Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized of Assessing the Financial SustainabilitY 74547 Jamaica’s HIV Program  Assessing the Financial SustainabilitY of Jamaica’s HIV Program World Bank co-authors: Markus Haacker, Elizabeth Mziray and Shiyan Chao © 2013 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org.  Table of Contents Acknowledgments v Acronyms and Abbreviations vii Executive Summary 1 HIV/AIDS in Jamaica: the State of the Epidemic and of the National Response 2 Projecting the Cost of the National Response to HIV/AIDS 3 HIV Incidence and the Cost of the National Response 3 Sustainability of the National Response to HIV/AIDS 5 I. Introduction 7 II. The State of HIV/AIDS and of the HIV/AIDS Response 9 The State of the Epidemic 9 Costs and the Financing of the National Response to HIV/AIDS 11 III. Projecting the Costs of the National Response to HIV/AIDS 17 Epidemiological Projections 17 Assumptions on Unit Costs and Coverage Rates 20 Projected Costs of HIV/AIDS Program 20 Some Policy Lessons 23 IV. HIV Incidence and the Costs of the HIV/AIDS Response 25 The Costs Incurred by each Additional HIV Infection 25 Assessing the Cost-Effectiveness of HIV Prevention Interventions 26 Prevention Policies and the Drivers of HIV Incidence 28 V. Fiscal Evaluation of the Costs of the HIV/AIDS Program 29 VI. Sustainability of the National Response to HIV/AIDS 33 Implications of Reduced Availability of External Financing 33 Funding Treatment Costs 36 VII. Conclusions 37 iii Assessing the Financial Sustainability of Jamaica’s HIV Program References 39 Appendix on Methodology 41 Figures ES1. Mortality by Cause 2 ES2. Composition of Spending Under NationalResponse to HIV/AIDS, 2010/11 2 ES3. Costs of National Response to HIV/AIDS2010–2030 3 ES4. Jamaica: Costs of One Additional Infection 4 ES5. Jamaica: Fiscal Costs of HIV/AIDS, “Commitment Basis”, 2010–2030 4 ES6. Fiscal Costs of HIV/AIDS, Current and Commitments, Across Scenarios 5 1. The Course of the HIV Epidemic in Jamaica, 1980–2010 10 2. Mortality by Cause, 2008 11 3. Mortality by Age and Sex, 2005–10 12 4. Composition of Spending Under National Response to HIV/AIDS, 2010/11 15 5. People Living with HIV/AIDS, 2010–2030 18 6. State of Epidemic Across 3 Scenarios, 2010–2030 19 7. Cost of National Response to HIV/AIDS, 3 Scenarios, 2010–2030 21 8. Costs of National Response to HIV/AIDS, 2010–2030, Across Scenarios 22 9. Costs of One Additional Infection, 2012 26 10. Fiscal Costs of HIV/AIDS, Commitment Basis, 2010–2030 30 11. Fiscal Costs of HIV/AIDS, Current and Commitments, Across Scenarios 31 12. Composition of Financing of National Response to HIV/AIDS, 2010 34 13. Role of External HIV/AIDS Financing, 2009 34 14. Domestic Financing Need and Rate of External Support 35 Tables 1. Summary Health Indicators and Impact of HIV/AIDS 11 2. Costs and Financing of National Response to HIV/AIDS, 2009/10–2010/11 13 3. Costs Incurred by HIV Infections, by Population Group 28 A1. Costs of Jamaica’s National Response to HIV/AIDS 43 Boxes 1. Macroeconomic and Fiscal Context of National Response to HIV/AIDS 13 2. Health Spending and Access to Health Services 14 3. Macroeconomic Impact of HIV/AIDS 22 4. Estimating the Financial Costs of New Infections 27 iv Acknowledgments Acknowledgments This study was undertaken at the request of the Government of Jamaica to inform its future policies. Led and financed by the World Bank, it was conducted in collaboration with the Government of Jamaica and the Joint United Nations Programme on HIV/AIDS (UNAIDS). The task team was led by Elizabeth Mziray and included Markus Haacker, Shiyan Chao, and Mario Mendez. Markus Haacker was the lead author for the report, and Mario Mendez provided administrative support. The study was overseen by a local study steering committee chaired by the Planning Institute of Jamaica and included representatives from the Ministry of Health, Ministry of Finance, UNAIDS and the World Bank. The team would like to thank the Honorable Dr. Fenton Douglas, Minister of Health, Dr. Jean Dix- on, Dr. Kevin Harvey, Dr. Nicola Skyers, and Ms. Barbara Scott from the Government of Jamaica, and Dr. Ernest Massiah and Dr. Pierre Somse from UNAIDS, for their guidance and support. The study has benefitted from input and suggestion from numerous government officials, including Dr. Sharlene Jarrett, Suzanne Watson, Terry Ann Frith, Sannia Sutherland, Edgar Watson, and Andrew Brown from the National HIV/STI Programme; Michael Maragh, Sandra Graham, and Evelyn Bloomfield from Ministry of Health; Walter James, Andrea Sheppard, Saskia Frater-Smith, Easton Williams, and Hugh Morris from Planning Institute of Jamaica; Othneil Hemans, Barbara Hew, and Carlene Momsas from Ministry of Finance, and Dunstan E. Bryan from Ministry of Labor and Social Security. Additionally, the study has also benefitted from contributions by the World Bank Prevention Science and Mathematical Modeling Reference Group, Dr. Timothy Hallett (Imperial College), Dr. Laith Abu-Raddad (Cornell University), Erva-Jean Stevens (UNAIDS), Cary Reid and Katherine Guerra (Clinton Health Access Initiative), Andrew Fearon, Dr. Geoffrey Barrow; as well as a discussion with Dr. Peter Figueroa; and com- ments received at the presentation of the draft report to government officials and development partners in April 2012. The team would like to thank the peer reviewers from the World Bank, Christine Pena, Aparnaa Somanathan, and Ajay Tandon; and acknowledge the support and guidance received from Ni- cole Klingen, David Wilson, Keith Hansen, Joana Godinho, Leslie Elder, and Jody Kusek. Finally, the team thanks Gregory Wlosinski, Printing and Multimedia, General Services for graphic design of the report. v Acronyms and Abbreviations Acronyms and Abbreviations AIDS Acquired immunodeficiency syndrome CDA Child Development Agency FY Financial year GDP Gross domestic product Global Fund Global Fund to Fight AIDS, Tuberculosis and Malaria HIV Human immunodeficiency virus MoH Ministry of Health MOT Modes of transmission NASA National AIDS Spending Assessment NGO Non-governmental organization NHF National Health Fund NHP National HIV/STI Programme NSP National strategic plan PIOJ Planning Institute of Jamaica PMTCT Prevention of mother-to-child transmission STI Sexually transmitted infection UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNGASS United Nations General Assembly Special Session (on HIV/AIDS) vii Executive Summary Executive Summary Jamaica has made many notable achievements in the fight against HIV/AIDS, which include a robust treatment program and improved HIV prevention programs that increasingly focus on the key drivers of the HIV epidemic and which are based on evidence. These attainments have resulted in a sustained decline in the estimated incidence of HIV and in a reduction in AIDS mortality. The national response to HIV/AIDS in Jamaica is currently financed by the Government as well as by sev- eral external sources, including the World Bank, the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) and the United States government. It is expected, however, that external financing will cease or be significantly reduced in the next two years. As a result, a substantial increase in domestic financing for the national HIV/AIDS response will be needed. However, public debt levels are high, and the country is feeling the repercussions of the global financial crisis, thus the availability of domestic re- sources is and will be very tight. Any shortfall in financing whether domestic, external or both will have serious implications for the delivery of HIV services. The Government of Jamaica requested this study so as to inform its future HIV/AIDS policy response. This study is one input in a series of actions that the Government will undertake to formulate a future sustainability plan and investment framework for the National HIV Program. This study was led and financed by the World Bank and conducted in collaboration with the Government of Jamaica and UNAIDS. The study aimed to assess the sustainability of Jamaica’s National HIV Program from a fiscal perspective. Specifically, the purpose of the study was to: 1. review current spending on HIV/AIDS and the sources of financing; 2. estimate the fiscal burden of the national HIV/AIDS response and assess the outlook for external fi- nancing of the HIV program; 3. project how the epidemic will unfold as well as what the costs would be under different potential scenarios; and 4. provide recommendations to inform policy decisions. 1 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure ES1. Mortality by Cause Figure ES1.A. Jamaica: Composition of Mortality, 2008 Figure ES1.B. Jamaica: Mortality by Age and Sex (Ages 15–59) 2005–2010 1.8 Annual Mortality (Percent) 1.6 Male (Actual) Total: 1.4 Female (Actual) 5,850 deaths, crude mortality 1.2 Male (no AIDS) HIV/AIDS at 0.36 percent. Injuries (24 percent) 1.0 (29 percent) Female (no AIDS) 0.8 0.6 Communicable, excl. HIV/AIDS 0.4 (7 percent) 0.2 Noncommunicable 0.0 (38 percent) Maternal and peri-natal conditions, and 0 10 20 30 40 50 60 nutritional de ciencies Age (2 percent) Source: WHO, 2011. Source: Author’s caluculations, based on United Nations Population Division, 2011. HIV/AIDS in Jamaica: the State of the Epidemic and of the National Response It is estimated that 1.7 percent of the population aged between 15 The national response to HIV/AIDS took off in 1988, when the Na- and 49 years are HIV-positive. However, the prevalence rate is much tional AIDS Committee was established and the first national HIV plan higher among men who have sex with men (32 percent) and among was launched. From 2009/10 to 2010/11, the cost of the national commercial sex workers (five percent). UNAIDS (2010a,b) estimates response increased from J$ 1,680 million (US$ 19.3 million) to J$ that 32,000 people were living with HIV in 2009 (i.e., 1.1 percent 1,727 million (US$ 20.4 million). of the total population of 2.8 million). HIV prevalence has declined from the peak of 2.2 percent (among those aged 15–49 years) it had The largest component of HIV/AIDS spending is on prevention (J$ 676 reached in 1995. million in 2009/10, and J$ 625 million in 2010/11, see Figure ES2). A large share of prevention spending was geared towards “communica- HIV/AIDS has reversed the health gains that had been achieved in tion for social and behavioral change” targeting the general or young Jamaica over many years. Whereas life expectancy in the country population. In addition, the program also included measures target- grew at an average rate of 0.2 years annually between 1955/60 and ing sex workers and their clients, men who have sex with men and 2005/10, life expectancy at birth was 1.6 years lower in 2005/10 as drug users (about J$ 10 million—equivalent to US$ 120,000—each a result of HIV/AIDS. However, such aggregate estimates mask the in 2010/11, according to the National AIDS Spending Assessment, impact of HIV/AIDS among specific population groups. For example, or NASA, see Ministry of Health and UNAIDS (2012)). Treatment and HIV/AIDS accounts for one-quarter of deaths between those aged 15 care absorbed J$ 469 million (US$ 5.5 million) in 2010/11, the bulk to 59 years and is the leading cause of death among males between of which is accounted for by the cost of antiretroviral therapy (19 the ages of 30 and 44 years (Figure ES1). percent of total spending in 2010/11). Figure ES2. Composition of Spending Under National Response to HIV/AIDS, 2010/11 (Percent of total) Total: J$ 1,727 million, equivalent to Program Management 0.15 percent (19.6 percent) of GDP or Prevention US$ 20.4 million Social (36.2 percent) Mitigation (17.1 percent) Treatment, Care, and Support (27.7 percent) Source: Ministry of Health and UNAIDS (2012), and author’s calculations. 2 Executive Summary Figure ES3. Costs of National Response to HIV/AIDS 2010–2030 (declining HIV incidence) 3,500 Other program expenses J$ Millions 2010 Prices 3,000 Mitigation Treatment and care 2,500 Prevention 2,000 1,500 1,000 500 0 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. Projecting the Cost of the National Response to HIV/AIDS Many factors influence the course of the HIV/AIDS epidemic. There- between 24,900 (Scenario 1) and 30,600 (Scenario 3) in 2030. fore, this study makes some assumptions, based on a set of key fac- The number of people receiving treatment, however, barely differs tors, to project the future course of the epidemic in Jamaica. This across the scenarios for the first 10 years of projections. This is be- analysis is reflected in three scenarios: cause in the case of HIV there is a long lag between the time of infection and the need for treatment. • Scenario 1 (declining HIV incidence): The number of new adult infections declines from 2,000 in 2010 to around 1,680 by Largely as a result of the increasing cost of treatment, the spending 2018 (i.e., by 2.2 percent annually) and thereafter declines at needs for the country rises steeply in all the scenarios—from J$ 1.7 a slower rate of one percent annually. In this scenario, HIV in- billion (US$ 20.4 million) in 2010 to between J$ 3.1 billion (US$ 36 cidence declines from 0.14 percent (ages 15–49) in 2010 to million) and J$ 3.4 billion (US$ 40.2 million) in 2030 (in constant 0.11 percent in 2030. 2010 prices). This is illustrated for Scenario 1 in Figure ES3. However, relative to GDP, this cost increase is less pronounced because GDP is • Scenario 2 (constant HIV incidence): The number of new infec- also projected to increase over this period. The spending needs rise tions remains constant at 2,000 from 2010. HIV incidence rates from 0.15 percent of GDP to between 0.20 and 0.22 percent of GDP. also remain broadly constant at about 0.14 percent. Comparing the scenarios, the difference in cost is remarkably small— • Scenario 3 (increasing HIV incidence): The number of new by 2030, the total cost of the national response to HIV/AIDS is only infections increases from 2,000 in 2010 to 2,300 by 2018 10 percent lower in Scenario 1 than in Scenario 3. The difference in and thereafter grows more slowly by one percent annually to cost (11 percent) is much smaller than the differences in the number 2,600. HIV incidence increases from 0.14 percent (ages 15–49) of people receiving treatment (19 percent) or the number of new in 2010 to 0.19 percent by 2030. infections (43 percent). This small difference in costs reflects the long lag between the time of HIV infection and the time when people In each of the scenarios, the number of people living with HIV/ AIDS need treatment. While increasingly important over the projection pe- increases over the period 2010 to 2030 (to 34,000 in Scenario 1; to riod, treatment need may occur well over a decade after infection. In 39,500 in Scenario 2; and to 45,200 in Scenario 3). This increase— the case of second-line treatment (i.e., when the patient no longer even in the scenarios with declining or constant HIV incidence— responds to first-line treatment and the drug regimen needs to be results from a steep increase in the number of people receiving switched to more expensive drugs) this lag is especially long. treatment and, therefore, living longer: from 7,900 in 2010 to HIV Incidence and the Cost of the National Response The cost of the national response to HIV/AIDS is projected to rise person living with HIV/AIDS. The cost rises to about J$ 40,000 (US$ steeply over the coming years, largely as a consequence of the in- 460) annually over the 15 years following an infection, mainly owing creasing numbers of people living with HIV/AIDS requiring treatment. to the increasing probability that the person will require and receive In the short-term, this rise in costs is not very responsive to changes (first-line) treatment. Subsequently, the expected cost declines owing in HIV incidence. This apparent disconnect is a result of the long lag to declining survival. This decline is moderated, however, by the in- between infection and treatment need and also because, as a per- creasing number of people who switch to more expensive second-line sistent health condition, HIV/AIDS requires ongoing treatment. Fig- treatment, which means that the average unit cost of treatment in- ure ES4 summarizes the expected cost of providing services to one creases for people surviving longer. Calculating the projected total 3 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure ES4. Jamaica: Costs of One Additional Infection Figure ES5. Jamaica: Fiscal Costs of HIV/AIDS, (J$, 2010 prices) “Commitment Basis”, 2010–2030 (Percent of GDP) 45,000 0.25 40,000 Total 35,000 0.20 Treatment Actual Expenditures 30,000 Total Costs, 25,000 0.15 “commitment basis” Current costs of 20,000 meeting demand Costs incurred by 15,000 0.10 by people living new infections with HIV/AIDS 10,000 0.05 Population-based spending 5,000 0 0.00 2012 2017 2022 2027 2032 2037 2042 2047 2010 2015 2020 2025 2030 Source: Author’s calculations. Source: Author's estimates and projections. Figure is based on Scenario 1, in which HIV incidence declines. costs incurred over time by each infection is relevant for assessing to current spending on people living with HIV/AIDS, at 0.08 percent of how cost-effective HIV prevention interventions will be. GDP. Additionally, population-based spending absorbs 0.06 percent of GDP. Because of declining HIV incidence, the cost incurred by new To assess the cost-effectiveness of a prevention intervention from infections comes down to 0.05 percent of GDP by 2030. While current a fiscal perspective, one needs to relate the outcome (HIV preven- spending increases, largely as a consequence of increasing treatment tions averted, or the resulting health gains) to the cost of achieving coverage and a build-up in the demand for treatment of past HIV it, as well as the financial savings from the reduced demand for HIV/ infections, the new demands for fiscal resources for the national re- AIDS-related services such as treatment. To achieve this, it is neces- sponse to HIV/AIDS (to meet the demand for services caused by new sary to evaluate the total cost caused by one infection over time (the HIV infections) declines over the projection period. lifetime of the infected individual). As most of the costs of treatment and other services in the case of HIV/AIDS will only become payable The differences in the fiscal costs involved between the three sce- at a later stage in the person’s infection, most of this initial outlay narios are illustrated in Figure ES6. By 2015, when actual spending could be invested (or not borrowed) for an extended period. The ex- barely begins to increase, the costs incurred by new infections are pected future cost may, therefore, be discounted by an appropriate over J$ 100 million (US$ 1.3 million, about 0.01 percent