• Introduction
• Linkage to development objectives
• Analysis of alternatives
• Affordability
• Efficiency

---

This paper examines four questions posed by economic analysis to help set priorities for government involvement with antiretroviral therapy for people living with HIV/AIDS. The main focus is on making decisions about policy relating to ARVs in developing countries where the needs are greatest and resource constraints are most binding. While the answers may vary considerably depending on individual country circumstances, the analytical foundations for setting priorities are the same. First, how does antiretroviral therapy link to broader health sector and country development objectives? Second, what other interventions need to be considered including the various ARV treatments? Third, which of these alternatives are realistically affordable given the country’s resource constraints? And fourth, which of the affordable alternatives are most efficient in achieving a favourable development impact?

Any consideration of government involvement with antiretroviral treatment for HIV/AIDS needs to start with a clear view of how such treatment links to broader health sector strategy and country development goals. Despite its terrible human toll, HIV/AIDS is only one of many health problems facing developing countries. And health improvement is only one of many pressing development challenges involved in raising people’s living standards -- from reducing illiteracy, to providing infrastructure and expanding job opportunities. Looking at this linkage just in terms of improving health requires an epidemiological assessment of the magnitude of ill-health due to HIV/AIDS relative to other health problems -- not only its importance today but also its significance tomorrow as morbidity and mortality from the epidemic continues to grow. A suggestive indicator of these relative magnitudes is the burden of disease due to disability and premature death from different causes. Recent estimates of the global burden of disease expressed in terms of the number of years of healthy life lost (Murray & Lopez, 1996) suggest that for developing countries:

These global estimates illustrate the magnitude of the health problem which antiretroviral treatment aims to address, while at the same time emphasizing the likely tradeoffs with other health problems of arguably greater magnitude. How much difference antiretroviral treatment could make to improving overall health therefore depends on the relative contribution of HIIV/AIDS to ill-health in specific country circumstances. Defining the strength of this link obviously requires a quantitative assessment of epidemiological magnitudes -- numbers of people who are asymptomatic, symptomatic and living with AIDS -- in the particular countries concerned. This in turn sets the stage for analyzing alternatives for government involvement with antiretrovirals.

Comparing alternative ways of addressing the HIV/AIDS problem -- different types of antiretroviral treatment, along with other preventive interventions -- will help decision makers choose the option most likely to achieve development objectives. It is not enough to think of the choice set just in terms of the technical options -- in practice the consideration of alternatives must also examine alternative target groups, as well as public or private sector alternatives in both provision and financing.

First, technical alternatives -- innovations in antiretroviral treatment have greatly expanded the choice set to include a wide range of different drug regimes involving reverse transcriptase and protease inhibitors, that can be given singly or as combination therapies, with widely different implications for unit costs and clinical outcomes.

Second, target groups are also different even with the same technical option -- ranging from people with AIDS, to symptomatics or asymptomatics, as well as infected pregnant women. The different population sizes involved have major implications for the aggregate costs of treatment, as well as its likely effectiveness in reducing the burden of disease. For example, Thailand’s population of 58 million was estimated to have a nationwide pool of some 746,000 infected adults in 1996 -- including a caseload of around 47,000 adults with AIDS, plus another 61,000 symptomatics, and 22,000 infected pregnant women.

Third, whether antiretrovirals are or could be provided by the private sector instead of the public sector needs to be examined carefully. Is there a private market for antiretrovirals, or is there a private sector delivery system that could provide them? If so public provision of antiretrovirals might induce substitution of patients seeking treatment out of the private sector, thus lowering the net impact of public sector provision.

Finally, is there any rationale for public financing of antiretroviral treatment instead of relying on private payments? One important justification for public subsidies involves interventions that generate epidemiological externalities -- where antiretroviral treatment prevents transmission of infection to others. The externality argument clearly does apply to preventing maternal-child transmission (MC1T). But it does not seem to apply to using antiretrovirals to treat other infected groups -- unless viral load is eliminated, as has been suggested with protease inhibitors.

A different justification for subsidizing antiretroviral treatment is that government policy generally does provide subsidies to help insure against catastrophic financial risks due to ill-health. But then governments should be consistent in applying their particular national price/subsidy policy across all diseases -- without discriminating in pricing policy between people with HIV and people with other conditions. Granting preferential subsidies to patients with HIV/AIDS risks crowding out uninfected patients that might get more health benefits from the fixed amount of available subsidies.

Analyzing the financing options for antiretrovirals leads straight into the affordability question that is of such importance in low and middle-income developing countries. Policymakers need to reject alternatives for government involvement that are not financially affordable -- otherwise they will get locked into unsustainable policies that undermine the objectives of getting involved in the first place. This means focusing on the subset of antiretroviral alternatives that are affordable from the point of view of those who have to finance the relevant costs -- government and households.

