Section I: Informing Scale-up and Resource Allocation
Universal access to HIV treatment in developing countries: going beyond the misinterpretations of the ‘cost-effectiveness’ algorithm
Moatti, Jean Paula,b; Marlink, Richardc,d; Luchini, Stephaneb,e; Kazatchkine, Michelf
From the aUniversity of the Mediterranean, Marseilles, France
bINSERM/IRD/University of the Mediterranean Research Unit 912 ‘Economic and Social Sciences, Health Systems and Societies’, Marseilles, France
cHarvard School of Public Health, Boston, Massachusetts, USA
dElizabeth Glaser Pediatric AIDS Foundation, Santa Monica, California, USA
eGREQAM, Research Group on Quantitative Economics, CNRS/EHESS, Marseilles, France
fGlobal Fund for AIDS, Tuberculosis and Malaria, Geneva, Switzerland.
Correspondence to Jean Paul Moatti, INSERM U379, 23 Rue Stanislas Torrents, 13006 Marseille, France. Tel: +33 612282000; e-mail: firstname.lastname@example.org
Background: Economic cost-effectiveness analysis (CEA) has been proposed as the appropriate tool to set priorities for resource allocation among available health interventions. Controversy remains about the way CEA should be used in the field of HIV/AIDS.
Methods and objectives: This paper reviews the general literature in health economics and public economics about the use of CEA for priority setting in public health, in order better to inform current debates about resource allocation in the fight against HIV/AIDS.
Results: Theoretical and practical limitations of CEA do not raise major problems when it is applied to compare alternatives for treating the same medical condition or public health problem. Using CEA to set priorities among different health interventions by ranking them from the lowest to the highest values of their cost per life-year saved is appropriate only under the very restrictive and unrealistic assumptions that all interventions compared are discrete and finite alternatives that cannot vary in terms of size and scale. In order for CEA to inform resource allocation compared across programmes to fight the AIDS epidemic, a pragmatic interpretation of this economic approach, like that proposed by the Commission on Macroeconomics and Health, is better suited. Interventions, like a number of preventive strategies and first-line antiretroviral treatments for HIV, whose marginal costs per additional life-year saved are less than three times the gross domestic product per capita, should be considered cost-effective.
Conclusion: Because of their empirical and theoretical limitations, results of CEA should only be one element in priority setting among interventions for HIV/AIDS, which should also be informed by explicit debates about societal and ethical preferences.
Money targeted at improving public health in developing countries must be spent where it can yield the highest returns. Since the advent of HAART in 1996, controversy has persisted about the economic rationale of scaling up access to these effective treatments of HIV infection in resource-limited settings. Some experts have argued that antiretroviral treatment (ART) was not a rational investment because it was not cost-effective compared with HIV prevention measures that would ‘save more lives’ [1–4]. By contrast, authors of empirical studies based on data collected alongside actual programme scale-up in developing countries have claimed that ART can be made cost-effective in such contexts [5–11]. Each side of the controversy refers to different interpretations of the cost-effectiveness criterion. The former tend to consider that, in order to maximize total health benefits under the constraint of limited resources, the selection of priorities between various health interventions should be strictly based on the ranking of their respective marginal costs per life-year saved [or any other standardized unit of outcome that can be used such as quality-adjusted life-years (QALY) or disability-adjusted life years], starting with those interventions with the lowest marginal cost until the total budget constraint is met. The latter rather propose a flexible use of the cost-effectiveness tool in which ratios of marginal costs per unit of outcome only help identify interventions that are worth taking into consideration (all those who remain below certain thresholds), and cost-effectiveness is only one consideration in allocating resources between specific diseases and interventions.
In June 2006, the 192 member states of the United Nations General Assembly committed themselves to the goal of ‘universal access to comprehensive prevention programmes, treatment, care and support’ for HIV/AIDS by 2010 , seemingly putting an end to this controversy. Some experts have, however, given recent impetus to the debate about the cost-effective use of healthcare resources in developing countries by arguing that targeting funds towards high profile diseases, such as AIDS, would fail to improve global health and could push poor societies into deeper trouble [13,14]. We hope to shed new light on this ongoing controversy.
The first section of this paper briefly summarizes the literature on health economics and public economics in order to show that the ‘cost-effectiveness maximization algorithm’ for allocating resources among health programmes is often misinterpreted, as it is in the case in which HIV prevention and treatment are opposed to each other on cost-effectiveness grounds. The second section argues that, even when it is correctly interpreted, this algorithm does not provide an appropriate basis for policy decisions. A more flexible and pragmatic use of the cost-effectiveness criterion is more consistent with both welfare economics and practical constraints faced by decision makers, and clearly gives an economic rationale in favour of ART for HIV with first-line drugs. The conclusion suggests that other economic tools in addition to cost-effectiveness analysis (CEA) should be mobilized in operational research applied to HIV prevention and care in developing countries.
