 1999, Cost Effectiveness of Single Dose NVP for Mothers and Babies to Decrease Vertical HIV-1 Transmission in Sub-Saharan Africa., The Lancet 354 (9181), pp. 803-809. Soderlund, N., Zwi, K., Kinghorn, A., Gray, G., 1999, Prevention of Vertical Transmission of HIV: Analysis of Cost Effectiveness of Options Available in South Africa., British Medical Journal 318, pp. 1650- 1656 (19 June). Hensher, M., 2000, Confidential Briefing: The costs and effectiveness of using NVP or AZT for the prevention of mother-to-child transmission of HIV – current best estimates for South Africa., Health Financing & Economics. 12