The distribution of treatment costs between these parties depends on two factors -- the choice of pricing policy, and its interaction with the relevant budget constraints. Estimates from Thailand illustrate these factors in the context of a middle-income country with a per capita income of USD 2,410 in 1994 (Prescott et al., 1996). Figure 1 shows the estimated unit costs of treatment expressed in Thai baht (approx. 25 baht per USD), together with the associated prices/subsidies for a range of ten treatment options for HIV/AIDS. Eight of these regimes involve administration of antiretroviral drugs, not including the new protease inhibitors:

• Model Al: adult AIDS -- universal coverage of optimal 0I treatment
• Model A2: adult AIDS -- zidovudine (ZDV) monotherapy
• Model A3: adult AIDS -- ZDV/ddI or ddC sequential monotherapy
• Model A4: adult AIDS -- ZDV/ ddI or ddC combination therapy
• Model AS: adult symptomatic HIV -- ZDV monotherapy
• Model A6: adult symptomatic HIV -- ZDV/ddI or ddC sequential monotherapy
• Model A7: adult symptomatic HIV -- ZDV/ddI or ddC combination therapy
• Model A8: adult asymptomatic HIV: ZDV monotherapy
• Model Cl: MCT: formula feeding
• Model C2: MCT: ZDV plus formula feeding

Expressed in equivalent USD the unit costs of treatment for adults with AIDS range from about USD 2,000 per patient-year with zidovudine monotherapy (model A2) to USD 4,000 per patient-year with zidovudine/ddI or ddC combination therapy (model A4). The annual unit costs of treatment for symptomatic and asymptomatic target groups appear lower because these earlier stages in the life cycle of treatment involve fewer provider visits and 0I drugs -- but later progression to AIDS will raise unit costs. Meanwhile the unit cost of treating pregnant women following the ACTGO76 protocol is only about USD 500 per mother-child pair (model C2). The pricing policy regime superimposed on these costs provides free -- fully subsidised -- antiretrovirals to patients, while requiring around one-third cost recovery for 0I drugs and inpatient/outpatient services. This subsidy structure reduces patient costs during AIDS to less than USD 500 per year.

Whether such a subsidy policy is affordable from the government’s point of view depends on the ratio of fiscal requirements to the projected budget constraint. Here the relevant budget constraint is assumed to be equal to the entire budget allocated to the National AIDS Programme. Options with index values greater than 100 are not affordable. Figure 2 shows the simulations assuming maximum utilization generated by universal coverage and perfect compliance. Under the assumed combination of pricing policy and fiscal constraints all the treatment options involving antiretrovirals appear unaffordable -- with the notable exception of preventing perinatal transmission, which has an affordability index of less than 20%. Just treating OIs (model Al) for the projected number of AIDS patients would exhaust the whole budget, without adding any antiretroviral drugs.

To assess affordability of this pricing policy from the patients’ point of view, it is important to think in terms of the distribution of incomes -- or its proxy, per capita consumption expenditure (PCE) -- instead of the average level of income in the whole population. Figure 3 shows the cumulative distribution function for three different categories of PCE -- medical care, nonfood expenditure and total consumption. The vertical axis shows the annual amount of PCE, while the horizontal axis shows the proportion of the total population. Suppose, for example, that average nonfood expenditure represents a realistic budget constraint on patients’ ability to pay for HIV/AIDS treatment. Then the subsidised price of antiretroviral treatments for adults with AIDS -- around 9,000 baht or USD 500 from Figure 1-- seems unaffordable for around 50% of the whole population, while still being affordable for the rest.

These empirical examples from Thailand -- a middle-income developing country -- suggest that large-scale programmes of antiretroviral treatment may not be easily affordable either from the fiscal standpoint of governments or by many private households, except for prevention of perinatal transmission.

Efficiency considerations point in the same direction as affordability. Figure 4 shows the simulations of effectiveness based on optimistic estimates of clinical efficacy, and assuming perfect patient compliance and universal coverage -- in practice, however, one might expect programme coverage and compliance rates to be much less than 100%. Effectiveness is measured in terms of the proportionate reduction in the burden of disease due to HIV/AIDS -- or gain in quality-adjusted life years (QALYs) -- that can be attributed to different treatment options. The simulations suggest, first, that the potential effectiveness of antiretroviral treatment options -- excluding protease inhibitors -- is clustered around 10-15% of the baseline disease burden. And second, the effectiveness of preventing perinatal transmission may be nearly as high as treating adults.

Putting these effectiveness estimates together with the cost figures makes it possible to compare the cost-effectiveness of antiretroviral alternatives (Figure 5). The adult antiretroviral regimes generate around 30 QALYs per million baht of expenditure. In contrast, preventing perinatal transmission appears to produce more than 600 QALYs per million baht -- a twentyfold difference.

AIDS Economics Home Page

ARV Treatment: Special Focus Back to Main Page