This issue includes an article on the average cost of antiretroviral therapy (ART) in five countries that receive bilateral support from the United States President's Emergency Plan for AIDS Relief (PEPFAR) . The article is pathbreaking as the first peer-reviewed publication of estimates of the unit costs of delivering ART in a range of countries through the US bilateral program. The US government assistance for the global HIV epidemic has risen from US$ 595 million in 2001 to a proposed US$ 5.6 billion in the 2012 budget request (http://www.kff.org/hivaids/upload/7029-07.pdf, accessed 27 May 2011), this does not include contributions to the Global Fund to Fight AIDS, Tuberculosis and Malaria: a request of US$ 1.3 billion was made for 2012 of which about US$ 0.7 billion may go to HIV based on past allocations). However, after years of geometric increases in donor financing for HIV treatment, funding from the US government has leveled off and from European donors actually decreased. Also, at the end of 2009, nearly 15 million people were estimated to be in need of ART, whereas only 5.2 million people were receiving it. Clearly, more resources are needed, but in today's economic climate funding gaps are expected. Now more than ever, it is absolutely essential that our limited resources be used as efficiently and effectively as possible. So the cost estimates in this article contribute to an active ongoing international effort at finding innovative ways to reduce the cost of ART, while improving its quality, so that existing funds can sustain more patients for more years. In this editorial, we comment on the Menzies et al.  study in light of this overarching context and raise the question of whether this cost study, or cost studies in general, can improve the performance of donor-financed HIV treatment.
The on-site mean annual costs for ART patients were US$ 896 (2009 US$) among a sample of 34 sites in four countries (Ethiopia, Nigeria, Uganda and Vietnam – the article explains that the unavailability of data on drug costs from Botswana forces the exclusion of the nine Botswana sites from the calculations of total average on-site cost). Although the article presents on-site average costs in each country, owing to its small sample size and purposive sampling strategy, it cannot conclude that costs differ by country to a statistically significant degree. However, although the annual costs clustered around US$ 600–999 for ART patients (still a difference of a factor of 1.7), the reported range in average annual cost at the level of the clinic was US$ 303–1979 (a difference of a factor of 6.5). These data clearly suggest that there are opportunities for significant efficiency improvements in HIV treatment programs.
The authors point to several factors that caused the observed variation. What seemed to matter most was the maturity of the site, that is how long it had been treating patients. Clearly, two phenomena may be occurring: an increase in service volume and learning effects. They also show evidence that patient mix matters as care for newly initiated ART patients cost 15–20% more than for established patients. Finally, they also hypothesized that the variation may reflect differences in factor input prices and in the package of services provided.
However, to what extent were the higher costs of the more expensive sites in Menzies et al.  due to factors that might legitimately increase the cost of treatment such as the complexity/severity of recruited patients (as measured by their stage of illness or their baseline CD4 cell counts), the quality of treatment they receive (as proxied by a facility's loss to follow-up or the average 1-year improvement in CD4 cell counts) or the cost of good management by on-site and off-site supervisors? To what extent were higher costs or lower quality due to inadequate scale, scope (i.e. integration with other health services), training, experience or governance? Of these determinants of efficiency and effectiveness, governance may be the most difficult to measure in observational data or to control and predict during future scale-up. Studies by McPake et al.  and more recently by Chaudhury et al.  and Leonard  foreshadowed the recent scandal at the Global Fund to Fight AIDS, Tuberculosis and Malaria that led to the suspension of fund agreements in Uganda and more recently in Djibouti, Mali, Mauritania and Zambia. (http://www.theglobalfund.org/en/pressreleases/?pr=pr_110124, accessed 16 May 2011). Future studies of the average cost of ART should investigate the degree to which aspects of institutional structure and employee incentives affect the quality of governance and should include the cost of deploying these management tools as a component of average cost. The plethora of plausible determinants of average costs means that an observational study such as this would need perhaps three to five times as large a sample randomly selected from a wider range of countries and institutional settings to answer these questions.
Although it is clear that opportunities for optimizing HIV treatment programs exist, we as a community have focused more on expanding access and cutting drug prices than on maximizing value in HIV treatment, resulting in minimal effort dedicated to understanding the causes and magnitude of inefficiencies. Historically, cost data on treatment program costs have been poorly tracked and have focused mostly on drug procurements. The increasing reliance on generics has achieved significant cost savings; although ART programs expanded significantly from 2005 to 2008, drug expenditures did not increase proportionally, resulting in savings of over US$ 300 million between 2005 and 2008 . However, we must build upon the work presented by Menzies et al.  to analyze and identify solutions to achieve similar clinical outcomes with fewer resources, or ‘efficiently effective’ treatment. We urgently need to identify best practices currently being implemented and drive other countries, programs and sites toward those best practices, whether they are in drug procurement, task shifting or reduction in unnecessary laboratory tests. For example, recent data suggests that in certain cases less intensive monitoring of routine HIV care is possible without an important decrement in clinical outcomes .
Future work by the US Office of the Global AIDS Coordinator (OGAC), other donors and governments and other independent researchers, using larger, more representative samples of HIV treatment facilities, should build on this pathbreaking article with the objective of estimating a function that relates the cost of complexity/quality-adjusted patient-years of treatment to its most important determinants, especially including policy instruments under potential government control. Such a function would provide immediate guidance to PEPFAR, other donors and recipient governments on maximizing the health benefits from their HIV treatment expenditures and will suggest hypotheses for implementation research projects designed to yield even greater efficiency improvements.
Conflicts of interest
The Bill & Melinda Gates Foundation's HIV Strategic Program Team includes an ‘Efficiency and Effectiveness’ initiative which aims to increase the impact per dollar spent on HIV programs. As part of this initiative, the foundation is supporting a study to estimate the cost of delivering HIV treatment in five sub-Saharan African countries. The Center for Global Development has been contracted by the foundation to provide Mead Over's technical support to this study. The views expressed herein are those of the authors and do not necessarily reflect the official policy or position of the Bill & Melinda Gates Foundation.
1. Menzies NA, Berruti AA, Berzon R, Filler S, Ferris R, Ellerbrock TV, Blandford JM. The cost of providing comprehensive HIV treatment in PEPFAR-supported programs. AIDS
2. McPake B, Asiimwe D, Mwesigye F, Ofumbi M, Ortenblad L, Streefland P, Turinde A. Informal economic activities of public health workers in Uganda: implications for quality and accessibility of care
. Soc Sci Med
3. Chaudhury N, Hammer J, Kremer M, Muralidharan K, Rogers FH. Missing in action: teacher and health worker absence in developing countries
. J Econ Perspect
4. Leonard KL. The cost of imperfect agency in healthcare: evidence from rural Cameroun
. J Dev Econ
5. Holmes CB, Coggin W, Jamieson D, Mihm H, Granich R, Savio P, et al. Use of generic antiretroviral agents and cost savings in PEPFAR treatment programs
6. DART Trial Team. Routine versus clinically driven laboratory monitoring of HIV antiretroviral therapy in Africa (DART): a randomized noninferiority trial. Lancet