What the cost-effectiveness algorithm should not be used for in resource allocation for HIV/AIDS
CEA and its derivative, cost–utility analysis (CUA), are methods of summarizing information on the relationship between resources expended on health interventions measured in monetary amounts, and the health outcomes measured in numerical units, in terms of the change in health or functional status resulting from these interventions . The resulting ratios are commonly expressed in CEA as cost per statistical life-year saved in, and in CUA as cost per QALY saved. Although still a matter of debate in some academic circles, CEA can be described as a truncated form of the standard tool for economic evaluation of public investments, cost–benefit analysis (CBA). In CEA, the measure of the health outcome is separated from other benefits and costs simply because assigning a dollar figure to human life-years or QALY may often in practice seem difficult and controversial [16–18].
Implicit to this approach, however, are some debatable assumptions that depart from the standard framework of welfare economics. In CBA, the measure of an individual's overall ‘utility’ is a function of multiple attributes [19,20]. In practice, CEA and CUA imply that individuals only derive utility from two attributes, their state of health and their consumption of goods and services. Such simplifications do not raise major methodological problems as long as CEA and CUA are applied to compare mutually exclusive strategies, i.e. alternative ways for treating the same medical condition or the same public health problem , such as an economic assessment associated with a clinical trial comparing two different ART regimens. The vast majority of CEA and CUA papers published in the health economics literature deal with such mutually exclusive strategies.
When CEA is extended to compare mutually compatible health interventions, i.e. to allocate resources across different interventions that could be implemented during the same time period, its underlying premise remains the same. The comparative analysis of various interventions still aims to maximize the total aggregate health benefits obtained from a given level of resources, or else to minimize the cost of achieving a particular health goal [22–24]. More complex methodological limitations occur, however, when results of CEA are used to inform priority decisions in budget allocation among various health interventions in general, among various interventions for HIV prevention and care, or between vertical programmes targeting specific diseases as opposed to improvements of the basic public health infrastructure. Because of such limitations, most economists no longer advocate the use of standardized ‘league’ tables of cost per QALY (or cost per life-year) ratios as the basis for budget allocation among health interventions in developed countries. Public health proponents, however, still present the cost-effectiveness ‘maximization algorithm’ as a straightforward tool to determine priority ranking among health interventions for developing countries [25–31]. In addition to the problem of dealing with developed and developing countries on different terms, this approach is based on very restrictive assumptions that do not apply to most practical situations for resource allocation.
A usual presentation of the cost-effectiveness algorithm in the public health literature, either explicit or implicit, can be summarized through the following simple example. Suppose a decision maker has to allocate a limited budget (BC for budget constraint) between six different interventions (Ii) that are mutually compatible, i.e. each intervention is feasible and the costs and health outcomes of any of them are independent of whether or not the other interventions are adopted. Suppose that for each Ii, a given cost (Ci) and an incremental gain in the aggregate health status of the targeted populations (ΔIi) can be measured relative to the costs and health outcomes associated with the status quo (defined as the prevailing situation if Ii is not implemented). Ranking the interventions from the lowest to the highest values of their cost-effectiveness ratios (Ci/ΔIi), and going down the list until BC is just met, may seem the only way to maximize the total health increment obtained by using available resources. This is illustrated in a ‘virtual’ league table whose use may seem very simple (Table 1). If, for example, BC is determined to total US$100 000, Table 1 clearly shows that I1, I2 and I3 should be selected because no alternative combination of Ii can provide a higher aggregate gain (1250 life-years saved) for this level of funding. If an additional US$55 000 becomes available, Table 1 recommends that I4 should be added. If the budget remains US$100 000, but a new intervention I7, whose total cost is similar to I3, is shown to have a lower cost-effectiveness ratio, then I7 should be preferred to I3. When applied to the various available interventions for fighting the AIDS epidemic, this algorithm would ineluctably lead to the conclusion that ‘a strong economic case exists for the prioritization of prevention services’, such as selective blood safety measures, targeted condom distribution, cotrimoxazole prophylaxis in HIV-positive infants and adults, and prevention of mother-to-child transmission that exhibit the lower average cost per life-year saved, rather than scaling up access to ART .
