UNAIDS prevalence estimates formed the basis of the missed "3 by 5" target, and so the news that they were so highly inflated is particularly timely coming on the heels of the WHO report on the initiative, which (among other things) illustrates the perverse relationship between disease estimates, advocacy targets and pharmaceutical markets. The "3 by 5" strategy was based on an analysis of resource needs which determined that "with optimal funding and technical capacity, access to lifesaving antiretroviral therapy could be expanded to reach 3 million people globally - half of those estimated to need it - by the end of 2005" (emphasis mine). It was explicitly designed as an ambitious advocacy target to mobilize resources, based almost entirely on the artificially high disease-burden estimates and constrained only by the absorptive capacity of recipient countries. It emphatically was not a realistic forecast of demand for antiretrovirals that could effectively inform the pharmaceutical industry's investment and manufacturing decisions. In fact, the WHO report implicitly defines "demand" as the number of people seeking treatment, completely divorced from their (or donors') willingness and ability to pay:
Most countries with a generalized epidemic have linear growth in the number of people seeking antiretroviral therapy. This indicates that, although stigma and lack of perceived benefits of treatment may slow down the uptake of antiretroviral therapy, especially in the early stages of scale-up, demand is not the limiting factor in scale up. Instead, the rate of increase is determined primarily by supply-side factors such as drug supply, funding, identifying HIV status and human resource capacity.
This definitional misunderstanding blurs the distinction between need and demand and indirectly contributes to pharmaceutical shortages and price constraints; it also represents a conflict of interest between the WHO's promotion of the advocacy targets and its provision of credible demand forecasting data for suppliers. The international community ultimately fell far short of the "3 by 5" target partly because of this disconnect between the campaign goals and the market realities, with the side effect of damaging prospects for future advocacy efforts.
With 1.3 million people now receiving treatment up from just 400,000 two years ago, the "3 by 5" initiative should not be regarded as a complete failure. It does, however, highlight some of the problems stemming from the use of 'outcome' targets as the centerpiece of resource mobilization campaigns, and points to some lessons about what to do differently in the future (although it is not yet clear if WHO itself has learned them). Going forward, the international community should take several steps if it is truly serious about increasing access to ARVs and other pharmaceutical products:
- Improve 'needs' estimates related to the disease burden and affected population, both by carefully selecting proxies and improving the 'measurability' of those indicators through better data collection at the country level. This area has already received increased attention with the recent establishment of the Health Metrics Network, and is likely to receive even more in light of the UNAIDS overestimates.
- Improve resource tracking of financial flows for different health priorities to inform both resource allocations and advocacy efforts, and increase the predictability of that funding where possible.
- Develop a framework to forecast global demand for health products based on information-sharing, accountability and improved methodologies, allowing suppliers to make informed investment and manufacturing decisions and to share the risk of those uncertainties with the purchasers.
The upshot of these policy actions would be greater efficiency in the supply chain, less risk and uncertainty, and ultimately lower prices. In combination with reliable estimates of donor support, this would enable the development of realistic and achievable treatment targets. Although these numbers may not be quite as 'sexy' for campaigning, they are far less likely to fail.
The WHO recognizes some of these pitfalls in its retrospective analysis of the "3 by 5" strategy, while still endorsing the use of future targets:
National targets have been shown to play a valuable role in mobilizing action and increasing accountability among stakeholders, including international technical agencies, donors and governments. Nevertheless, it is clear that treating half of those in need by the end of 2005 was not realistic for all countries, particularly those with very weak health infrastructure and a very high burden of disease. Future country level targets will need to be sufficiently ambitious to mobilize action, will need to be country-driven and should take into account factors other than burden of disease, such as local capacity. Targets for treatment need to be complemented by achievable targets for the other elements of a comprehensive response to HIV/AIDS, including prevention and impact mitigation.
Others don't seem to have taken even those lessons to heart. In his response to the report, Paul Zeitz, director of the Global AIDS Alliance, said:
Now the world needs to focus on a clear, numerical goal once again, which is clearly linked to people's lives. This goal should be 10 million people on treatment by 2010, or "10 by 10."
My point is not that outcome targets are always or necessarily bad, but rather that there are significant downsides which must be weighed against the potential benefits of additional resources in each case. Personally, I think that the Economist said it best: "give 'em a number or a date, but not both."