Why Numbers Matter & What Went Wrong

In a recent article, How AIDS in Africa Was Overstated, Washington Post correspondent Craig Timberg presents evidence that global HIV/AIDS rates have been systematically overestimated by UNAIDS based on the long-held (and erroneous) assumption that infection rates among pregnant women attending antenatal clinics can serve as a rough proxy of the general adult prevalence in that country. However, new Demographic and Health Survey data show that those earlier projections were skewed in favor of young, urban, sexually-active women - a group that tends to have a much higher HIV rate than the general population - and so were off by up to a scale of magnitude, directly challenging UNAIDS credibility going forward.While this situation is certainly linked to problems of data quality and availability, information management, and 'measurability' of health outcomes, it also raises a larger question: "What are we measuring?" In this case, UNAIDS was measuring one thing and calling it another, with a direct impact on the funding priorities of national governments and donor agencies. As the Washington Post put it in a subsequent editorial:

There has long been a concern that AIDS, while underfunded relative to the number of people in need of treatment and education, is arguably overfunded relative to other diseases. Prolonging the life of an HIV-positive adult is expensive and difficult; preventing malaria with low-tech mosquito nets may be more cost-effective. The revisions of the AIDS data should prompt a reopening of such questions.

It is unclear whether UNAIDS fully recognizes the importance of this question in resource allocation decisions, as evidenced by their response (PDF):

Comparing the disadvantages of providing treatment to a person with AIDS with treating someone suffering from malaria, creates a dangerous dichotomy. Both deserve treatment and any health care delivery solution at country level must recognize that there should not be tradeoffs based on cost and simplicity of treatment.

Whether or not there should be tradeoffs in global health priorities, the reality is that there are - and given that, I cannot think of a much better determinant than cost-effectiveness. More than anything, this demonstrates the extent to which UNAIDS and the rest of the international health community have shifted towards advocacy over accuracy.