of GDP) interest rate until it must be paid in reality. Applying an interest rate of lower in the scenario with reduced HIV incidence (Scenario 1) when five percent to discount the expected future cost, the estimated cost compared to Scenario 2 with constant HIV incidence, and by over J$ caused by one adult HIV infection is equal to J$ 500,000 (equivalent 200 million (US$ 2.4 million) compared to Scenario 3 with increasing to US$ 5,800), which is about the level of GDP per capita in Jamaica. HIV incidence (at constant 2010 prices). In Scenario 2 (constant HIV This means that preventing one HIV infection at a cost of less than incidence), this difference (relative to Scenario 1) increases to J$ 196 this, reduces the financial cost of the national response to HIV/AIDS, million (US$ 2.3 million) by 2020 and to J$ 294 million (US$ 3.5 and is a good financial investment, besides the positive health bene- million) by 2030, while in Scenario 3 with increasing HIV incidence it fit. For an HIV infection averted at a higher cost, the financial saving increases to J$ 378 million (US$ 4.5 million) by 2020 and to J$ 623 of J$ 500,000 needs to be subtracted from the overall cost to assess million (US$ 7.3 million) by 2030. the cost-effectiveness of the intervention. These findings are relevant for the design and evaluation of HIV/AIDS The estimates of the costs incurred by one infection also yield a clearer policies because they provide: picture of the evolving fiscal burden of HIV/AIDS. HIV/AIDS spending in one year can be summarized as the sum of all “population-based” 1. Estimates of the financial implications of alternative HIV/AIDS spending (largely prevention spending) that cannot be directly allo- policies, including the implications of changes in HIV incidence cated to people living with HIV/AIDS and the cost incurred by all new for spending needs. infections (i.e. the number of new infections multiplied by the financial 2. A tool that helps to rank alternative policies in terms of their costs per HIV infection). To visualize this, Figure ES5 compares current cost-effectiveness. HIV/AIDS spending and new costs or spending commitments caused by HIV infections in the case of Scenario 1 (declining incidence). Ac- 3. The consequences of higher or lower effectiveness of HIV pre- tual spending increases throughout the projection period. Initially, the vention interventions. financial commitments caused by new HIV infections are about equal 4 Executive Summary Figure ES6. Fiscal Costs of HIV/AIDS, Current and Commitments, Across Scenarios (Percent of GDP) 0.25 Actual Spending 0.20 0.15 Spending on 0.10 Scenario 1 Commitment Basis Scenario 2 0.05 Scenario 3 0.00 2010 2015 2020 2025 2030 Source: Author’s estimates. The analysis supports the design of cost-effective HIV/AIDS policies is taken into account, the financial consequences of new infections by linking: (1) the costs of alternative HIV prevention interventions, are much larger than the cost incurred directly by one infection. The (2) their outcomes for the course of the epidemic and (3) the conse- potential for downstream infections differs widely across the popu- quences, assessed in terms of the costs of the resulting demand for lation according to sexual risk behavior. For individuals characterized HIV/AIDS services. by “low-risk” sexual behavior, “downstream” infections add about 60–80 percent to the cost incurred by one infection. For female sex One of the most important consequences of new infections is the workers, on the other hand, the expected cost of “downstream” in- potential to cause further HIV infections (i.e., once infected an indi- fections is four times higher than the direct costs attributable to a vidual is able to infect others, who in turn are able to infect further single infection; for men who have sex with men, this cost is ten individuals). This is an important issue in Jamaica, where the epidem- times higher. ic is concentrated. When the cost of such “downstream” infections Sustainability of the National Response to HIV/AIDS Whether the national HIV/AIDS response is financially sustainable (as there is a lag in the need for treatment after each incidence depends on how the country deals with the challenges that arise pri- of infection). Nevertheless, in the long run successful prevention marily from: (1) the increasing cost of providing HIV/AIDS services to efforts will reduce government spending, which will play an import- the greatly expanding number of people eligible for and receiving ant role in helping to ensure that funding of the HIV/AIDS program treatment; and (2) the uncertain outlook of external funding. can be sustained. In 2010/11, the Government of Jamaica met almost one third of total Taking into account the new eligibility criteria for funding from the spending (31 percent). External financing, meanwhile, accounted for Global Fund alone, the share of domestic financing would increase 69 percent of the cost of the national response to HIV/AIDS. By far the to at least 50 percent of total spending over the next few years. This most substantial external financial contribution was from the Glob- comes in addition to the increasing total cost of the national response al Fund (49 percent of total spending), followed by the World Bank to HIV/AIDS. Given the context of scarce fiscal resources, all demands (9 percent of total spending) and the United States Government for additional funding will need to be well founded. To this end, the (2 percent of total spending). analysis presented in this report provides insights regarding the driv- ers of the costs of the national response to HIV/AIDS (in terms of the The external financing from all the main donors is set to end over dynamics of HIV infections, and the link between HIV infections and the next few years, and the funding outlook is uncertain beyond this the demand for HIV/AIDS-related services), and the cost-effectiveness period. In the face of rising spending needs, a decline in the rate of of HIV/AIDS interventions. external support will result in an even steeper increase in domestic funding needs. For example, under Scenario 1 (declining HIV inci- HIV/AIDS spending in the health sector (largely treatment cost) is dence), if the rate of external support was to decline from 69 to projected to remain at about two to three percent of public health 45 percent, then domestic funding needs will need to double by spending, which appears a sensible amount given the burden of dis- 2015—from J$ 539 million (US$ 6.3 million) in 2010 to J$ 1,144 ease (accounting for nine percent of all deaths in 2005/10 and for 24 million (US$ 13.5 million) in 2015—and will need to increase to percent of those aged between 15 and 59 years). More specifically, J$ 1,353 million (US$ 15.9 million) by 2020 and J$ 1,683 million the annual cost of providing treatment—which represents about 10 (US$ 19.8 million) by 2030. The consequences of a decline in exter- percent of GDP per capita for first-line treatment and 20 percent for nal funding are fairly even across the three scenarios due to the fact second-line treatment—falls well within a range that would normally that changes in HIV incidence do not affect spending immediately be supported by the public health system. 5 Assessing the Financial Sustainability of Jamaica’s HIV Program An additional consideration for policy makers to take into account To ensure the sustainability of the national HIV response, the study is access to health services. The Jamaica Survey of Living Conditions points to three priority areas for the Government: 2009 points to the presence of social barriers in access to health services, both in terms of utilizing services and in the financial con- • Develop a long-term sustainability plan with different financing sequences of seeking care. HIV/AIDS, as a persistent and expensive options that defines the roles of different national agencies (as health condition, potentially exacerbates such inequities. well as donor agencies). This plan would build on the cost esti- mates from this study as well as a public expenditure review to In summary, the rationale for sustaining broad access to treatment quantify government health spending. through the public health system appears sound from a fiscal and public health perspective (even before taking into account the social • Assess the efficiency of funding allocations in the national re- repercussions). sponse to HIV/AIDS, taking into account not only the imme- diate health consequences but also the epidemiological and The study also offers tools to assess the epidemiological as well as financial repercussions of alternative HIV/AIDS interventions. the financial consequences of investments in prevention. It finds that there are large differences in the risk, and therefore cost, of onward, • Assess the allocative efficiency between prevention and treat- additional infections across population groups. Specifically, because ment, and between specific programs within prevention, to of the disease dynamics, the cost caused by an additional infection guide targeted prevention efforts, particularly for most at risk (and saved by an infection prevented) is significantly higher for most- populations; effective prevention in the short term will result in at-risk populations, such as men who have sex with men and female lower spending on treatment in the long term. sex workers. To improve the cost-effectiveness of prevention efforts, investments in prevention will need to be tested against these cost savings—in addition to the health outcomes. 6 I. Introduction I. Introduction Jamaica faces increasing financing needs over the coming years to respond to its HIV/AIDS epidemic, largely reflecting the increased need for treatment. At the same time, the financing of its national HIV/ AIDS response has come under pressure from a domestic fiscal adjustment as well as from shrinking ex- ternal resources. The response in Jamaica is currently financed by the Government of Jamaica and from other sources, such as the World Bank, the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) and the United States Government. Due to the high level of public debt and the repercussions of the global financial crisis, the availability of domestic resources is and will continue to be very tight. It is expected that during the next two years external financing will cease or be significantly reduced, result- ing in a need for a substantial increase in domestic financing for the country’s HIV/AIDS response. Any shortfall in financing—whether domestic, external or both—will have serious implications for the delivery of HIV services. The objectives of the study are to: • Review current spending on HIV/AIDS and the sources of financing (domestic and external). • Estimate the fiscal burden of the national HIV/AIDS response and assess the outlook for external fi- nancing of the HIV program. • Project how the epidemic will unfold as well as what the costs would be under different scenarios. • Provide recommendations to inform policy decisions. To meet these objectives, the assessment utilizes an epidemiological model to project the number of new infections; a costing framework to project the cost of the national response and an economic framework to translate the number of new infections to fiscal costs. This report provides an assessment of and guid- ance on the financial sustainability of Jamaica’s national response to HIV/AIDS. The analysis is organized in four major sections: • Section II describes the state of the epidemic and of the national response. It summarizes the available data on the state and course of the epidemic thus far; discusses the evolving national response, and its costs and financing; and reviews the state of public finance and overall trends in health financing and access to health services 7 Assessing the Financial Sustainability of Jamaica’s HIV Program • Section III projects the state of the epidemic and the costs of the only with a long lag. This is because HIV/AIDS is a persistent national response from 2010 to 2030. The uncertainties regarding health condition that requires treatment for a long period. In the demand for HIV/AIDS services and the consequences of the addition, an intervention may affect the disease dynamics, with National HIV/AIDS Strategic Plan (NSP) being developed are il- effects spread over a long time. Section IV develops tools to lustrated by way of three envisaged scenarios: (1) A scenario with assess the fiscal costs arising from one additional HIV infection declining HIV incidence, broadly in line with the NSP; (2) a scenario (or the savings achieved by an infection prevented), Section with constant HIV/AIDS incidence; and (3) a scenario with increas- V uses these tools to assess the fiscal costs of the national ing HIV incidence, which could result from a reversal of recent response to HIV/AIDS. gains or as a consequence of ineffective prevention programs. • Section VI considers policy lessons and challenges. This relates • Sections IV and V analyze the links between HIV incidence and to the implications of declining external funding for domestic the costs of the national response. In particular, they develop funding needs, the cost-effectiveness of HIV/AIDS interven- an economic framework to assess the fiscal burden of HIV/AIDS tions, the allocative effectiveness of the HIV/AIDS program, and and the cost-effectiveness and allocative effectiveness of HIV/ funding of the increasing demand for treatment. AIDS interventions. The assessment of the cost-effectiveness of HIV/AIDS interventions is complicated, because the demand for Finally, the Conclusions provide key messages from the analysis and treatment and other HIV/AIDS services follows an HIV infection suggest priority next steps for the national HIV/AIDS response. 8 II.  The State of HIV/AIDS and of the HIV/AIDS Response II.  The State of HIV/AIDS and of the HIV/AIDS Response To prepare the ground for the financial analysis, the present section describes the background of the evolving national response to HIV/AIDS. This is provided in two parts: • The state of the epidemic: this reviews the epidemic thus far—including the latest evidence on the current state of the epidemic and the drivers of HIV incidence. • Costs and the financing of the national response to HIV/AIDS: an estimate of the costs of financing the evolving national response to HIV/AIDS. Included in this subsection are two boxes, which describe: • The macroeconomic and fiscal context, in Box 1. • Overall health spending and access to health services across the population, in Box 2. The State of the Epidemic The estimated HIV prevalence rate in Jamaica is 1.7  percent among the adult population (aged 15–49 years). Higher prevalence rates are reported for the most-at-risk populations (32 percent among men who have sex with men and five percent among commercial sex workers). According to the Ministry of Health, HIV incidence attained a peak of 5,500 new infections among adults and children in 1994 (Figure 1a) and subsequently declined to about 2,500 by 2005, and con- tinued to decline (but more slowly) until 2010.1 AIDS mortality increased later and more gradually than HIV incidence, peaking at about 3,300 deaths in 2004. Around this year, the scaling-up of antiretroviral treatment also took off (from 180 persons receiving treatment in 2004 to 8,150 persons in 2010), which contributed to a steep decline in AIDS mortality to 1,800 deaths by 2010. It should be noted, however, that these estimates are subject to considerable uncertainty; they are generated by models with the result that people known to be HIV-positive account for only a subset of people living with HIV/AIDS and direct observations on HIV incidence are unavailable. Estimates presented in this section have been made by the National HIV/AIDS Program, using the UNAIDS EPP model and Spectrum and were provided by Dr. 1  Sharlene Jarrett. Earlier estimates of the state and course of the epidemic are reported in Duncan and others (2010). 9 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 1. The Course of the HIV Epidemic in Jamaica, 1980–2010 Figure 1a. Jamaica: HIV Incidence and Deaths, Figure 1b. Jamaica: People Living with HIV/AIDS 1980-2010 1980-2010 6,000 40,000 HIV incidence 35,000 Second-Line Treatment 5,000 HIV mortality 30,000 First-Line Treatment 4,000 25,000 Eligible, Not Receiving Treatment 3,000 20,000 Not Eligible for 15,000 Treatment 2,000 10,000 1,000 5,000 0 0 1980 1985 1990 1990 2000 2005 2010 1980 1985 1990 1995 2000 2005 2010 Source: Data provided by Minitry of Health/National HIV/AIDS Program Source: Data provided by Minitry of Health/National HIV/AIDS Program The number of people living with HIV/AIDS increased until about 2000 other areas achieved over eight years. HIV/AIDS-related mortality is when it reached 37,000 (Figure 1b). The number then declined for a concentrated among the adult population—HIV/AIDS accounted for number of years, as increasing mortality among people living with HIV/ about nine percent of all deaths in 2005–10, but only 1.5 percent of AIDS outstripped the number of new infections. After 2004, however, infant deaths. increasing access to treatment resulted in a steep decline in mortality. As a consequence, people receiving treatment account for an increasing Figure 2 provides an analysis of the contribution of HIV/AIDS to the share of people living with HIV/AIDS (rising to 27 percent by 2010).2 burden of disease. For the overall population (Figure 2a), non-com- municable diseases are the dominant cause of mortality, accounting According to the Ministry of Health/Jamaica National HIV/STI Pro- for over two-thirds of deaths. Within this category, cardiovascular gramme (MoH/NHP) (2010a), there were 378 AIDS deaths reported diseases account for 32 percent of total deaths. HIV/AIDS (eight per- in 2009 (down from 692 in 2002), and about 900 new cases of cent of all deaths) accounts for almost one-half of all deaths from HIV were reported for that year (down from about 1,200 annually communicable diseases (17 percent). Owing to the age profile of HIV/ in 2004–2006).3 As of the end of 2010, there were 7,560 adults AIDS-related deaths, population averages mask somewhat the health (and 487 children) receiving antiretroviral treatment, corresponding consequences of the epidemic. This is evident from Figure 2b, which to an antiretroviral treatment coverage rate among adults of 56 per- shows the main contributors to mortality for the 15–59 population. cent.4 Duncan and others (2010) estimate that about one quarter Here, HIV/AIDS accounts for almost one-quarter of deaths. of people living with HIV/AIDS are men who have sex with men, three percent are female sex workers, five percent are male clients These summary indicators of health outcomes also offer some insights of sex workers, 0.6 percent are crack/cocaine users, 0.4 percent are on the consequences of the scaling-up of treatment. According to UN- prison inmates, 36 percent are “remaining males,” and 30 percent AIDS (2010a,b), the number of AIDS deaths has dropped from 2,600 in are “remaining females.” 