A more careful examination of this algorithm, however, reveals several drawbacks that strongly limit its potential use for the optimal allocation of resources among interventions. First, the algorithm does not tell us anything about the optimal amount of the total budget from a societal point of view. In order to maximize social benefits, it may be more appropriate to spend more (US$485 000 rather than US$100 000) and thus implement the whole package of six interventions in Table 1. Second, the algorithm does not protect against the fact that some potentially more cost-effective interventions may have been ignored (as would be the case if the above-mentioned I7 actually existed) and is therefore sensitive only to the scope of alternatives considered. Third, the algorithm only works if total incremental costs and benefits have been properly measured in a similar standardized way for all interventions. Such standardization is often lacking in league tables comparing ratios that come from different sources, notably because values of incremental costs and benefits are very sensitive to the choice of the reference ‘base case’ scenario that has been used to figure the status-quo situation in each empirical CEA [32,33]. Fourth, the application of the algorithm may be sensitive to the specific size of each intervention, leading to biases caused by the ‘lumpiness’ of interventions . For example, imagine that BC is reduced to US$80 000. In choosing between interventions in Table 1, it will follow that I3 will have to be eliminated because its implementation would exceed the available budget and I5 will be chosen by the cost-effectiveness algorithm. This choice would not, however, correspond to the maximization of benefit and, indeed, for that same amount of US$80 000, I2 and I3 would provide greater benefit than the combination of I1, I2 and I5.
This so-called ‘lumpiness’ problem calls attention to a more general limitation of the algorithm presented at the beginning of this section: it does not allow for the fact that the cost-effectiveness ratio of most health interventions varies according to their scope and scale. Many types of health interventions can be expanded or contracted either by applying the same intervention to a larger (inversely a smaller) pool of individuals or by varying the intensity of the intervention per individual (for example, different antiretroviral drug regimens for first-line HIV treatment may have different cost-effectiveness ratios in patients presenting similar clinical and immunovirological characteristics). The usual cost-effectiveness algorithm in the public health literature only works under the very restrictive and unrealistic assumption that all compared interventions are discrete and finite alternatives that cannot vary in terms of their size and scale . If we forego this assumption, it cannot be inferred from Table 1, as some experts sometimes conclude, that I6 (for example, ART for HIV care) would always be less cost-effective than I1 (for example, social marketing of condoms). This would be the case only if any increase in scale of I1 has no impact on the incremental cost-effectiveness ratio. In particular, this would mean that average costs of I1 and marginal costs per additional unit of outcome remain equal to each other, i.e. that returns of the intervention are always constant. This situation is quite unlikely, however, because returns of each intervention vary according to various scales of implementation. The marginal cost-effectiveness ratio (MCER) may be below or above the overall cost-effectiveness ratio depending on the scale that is chosen.
In the real world, health interventions often follow the law of diminishing returns. Successive equal unit addition of inputs will result, from some point on, in additions of output at a diminishing rate . For example, although the unit cost of condoms is low, increased efforts are needed to promote their use in groups in which social and cultural barriers are difficult to overcome. The cost per life-year saved may increase exponentially to the point that the MCER of condom promotion (I1 in our example) would become higher than that of ART in some groups of patients (I6). The notion that ART would never be cost-effective in developing countries assumed that the implementation of other strategies for HIV care and prevention, whatever their decreasing returns, will always dominate even the most cost-effective strategies using ART (see Hypothesis 1 in Fig. 1, in which the curve of the MCER for ART is always dominated). In fact, it is certain that ART will become more cost-effective at some levels of respective scales of alternative interventions (as shown with Hypothesis 2 in Fig. 1, in which two MCER curves intersect).
How cost-effectiveness analysis should be appropriately used in resource allocation for HIV/AIDS
In order to maximize the total health benefit under budget constraint effectively when choosing among various interventions that may vary in scope and scale, and to determine the optimal package of interventions, a correct cost-effectiveness algorithm implies that the magnitude of each intervention is defined in a way that equalizes the MCER of all compared interventions [18,21]. Intuitively, that means that I1 in Table 1 should be extended in scale until its MCER becomes equivalent to that of I2 and so forth until the MCER of each of the six interventions converge to an identical value.
In practice, however, the scale of any intervention may be influenced by a set of different parameters. The scale of an ART programme will be characterized not only by the quantity of available drugs, but also by the number of medical and health personnel, the laboratory equipment available for monitoring, the distance between the homes of the targeted population and the healthcare delivery facility, and so on. If a number of relevant dimensions affect scale (x1, x2,…xk), then total costs (TC) and total benefits (TB) will be a function of the value of each of these dimensions: TC(x1, x2,…xk) and TB (x1, x2,…xk). The MCERi for any given value of x1, x2,…xk and for each scale dimension i, can be defined as the derivatives of the two functions: (δTC/δ xi)/(δTB/δ xi). The ratio between these two terms represents the change in costs divided by the change in benefits that result from a one-unit change in scale dimension i.
