The Post’s Criticism of PEPFAR Is No Better Founded than Its Praise

An editorial in Thursday’s Washington Post uses President Obama’s nomination of Dr. Eric Goosby to head PEPFAR as an occasion to urge increased PEPFAR attention to HIV prevention. This is a worthy objective. PEPFAR’s inability to bring concrete results-based performance measures to bear on HIV prevention has been criticized by the Institute of Medicine’s 2007 evaluation, in my own analysis of PEPFAR and by other observers.

However, the Post’s editorial confounds the issue, first with its subtitle,

Related Experts

Mead Over

Senior Fellow Emeritus

“Reducing the rates of HIV infection would indicate clear progress toward beating AIDS”,

and then when it states: “And this brings us to one failing of PEPFAR. According to the Annals of Internal Medicine study, the rate at which people were becoming infected with HIV remained unchanged.”

The subtitle could have been fixed with a single word; it should have read: “Reducing the rates of new HIV infection would indicate clear progress toward beating AIDS.” This is important because PEPFAR’s greatest success has surely been in its massive expansion of AIDS treatment – which increases the number of people living with AIDS. The confounding of the rate of existing infections (the “prevalence rate”) with the rate of new infections (the “incidence rate”) is not unique to the Washington Post. For example, the official indicator for Target 6.A of the Millennium Development Goals is the reduction of “HIV prevalence among population aged 15-24 years”. This indicator is better than using prevalence of older age groups, but is still incompatible with the objective of expanding AIDS treatment.

Because the HIV incidence rate is hard to measure and absent from the UNAIDS data, the Annals of Internal Medicine article cited by the Post instead tried to explain the HIV prevalence rate. Using the same fancy statistical techniques applied to UNAIDS’ projected AIDS mortality rates, the authors concluded that there was no statistically significant difference in HIV prevalence between PEPFAR focus countries and the others. Their analysis of HIV prevalence is not subject to the same criticisms that I leveled at their analysis of mortality data, because unlike its mortality data, UNAIDS’ prevalence data is based on real observations of prevalence at ante-natal clinics and in population-based surveys. However, a finding of no statistical difference in prevalence does not support the Washington Post’s claim that PEPFAR has been doing too little on prevention. For prevalence to remain unchanged despite large reductions in mortality, incidence would have had to decline. (See John Stover’s remarks)

This takes us back to my previous criticism of this paper. We unfortunately do not know that AIDS mortality has declined more in the PEPFAR countries than in the non-PEPFAR countries. Therefore we have no way of using the unchanged prevalence to infer a decline in incidence.

So I agree with the Post’s editors that PEPFAR has paid too little attention to prevention – but I can’t support my opinion with the UNAIDS data or with the Annals of Internal Medicine article that analyzed them.

If you have arrived here, you are reading more of CGD’s blogs than are the Post’s editors. Maybe they should subscribe to our RSS feed!