2005 to 1,200 in 2009, largely as a consequence of improved access to treatment. This represents a decline in crude mortality among the The principal drivers of the HIV epidemic have been multiple part- population overall from 0.81 percent to 0.75 percent (with the share nerships, early sexual debut, high levels of transactional sex and in- from HIV/AIDS declining from 12 percent to six percent) with a decline adequate condom use (MoH/NHP (2010b)). According to the 2008 “HIV/AIDS Knowledge Attitudes and Behavior Survey” (Hope Enter- prises, 2008), 53 percent of males and 21 percent of females engaged 2  In Fig. 1b, the discontinuity in the numbers of people not eligible for treat- ment (based on current treatment guidelines) and of those eligible but not in transactional sex, and the rate of condom use in this group was receiving treatment between 2005 and 2006 reflects a shift in eligibility cri- around 60 percent. Of respondents to the survey, 39 percent reported teria for treatment incorporated in the estimates by the Ministry of Health/ having multiple partners in the last 12 months (62 percent of males, National HIV/STI Program (MoH/NHP)—from a CD4 count below 200 to a and 17 percent of females). Twenty percent of male respondents, and CD4 count below 350. 34 percent of female respondents knew their HIV status (i.e., received 3  The reported number of people living with HIV/AIDS in 2009 is lower than an HIV test in the last 12 months and knew the results).5 Among the estimated number of people living with HIV/AIDS, as many people living persons living with HIV/AIDS, 23  percent reported having sex with with HIV/AIDS may not know their status, and not all HIV cases are reported. a sex worker, and more than 80 percent reported multiple partners 4  According to the draft National HIV/AIDS Strategic Plan. (MoH/NHP (2010b)). 5  The 2012 “HIV/AIDS Knowledge Attitudes and Behavior Survey” had not been released at the time of writing, but preliminary findings were reported The impact of HIV/AIDS is reflected in recent trends in summary health by MoH/NHP (2012). According to these findings, the number of people re- porting sexual intercourse with more than one partner in the past 12 months indicators (Table 1). HIV/AIDS has reduced life expectancy at birth by had declined to 28.1 percent; among these, 56.9 percent reported use of a 1.6 years (2.1 years for males, 1.0 years for females) as of 2005–10. condom at their last sexual intercourse. The reported percentage of women This is a substantial reversal. Between 1955–60 and 2005–10, life and men aged 15–49 “who received an HIV test in the past 12 months and expectancy grew at an average rate of 0.2 years annually. The impact know their results” had increased steeply to 58.7 percent (47.2 percent for of HIV/AIDS is therefore equivalent to a reversal of health gains in men, 67.7 percent for women) 10 II.  The State of HIV/AIDS and of the HIV/AIDS Response Figure 2. Mortality by Cause, 2008 Figure 2a. Jamaica: Composition of Mortality, 2008 Figure 2b. Jamaica: Composition of Mortality, 2008 (All Ages) (Ages 15-59) HIV/AIDS Injuries (8 percent) (11 percent) Communicable, Total: excl. HIV/AIDS Total: 20,834 deaths, (9 percent) 5,850 deaths, crude mortality crude mortality HIV/AIDS at 0.77 percent. at 0.36 percent. Injuries (24 percent) 1/MPN (4 percent) (29 percent) Communicable, excl. HIV/AIDS (7 percent) Noncommunicable 1/MPN: (68 percent) Maternal and Noncommunicable (38 percent) MPN (See Fig 2a) perinatal (2 percent) conditions, nutritional Source: WHO, 2011. deficiencies. Source: WHO, 2011. Table 1. Jamaica: Summary Health Indicators and Impact of HIV/AIDS 2005–10 1990–95 Without Impact of Actual Actual HIV/AIDS HIV/AIDS Life expectancy at birth (Years) Male 67.9 69.6 71.7 –2.1 Female 73.0 75.0 76.0 -1.0 Remaining life expectancy at age 15 (Years) Male 55.8 57.1 59.1 –2.1 Female 60.8 62.3 63.3 –1.0 Crude death rate (Percent) 0.76 0.74 0.67 0.06 Infant mortality (Per 1,000) 28.8 24.4 24.0 0.4 Under-five mortality (Per 1,000) 34.6 28.9 28.1 0.8 Source: United Nations Population Division, 2011. The impact of HIV/AIDS relates to the estimates for 2005–10. Estimates of actual health indicators for 1990–95 already reflect impact of HIV/AIDS. in mortality of those between ages 15 and 59 years from 0.42 to 0.34 ity of a scenario without AIDS. The impact on women is less pro- (with the share from HIV/AIDS declining from 36 percent 20 percent). nounced—HIV/AIDS increases mortality by about two-thirds between ages 30 and 39 years. As a consequence of increased premature mor- AIDS-related mortality is highly concentrated among mid-age adults tality, survival to older ages declines; for example, the probability of (Figure 3). In the case of males between ages 30 and 45 years, HIV/ reaching age 50 declines by five  percentage points for males, and AIDS is the dominant cause of death (Fig. 3a)—double the mortal- two percentage points for females. Costs and the Financing of the National Response to HIV/AIDS The national response to HIV/AIDS took off in 1988 when the Na- sponse is guided by the “HIV and AIDS in Jamaica: NSP 2007–2012” tional AIDS Committee was established and the first national HIV plan (Government of Jamaica, 2008). A new NSP for 2013–18 is being was introduced.6 The response is managed by the National HIV/STI prepared, but has not been finalized yet. Programme in the Ministry of Health, which is also the counterpart for funding from major donors. The national response to HIV/AIDS Notable achievements of the national response to HIV/AIDS include has significantly benefitted from external financing, notably from behavioral changes such as increased condom use, the expansion in the World Bank (two projects approved in 2001 and 2008) and the treatment under the public access treatment program supported pri- Global Fund (2004 and 2008). External financing was instrumental in expanding prevention programs and extending access to antiretroviral treatment through public health services. (See also the discussion on See Figueroa and others (2008) for a comprehensive discussion of Jamaica’s 6  HIV/AIDS spending and financing below.) Currently, the national re- national response to HIV/AIDS. 11 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 3. Mortality by Age and Sex, 2005–10 Figure 3a. Jamaica: Mortality by age and sex Figure 3b. Jamaica: Survival Rates from Age 0 by sex, 2005-2010 2005-2010 1.8 100 Annual Mortality (Percent) 1.6 Male (Actual) 90 Survivors (Percent) 1.4 80 Female (Actual) 70 1.2 Male (no AIDS) 60 Male 1.0 Female (no AIDS) 50 Female 0.8 40 0.6 30 Male (no AIDS) 0.4 20 Female (no AIDS) 0.2 10 0.0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 70 80 Age Age Source: Author’s caluculations, based on United Nations Population Division, 2011. Source: Author’s caluculations, based on United Nations Population Division, 2011. marily by the Global Fund, a considerable reduction in HIV prevalence becoming orphans and for the need of social support among among female sex workers (from 12 percent in 1989 to five percent young people, it is appropriate to include part of the budget of in 2008), and increased access to and uptake of HIV testing. Accord- the CDA under the costs of the national response to HIV/AIDS. ing to model-generated estimates by the Ministry of Health, HIV inci- dence has declined from a peak of over 5,000 HIV infections in 1995 Estimates on the costs and sources of financing of the national to around 2,600 in 2002, and to 2,400 by 2010, while the rate of response to HIV/AIDS are summarized in Table 2. From 2009/10 transmission of HIV from mother-to-child has declined from 25 per- to 2010/11, the costs of the national response to HIV/AIDS in- cent before 2000 to less than five percent since 2007. creased from J$ 1,681 million (US$ 19.3 million) to J$ 1,727 million (US$ 20.4 million). More than two-thirds of total spending for the na- However, numerous challenges persist, that act as barriers to further tional HIV response was financed externally, notably through a grant reductions in HIV incidence. These include a lack of awareness of HIV from the Global Fund of J$ 842 million (US$ 10 million) in 2010/11 status (close to 50 percent of people living with HIV/AIDS), the strong and a loan from the World Bank of J$ 161 million (US$ 1.9 million) stigma against men who have sex with men combined with very in 2010/11.7 high HIV prevalence in this population group (Duncan and others, 2010), and multiple partners and transactional sex in the general HIV/AIDS spending financed from the national budget (J$ 416 million population (Figueroa and others (2008), Hope Enterprises (2008)). in 2009/10 and J$  535  million in 2010/11) covers not only some Consequently, most HIV infections now occur among men who have of the operating costs of the NHP, but also the estimated costs of sex with men (53 percent, according to preliminary estimates of an support for children affected by HIV/AIDS delivered through the CDA ongoing study into modes-of-transmission) and their female partners and some of the costs of delivering HIV/AIDS services not captured in (11 percent). The national response to HIV/AIDS has made little in- specific HIV/AIDS budget line items (human resources, facilities). The roads so far in containing the spread of the epidemic among these bulk of HIV/AIDS services were delivered through the public sector population groups. (77  percent and 74  percent of the total spending in 2009/10 and 2010/11, respectively) and through civil society organizations, while Data on the costs of the national response to HIV/AIDS were estimated international agencies and private providers played a small role. Total based on information obtained from the following three sources: HIV/AIDS spending was equivalent to 0.45 percent of public spending in 2010/11, and domestically financed HIV/AIDS spending absorbed • Actual spending recorded under the national response to HIV/ 0.14 percent of public spending in that year. AIDS, as compiled by the National AIDS Spending Assessment (NASA) for the fiscal years 2009 and 2010. Figure  4 shows the composition of national response HIV/AIDS spending in 2010/2011. The largest component of HIV/AIDS spending • Data on unit costs of specific HIV/AIDS-related services, such as is prevention efforts (J$ 676 million in 2009/10 and J$ 625 million the estimates of the unit costs of treatment and related tests in 2010/11). A large share of prevention spending has been geared compiled by the Clinton Health Access Initiative. towards “communication for social and behavioural change” target- ing the general or young population but the program also included • Data on government spending that does not directly target measures targeting sex workers and their clients, men who have people living with HIV/AIDS or those affected by the epidemic sex with men, and drug users (about J$  10  million—equivalent to but which addresses conditions related to both the impact of US$ 120,000—each in 2010/11). HIV/AIDS and to other factors. For example, HIV/AIDS results in higher mortality among young adults and increases the number of orphans and vulnerable children, but is not the only cause of children becoming orphans. In Jamaica, support for orphans The NASA shows spending supported by the World Bank of J$207 million. 7  and vulnerable children is provided by the Child Development This probably includes a contribution from the Government of Jamaica of J$ Agency (CDA). As HIV/AIDS is a principal cause of children 46 million, which is attributed to domestic financing here. 12 II.  The State of HIV/AIDS and of the HIV/AIDS Response Box 1. Macroeconomic and Fiscal Context of the National Response to HIV/AIDS GDP per capita in Jamaica is estimated at about US$ 5,400 in 2011 (IMF, 2011). Compared to other countries in the Caribbean with elevated levels of HIV prevalence, GDP per capita is on about the same level as in the Dominican Republic (US$ 5,400), much lower than in The Bahamas (US$ 23,200) and Trinidad and Tobago (US$ 16,700), but much higher than in Haiti (US$ 740). Economic growth has been low in recent years. Between 1997 and 2007, real GDP grew by an average annual rate of 1.4 percent, and real GDP per capita by 0.7 percent. Owing to the global financial crisis, GDP growth was negative in 2008–2010 and dropped to minus 3.0 percent in 2009. For the fiscal year 2011/12 (April to March), the Ministry of Finance estimates GDP growth at 1.3 percent (Phillips, 2012). The fiscal situation is challenging. The Ministry of Finance and Planning estimates that public debt stood at 128 percent of GDP at the end of March 2012, of which external debt accounted for 58 percent of GDP (Phillips, 2012).Reflecting the high level of public debt, more than one-half of public expenditures were absorbed by debt service (interest and amortization) in 2010/11. The government intends to reduce the budget deficit (6.2 percent of GDP in 2011/12) to 3.8 percent of GDP in 2012/13 and aims at a balanced budget by 2015/16 (Phillips, 2012). Table 2. Costs and Financing of Jamaica’s National Response to HIV/AIDS 2009/10 2010/11 J$ US$ % of J$ US$ % of 0millions millions Total millions millions Total Total 1,680.6 19.3 100.0 1,727.2 20.4 100.0 By source Government of Jamaica 416.0 4.8 24.8 535.0 6.3 31.0 HIV/AIDS spending 173.6 2.0 10.3 259.1 3.1 15.0 Share of CDA budget 1 242.4 2.8 14.4 275.9 3.3 16.0 External sources 1,260.5 14.5 75.0 1,188.2 14.0 68.8 o/w Global Fund grant 868.8 10.0 51.7 842.1 10.0 48.8 o/w World Bank loan 111.2 1.3 6.6 161.2 1.9 9.3 Private sources 2 4.0 0.0 0.2 4.0 0.0 0.2 By provider Public sector 1,299.3 15.0 77.3 1,283.3 15.2 74.3 International agencies 65.4 0.8 3.9 28.4 0.3 1.6 Civil Society 300.5 3.5 17.9 386.4 4.6 22.4 Private 15.4 0.2 0.9 29.0 0.3 1.7 By spending category Prevention 675.7 7.8 40.2 625.4 7.4 36.2 Treatment, care, and support 411.5 4.7 24.5 469.0 5.5 27.2 o/w: Antiretroviral therapy 257.2 3.0 15.3 334.2 3.9 19.3 Social mitigation 251.8 2.9 15.0 295.0 3.5 17.1 Program management 341.5 3.9 20.3 337.8 4.0 19.6 Memorandum items (In units indicated) (In units indicated) Total costs (percent of GDP) 0.15 0.14 Total costs (percent of public expenditures) 3 0.40 0.45 Domestically financed HIV/AIDS spending 0.10 0.14 (percent of public expenditures) 3 Total costs (percent of public health exp.)3 5.5 7.2 Treatment, cane and support (percent of public health exp.)3 1.4 2.0 Source: Preliminary NASA estimates and author’s estimates. 1 Share of budget of Child Development Agency (CDA), in line with estimated burden of HIV/AIDS. 2 Private donations to national HIV/AIDS program. 3 Not all HIV/AIDS spending is included in public accounts. HIV/AIDS expenditures are shown relative to total public spending or public health spending to illustrate the relative magnitudes, but not to attribute part of public spending to the HIV/AIDS program. 13 Assessing the Financial Sustainability of Jamaica’s HIV Program Box 2. Health Spending and Access to Health Services Total health spending accounted for about Jamaica: Health Expenditures by Source of Funding, 5.1 percent of GDP in 2009 (see box figure), equiva- 1995-2009 (Percent of GDP) 7.0 1.20 lent to US$ 228 per capita. The most recent increase in public health spending partly reflects the abolition 6.0 1.00 of user fees in 2008. Private health spending (in rela- Private Insurance 5.0 tion to GDP) has increased since 2005 (and may have 0.80 increased faster in real terms). The costs of health ser- 4.0 Private—Out of Pocket 0.60 vices in the consumer price index have increased at 3.0 a much lower annual rate (about four percent) than NGOs 0.40 2.0 overall inflation (about 14  percent) between 2006 Public 1.0 0.20 and 2010. Externally Financed (right scale) However, health spending and access to health ser- 0.0 0.0 vices is not distributed evenly across the population. 1995 1997 1999 2001 2003 2005 2007 2009 Health spending increases more than proportionally Source: WHO (2011). with income (see box table), and the data reported by STATIN and PIOJ (2010) also suggest an urban/ rural divide. Access to health insurance is higher in the richest quintile than in the poorest quintile (20.8 percent vs. 13.6 percent). Care-seeking poor households predominantly rely on public health facilities (63 percent), while the wealthiest households rely on private providers (76.6 percent) Health Spending and Health Seeking Across Population Groups, 2009 Annual consumption per capita Having health Reporting Purchasing Health insurance illness Seeking care medication Total J$ J$ % of total % of individuals % of ind. reporting illness By Consumption Quintile Poorest 64,258 1,031 1.6 13.6 9.6 64.6 59.3 2 106,240 2,175 2.0 16.3 11.9 71.7 72.8 3 148,214 3,426 2.3 12.7 10.2 76.7 72.5 4 206,917 4,695 2.3 16.9 10.9 77.6 70.4 Richest 419,025 12,020 2.9 20.8 11.1 81.7 84.0 By Region Kingston MA 252,835 6,705 2.7 27.7 8.7 84.6 81.9 Other Urban 232,136 5,993 2.6 22.2 10.2 82.8 79.3 Rural 158,474 3,720 2.3 12.8 12.3 70.9 68.5 Jamaica 205,693 5,201 2.5 19.8 10.6 77.1 74.4 Source: STATIN and PIOJ, 2010. Available household data (STATIN and PIOJ, 2010) also point to differences in the economic consequences of injury or episodes of illness across income groups. For poor households seeking care, the costs average at 26 percent of their monthly income per capita, compared to nine percent for the wealthiest households. Similarly, the average costs of purchasing medication are 29 percent for the poorest wealth quintile, but only 10 percent for the richest group. 14 II.  The State of HIV/AIDS and of the HIV/AIDS Response Figure 4. Composition of Spending Under National Response to HIV/AIDS, 2010/11 (Percent of total) Total: J$ 1,727 million, equivalent to Program Management 0.15 percent (19.6 percent) of GDP or Prevention US$ 20.4 million Social (36.2 percent) Mitigation (17.1 percent) Treatment, Care, and Support (27.7 percent) Source: Ministry of Health and UNAIDS (2012), and author’s calculations. Treatment and care absorbed J$  411  million in 2009/10 and Social mitigation spending is dominated by the support given to J$ 469 million in 2010/11. The costs of providing antiretroviral ther- orphans and vulnerable children through the CDA. The national re- apy accounted for an increasing share of total spending (15 percent sponse to HIV/AIDS also supports measures to create an “enabling in 2009/10 and 19 percent in 2010/11). The estimated costs of treat- environment” for people living with HIV/AIDS. Program manage- ment and care also include some provisions to capture the costs of ment (including program administration and monitoring and evalu- health services other than antiretroviral therapy, which are not fully ation) accounts for 20 percent of the costs of the national response captured by recorded HIV/AIDS spending. to HIV/AIDS. 15 III.  