The practical implementation of the theoretical cost-effectiveness maximization algorithm to optimize resource allocation across interventions therefore faces two major issues, one technical, one ethical.
In theory, the analyst should be able to estimate, for each compared intervention, a production function that relates the amounts invested to the associated changes in health outcomes. League tables usually provide a single point estimate and often do not specify the exact scale at which the measure has been performed [37–40]. Estimations of such functions for all interventions may, however, require considerable time and resources and are rarely available in the literature. Moreover, empirical observations may unwittingly introduce biases in estimating the production functions that would be appropriate for CEA. The way professionals treat the same clinical condition or address a public health problem may vary widely among and within countries [41–43]. Econometric analysis shows that variations in costs of HIV care are not entirely explained by individual clinical differences . An ongoing research project reveals that unit costs of HIV testing and counselling per client may vary by a factor of 10 to sometimes 100 between different developing countries, and between different sites within countries . These variations may actually reflect differences in the epidemiological mix of treated patients, as well as differences in economies of scale and economies of scope . The existence of economies of scale implies that there are efficiency gains to expanding the size of an activity: when the number of treated patients increases in a single facility, the unit cost per patient tends to decrease if there are economies of scale. The concept of economies of scope relates to the efficiency implications of diversifying the product line of the same facility: economies of scope will indeed exist if complementarities in the production of different outputs (such as integrating tuberculosis and HIV treatment in the same facility) lead to a decreased unit cost for all the outputs concerned; it is precisely the role of a production function to take those differences into account. A problem occurs, however, if these variations are also caused by differences in logistic and management capacities, or the ways providers and managers react to inappropriate economic incentives, tariffs and institutional arrangements . In this latter case, the cost-effectiveness ratio of some interventions may be unduly penalized by factors that are not intrinsically related to the intervention being evaluated, but are caused by external variations in the economic and managerial environment. Unbiased comparison of programme alternatives for economic evaluation implies that each alternative is given the same chance to succeed, i.e. in conditions of optimal productivity, but empirical data about an intervention may be based on observations that do not correspond to such maximum productivity. In-depth econometric analysis is clearly needed to separate the two types of effect, those related to the objective, intrinsic conditions of production, and those externally driven effects resulting from the ineffective design of health systems. Health economics research on healthcare reforms in developing countries certainly helps remedy the latter type of factors, but in the meantime system problems should not interfere with CEA of alternative interventions.
The ethical issue in correctly implementing the cost-effectiveness algorithm relates to the elicitation of individual and collective preferences that should be at the core of evaluating public investments, grounded in the framework of welfare economics . The equalization of marginal cost per life-year saved or per QALY in all circumstances may not correspond to societal preferences that may instead favour some degrees of difference between the marginal costs per unit of outcome of various health interventions. It can sometimes be argued that such preferences are fuelled by individual ‘biases’ in probabilistic risk assessment and by incomplete information that should be disregarded when making public choices . These differences may, however, also express legitimate preferences arising from equity considerations. Issues of vertical equity that are at the heart of egalitarian theories of social justice may lead to preferences for higher marginal costs per life-year for interventions that prioritize the most deprived sectors of the population in order to reduce the gaps between them and the rest of society [50,51]. They may also come from preferences related to the specific character of the health problem at stake : for example, society may express a stronger risk aversion for catastrophic losses (such as the simultaneous loss of 100 life-years because of a pandemic) than for the same absolute number of life-years lost from a chronic disease spread across time and space, and therefore call for higher costs per life-year saved at the margin in the former situation [53,54].
In recent years, alternative methodologies to the use of incremental cost-effectiveness ratios have been proposed in order to take into account societal preferences when implementing an economic assessment of health interventions. This includes: selecting programmes on the basis of a net benefits approach, whose objective is to maximize the total net benefit achieved, in which the total net benefit is the sum of the average benefits of the implemented programmes [55,56], as well as cost–value analysis that proposes technical solutions for the integration of distributive concerns into economic evaluation [57,58].