Projecting the Costs of the National Response to HIV/AIDS III.  Projecting the Costs of the National Response to HIV/AIDS The following sections provide a forward-looking analysis (1) to quantify the fiscal costs of HIV/AIDS and of the national response to it (based on estimates of the state and course of the epidemic and on the objectives and targets of the national response); and building on that goal (2) to provide an analysis that supports decision making on policy and financing options in respect of the national HIV/AIDS program. To achieve these objectives, the current section provides projections of the costs of the national response to HIV/AIDS through 2030. The uncertainty about the course of the epidemic is reflected in three scenarios with different assumptions about the course of HIV incidence. The analysis includes: • Epidemiological projections. • Unit costs and coverage rates of HIV/AIDS services. • Projected costs of the HIV/AIDS program, based on estimates and projections of the state of the epi- demic, the coverage rates of HIV/AIDS services and unit costs of providing them. • Some policy lessons. In addition, the macroeconomic impact of HIV/AIDS is set out in Box 3. Epidemiological Projections The epidemiological projections are based on estimates and projections produced by the Ministry of Health, using the Spectrum package. An earlier version of these estimates is described by Duncan and others (2010). For the analysis offered in this report, the estimates were transformed into a spread- sheet-based model, to better capture the links between HIV incidence, program objectives, and the costs of the HIV/AIDS program. The assumptions on the coverage rates of HIV/AIDS-related services are based on the NSP and the projections provided by the Ministry of Health. Most importantly, the coverage rate of antiretroviral treatment is assumed to rise from 73  percent of people newly requiring first-line treatment in 2011 to 90 percent of people newly requiring first-line treatment by 2015, and then to remain at this rate. 17 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 5. People Living with HIV/AIDS, 2010–2030 Ninety percent of people reaching the need for second-line treat- ment are assumed to initiate it.8 Other important assumptions (de- tailed below) regard the number of HIV tests conducted, access to services for the prevention of mother to child transmission of HIV, treatment of opportunistic infections, palliative care, inpatient care Figure 5a. People Living with HIV/AIDS (Ages 15+) and support to children orphaned by HIV/AIDS. 2010-2030 (declining HIV incidence) The uncertainty about the future course of the epidemic is reflected 50,000 Second-Line Treatment 45,000 First-Line Treatment in the following three scenarios: 40,000 Eligible, Not Receiving Treatment 35,000 Not Eligible for Treatment • Scenario  1 (declining HIV incidence): In this scenario 30,000 (broadly modeled in line with the objectives of the draft Na- 25,000 tional Strategic Plan and based on the projections of the Min- 20,000 istry of Health), the number of new adult infections declines 15,000 from 2,000 in 2010 to about 1,680 by 2018 (i.e., by 2.2 per- 10,000 cent annually) and more slowly thereafter at one  percent 5,000 annually. The HIV incidence rate declines from 0.14 percent 0 2010 2015 2020 2025 2030 (ages 15–49 years) in 2010 to 0.11 percent in 2030. Source: Author’s estimates and projections. • Scenario 2 (constant HIV incidence): The number of new infections remains constant at 2,000 from 2010. HIV inci- dence rates also remain broadly constant at about 0.14 per- cent, as the size of the population aged 15–49 years barely changes over the projection period (UNPD, 2011). Figure 5b. People Living with HIV/AIDS (Ages 15+) 2010-2030 (constant HIV incidence) • Scenario  3 (increasing HIV incidence): The number of 50,000 new infections increases from 2,000 in 2010 to 2,300 by 2018 Second-Line Treatment 45,000 First-Line Treatment and grows more slowly by one percent annually thereafter (to 40,000 Eligible, Not Receiving Treatment 2,600 or 0.19 percent of the population aged 15–49 years by 35,000 Not Eligible for Treatment 2030). Such an increase may reflect a reversal in sexual risk 30,000 behavior (as suggested by recent behavioral data) data or a 25,000 reduction in funding for HIV prevention activities. 20,000 15,000 Figure 5 summarizes the projections of the course of the epidemic 10,000 under the three scenarios. At first sight, these projections look 5,000 0 fairly similar. The number of people living with HIV/AIDS increases 2010 2015 2020 2025 2030 to 34,000 in Scenario 1 (declining HIV incidence), to 39,500 in Source: Author’s estimates and projections. Scenario  2 (constant HIV incidence), and to 45,200 in Scenar- io  3 (increasing HIV incidence). The principal factor behind this increase is the rising number of people receiving treatment, in- creasing from 7,900 in 2010 to between 24,900 (Scenario 1) or 30,600 (Scenario 3). Figure 5c. People Living with HIV/AIDS (Ages 15+) To allow comparisons between the scenarios, Figure 6 shows pro- 2010-2030 (increasing HIV incidence) jections across the scenarios of the number of new infections, the 50,000 Second-Line Treatment number of people living with HIV/AIDS not yet eligible for treat- 45,000 First-Line Treatment ment (based on national treatment guidelines) and the number 40,000 Eligible, Not Receiving Treatment of people receiving treatment. The projected numbers of new 35,000 Not Eligible for Treatment 30,000 infections (Figure  6a) diverge noticeably over the first few years, 25,000 differing by up to 18 percent by 2015. This spread—representing 20,000 a decline in infections in Scenario 1 relative to Scenario 3—subse- 15,000 quently grows to 30 percent (by 2020) and 43 percent (by 2030). 10,000 5,000 0 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. 8  A rate below 100 percent does not imply that patients reaching the need for second-line treatment are denied it; it also reflects loss to follow-up and to those whose treatment need is not diagnosed correctly. 18 III.  Projecting the Costs of the National Response to HIV/AIDS The changes in HIV incidence are—with a lag of a few years— Figure 6. State of Epidemic Across 3 Scenarios, closely reflected in the number of people not yet eligible for treat- 2010–2030 ment (Figure 6b).9 Between 2010 and about 2015, the number of HIV-positive people not yet eligible for treatment declines in each of the scenarios because HIV incidence had been coming down before 2010 (compare Figure 1). Subsequently, the number of peo- Figure 6a. New Adult HIV Infections, 2010-2030, ple not yet eligible for treatment follows the trends in HIV inci- 3 Scenarios 3,000 dence shown in Figure 6a. By 2020, the difference in the number of people not yet eligible for treatment between Scenarios 1 and 3 2,500 reaches 21 percent, and rises to 38 percent by 2030. 2,000 The number of people receiving treatment, however, barely differs 1,500 across the scenarios for the first 10 years of projections (Figure 6c). Scenario 1 By 2020, the gap between Scenarios 1 and 3 reaches five percent. 1,000 Scenario 2 By 2030, the gap rises to 19 percent, which is much smaller than the difference in the underlying trends in HIV incidence. This low 500 Scenario 3 variation reflects two factors—the long and variable time period 0 between an HIV infection and the need for treatment, and the long 2010 2015 2020 2025 2030 survival rates among people receiving treatment. Source: Author’s estimates. For the analysis of the fiscal sustainability of Jamaica’s HIV/AIDS program, two important outcomes arise from this analysis: • Successful prevention measures do not result in reduced spending on HIV/AIDS services in the short term. Figure 6b. HIV+ People Not Eligible for Treatment, 2010-2030, 3 Scenarios 14,000 • However, effective prevention measures that result in a de- cline in HIV incidence Scenario  1 do have an effect on fi- 12,000 nancial resource needs in the long-term. In Scenario  1 for 10,000 example, the number of HIV-positive people not yet eligible for treatment is reduced by 38 percent by 2030. As a con- 8,000 sequence, fewer people will require treatment—and absorb 6,000 Scenario 1 fiscal resources—over the following years. 4,000 Scenario 2 In closing this preliminary analysis, it is useful to point to some lim- 2,000 Scenario 3 itations of these scenarios. The emphasis is on the implications of 0 more or less successful HIV prevention policies for the state of the 2010 2015 2020 2025 2030 epidemic and the financial costs of the national response to HIV/ Source: Author’s estimates. AIDS, whereas other variables, notably access to treatment, is the same across the scenarios. This approach does not capture some of the possible interactions between treatment access and the success of prevention efforts. To attain a treatment coverage rate of 90 per- cent, it is necessary to reach widely into most-at-risk populations. In Figure 6c. People Receiving Treatment 2010-2030 this sense, there is a complementarity between successful preven- 3 Scenarios tion efforts and attaining high treatment coverage. 35,000 30,000 Scenario 1 Additionally, the analysis does not capture the potential implica- 25,000 Scenario 2 tions of treatment access on the health consequences of HIV/AIDS and the state of the epidemic if the coverage target is not reached Scenario 3 20,000 or sustained. A lower fiscal burden would then mask the deteriorat- 15,000 ing health consequences of HIV/AIDS. Additionally, improved treat- ment access and prevention outreach across population groups 10,000 could feed back into HIV incidence through channels that could be 5,000 assessed more explicitly. 0 2010 2015 2020 2025 2030 Source: Author’s estimates. Following an HIV infection, individuals progress to treatment need (at a 9  CD4 eligibility criterion of 350) after around 6 years. 19 Assessing the Financial Sustainability of Jamaica’s HIV Program Assumptions on Unit Costs and Coverage Rates As explained in the discussion on the costs of the national response Initiative, with the addition of allowances for the costs of human to HIV in 2009/10 and 2010/11 (above), the analysis of the costs of resources and facilities. For example, the annual costs of first-line the national response to HIV/AIDS draws from data on actual HIV/ treatment for adults were set at J$ 35,464 as of 2012, which AIDS spending (as compiled in the NASA), estimates of the unit costs comprises the costs of drugs (J$ 14,181), test kits (J$ 2,282) and of providing certain HIV/AIDS services, and public spending that partly human resources and facilities (J$ 19,000). The costs of drugs addresses the consequences of HIV/AIDS but is not captured by any for adult second-line treatment were set at J$ 76,207. The costs specific HIV/AIDS budget line items. The most important assumptions of drugs for pediatric treatment were higher, at J$ 18,526 for underlying the projections are: first-line treatment and J$ 88,104 for second-line treatment as of 2012. Additionally, allowances for the costs of non-antiretro- • Prevention. Data in the draft NSP suggest that about 300,000 viral treatment, palliative care, and inpatient care were included HIV tests were conducted in 2010. Based on this, the projections in the estimates, which items account for about 20 percent of assume that 15 percent of people aged 15–49 years received the total costs of treatment and care. an HIV test in 2010 and that this share will increase to 20 per- cent by 2015.10 Unit costs of HIV tests (J$ 255) are based on • Social mitigation. The NASA identifies only a small amount local estimates provided by the Clinton Health Access Initiative. of social mitigation spending.11 However, some social support Additionally, an allowance for the costs of human resources provided by the government is not HIV-specific, even though the and facilities (J$ 425) has been included in the cost estimates. needs that are supported are partly caused by HIV/AIDS. Based For the prevention of mother-to-child transmission (PMTCT) of on the assumption that one-half of the budget of the CDA is HIV, the coverage of screening was set at 95 percent, in line with allocated to orphans and vulnerable children, an average annual the National Strategic Plan 2008–12. Based on Ministry of Health cost of J$ 10,741 per orphan was included in the projections.12 estimates and projections of the numbers of mothers receiving PMTCT, the coverage rate of PMTCT was set at 73 percent in • Administration of HIV program. Based on the NASA, the costs 2010, rising to 95 percent by 2015 and remaining at that level. of operating the HIV/AIDS program, monitoring and evaluation, The costs of other prevention measures included in the projec- and all activities not included elsewhere, were set at 20 percent of tions are extrapolated from current spending and assumed to the costs of prevention, treatment and care, and social mitigation. increase in line with population growth. • Change in costs over time. The costs of drugs and test kits • Treatment and Care. The costs of antiretroviral treatment are remain constant. Most other services are intensive in human based on local estimates provided by the Clinton Health Access resources. Unit costs therefore rise in line with GDP per capita, to capture the effects of rising wages. Projected Costs of HIV/AIDS Program Based on the epidemiological projections (as well as underlying as- Almost all of this increase is accounted for by the increasing costs of sumptions on the coverage rates of treatment and other HIV/AIDS treatment and care, rising from J$ 470 million in 2010 to between services) and the estimates of unit costs, it is possible to project the future costs of the national response to HIV/AIDS. This analysis is in- tended to address two issues: 10  Preliminary data from the 2012 “HIV Knowledge Attitudes and Behavior Survey,” quoted by the Ministry of Health and the National HIV/STI Pro- • Projected costs and funding needs. To appreciate the fis- gramme (2012), which show a steep increase in access to testing, suggest cal burden, evaluate program alternatives and consider financ- that this rate may have to be adjusted upwards. ing options, it is necessary to understand how the costs of the 11  For example, in 2010/11, the NASA reports J$ 2.5 million for orphans and vulnerable children, J$ 7.1 million for “social protection through provision HIV/AIDS program evolve of social services,” and J$ 11.3 million for HIV-specific income generation projects. • Impact of prevention outcomes. To assess how epidemi- 12  The costs per orphan actually receiving care through public services are ological outcomes based on alternative assumptions on HIV higher. In the absence of sufficient data about the number of orphans sup- incidence translate into spending needs ported through the public sector, estimates are based on the average cost per orphan to the public sector (public spending on orphans, divided by to- Figures 7 and 8 summarize the projections over the period tal number of orphans in the country). The costs are then attributed to the 2010–30 (see also Appendix Table 1). In all scenarios, the costs national response to HIV/AIDS in line with the estimated share of children rise steeply—from J$ 1.7 billion in 2010 to between J$ 3.1 billion orphaned by HIV/AIDS among the total number of orphans. (Scenario  1) and J$  3.4  billion (Scenario  3) in 2030 (in constant 13  There is a small discrepancy in the costs relative GDP estimated here for 2010 prices). As GDP is projected to increase over this period, the 2010, and the estimates of HIV/AIDS spending reported above for fiscal year increase relative to GDP is less pronounced, with spending needs 2010/11. This discrepancy (0.15  percent of GDP vs. 0.14  percent of GDP) rising from 0.15 percent of GDP to between 0.20 percent of GDP arises because the estimates and projections are set up on a calendar year basis, whereas the spending figures relate to the fiscal year. and 0.22 percent of GDP.13 20 III.  Projecting the Costs of the National Response to HIV/AIDS Figure 7. Cost of National Response to HIV/AIDS, 3 Scenarios, 2010–2030 Figure 7.1.a. Costs of National Response to HIV/AIDS Figure 7.1.b. Costs of National Response to HIV/AIDS 2010-2030 (declining HIV incidence) 2010-2030 (declining HIV incidence) 3,500 Other program expenses Other program expenses Mitigation J$ Millions 2010 Prices 3,000 Mitigation 0.20 Treatment and care Percent of GDP Treatment and care Prevention 2,500 Prevention 0.15 2,000 1,500 0.10 1,000 0.05 500 0 0.00 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. Source: Author’s estimates and projections. Figure 7.2.a. Costs of National Response to HIV/AIDS Figure 7.2.b. Costs of National Response to HIV/AIDS 2010-2030 (constant HIV incidence) 2010-2030 (constant HIV incidence) 3,500 Other program expenses Other program expenses Mitigation J$ Millions 2010 Prices 3,000 Mitigation 0.20 Treatment and care Percent of GDP Treatment and care Prevention 2,500 Prevention 0.15 2,000 1,500 0.10 1,000 0.05 500 0 0.00 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. Source: Author’s estimates and projections. Figure 7.3.b. Costs of National Response to HIV/AIDS Figure 7.3.a. Costs of National Response to HIV/AIDS 2010-2030 (increasing HIV incidence) 2010-2030 (increasing HIV incidence) 3,500 Other program expenses Figure 7.3.b. Costs of National Response to HIV/AIDS Other program expenses Mitigation J$ Millions 2010 Prices 3,000 2010-2030 Mitigation (reversal in HIV incidence) 0.20 Treatment and care Prevention Percent of GDP Other expenses programand Treatment care 2,500 Mitigation 0.20 Prevention 0.15 2,000 Treatment and care Percent of GDP Prevention 0.15 1,500 0.10 1,000 0.10 0.05 500 0.05 0 0.00 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. 0.00 Source: Author’s estimates and projections. 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. 21 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 8. Costs of National Response to HIV/AIDS 2010-2030 (across scenarios) 3,500 3,000 2,500 2,000 Scenario 1 1,500 Scenario 2 1,000 Scenario 3 500 0 2010 2015 2020 2025 2030 Source: Author’s estimates. J$ 1.5 billion and J$ 1.8 billion by 2030. The increase in the costs of the number of people receiving treatment (19 percent) or the number treatment is steeper than the increase in the number of people receiv- of new infections (43 percent). Nevertheless, the difference in costs of ing treatment, because an increasing share of people on treatment 11 percent between the two scenarios will not be insignificant when receives more expensive second-line treatment. it actually comes to financing of the national HIV program. For example, in Scenario 1, the number of people receiving treatment This low variability reflects the long lag between HIV infections and rises from about 8,000 in 2010 (of which eight percent receive sec- treatment need. This applies, in particular, to the costs of second-line ond-line treatment) to about 20,000 in 2020 (of which 24 percent are treatment, which become increasingly important over the projection on second-line treatment) and to 25,000 in 2030 (of which 38 per- period, and for which transition from infection (via first-line treatment) cent are on second-line treatment. Largely as a result of this compo- to the need for second-line treatment can take well over a decade. sition effect, the average annual unit costs of antiretroviral treatment Additionally, it is assumed that prevention spending remains the same rise from J$  42,400 to J$  59,200 over this period, accounting for across scenarios, so that total spending changes less across scenarios J$ 13,900 of the total increase of J$ 16,800 (with the balance reflect- than the costs of treatment. ing increasing wage costs).14 Comparing the scenarios (Figure 8), the differences in costs are re- 14  In Scenario 3, with people receiving treatment increasing to 30,600, the markably small—by 2030, the total costs of the national response share of people receiving second-line treatment by 2030 is somewhat low- to HIV/AIDS in the scenario with declining HIV incidence (which is er (34.4 percent). This is owing to the fact that many of the additional HIV broadly in line with the objectives of the draft NSP) are only 11 per- infections in this Scenario occur relatively late, and many of the additional cent lower than for the scenario envisaging a substantial increase in individuals infected do not proceed to second-line treatment by the end of HIV incidence. This difference is much smaller than the differences in the projection period. Box 3. Macroeconomic Impact of HIV/AIDS The economic framework applied in this study includes a macroeconomic model, for two reasons. First, the projections of the costs of the HIV program extend over two decades. To assess the magnitude of these costs it is necessary to take into account the fact that GDP (and the government’s domestic revenues) is expected to increase substantially over this period. Second, the negative impact of HIV/AIDS on economic growth may undermine the government’s capacities to finance the response to HIV/AIDS. The macroeconomic projections utilize a simple neoclassical growth model. The projections on the growth of the working age pop- ulation were obtained from UNPD (2011). Labor productivity is assumed to increase after 2016 at a rate of 1.5 percent annually; until then, the parameters are set to replicate growth projections from the IMF (2011b) and the savings rate is set at 20 percent, broadly in line with national accounts. HIV/AIDS is assumed to affect GDP growth in several ways: (1) AIDS-related mortality affects productivity (one percent additional mor- tality reduces productivity by 0.5 percent); (2) HIV/AIDS slows down the growth of the working-age population; (3) increased spending on HIV/AIDS reduces national savings proportionally (at a savings rate of 20 percent, J$ 1 million of additional HIV/AIDS spending crowds out J$ 200,000 of investment and J$ 800,000 of consumption spending). This model suggests that HIV/AIDS does have a small impact on GDP growth, largely because increased mortality reduces the growth rate of the working-age population and that the impact on GDP per capita is very small. The model predicts that GDP will be 2.8 percent smaller in 2030 than in the absence of HIV/AIDS. This is largely accounted for by the smaller size of the working-age population (minus 2.4 percent, due to the accumulated effect of higher mortality), whereas GDP per capita is reduced by 0.3 percent. 22 III.  Projecting the Costs of the National Response to HIV/AIDS Some Policy Lessons There are a number of lessons for HIV/AIDS policies and the financing spread over several decades, progress in reducing HIV inci- of the HIV/AIDS program that can be drawn from the analysis of the dence does not immediately reduce the costs of care, treat- projected costs of the HIV/AIDS program, namely: ment, and social support. • The most important driver of the rising costs of • The projected costs, however, understate the differ- the HIV program is the cost of treatment and, ulti- ences in the fiscal consequences across scenarios. mately, the number of HIV infections. Treatment costs Spending in Scenario 1 (with declining HIV incidence) is only are a consequence of earlier HIV infections, although long 11 percent lower in 2030 than in Scenario 3 (with an increase lags occur between infection and need of treatment. To bet- in HIV incidence). At that time, the number of HIV-positive ter understand the cost-effectiveness of HIV interventions, people not yet eligible for treatment (based on treatment the next section will provide explicit estimates of the con- guidelines), but who are in line to initiate treatment over the sequences of new HIV infections in terms of the resulting next couple of years, is reduced by one-third (and the number spending needs of new infections by 43  percent). Therefore, spending com- mitments to serve this impending need are greatly reduced • Improved prevention efforts reduce demand for HIV/ between these scenarios. The next section will provide tools AIDS services in the long-term, but do not provide a to measure the fiscal consequences of alternative HIV/AIDS short-term fix to contain the funding needs of the HIV/ policies more precisely, taking into account these spending AIDS program. As the consequences of new HIV infections— commitments. in terms of an increased demand for HIV/AIDS services—are 23 IV.  HIV Incidence and the Costs of the HIV/AIDS Response IV.  HIV Incidence and the Costs of the HIV/AIDS Response The demand for HIV/AIDS services and the costs of the national response are ultimately driven by HIV infections, and a large share of the budget of the National HIV/AIDS Program is dedicated to HIV preven- tion efforts. The present section analyzes the link between HIV incidence and the costs of the HIV/AIDS program in three stages: • The costs incurred by each additional HIV infection. These costs are estimated as the actuarially fair value of providing HIV/AIDS services to a person living with HIV/AIDS from the time of infection to the time of death, in line with the epidemiological model and treatment coverage rates targeted by the National HIV/AIDS Program. • Cost-effectiveness of HIV prevention interventions. Based on the estimates of the costs incurred by an addi- tional HIV infection, the cost-effectiveness of HIV prevention interventions can be assessed. • Prevention policies and the drivers of HIV incidence. Each infection that is prevented means that an individual not infected will also not pass on HIV to. This aspect is particularly important among high-risk groups, where high-risk sexual behavior is also associated with a higher probability of onward transmission of HIV, including to people outside the respective risk group. The Costs Incurred by each Additional HIV Infection An HIV infection results over time in additional demand for HIV/AIDS-related services and the consequent additional fiscal costs of providing them in line with the objectives of the national response to HIV/AIDS. The overall costs of meeting the demand for HIV/AIDS services for all people living with HIV/AIDS were estimated in the previous section. The same framework can be used to estimate the costs that are incurred by one addi- tional infection. Figure 9 shows such estimates, which were obtained by adding one HIV infection occurring in 2012 to the projections and calculating how this additional infection subsequently increases the demand for HIV/AIDS services and the costs of the national response to HIV/AIDS.15 The analysis proceeded in three steps: (1) the projected costs of the national response were recorded; (2) one additional HIV infection was added in 2012, 15  and the new projected costs were obtained; and (3) the costs incurred by the additional HIV infection are obtained as the difference between the program-level costs obtained in steps (1) and (2). As the consequences of an HIV infection differ somewhat between males and females, the results shown are a weighted average of the costs incurred by an additional HIV infection among males and females, respectively. 25 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 9. Jamaica: Costs of One Additional Infection, 2012 (J$, 2010 prices) 45,000 40,000 Total 35,000 30,000 Treatment 25,000 20,000 15,000 10,000 5,000 0 2012 2017 2022 2027 2032 2037 2042 2047 Source: Author’s calculations. The expected costs caused by an additional HIV infection in 2012 After around 15  years, expected costs decline, primarily reflecting rise to about J$ 40,000 annually over the 15 years following an in- declining survival. At the same time, there is an increasing share of fection. This build-up in costs is mainly accounted for by treatment people receiving treatment who switch to more expensive medica- costs, reflecting the increasing probability that the person will require tion, so that the average unit costs of treatment increase, and the and receive (first-line) treatment. While most people deemed eligible expected costs caused by one infection decline over time more slowly will receive treatment (assumed coverage rate of 90 percent), the ex- than survival does.16 pected costs are lower than the unit costs of treatment because of the high rate of emigration (especially in the age groups in which HIV These costs incurred over time by an HIV infection are relevant for the incidence is also high), and because of the fact that the time from assessment of the cost-effectiveness of HIV prevention interventions, infection to the need for treatment differs across people living with as the costs of the prevention intervention are at least partly offset HIV/AIDS. The other cost drivers are social mitigation (mainly orphan by the subsequent financial savings owing to the diminished need for support) and the management of the HIV program (overhead). HIV/AIDS-related services. This point is developed more fully in the next section. Assessing the Cost-Effectiveness of HIV Prevention Interventions It is challenging to assess the cost-effectiveness of HIV prevention equal to J$ 500,000 (rounded to the nearest J$ 1,000), or US$ 5,800, interventions, because it is necessary to take into account the im- roughly the level of GDP per capita. proved health outcomes, the costs of the intervention and the finan- cial savings from the reduced demand for HIV/AIDS services which are This means that preventing one HIV infection at a cost of less than spread over several decades (as discussed in the previous section). In J$ 500,000 (US$ 5,800) reduces the financial costs of the national re- order to establish the net costs of the intervention, i.e., the costs of sponse to HIV/AIDS—it is a good financial investment, even without the intervention minus the financial savings which result from it, the taking into account the positive health effects. For an HIV infection financial savings—which are spread over many years—need to be averted at a higher cost, the net costs are obtained by subtracting the translated into an equivalent one-off amount to arrive at the actu- financial savings of J$ 500,000 from the overall cost. For example, if arially fair value of the financial savings, namely the one-off amount an intervention prevents one HIV infection at a cost of J$ 600,000, that is equivalent to the savings occurring over time. Equivalently, it the health gains arrive at a net cost of J$ 100,000, and if it costs less is the one-off amount that the government would need to set aside than J$ 500,000, it saves money. at the time of infection to pay for the costs of HIV/AIDS services of In many cases, the effects of an HIV intervention cannot be inter- one patient during the course of the disease. As this one-off amount preted as an infection prevented forever, as the beneficiaries of a would yield interest until it is used to pay for HIV/AIDS services, it is prevention intervention may contract HIV in the future. The estimates necessary to discount the expected future costs at the interest rate of the costs incurred by one additional infection can also be used in relevant for the government. The actuarially fair value of the costs incurred by one infection (or savings from one infection prevented) is therefore obtained by adding up the discounted expected costs in- 16  One important issue that this study does not cover in detail is the interac- curred by one infection over time. tion between treatment access and the quality of HIV/AIDS-related services and disease progression. Improved management and delivery of treatment For reasons explained in Box 4, the estimates offered here are based may prolong the duration of first-line treatment, and delay onset of treatment on an interest rate of five percent, which is at around the level of the failure and transition to more expensive second-line treatment. Thus, there is real interest rate that the government pays for its public debt. At this a potential for improving survival rates, while reducing the expected costs of interest rate, the estimated costs incurred by one HIV infection are second-line treatment that is not explored in this report. 26 IV.  HIV Incidence and the Costs of the HIV/AIDS Response Box 4. Estimating the Financial Costs of New Infections The estimates of the cost incurred by new infections and the fiscal space committed by the HIV program used in this report are based on an actuarially fair estimate of the expected cost of meeting the demands for HIV/AIDS services caused by a new infection. This is the amount that the government would hypothetically agree to pay to some agency at the time an infection occurs to cover the cost of meeting the demand for HIV/AIDS services for that patient. This actuarially fair estimate discounts future HIV/AIDS spending at the interest rate relevant for the government. As the cost projec- tions are specified in real terms (at constant prices), the relevant interest rate is the real interest rate on public debt. To calculate this actuarially fair estimate, it is necessary to choose the appropriate interest rate. The real interest rate on public debt has averaged 7.9 percent for debt denominated in Jamaican dollars and 4.1 percent for debt denominated in foreign-currencies—5.8 per- cent overall. These averages may not be representative, however, as they include a spike in the domestic interest rate in 2009/10. The projections offered in this report are based on a real interest of five percent. This might be too high if the fiscal situation improves and the risk premium on Jamaica’s debt (5.2 percent in late 2010, according to IMF (2011a)) improves. On the other hand, global interest rates are low at present and might rise over the projection period, also raising the interest rate on Jamaica’s public debt. The choice of interest rate has implications for the estimated costs incurred by additional infections. At an interest rate of five percent, this is estimated at J$ 500,000, or US$ 5,800. At a lower interest rate of three percent, the estimated costs would be higher (as future spending is valued higher), at J$ 691,000. At an interest rate of seven percent, the costs incurred by a new infection are equivalent to J$ 377,000. Public debt and interest 2008/09 2009/10 2010/11 Public debt (percent of GDP) 108.9 119.1 118.6 Denominated in J$ 45.9 51.5 50.1 Denominated in foreign currency 63.0 67.6 68.5 Nominal interest rate on public debt Domestic 18.3 26.7 15.2 Foreign 6.1 6.0 5.0 Inflation (consumer price index) Jamaica 12.4 9.7 11.4 Foreign (United States) –0.4 2.3 2.7 Implied real interest rate Domestic 5.3 15.5 3.4 Foreign 6.5 3.6 2.3 Data sources: Jamaican authorities, as reported by IMF (2011a). The interest rate on public debt was calculated based on annual interest payments. U.S. inflation was obtained from U.S. Bureau of Labor Statistics (http://www.bls.gov/cpi/). situations in which an HIV infection is delayed rather than prevented screening of pregnant women for HIV and providing treatment for forever. A delay in an HIV infection means that the expected costs of the prevention of mother-to-child transmission to women requiring providing HIV/AIDS services occur later and the amount that would it) are estimated at around J$  33  million in 2012, which results in need to be put aside now to meet these future costs declines (as the about fifty HIV infections averted in that year. Preventing these fifty money put aside could earn some interest for longer). For example, HIV infections results in a financial saving equivalent to J$ 24.2 mil- a delay in an HIV infection by one year results in a financial saving lion. The net costs of the PMTCT program in that year therefore equals equivalent to J$ 25,000, or about US$ 290.17 J$ 8.8 million, and the net costs per pediatric HIV infection averted is equal to J$ 176,000. To illustrate how these methods can be used in practice, consider the cost-effectiveness of Jamaica’s program for the prevention of moth- er-to-child transmission (PMTCT) of HIV. As disease progression and 17  As the expected costs are also delayed by one year, all costs are addition- unit costs of treatment differ from adults, the cost savings of one HIV ally discounted by one year’s interest. The financial savings from a one-year infection prevented at or around birth is also different—estimated delay are therefore calculated as one year’s interest rate (here: five percent) at J$ 483,000 (equivalent to US$ 5,600). The total costs of the pro- times the costs incurred by one infection (J$  500,000), which is equal to gram for the prevention of mother-to-child transmission of HIV (i.e., J$ 25,000. 