Because of the technical and ethical issues as well as other methodological problems associated with CEA, the idea that a unique MCER value threshold should guide the allocation of resources between the healthcare and public health sectors has never been considered feasible and has essentially been abandoned in developed countries . Instead, regulatory agencies charged with informing reimbursement decisions for drugs and medical procedures use a set of different MCER values . On the one hand, experts from high-income Organization for Economic Cooperation and Development countries usually agree that medical innovations should be adopted whose marginal costs per additional life-year saved are under US$50 000 [approximately twice the gross domestic product (GDP) per capita], whereas those with costs above US$150 000 (six times the GDP per capita) are too expensive for general use . Choices about interventions with MCER falling in the intermediary values require a more in-depth debate to elicit the specific societal preferences involved. On the other hand, the Commission on Macroeconomics and Health, as well as the World Health Organization–Choice working group, have recommended labelling interventions that have cost-effectiveness ratios of less than three times the GDP per capita as ‘cost-effective’ and those that cost less than one times the GDP per capita as ‘very cost-effective’ [62,63]. According to this latter criterion and based on evidence from HIV treatment programmes [5–11], first-line ART regimens should clearly be considered cost-effective in developing countries, even in those with the lowest GDP. The cost per additional life-year saved when switching patients to ART second-line regimens, however, remains higher than three times the GDP per capita of most developing countries because of the high price of second-line drugs. Efforts to decrease these prices would help make access to second-line regimens cost-effective everywhere. According to this pragmatic GDP criterion, incremental costs related to CD4 cell count monitoring of ART already appear to be economically acceptable in low-income countries, whereas viral load monitoring costs are still too high to be put to general use .
This pragmatic approach proposed for the use of cost-effectiveness ratios is, of course, open to criticism because there is still considerable debate about the exact threshold values that should be used in practice, and as it may not provide a sufficient answer to what health interventions should be prioritized and what should benefit from various forms of public subsidies (the number of programmes that would be considered ‘cost-effective’ may indeed exceed the total funding capacities of most public health systems in developing countries). It, however, corresponds to an interactive model influenced by ‘positive economics’ that seems best suited to the complex trade-offs involved in the decision-making process of the real world . It should also be noted that an extended set of cost-effectiveness values for determining which health programmes are socially acceptable better takes into account the theoretical and practical limitations that analysts encounter when they implement the cost-effectiveness maximization algorithm. In addition, attempts to articulate between the microeconomic level of analysis applied in CEA and CUA on the one hand, and the macroeconomic global impact of any health investment on the other  can give an economic theoretical rationale to some of the proposed threshold values. It has been shown that saving one additional life-year for less than once (or twice) the GDP per capita would correspond to a marginal benefit for the economy as a whole that is superior to its marginal cost, whereas spending more than six to eight times the GDP per capita to save one additional life-year definitely implies, through the opportunity cost of this investment, a loss of a higher number of statistical life-years in the rest of society .
Going beyond the cost-effectiveness algorithm
In conclusion, transferring economic evidence to operational decision making implies a number of challenges that include improvements in methodological validity and relevance of empirical applications of economic analysis. These challenges, however, go far beyond pure methodological debates because they have to deal with the incorporation of societal and political constraints as well as collective preferences in the process of priority setting for healthcare . As mentioned above, alternative economic methods to CEA may facilitate such incorporation, but a politico-economic analysis of funding decisions, as well as greater public and community participation in the decision-making process, are also clearly needed .
In the field of HIV/AIDS, a normative and restrictive application of cost-effectiveness league tables tends to oppose prevention to treatment rather than emphasize their synergies and complementarities . The pragmatic approach, based on a set of cost-effectiveness threshold values, that is already de facto implemented in developed countries and that has been recommended by the Commission on Macroeconomics and Health for developing countries, gives an economic rationale for including both a number of preventive and ART strategies in the optimal package of health services for HIV/AIDS. Such an approach clearly implies that results from CEA and CUA remain only one element in the decision-making process for setting health priorities that has to be balanced with other considerations of epidemiological burden, equity and political and social acceptability . Adopting this alternative pragmatic approach still leaves plenty of room for adequate CEA and CUA in HIV care.
Success in making ART cost-effective will depend on whether delivery of treatment simply duplicates the standards established in developed countries or creatively adapts them to the constraints of low-resource settings. World Health Organization guidelines for ART in resource-limited settings already advocate the use of the least expensive options both for first and second-line treatment and for laboratory tests to monitor viral load and CD4 cell counts . CEA and CUA can make future useful contributions by focusing on what they are really appropriate for: help in determining the optimal strategies between mutually exclusive options in terms of criteria for initiating treatment, biological monitoring protocols, drug regimens and modes of delivery of HIV services. These tools will, however, not be able to tackle the whole scope of economic and ethical issues related to scale up.
The proper implementation of programmes for scaling up and the achievement of universal access to ART and HIV care face other major challenges, all related to the long list of structural problems in the functioning of healthcare systems and public health policies of developing countries. These issues include lack of coordinated efforts among donors, poor healthcare infrastructure, largely insufficient primary care delivery at the decentralized level, lack of qualified health workforce (aggravated by brain drain to developed countries and AIDS-related mortality among health professionals), poor procurement and distribution mechanisms for drugs and other supplies, and a lack of long-term, sustainable domestic and foreign funding for healthcare [73,74]. Disease-targeted programmes, like ART scale up, will succeed only insofar as they effectively demonstrate positive effects on health systems and global health improvements and show that they do not jeopardize other public health efforts. Accordingly, future operational research projects need to measure the positive (or negative) impacts of HIV programmes on economies of scale and scope in the whole health system and to determine whether these programmes may (or may not) favour global reforms in financing and organizing healthcare delivery systems. To fulfill this purpose, operational research projects will have to borrow tools from a variety of fields in economics and management sciences that go far beyond CEA and CUA.