27 Assessing the Financial Sustainability of Jamaica’s HIV Program Prevention Policies and the Drivers of HIV Incidence The analysis so far has focused on the consequences of an HIV infec- Table 3. Costs Incurred by HIV Infections, by tion in terms of the costs of providing HIV/AIDS-related services for Population Group the individual infected. However, to assess the cost-effectiveness of (in J$) (in US$) HIV/AIDS policies, it is necessary to understand the implications of an intervention on the population level. This distinction is important—an Direct costs incurred by HIV infection 500,000 5,800 individual contracting HIV may infect further people, augmenting the Additional costs from “downstream” consequences of the initial HIV infection. This point is particularly rel- infections evant in Jamaica, which is facing a concentrated HIV epidemic, and Low-risk individual (male or female) 361,000 4,200 where certain population groups play a disproportionate role in the spread of the epidemic. Casual sex (male) 407,000 4,700 Casual sex (female) 303,000 3,500 The ongoing “modes-of-transmission” (MOT) study18 offers an op- Female sex worker 2,115,000 24,600 portunity to analyze the consequences of additional HIV infections across population groups, including the risk that an infected individ- Men who have sex with men 4,832,000 56,300 ual passes on the virus. The MOT study differentiates between female Source: Author’s calculation. sex workers, their male clients and the female partners of their clients; men who have sex with men and their female partners; individuals engaging in casual heterosexual sex and their partners, low-risk het- ulations, the increased risk of further HIV infections adds between erosexual individuals and no-risk individuals who are not sexually J$ 303,000 and J$ 407,000, or an additional 60 percent to 80 per- active. HIV incidence, and the probability that an individual passes on cent, to the costs directly incurred by a new HIV infection. For these the virus, differ vastly across these groups. For example: groups, the total financial savings from one HIV infection prevented are therefore between J$ 803,000 to J$ 907,000. This is the value • HIV incidence among (non-infected) men who have sex that should be applied in assessing the cost-effectiveness of HIV in- with men is estimated at 2.7 percent. HIV-positive men who terventions targeting the general population. have sex with men, over the course of their infection, infect 1.8 additional men who have sex with men and 0.4 of their For female sex workers, the cost of downstream infections are es- female partners. Coverage of preventive behavior is estimated timated at around J$ 2.1 million (US$ 24,600), four-fold the direct at 50 percent. costs. This reflects the risk that a female sex worker infects her clients, who may pass on the virus to more female sex workers (and so on) • HIV incidence among female sex workers, where condom and to their regular partners. These high costs already reflect an es- coverage is believed to be high (about 90 percent) is estimated timated coverage rate of condom use of 90 percent; without it, the at 0.3 percent. According to the MOT study, each female sex costs of onward infections would be higher. worker may infect 2.2 clients, and each client would pass on the virus to 0.6 of their female partners and to 0.2 sex workers. Estimating the costs on “downstream” infections for men who have sex with men is challenging because the draft MOT study estimates • Individuals engaging in casual heterosexual sex are ex- that HIV incidence in this group is on an explosive path (with each pected to infect 0.5 others through sexual activity. Additionally, new infection resulting in 1.8 infections in the next round, and so male individuals infect 0.4 female partners from the low-risk on). This rate would come down as the HIV prevalence among men population. who have sex with men rises and fewer remain to be infected. Even- • Low-risk individuals face a low-risk of contracting HIV tually, HIV prevalence among men who have sex with men stabilizes (0.1 percent), but if infected are likely to pass the virus on to at a level where each man who has sex with men infects about one their partners, as condom use in this population group is low. other. To estimate the expected costs of downstream infections, it is HIV incidence in the low-risk population is about the same as assumed that each infection among men who have sex with men now among the population engaging in casual sex, owing to lower directly results in 1.8 “downstream” infections but that this rate will rates of condom use (17 percent, compared to 65 percent in come down to one in later rounds (as HIV prevalence has risen in be- the casual-sex population). tween). On this basis, the costs from “downstream” infections among men who have sex with men add up to J$ 4.8 million (US$ 56,300)— Table 3 summarizes the estimated costs incurred by new infections, almost ten-fold the direct costs. including allowances for the costs resulting from “downstream” in- fections. The latter are calculated based on the expected number of people to whom an individual newly infected passed the virus on The MOT study calculates the expected number of new infections per year 18  to, and so on, and the expected costs of providing HIV/AIDS-related on the basis of the current distribution of infections and patterns of risk within services to the people infected, applying (as elsewhere in this study) a population. It uses data from different studies to generate an estimate of the a discount rate of 5  percent. For the low-risk and casual-sex pop- specific populations contributing to HIV incidence. 28 V.  Fiscal Evaluation of the Costs of the HIV/AIDS Program V.  Fiscal Evaluation of the Costs of the HIV/AIDS Program The discussion so far has focused on two different perspectives on the fiscal burden of the national re- sponse to HIV/AIDS, namely: (1) the projected costs of the national response to HIV/AIDS over the period 2010 to 2030 and (2) the link between HIV incidence and the subsequent costs of providing HIV/AIDS services. The current section merges these approaches, attributing the costs of the national response to HIV/AIDS explicitly to HIV infections. Even though most of the demand for HIV/AIDS-related services is ultimately caused by HIV infections, the long lags between infections and need for treatment or other services means that the costs of the national response to HIV/AIDS are barely responsive to changes in HIV incidence in the short-term. In the three scenarios discussed earlier, a large change in HIV incidence (43 percent by 2030) resulted in a reduction of spending of only 11 percent in 2030. An alternative view of the fiscal costs of the response to HIV/AIDS is developed in this section that attrib- utes most HIV/AIDS spending directly to the underlying HIV infections. This helps to resolve the apparent inconsistency that most HIV/AIDS spending is caused by HIV infections, but that HIV/AIDS spending barely budges as the number of HIV infections changes, and to gain a better understanding of the fiscal conse- quences of alternative HIV/AIDS policies. This analysis is based on the discussion in the previous section on the costs incurred by each individual infection, which provided estimates of the actuarially fair costs incurred by one additional infection. If each infection in 2012 incurs a cost of J$ 500,000 and there are 2,000 new HIV infections, then the total fiscal costs incurred by new infections in 2012 is J$ 1,000 million. This is the amount the government would have to set aside in 2012 to cover the anticipated costs of providing HIV/AIDS services to all people who got infected in 2012. Thus, the policy commitments of the National HIV/AIDS Program are interpreted as spending commitments. For Scenario 1, the costs incurred by new infections in 2012 are estimated at J$ 970 million (1,940 infec- tions at J$ 500,000 each). Additionally, it is necessary to take into account the spending that does not di- rectly meet the needs of people living with HIV/AIDS and cannot therefore be attributed to the underlying 29 Assessing the Financial Sustainability of Jamaica’s HIV Program HIV infections. This population-based spending—largely prevention After 2015, treatment coverage rates are assumed to be constant spending targeting the general population or sub-populations, but (and the same across scenarios), so that the changes in the costs not necessarily people living with HIV, as well as certain administrative incurred by new infections more directly reflect the course of HIV in- expenses—is estimated at J$ 805 million (US$ 9.5 million) in 2012. cidence. In Scenario 1, spending commitments caused by new infec- The total costs of the National HIV/AIDS Program on a commitment tions decline relative to GDP from 2015 to 2030, broadly in line with basis therefore adds up to J$ 1,775 million (US$ 20.9 million) under declining HIV incidence (the other factors at play are GDP growth and Scenario 1. Looking ahead, the total financial commitments under the changes in unit costs of HIV/AIDS services, some of which include hu- HIV/AIDS program can be calculated similarly. The costs over time of man resource costs and are linked to GDP per capita). In this scenario, serving all people living with HIV/AIDS who are currently alive add up the costs incurred by new HIV infections reach 0.05 percent of GDP by to 3.4 percent of GDP.19 This number has been calculated in the same 2030, and the total costs of the HIV program (on a commitment basis) way that the costs incurred by a new infection were calculated, as the decline to 0.12 percent of GDP over this period. actuarially fair value that the government would have to set aside now to cover the future costs of HIV/AIDS services to people currently In Scenario  3, by contrast, the costs incurred by new infections living with HIV/AIDS, plus population-based spending. remain higher, at about 0.09  percent of GDP over the projection horizon,21 and consequently the total costs of the HIV program Similar to the earlier discussion on the cost-effectiveness of specific (on a commitment basis) remain at 0.17 percent of GDP. The cost prevention interventions, the current analysis can be used to evaluate savings from the decline in HIV incidence between Scenario 3 and the cost-effectiveness of alternative HIV/AIDS programs. Figure  10 Scenario 1 thus result in fiscal savings which rise to 0.04 percent shows current HIV/AIDS spending and new spending commitments of GDP (about one-quarter of the costs of the HIV/AIDS program) caused by HIV infections for Scenario  1 (declining HIV incidence) by 2030. and Scenario 3 (increasing HIV incidence). Actual spending increases throughout the projection period (from 0.15 percent of GDP in 2010 Figure 11 illustrates the differences between the scenarios more di- to 0.20 percent of GDP in 2030 in Scenario 1, and 0.22 percent of rectly. By 2015, when actual spending barely budges, the costs in- GDP in Scenario 3). Coincidentally, the financial commitments caused by new HIV infections are initially about equal to current spending on people living with HIV/AIDS, at 0.08 percent of GDP.20 19  For some purposes, it is useful to also include the estimated costs incurred by projected future infections, and not only HIV infections which have already From this time on, however, the financial commitments caused by occurred, in the estimate of fiscal space absorbed by the national response new infections diverge between Scenario 1 and Scenario 3. Between to HIV/AIDS. This wider estimate of the fiscal costs of HIV/AIDS adds up to 2010 and 2015, they remain broadly constant in Scenario 1, reflect- 4.8 percent, of which 1.4 percent of GDP represent the expected costs of ing two opposing forces—treatment coverage rates are assumed to projected infections. increase in this period (so that the expected costs caused by each HIV 20  This reflects two factors: (1) Without changes in underlying factors (HIV infection increase), while HIV incidence declines. In contrast, the costs incidence, coverage rates) the costs incurred by new infections would be lower incurred by new HIV infections increase in Scenario 3, and by 2015 than current spending, because the former involve discounting. (2) Because of the ongoing scaling-up of treatment, people newly infected have a bet- the additional costs implied by the higher number of HIV infections ter chance of obtaining treatment, while cohorts who acquired HIV earlier (relative to Scenario 1) amount to J$ 205 million (US$ 2.4 million)— have been depleted because of a relative lack of treatment (reducing current in this scenario. The costs of the HIV/AIDS program, taking into ac- spending). count the spending commitments implied by new HIV infections, is 21  While the costs incurred by new infections increase in absolute terms in thus 11 percent higher by 2015, compared to Scenario 1 in which HIV Scenario 3, in line with the increasing HIV incidence, GDP also grows over incidence declines. the projection period, and the costs remain broadly constant relative to GDP. Figure 10. Fiscal Costs of HIV/AIDS, Commitment Basis,2010–2030 Figure 10a. Jamaica: Fiscal Costs of HIV/AIDS, “Commitment Figure 10b. Jamaica: Fiscal Costs of HIV/AIDS, “Commitment Basis”, 2010-2030 (Percent of GDP)—Scenario 1 Basis”, 2010-2030 (Percent of GDP)—Scenario 3 0.25 0.25 0.20 0.20 Actual Expenditures Actual Expenditures Total Costs, 0.15 0.15 Current costs of Total Costs, “commitment basis” meeting demand by Costs incurred by Current costs of new infections “commitment basis” meeting demand by people living with 0.10 Costs incurred by 0.10 HIV/AIDS people living with new infections HIV/AIDS 0.05 Population-based spending 0.05 Population-based spending 0.00 0.00 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. Source: Author’s estimates and projections. Figure is based on Scenario 1 in which HIV incidence declines Figure is based on Scenario 3 in which HIV incidence increases 30 V.  Fiscal Evaluation of the Costs of the HIV/AIDS Program Figure 11. Fiscal Costs of HIV/AIDS, Current and Commitments, Across Scenarios (Percent of GDP) 0.25 Actual Spending 0.20 0.15 Spending on 0.10 Scenario 1 Commitment Basis Scenario 2 0.05 Scenario 3 0.00 2010 2015 2020 2025 2030 Source: Author’s estimates. curred by new infections in the scenario with reduced HIV incidence • Similar to the analysis of the costs incurred as a consequence are already reduced by over J$  100  million (0.01  percent of GDP) of individual HIV infections, which provides the basis for an compared to the scenario with constant HIV incidence, and by over analysis of the cost-effectiveness of specific HIV interventions, J$  200  million compared to the scenario envisaging an increase the program-level analysis provides a tool (together with the in HIV incidence (at constant 2010 prices). This gap increases to resulting health outcomes) to rank alternative policies in terms J$ 196 million by 2020 and J$ 294 million by 2030 (relative to the of their cost-effectiveness. scenario with constant HIV incidence), or to J$ 379 million by 2020 and J$ 623 million by 2030 (relative to the scenario envisaging an • The present analysis merely highlights the consequences of increase in HIV incidence). higher or lower effectiveness of HIV prevention interventions but does not specify where this derives from. Together with a These findings are relevant for the design and evaluation of HIV/AIDS more specific policy framework, and more explicit epidemio- policies in various ways: logical modeling, the analysis supports the design of cost-ef- fective HIV/AIDS policies, linking the costs of alternative HIV • The analysis provides accurate estimates of the financial impli- prevention interventions, their outcomes for the course of the cations of alternative HIV/AIDS policies, including the implica- epidemic, and the consequences in terms of the costs of the tions of changes in HIV incidence for spending needs. resulting demand for HIV/AIDS services. 31 VI.  Sustainability of the National Response to HIV/AIDS VI.  Sustainability of the National Response to HIV/AIDS The financial sustainability of the national response to HIV/AIDS in Jamaica faces challenges that arise primarily from two sources: • The projected increase in the costs of providing HIV/AIDS services, largely because the number of peo- ple eligible for and receiving treatment expands greatly over the coming years. • The outlook for Jamaica in obtaining external funding is uncertain. As availability of external funding is expected to decrease, the need for domestic financing to sustain the HIV/AIDS program and meet the increased demand for services will increase. These challenges occur against a background of tight public finances, so that claims for additional gov- ernment funding will be closely scrutinized. The analysis thus far makes a contribution to dealing with these challenges (1) by focusing on quantifying the financing challenges, offering projections of the costs of the national response to HIV/AIDS until 2030, and (2) by describing some tools for assessing the cost-effectiveness and allocative effectiveness of HIV/AIDS interventions and HIV/AIDS programs. The current section pulls together some of the lessons regarding the current and impending policy challenges, looking at: • The implications of reduced availability of external financing. • The cost-effectiveness of HIV/AIDS interventions. • The allocative effectiveness of HIV/AIDS interventions, especially of investments in HIV prevention. • Financing of expanding treatment needs. Implications of Reduced Availability of External Financing External support has played a large role in the financing of Jamaica’s national response to HIV/AIDS. Ac- cording to the latest available data (Figure 12, also recall the discussion in Section 2), external financing accounted for around two-thirds (68.8 percent) of the costs of the national response to HIV/AIDS. By far the most substantial financial contribution was from the Global Fund (48.8 percent of total spending, 33 Assessing the Financial Sustainability of Jamaica’s HIV Program Figure 12. Composition of Financing of National Response to HIV/AIDS, 2010 (Percent of Total Public HIV/AIDS Spending) World Bank (10.0%) US Goverment (0.8%) Other External (4.2%) Government of Jamaica (32.3%) Total: J$ 1,606 million, Global Fund equivalent to (52.4%) 0.13 percent of Private GDP or (0.2%) US $19.0 million Source: Draft NASA and author’s calculations and 71  percent of external support), followed by the World Bank The changing global funding environment will already have a sig- (9.3  percent of total spending) and the United States Government nificant impact on the financing of Jamaica’s HIV/AIDS program over (1.4 percent of total spending).22 The Government of Jamaica accounted the next few years. Contributions from the Global Fund will be much for almost one-third of total spending (31.0 percent), while private do- reduced according to the new funding criteria introduced in 2011; it mestic spending does not play a role in recorded HIV/AIDS spending.23 is also by no means certain that Jamaica will obtain any new funding from the Global Fund at all. These criteria mean that Global Fund sup- Most external funding is not secured beyond the near future. The World port cannot exceed two-thirds of the budgetary resources committed Bank project will close in November 2012, the Global Fund grant will by the Government of Jamaica.24 end in 2013 and the bulk of the support from the United States Gov- ernment will terminate in 2014. These uncertainties are compounded To illustrate the implications in this change in funding criteria, the by a stagnating or negative global outlook for HIV/AIDS funding. Most new criteria are applied retroactively to HIV/AIDS spending and fund- significantly, the Global Fund has recently postponed its new call for proposals (“Round 11”)—for which the Government of Jamaica had 22  The NASA did not clearly identify collateral financing by the Government prepared a submission—and has introduced tighter eligibility criteria. of Jamaica of projects supported by the World Bank (J$ 40 million) and the United States Government (J$ 15 million). In Figure 11, the amounts spent by Compared to other countries, the extent of external support for Jamai- the Government of Jamaica are netted out from spending supported by the ca’s HIV/AIDS program has been relatively strong. Figure 13 summa- World Bank and the United States Government. rizes the available data on the external contributions to the financing 23  This is also a consequence of data limitations, as there are no estimates of HIV/AIDS programs for 2009 (from UNAIDS 2010), based largely available of the private costs of accessing HIV/AIDS services. Data from house- on data from NASAs and UNGASS reports. The extent of external hold surveys (compare Section 2) suggest that wealthier quintiles predomi- nantly use private health services, but it is possible that this is not the case for support in Jamaica is among the highest compared to countries at a specialized HIV/AIDS services. similar level of economic development, but it is well within the range 24  For upper-middle-income countries (according to World Bank definition) like enjoyed by countries with a similar level of HIV prevalence. However, Jamaica, the Global Fund demands that proposals meet a “minimum thresh- those countries with a similar level of HIV prevalence but which enjoy old for counterpart financing” of 60 percent, i.e. total HIV/AIDS funding from a higher rate of external support (Guinea, Djibouti, and Guinea-Bis- the domestic budget needs to account for at least 60 percent of the sum of sau) are characterized by much lower level of GDP per capita. total domestic budgetary funding and Global Fund support. Figure RoleRole 13. 13a. Figure of External HIV/AIDS of External HIV/AIDSFinancing, Financing 2009 (Percent of total Figure spending) 13b. Role of External HIV/AIDS Financing (Percent of total spending) (Percent of total spending) 120 Guinea Djibouti 100 Fiji 100 Guinea-Bissau 80 Jamaica 80 Jamaica 60 Belize Belize 60 40 40 20 20 0.00 0.00 0 2,000 4,000 6,000 8,000 10,000 0 5 10 15 20 GDP per capita (U.S. dollars) HIV prevalence, ages 15-49 (percent) Source: Author’s calculations, based on data from UNAIDS (2010) and IMF (2010). Source: Author’s calculations, based on data from UNAIDS (2010) and IMF (2010). 34 VI.  Sustainability of the National Response to HIV/AIDS Figure 14. Domestic Financing Need and Rate of External Support (J$ millions, 2010 prices) 3,500 3,000 Support Total Costs 2,500 declining to zero 2,000 Support declining to 45 percent 1,500 1,000 Support continues at 2010 rate 500 (69 percent) 0.00 2010 2015 2020 2025 2030 Source: Author’s estimates and projections. ing in 2010, when 31.0  percent of the total costs of the national financing needs of the National HIV/AIDS Program. While these are response to HIV/AIDS were funded by the Government of Jamaica,25 slow to materialize (owing to long lags between HIV infection and and 48.8 percent of the total costs were contributed by the Global treatment), they play an important role for sustaining the funding of Fund. Under the new funding rule, the share of the national govern- the HIV/AIDS program in the longer run. ment would need to rise to 47.8 percent (i.e., from J$ 535 million to J$ 826 million). One aspect of attaining the financial sustainability of the National HIV/ AIDS Program is to ensure that HIV/AIDS interventions are delivered in The estimates of total spending used in the current study, however, a cost-effective manner and—where there is a choice between differ- include expenditures addressing the impact of HIV/AIDS which are ent measures—that the most cost-effective interventions are chosen not formally included in the national response to HIV/AIDS, notably and that the available funds are allocated in the most efficient manner. a share of the CDA budget (J$  275  million in 2010). If this is not accepted as counterpart financing by the Global Fund, then domestic The focus of the present study is on the fiscal costs of the HIV/AIDS financing would need to rise to J$ 937 million (54.2 percent of the program overall, rather than the analysis of specific HIV/AIDS inter- total costs). ventions and the design of the HIV/AIDS program. Nevertheless, the tools applied here—explicitly linking the costs of the HIV/AIDS pro- For the longer-run projections, assumptions regarding the availability gram to the underlying HIV infections—are suitable to address such of external funding are largely speculative at this stage. To illustrate issues as allocative efficiency in the HIV/AIDS program. the consequences of shifts in external funding, Figure 14 illustrates various possibilities (based on Scenario  1, with declining HIV inci- Specifically, the estimates of the costs incurred by one HIV infection, dence). With rising spending needs, a decline in the rate of external under the objectives of the National HIV/AIDS Program, provide a support will result in an even steeper increase in domestic funding yardstick for assessing the cost-effectiveness of HIV/AIDS interven- needs. For example, if the rate of external support declines from tions. Cost-effectiveness is then calculated relating the health out- 69  percent to 45  percent (broadly in line with the changed Global comes of interest (HIV infections averted, life years saved, etc.) to the Fund criteria), domestic funding needs will need to double by 2015, costs of the intervention, minus the financial savings implied by the from J$ 539 million (US$ 6.3 million) in 2010 to J$ 1,144 million (US$ reduced number of HIV infections. 13.5 million) in 2015, and increase to J$ 1,353 million (US$ 15.9 mil- lion) by 2020, and J$ 1,683 million (US$ 19.8 milllion) by 2030. In addition to the challenge of assessing and attaining cost-effective- ness, effective spending allocation involves comparisons and choices Across scenarios, the consequences of the decline in external funding between different types of interventions—for example, between are fairly even initially, because the assumed changes in HIV inci- prevention interventions that target different population groups. dence affect spending with a long lag. For example, if external sup- The contribution of this study to assessing the allocative efficiency of port drops to 45 percent of the costs of the national response to HIV/ HIV/AIDS interventions is a framework that allows for comparisons AIDS, domestic funding needs in 2015 differ by about J$  4  million to be made between the outcomes and financial consequences of (US$ 46,000) by 2015 between Scenario 1 (declining HIV incidence) the different interventions. To this end, it is necessary to (1) assess and Scenario 3 (increasing HIV incidence). However, as the assumed the impacts of an intervention at the population-level in addition to changes in HIV incidence feed through to HIV/AIDS spending, the difference in domestic funding needs between Scenario 1 and Sce- 25  Additionally, the loan from the World Bank may be counted as domestic nario 3 rises to J$ 33 million(US$ 390,000) by 2020, and J$ 183 mil- financing as it will need to be repaid by the Government of Jamaica, unlike lion(US$ 2.2 million) by 2030. the grant financing from the Global Fund. If spending supported by the World Bank loan can be counted as domestic funding, the share covered by the Thus, successful HIV prevention interventions—under a wide range of national government would need to rise to 44.1 percent (in addition to the assumptions regarding the extent of forthcoming external support— World Bank loan), or 50.5 percent if a part of the CDA budget is not accepted will have an increasing and potentially large impact on the domestic as counterpart financing. 35 Assessing the Financial Sustainability of Jamaica’s HIV Program its direct impacts (infections prevented in targeted group, and result- • As to the consequences of HIV infections across population ing health gains and financial savings), and (2) consider the conse- groups, the expected costs of providing HIV/AIDS services to one quences of such impacts at the population-level (“downstream” HIV person living with HIV/AIDS from the time of infection are esti- infections averted, and the health gains and financial consequences mated at J$ 500,000 (US$ 5,800). However, the HIV epidemic in resulting from these). Jamaica is concentrated, so the probability that an HIV-positive person passes HIV on to another person (and so on) differs widely Many such decisions require a more explicit epidemiological analy- across population groups. Taking into account these downstream sis than the one offered here. Nevertheless, the analysis offers some infections, the financial savings achieved by preventing one HIV lessons and examples for assessing the efficiency of spending alloca- infection is assessed at around J$ 800,000 to J$ 900,000 for the tions, namely: general population (“low risk” or “casual sex”), J$ 2.6 million • For the prevention of mother-to-child transmission of HIV, the for female sex workers and J$ 5.3 million for men who have sex analysis estimates that the cost of averting one pediatric in- with men. To assess the allocative efficiency of HIV/AIDS spend- fection was J$ 658,000. Each pediatric HIV infection averted, ing across prevention interventions, it is necessary to take into however, results in expected savings of J$ 483,000, so that the account the population-level impacts of the intervention (includ- net costs of averting one pediatric HIV infection are J$ 176,000. ing downstream infections prevented and the resulting health outcomes), as well as the resulting financial savings. Funding Treatment Costs The financing of treatment costs is one of the greatest financial chal- cost-effectiveness of health interventions. For example, the WHO lenges in sustaining the HIV/AIDS program over the coming years. would consider health interventions that save a life year at a cost that Treatment costs are expected to rise sharply because an increasing is lower than GDP per capita as “highly cost-effective” and the eco- number of people living with HIV/AIDS will receive treatment and the nomic literature adopts values well exceeding GDP per capita in the increasing need for second-line treatment will drive up average costs. cost-benefit analysis of health or environmental interventions (see, In addition, the (drug) costs of antiretroviral treatment are currently e.g., Murphy and Topel, 2006). Compared to these benchmarks, the supported by the grant from the Global Fund, and the outlook for costs of antiretroviral treatment in Jamaica (about 10 percent of GDP funding from this source after expiry of the current arrangement in per capita for first-line treatment, and 20 percent of GDP per capita 2013 is uncertain. for second-line treatment) are relatively small. Should external funding for the treatment program no longer be Beyond the immediate health outcomes, it is plausible that HIV/AIDS available (or be much reduced), the funding of the treatment program exacerbates social inequities in access to health. The Jamaica Survey would not pose unusual challenges, from a public health perspective. of Living Conditions 2009 (STATIN and PIOJ, 2010) highlighted the HIV/AIDS-related health spending (approximated by the costs of care fact that poorer households were less likely to seek medical services and treatment, the prevention of mother-to-child transmission, and in an episode of illness and that the costs for those households seek- counseling and testing) amounted to J$  800  million in 2010, cor- ing care were high relative to income. At a cost that is equivalent to responding to two to three  percent of public health spending. This over 60 percent of consumption per capita, even first-line treatment amount seems in line with the contribution of HIV/AIDS to the burden would be unaffordable for many members of the poorest consump- of disease—the epidemic accounts for about eight percent of deaths, tion quintile (especially considering that it is a persistent health con- one-half of the deaths from communicable diseases among the over- dition). all population, and one-quarter of deaths among the population aged 15–59 years (WHO, 2011). One policy proposal that is being considered is the funding of an- tiretroviral drugs through the National Health Fund (NHF), which More specifically, the annual costs of antiretroviral treatment (drugs, purchases and markets essential drugs at subsidized prices and is tests and human resources), based on current estimates provided by financed by the National Insurance Fund, which provides coverage the Clinton Health Access Initiative, are equal to nine percent of GDP to all residents. Among the advantages of providing antiretroviral per capita (J$ 41,000) and 21 percent of GDP per capita (J$ 88,000) drugs through the NHF is the possibility to solicit (income-dependent) for first- and second-line therapy, respectively. These costs may under- co-payments for the drugs, which could help contain the fiscal costs state the cost of extending life by one year (as patients would survive while addressing equity challenges. However, the costs of antiretro- without treatment for some period) but do not include financial sav- viral drugs are large relative to the budget of the NHF, which paid ings in other areas, e.g., reduced non-antiretroviral treatment costs, out benefits of about J$ 1.5 billion in FY 2007/08 (NHF, 2009). The savings in social mitigation expenses and a reduced risk of onward costs of antiretroviral drugs are estimated at about J$ 230 million in HIV transmission (“treatment as prevention”). 2012 and are projected to rise to J$  330  million by 2015, and to J$ 540 million by 2020 in Scenario 1 (J$ 560 million in Scenario 3). Setting aside these uncertainties, the costs of treatment neverthe- The NHF would therefore require additional funding to take on such less appear far below benchmarks commonly used in assessing the increased responsibilities. 36 VII. Conclusions VII. Conclusions The national response to HIV/AIDS faces considerable financing challenges, arising from the rising costs of the HIV/AIDS program (largely reflecting the increasing need for treatment), an uncertain and probably negative outlook for the availability of external funding and tight public finances. To inform the future policy response of the Government of Jamaica, this study set out to (1) review current spending on HIV/ AIDS and the sources of financing; (2) estimate the fiscal burden of the national HIV/AIDS response and assess the outlook for external financing of the HIV program; (3) project how the epidemic will unfold as well as what the costs would be under different scenarios; and (4) provide recommendations to inform policy decisions. Key findings include: • The costs of the national response to HIV/AIDS are estimated at about J$ 1.7 billion (US$ 20.4 million) as of 2010/11 (0.14 percent of GDP), of which about one-third is financed from the domestic budget and the rest from external sources. • Spending needs will rise steeply over the coming years to about J$ 2.5 billion (US$ 28 million) by 2020, largely as a consequence of the increasing number of cases needing treatment and the increasing need for the use of second-line drug treatment. Current gains in HIV prevention will reduce spending needs only very slowly—owing to the long and variable lags between HIV infection and treatment need. Improved prevention outcomes are therefore no quick fix for the funding needs of the HIV/AIDS Program. • The Government of Jamaica currently pays for about one-third of the costs of the national response to HIV/AIDS. Agreements with major funders (Global Fund, World Bank, U.S. Government) are set to expire over the next few years. The recent change in the Global Fund’s funding criteria would mean that the share financed by the Government of Jamaica would need to rise to at least one-half. • With total costs to the Government of Jamaica remaining around 0.2 percent of GDP over the com- ing years (minus any external support), the challenges for financing the national HIV/AIDS Program is manageable and fiscally sustainable provided HIV/AIDS remains a national development priority. 37 Assessing the Financial Sustainability of Jamaica’s HIV Program • Treatment costs appear to be in line with the burden of disease of government programs and the cost analysis of other main and health spending in general. Overall HIV/AIDS spending is disease programs and health spending in Jamaica. The sus- about three percent of public health spending, which appears tainability of the National HIV Program cannot be achieved in consistent with the burden of disease. Also, the annual costs isolation from other public expenditure programs. The sustain- of treatment (about 10 percent of GDP per capita for first-line ability plan will need to review possible sources of financing treatment and 20  percent of GDP per capita for second-line from both the public and private sectors and also institutional treatment) are within a range that could reasonably be sup- sustainability. The Government is moving in the right direction ported through public health spending. by (1) exploring the possibility of financing HIV/AIDS drugs through the NHF, which is funded by taxation; and (2) strength- • Moreover, public provision of treatment fulfills a broader so- ening the institutional sustainability of the National HIV Pro- cial role, as HIV/AIDS is a persistent health condition that is gram by integrating it with the Family Planning Board to create expensive to care for and treat, and that therefore is likely to a Sexual and Reproductive Health Authority for Jamaica. exacerbate social inequities in accessing health. 