Conflicts of interest: None.
1. Ainsworth M, Teokul W. Breaking the silence: setting realistic priorities for AIDS control in less developed countries. Lancet 2000; 356:55–60.
2. Creese A, Floyd K, Alban A, Guinness L. Cost-effectiveness of HIV/AIDS interventions in Africa: a systematic review of the evidence. Lancet 2002; 359:1635–1642.
3. Marseille E, Hofmann PB, Kahn KG. HIV prevention before HAART in sub-Saharan Africa. Lancet 2002; 359:1851–1856.
4. Canning D. The economics of HIV/AIDS in low-income countries: the case for prevention. J Econ Perspect 2006; 3:121–142.
5. Badri M, Cleary S, Maartens G, Pitt J, Bekker LG, Orrell C, Wood R. When to initiate highly active antiretroviral therapy in sub-Saharan Africa? A South African cost-effectiveness study. Antiviral Ther 2006; 11:63–72.
6. Goldie SJ, Yazdanpanah Y, Losina E, Weinstein MC, Anglaret X, Walensky RP, et al
. Cost-effectiveness of HIV treatment in resource-poor settings– the case of Cote d'Ivoire. N Engl J Med 2006; 355:1141–1153.
7. Over M, Revenga A, Masaki E, Peerapatanapokin W, Gold J, Tangcharoensathien V, Thanprasertsuk S. The economics of effective AIDS treatment in Thailand. AIDS 2007; 21(Suppl 4):S105–S116.
8. Hogan DR, Baltussen R, Hayashi C, Lauer JA, Salomon JA. Cost effectiveness analysis of strategies to combat HIV/AIDS in developing countries. BMJ 2005; 331:1431–1437.
9. Gold J, Heywood P, Over M, Gupta I, Subhah H. HIV/AIDS treatment and prevention in India. Modeling the costs and consequences. Washington, DC: World Bank Publications; 2004.
10. Laurent C, Ngom Gueye NF, Ndour CT, Gueye PM, Diouf M, Diakhate N, et al
, ANRS 1215/1290 Study Group. Long-term benefits of highly active antiretroviral therapy in Senegalese HIV-1-infected adults. J Acquir Immune Defic Syndr 2005; 38:14–17.
11. Diomande FV, Bissagnene E, Nkengasong JN, Maurice C, Monga B, Laga M, Nolan ML. The most efficient use of resources to identify those in need of antiretroviral treatment in Africa: empirical data from Cote d'Ivoire's drug access initiative. AIDS 2003; 17(Suppl. 3):S87–S93.
12. United Nations. High level meeting on HIV/AIDS.
Political declaration on HIV/AIDS. Resolution A/RES/60/262. Washington, DC: United Nations; 15 June 2006.
14. England R. Are we spending too much on HIV? BMJ 2007; 334:344.
15. Gold MR, Siegel JE, Russell LB, Weinstein MC, editors. Cost-effectiveness in health and medicine.
New York: Oxford University Press; 1996.
16. Mishan EJ. Cost–benefit analysis: an informal introduction. 4th edn London: Unwin Hyman; 1988.
17. Phelps C, Mushlin A. On the (near) equivalence of cost-effectiveness and cost–benefit analysis. Int J Technol Assess Health Care 1991; 7:12–21.
18. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. J Health Econ 1997; 16:1–31.
19. Keeney R, Raiffa H. Decisions with multiple objectives. New York: John Wiley and Sons; 1976.
20. Torrance GW, Furlong W, Feeny D, Boyle M. Multiattribute preference functions: health utilities index. Pharmacoeconomics 1995; 7:503–520.
21. Garber AM. Advances in cost-effectiveness analysis of health interventions. In: Culyer, A.J., Newhouse, J.P., editors. Handbook of health economics 2000
, edn 1, vol 1. Amsterdam: Elsevier Science BV; 2000. pp. 181–221.
22. Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and medical practices. N Engl J Med 1977; 296:716–721.
23. World Bank. World development report: investing in health.
Washington, DC: World Bank Publications; 1993.
24. Chapman RH, Stone PW, Sandberg EA, Bell C, Neumann PJ. A comprehensive league table of cost–utility ratios and a sub-table of “panel-worthy” studies. Med Decis Making 2000; 20:451–467.