2. Assess the allocative efficiency of the national response to • New infections eventually result in additional demands for HIV/AIDS, taking into account not-only the immediate health services which are expensive and may extend over several de- consequences but also the epidemiological and financial reper- cades. Overall, the fiscal cost incurred by one additional HIV in- cussions of alternative HIV/AIDS interventions. This report con- fection, under the objectives of the National HIV/AIDS Program, tributes to such an analysis of cost-effectiveness by estimating is estimated at J$ 500,000 (US$ 5,800). the costs incurred by new HIV infections over time (so that the resulting savings can be netted out from the costs of preven- • When the risk of onward infections is taken into account, the tion interventions) and by discussing the role of “downstream” costs incurred by new infections range from J$ 800,000 to infections (allowing an assessment of the cost-effectiveness of J$ 900,000 for the general population to J$ 2.6 billion(about prevention interventions across population groups). US$ 30,000) for female sex workers and J$ 5.3 billion (about 3. The allocative efficiency analysis needs to further assess US$ 60,000) for men who have sex with men. spending choices across prevention activities. Jamaica will To take the national response to HIV/AIDS further, the study points to need to make stronger effort in reaching the most at-risk pop- three priority areas for the Government: ulations, particularly men who have sex with men, to prevent HIV transmission. This report clearly indicates that the costs 1. Develop a long-term sustainability plan with different financing incurred by new infections from sex workers and men who options that defines the roles of different national agencies (as have sex with men are very high and that effective prevention well as donor agencies). This would build on the cost estimates in the short-term will result in lower spending on treatment in presented here, as well as upon the public expenditure review the long-term. 38 References References Duncan, Jacqueline, Sharlene Beckford Jarrett, and Kevin Harvey, ———, 2010c, “Getting to Zero: 2011–2015 Strategy” (Geneva: 2010, “Using Estimation and Projection Package and Spectrum UNAIDS). for Jamaica’s National HIV Estimates and Targets,” Sexually Lule, Elizabeth, and Markus Haacker, 2011, Fiscal Dimension of HIV/ Transmitted Infections, Vol. 86, Suppl. 2, pp. ii43–ii47. AIDS in Botswana, South Africa, Swaziland, and Uganda(Wash- Figueroa, J. Peter, and others, 2008, “A Comprehensive Response to ington DC: World Bank). the HIV/AIDS Epidemic in Jamaica – A  Review of the Past 20 Ministry of Finance, 2011, “Jamaica Memorandum on the Budget” Years,” West Indian Medical Journal, Vol. 57, No. 6, pp. 562–76. (Kingston: Ministry of Finance) Government of Jamaica (GoJ), 2008, “HIV and AIDS in Jamaica: Na- Ministry of Health (MoH) and Joint United Nations Programme on tional Strategic Plan 2007–2012, FINAL DRAFT, March 2008,” HIV/AIDS (UNAIDS), 2012, Jamaica – National AIDS Spending (Kingston: GoJ). Assessment (Kingston: Ministry of Health). Guthrie, Teresa, Nhlanhla Ndlovu, Farzana Muhib, Robert Hecht, and Ministry of Health/National HIV/STI Programme (MoH/NHP), 2012, Kelsey Case, 2010, “The Long-Run Costs and Financing of HIV/ “Country Progress Report Jamaica, Submitted March 30, 2012” AIDS in South Africa” (Washington DC: Results for Development (Kingston: Ministry of Health). Institute). ———, 2010a, “Country Progress Report to the Secretary General Hecht, Robert, and others, 2010, “Financing of HIV/AIDS Programme of the United Nations on the United Nations General Assembly Scale-up in Low-Income and Middle-Income Countries, 2009– Special Session – Reporting period: January 2008–December 31,” The Lancet, Vol. 376, pp. 1254–60. 2009” (Kingston: Ministry of Health). Heller, Peter S., 2005, “Understanding Fiscal Space,” IMF Policy Dis- ———, 2010b, “Jamaica HIV Epidemic Update, January to Decem- cussion Paper No. 05/4 (Washington DC: International Monetary ber 2009” (Kingston: Ministry of Health). Fund). Murphy, Kevin M., and Robert Topel, 2006, “The Value of Health and Hogan, Daniel R, Rob Baltussen, Chika Hayashi, Jeremy A. Lauer, Longevity,” Journal of Political Economy, Vol. 114, No. 5, pp. and Joshua A. Salomon, 2005, “Cost Effectiveness Analysis of 871–904. Strategies to Combat HIV/AIDS in Developing Countries,” British Phillips, Peter, 2012, “A New and Binding Covenant for Stability, Eq- Medical Journal, Vol. 331, pp. 1431–7. uitable Growth and Prosperity,” Budget Debate Presentation to Hope Enterprises, 2008, “Report of Findings from HIV/AIDS Knowl- Parliament, May 24 (Kingston: Ministry of Finance). edge Attitudes and Behavior Survey, Jamaica 2008” (Kingston: Schwartländer, Bernhard, and others, 2011a, “Towards an Improved Hope Enterprises Ltd.). International Monetary Fund (IMF), Investment Approach for an Effective Response to HIV/AIDS,” 2011a, “Jamaica: Third Review Under the Stand-By Arrange- The Lancet, Vol. 377, pp. 377, pp. 2031–41. ment – Staff Report” (Washington DC: IMF). Schwartländer, Bernhard, and others, 2011b, supplementary web ap- ———, 2011b, World Economic Outlook Database, September pendix to Schwartländer (2011a). 2011 Edition (Washington DC: IMF). Statistical Institute of Jamaica (STATIN) and Planning Institute of Ja- ———, 2010, “Jamaica: 2009 Article IV Consultation and Request maica (PIOJ), 2010, Jamaica Survey of Living Conditions 2009 for a Stand-By Arrangement – Staff Report” (Washington DC: (Kingston: STATIN and PIOJ). IMF). United Nations Children’s Fund (UNICEF), 2011, State of the World’s Joint United Nations Programme on HIV/AIDS (UNAIDS), 2010a, UN- Children 2011 (New York: UNICEF) AIDS Report on the Global AIDS Epidemic 2010 (Geneva: UN- United Nations Population Division (UNPD), 2011, World Population AIDS). Prospects: The 2010 Revision (New York: UNPD). ———, 2010b, “HIV Estimates with Uncertainty Bounds, 1990– World Health Organization (WHO), 2011, Global Health Observatory 2009,” obtained online at http://www.unaids.org/en/dataanaly- Data Repository (found at: http://apps.who.int/ghodata/), ac- sis/epidemiology/ (Geneva: UNAIDS). cessed August 2011 (Geneva: WHO). 39 Appendix on Methodology Appendix on Methodology The analysis offered in this report combines three building blocks: (1) an epidemiological model, (2) a costing framework for projecting the costs of the national response to HIV/AIDS, and (3) an economic framework to calibrate the costs incurred by new infections. Epidemiological Framework The epidemiological framework is modeled following the estimates produced by the Ministry of Health using the Spectrum model, updated from the analysis described by Duncan and others (2010). As some of the economic analysis offered here (e.g., analyzing the consequences of individual infections over time) exceed the capabilities of the Spectrum model, the estimates and projections provided by the Ministry of Health were transformed into a spreadsheet-based model, which is integrated in the economic frame- work, and which has been calibrated to replicate the estimates and projections provided by the Ministry of Health. Figure A1 (Counterfactual) Births Orphans HIV+ births Missing births (Counterfactual) Young Adult (no HIV) Death HIV Premature Death CD4>350 Treatment Unmet Treatment Treatment need 350 need failure failure Treatment Treatment Treatment need 200 (1st line) (2nd line) Time 41 Assessing the Financial Sustainability of Jamaica’s HIV Program The structure of the epidemiological module is summarized in • For individuals receiving second-line treatment, transition to Figure  A1. As an HIV infection occurs, adults move from state “no treatment failure and death is determined by a Weibull distri- HIV” to “HIV+, CD4>350” (HIV positive, with a CD4 count above bution with a mean transition time of 10 years. 350). Eventually, the CD4 count, in the absence of treatment, falls below 350 and 200, respectively. This distinction is important to rep- • Through 2015, the number of people initiating treatment is licate the estimates made by the Ministry of Health (which involve a based on the estimates and projections provided by the Min- switch in the treatment eligibility criterion from a CD4 count of 200 to istry of Health. Beyond 2015, 90 percent of people newly re- a CD4 count of 350 in 2005—see main text, Figure 1b). However, in quiring treatment are expected to receive it, in line with the line with current government policy, projections are based on a CD4 draft NSP. eligibility criterion of 350. • Through 2015, the transition to second-line therapy is based on As individuals reach the stage “treatment need,” they either initiate estimates and projections received from the Ministry of Health. treatment, or progress to “unmet need” and eventually to premature Thereafter, it is assumed that 90  percent of people receiving death. Those receiving first-line treatment eventually reach first-line first-line treatment need move on to second-line treatment (con- treatment failure and then progress either to initiate second-line sistent with the projections by the Ministry of Health). treatment or to premature death. Individuals on second-line treat- • For new infections at age 0 (the model does not differentiate ment eventually reach treatment failure and then premature death. between infections before, at, or after birth), it is assumed that Additionally, there is a module determining the impact of HIV/AIDS the probability of mother-to-child transmission is 24  percent on the young population, including mother-to-child transmission of (without interventions), that it is reduced by two-thirds through HIV and the impact of HIV on the number of orphans. measures to prevent of mother-to-child transmission of HIV, and Specifically, the transitions between the states are governed by the by two-thirds if the mother is receiving treatment. following assumptions: • The number of orphans is estimated and projected based on a • The transition from HIV infection to eligibility for treatment at module that builds on male and female adult mortality, and on a CD4 count of 350 is based on a Weibull distribution, with a the survival probabilities of children (depending on treatment mean transition time of six years for males, and seven years for access). females. From this point, transition to eligibility for treatment at One particular aspect of the epidemiology of HIV/AIDS in Jamaica is a CD4 count of 200 takes another three years (again based on the high rate of outward migration, especially of people in the age a Weibull distribution). In the absence of treatment, death oc- groups where HIV incidence and prevalence are high. This means curs on average three years after reaching a CD4 count of 200. that many people infected with HIV in Jamaica leave the country be- • The model can accommodate treatment initiation at a CD4 fore they require HIV/AIDS-related services. For example, 12 percent count of 350 or 200 (or a combination of the two possibilities). of men and 14 percent of women who become infected in Jamaica For individuals initiating first-line treatment at a CD4 count of emigrate before reaching first-line treatment need (at a CD4 count 350, transition to second-line treatment need takes 11 years, of 350). These outflows have been taken into account, e.g., by speci- while for individuals who start at a CD4 count of 200, it is eight fying an annual migration rate at different stages of the disease (e.g., years. Unless second-line treatment is initiated, death occurs on 2.4 percent for people above a CD4 count of 350). average three years after reaching second-line treatment need. Projected Costs of National Response to HIV/AIDS The costs of the national response to HIV/AIDS are projected based on estimated HIV/AIDS spending (as documented in the NASA re- on a straightforward costing model. The numbers of people receiving port), budget data (e.g., for the costs of support to orphans) or direct various types of HIV/AIDS-related services are based on the epide- estimates of unit costs (e.g., for treatment costs). These assumptions miological estimates and the relevant coverage rates (based on the are documented in the main text (section “Assumptions on Unit Costs estimates and projections by the Ministry of Health or on the draft Na- and Coverage Rates,” page 20). tional Strategic Plan 2013–18. To this, unit costs are applied, based Costs Incurred by New Infections One peculiar aspect of HIV/AIDS, greatly complicating the assessment living with HIV/AIDS are spread over a very long period of time, and of the financial implications of different interventions of program al- could span several decades if an individual obtains access to treat- ternatives, is the fact that the costs of services required by a person ment (which most do in the projections offered here, at a coverage 42 Appendix on Methodology rate of 90  percent). Current spending alone—without taking into that is directed at the general population or population sub-groups, account the state and course of the epidemic—is thus a very poor irrespective of their HIV status. To distinguish it from spending to ad- indicator of the fiscal burden of HIV/AIDS. dress the demands for services by people living with HIV/AIDS, this is referred to as “population-based” spending. To provide a more accurate analysis of the fiscal burden of HIV/AIDS, the report therefore offers estimates of the actuarially fair value of the For a short-run analysis, e.g., projecting the financing needs of a na- costs incurred by new HIV infections. For one new infection, this is the tional HIV/AIDS program over a few years, these different types of amount that the government would have to put aside at the time the spending are similar. For an analysis of the financial consequences infection occurs, in line with its policy commitments under the HIV/ of alternative HIV policy options, however, these spending categories AIDS program, in order to cover the expected life-time costs of provid- are very different in terms of the links between past HIV infections, ing HIV/AIDS-related services to the person who becomes infected. As current spending, and future resource needs. the amount put aside would earn some interest until it is used to pay for HIV/AIDS-related services, future costs are discounted at an inter- The current demand for HIV/AIDS-related services and—under the est rate, which is based on the rate the government pays on its public government’s policy commitments to provide certain HIV/AIDS-relat- debt (see Box 4—“Estimating the Financial Costs of New Infections”). ed services—the costs of meeting this demand are basically “locked The projections of the costs caused over time by one additional HIV in” by the number of infections which have occurred over the last infection, and estimates of the actuarially fair value of these costs, are years or decades. Spending in these categories is therefore bare- documented in the main text (Figure 9, and pages 25–26). ly responsive to any progress in reducing HIV incidence. To assess the fiscal consequences of alternative HIV/AIDS policies, the savings However, not all HIV/AIDS spending can be attributed to earlier HIV in- from reduced HIV incidence associated with an HIV/AIDS program fections in this way. This applies, in particular, to prevention spending are therefore estimated instead by the number of new HIV infections, Table A1. Costs of Jamaica’s National Response to HIV/AIDS 2010 2020 2030 J$ millions % of GDP J$ millions % of GDP J$ millions % of GDP Scenario 1 Total 1,727.2 0.15 2,519.9 0.19 3,420.1 0.22 Prevention 625.4 0.05 803.7 0.06 907.8 0.06 Treatment, care, and support 469.0 0.04 1,109.0 0.08 1,824.2 0.12 o/w: Antiretroviral therapy 334.2 0.03 1,045.4 0.08 1,744.0 0.11 Social mitigation 295.0 0.03 187.2 0.01 118.0 0.01 Program management 337.8 0.03 409.9 0.03 420.0 0.03 Scenario 2 Total 1,727.2 0.15 2,492.9 0.19 3,246.6 0.21 Prevention 625.4 0.05 803.6 0.06 907.6 0.06 Treatment, care, and support 469.0 0.04 1,088.2 0.08 1,687.5 0.11 o/w: Antiretroviral therapy 334.2 0.03 1,025.5 0.08 1,613.0 0.10 Social mitigation 295.0 0.03 185.6 0.01 110.4 0.01 Program management 337.8 0.03 415.5 0.03 541.1 0.03 Scenario 3 Total 1,727.2 0.15 2,519.8 0.19 3,393.6 0.22 Prevention 625.4 0.05 803.7 0.06 907.8 0.06 Treatment, care, and support 469.0 0.04 1,109.0 0.08 1,803.7 0.12 o/w: Antiretroviral therapy 334.2 0.03 1,045.4 0.08 1,724.4 0.11 Social mitigation 295.0 0.03 187.2 0.01 116.6 0.01 Program management 337.8 0.03 420.0 0.03 565.6 0.04 Source: Author's estimates. 43 Assessing the Financial Sustainability of Jamaica’s HIV Program multiplied by the expected costs incurred by one infection. This meth- AIDS. The two different types of spending are linked, however. As pre- od is illustrated on pages 29–31 which provides an interpretation of vention spending (most of which would fall under the category of pre- Jamaica’s HIV/AIDS program in terms of fiscal space. vention-based spending) reduces the number of HIV infections (and the associated spending commitments), the analysis of the costs incurred By contrast, as population-based spending cannot be attributed to by new infections is immediately relevant for the assessment of the previous HIV infections, and as current HIV infections do not cause cost-effectiveness of prevention interventions. This point is illustrated future population-based spending, this spending category is treated as in the main text on page 27 in a discussion on the costs and resulting current spending in the analysis of the evolving fiscal burden of HIV/ fiscal savings of the program to prevent mother-to-child transmission. Assessing Consequences of New HIV Infections across Population Groups The estimates of the costs incurred by one new HIV infection, togeth- sex workers over time results in an additional 1.5 infections among er with estimates of the modes of transmission of HIV/AIDS, can be female sex workers later on—0.6 (=2 clients infected per sex worker, used to estimate the financial consequences of HIV infections across each expected to infect 0.3 additional sex workers) in the first round, population groups, including the costs of “downstream” infections. 0.36 in the second, etc. Together with the initial infection, then, an The “modes of transmission” framework is first applied to estimate expected total of 2.5 female sex workers become infected, which the expected number of downstream infections caused by one new implies that a total of five clients (2.5 infections among female sex HIV infection. For example, a female sex worker may infect two cli- workers over time, times two clients infected per sex worker), and ents, and each client in turn infects 0.3 sex workers and 0.7 regular 3.5 partners of clients become infected (five clients, multiplied by 0.7 partners. This would mean that a new HIV infection among female partners becoming infected per client). 44 www.worldbank.org/lacaids www.worldbank.org/aids