25. Stone PW, Teutsch S, Chapman RH, Bell C, Goldie SJ, Neumann PJ. Cost–utility analyses of clinical preventive services– published ratios, 1976–1997. Am J Prev Med 2000; 19:15–23.
26. Vassall A, Compernolle P. Estimating the resource needs of scaling-up HIV/AIDS and tuberculosis interventions in sub-Saharan Africa: a systematic review for national policy makers and planners. Health Policy 2006; 79:1–15.
27. Currie CSM, Floyd K, Williams BG, Dye C. Cost, affordability and cost-effectiveness of strategies to control tuberculosis in countries with high HIV prevalence. BMC Public Health 2005; 5:130.
28. Harling G, Wood R, Beck EJ. Efficiency of interventions in HIV infection, 1994–2004. Dis Manag Health Outcomes 2005; 13:371–394.
29. Hogan DR, Baltussen R, Hayashi C, Lauer JA, Salomon JA. Achieving the millennium development goals for health– cost effectiveness analysis of strategies to combat HIV/AIDS in developing countries. BMJ 2005; 331:1431–1435.
30. Hutton G, Wyss K, N'Diekhor Y. Prioritization of prevention activities to combat the spread of HIV/AIDS in resource constrained settings: a cost-effectiveness analysis from Chad, Central Africa. Int J Health Plan Manag 2003; 18:117–136.
31. Pinkerton SD, Johnson-Masotti AP, Holtgrave DR, Farnham PG. Using cost-effectiveness league tables to compare interventions to prevent sexual transmission of HIV. AIDS 2001; 15:917–928.
32. Gerard K. Cost–utility in practice– a policymaker's guide to the state-of-the-art. Health Policy 1992; 21:249–279.
33. Paalman M, Bekedam H, Hawken L, Nyheim D. A critical review of priority setting in the health sector: the methodology of the 1993 World Development Report. Health Policy Plan 1998; 13:13–31.
34. Bartezzaghi E, Verganti R, Zotteri G. A simulation framework for forecasting uncertain lumpy demand. Int J Production Econ 1999; 59:499–510.
35. Karlsson G, Johannesson M. The decision rules of cost-effectiveness analysis. Pharmacoeconomics 1996; 9:113–120.
36. Fellner W. Modern economic analysis. New York: McGraw-Hill; 1960.
37. Zaric GS, Brandeau ML. Optimal investment in a portfolio of HIV prevention programs. Med Decis Making 2001; 21:391–408.
38. Johannesson M, Johansson PO, Kristrom B, Borgquist L, Jonsson B. Willingness-to-pay for lipid-lowering: a health production function-approach. Applied Econ 1993; 25:1023–1031.
39. Street A. How much confidence should we place in efficiency estimates? Health Econ 2003; 12:895–907.
40. Healey A, Mirandola M, Amaddeo F, Bonizzato P, Tansella M. Using health production functions to evaluate treatment effectiveness: an application to a community mental health service. Health Econ 2000; 9:373–383.
41. Westert GP, Groenewegen PP, Boshuizen HC, Spreeuwenberg PMM, Steultjens MPM. Medical practice variations in hospital care; time trends of a spatial phenomenon. Health & Place 2004; 10:215–220.
42. Gulliford MC, Mahabir D, Ukoumunne OC. Evaluating variations in medical practice between government primary care health centres. J Clin Epidemiol 2001; 54:511–517.
43. Davis P, Gribben B, Scott A, Lay-Yee R. The “supply hypothesis” and medical practice variation in primary care: testing economic and clinical models of inter-practitioner variation. Soc Sci Med 2000; 50:407–418.
44. Aracena B, Gutiérrez JP, Bertozzi SM, Gertler P. Cost of AIDS care in Mexico: what are its main individual predictors? Arch Med Res 2005; 36:560–566.
45. Kumaranayake L. The economics of scaling up: cost estimation for HIV/AIDS interventions. AIDS 2008; 22(Suppl. 1):S23–S33.
46. Preyra C, Pink G. Scale and scope efficiencies through hospital consolidations. J Health Econ 2006; 25:1049–1068.
47. Ellis RP, McGuire TG. Predictability and predictiveness in health care spending. J Health Econ 2007; 26:25–48.
48. Moatti JP. Ethical issues in the economic assessment of health care technologies. Health Care Anal 1999; 7:153–165.
49. Gaba A, Viscusi WK. Differences in subjective risk thresholds: worker groups as an example. Manag Sci 1998; 44:801–811.
50. Rawls J. A theory of justice. Cambridge, MA: Harvard University Press; 1971.
51. Bommier A, Stecklov G. Defining health inequality: why Rawls succeeds where social welfare theory fails. J Health Econ 2002; 21:497–513.
52. Pratt JW, Zeckhauser RJ. Willingness to pay and the distribution of risk and wealth. J Polit Econ 1996; 104:747–763.
53. Bastide S, Moatti JP, Pages JP, Fagnani F. Risk perception and social acceptability of technologies: the French case. Risk Anal 1989; 9:215–223.
54. Chilton S, Covey J, Hopkins L, Jones-Lee M, Loomes G, Pidgeon N, Spencer A. Public perceptions of risk and preference-based values of safety. J Risk Uncertainty 2002; 25:211–232.
55. Stinnett AA, Mullahy J. Net health benefits: a new framework for the analysis of unceretainty in cost-effectiveness analysis. Med Decis Making 1998; 18(Suppl 2):S68–S80.
56. Laska EM, Meisner M, Siegel C, Stinnett AA. Ratio-based and net benefit-based approaches to health care resource allocation: proofs of optimality and equivalence. Health Econ 1999; 8:171–174.
57. Nord E. The trade-off between severity of illness and treatment effect in cost–value analysis of health care. Health Policy 1993; 24:227–238.
58. Nord E, Pinto JL, Richardson J, Menzel P, Ubel P. Incorporating societal concerns for fairness in numerical valuations of health programmes. Health Econ 1999; 8:25–39.
59. Tsuchiya A, Williams A. Welfare economics and economic evaluation. In: Drummond M, McGuire A, editors. Economic evaluation in health care: merging theory and practice. Oxford and New York: Oxford University Press; 2001. pp. 22–45.
60. Devlin N, Parkin D. Does NICE have a cost-effectiveness threshold and what other factors influence its decisions? A binary choice analysis. Health Econ 2004; 13:437–452.
61. Towse A, Pritchard C, Devlin N. Cost-effectiveness thresholds: economic and ethical issues. London: King's Fund and Office of Health Economics; 2002.
62. Sachs JD, editor. Macroeconomics and health: investing in health for economic development.
Report of the Commission on Macroeconomics and Health (CMH) of the World Health Organization. Geneva: World Health Organization; 2001.
63. Hutubessy RCW, Baltussen RMPM, Torres-Edejer TT, Evans DB, WHO–CHOICE Working Group. WHO–CHOICE: choosing interventions that are cost-effective.
In: Murray, C.J.L., Evans, D.B., editors. Health systems performance assessment. Debates, methods and empiricism.
Geneva, Switzerland: WHO Editions, 2003. pp. 823–835.
64. Bishai D, Colchero A, Durack DT. The cost effectiveness of antiretroviral treatment strategies in resource-limited settings. AIDS 2007; 21:1333–1340.
65. Williams I, Bryana S. Understanding the limited impact of economic evaluation in health care resource allocation: a conceptual framework. Health Policy 2007; 80:135–143.
66. Jones-Lee MW, editor. The value of life and safety.
Amsterdam: North Holland Publishing Company; 1982.
67. Lutter R, Morrall JF, Viscusi WK. The cost-per-life-saved cut-off for safety-enhancing regulations. Econ Inquiry 1999; 37:599–608.
68. Cookson R, Hutton J. Regulating the economic evaluation of pharmaceuticals and medical devices: an European perspective. Health Policy 2003; 63:167–178.
69. Jan S, Dommers E, Mooney G. A politico-economic analysis of decision making in funding health service organizations. Soc Sci Med 2003; 57:427–435.
70. Moatti JP, N'Doye I, Hammer SM, Hale P, Kazatchkine M. Antiretroviral treatment for HIV infection in developing countries: an attainable new paradigm. Nat Med 2003; 9:1449–1452.
71. Laxminarayan R, Mills AJ, Brenan JG, et al
. Advancement of global health: key messages from the Disease Control Priorities Project. Lancet 2006; 367:1193–1208.
72. World Health Organization. Scaling up antiretroviral therapy in resource-limited ssettings. Guidelines for a public health approach. Geneva: World Health Organization HIV/AIDS Department; 2002.
73. Preker AS, Langenbrunner JC, editors. Spending wisely: buying health services for the poor
. Washington, DC: The World Bank; 2005.
74. Wester CW, Bussmann H, Avalos A, Ndwapi N, Gaolathe T, Cardiello P, et al
. Establishment of a public antiretroviral treatment clinic for adults in urban Botswana: lessons learned. Clin Infect Dis 2005; 40:1041–1044.
cost-effectiveness; welfare economics; low-resource settings; HIV prevention; antiretroviral treatment
© 2008 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.