Enter Uganda, whose strongly decreased AIDS prevalence prompted commentators from either side to claim credit.
So which group is right?
I was browsing the UN Department of Economic and Social Affairs, Population Division, when I came across their national estimates for women from 15 to 49 who are "married or in union" and currently using condoms, or whose partners are using condoms. (Click Contraceptive Prevalence here.) Results:
Uganda 1995: 0.8%
Uganda 2001: 1.9%
Uganda 2005: 1%
Uganda 2006: 1.7%
Now at its height, in 1991, AIDS prevalence ran at around 15% of all adults in Uganda. Yet these UN figures show that only four years later, in the middle of a prolonged decline in AIDS, less than 1% of women in relationships were using condoms.
So that's a puzzler. It seems unlikely that condom use could have reversed the rise of AIDS if hardly anyone was using them.
There are caveats: the UN statistics above might be stronger if they included men or women not in "marriage or union". Also a major factor for the improvements in Uganda is the grisy fact of high death rates among the afflicted: if people die before they get the chance to spread the disease then the disease naturally declines.
Also, the US Agency for International Development does show rising sales of condoms in Uganda, from less than one million in 1991 to around 23 million in 2000. Yet Uganda had a population of around 24 million in 2000, so even then the average person (if we assume only adults are buying condoms and Uganda has a young population) must only have bought a handful of condoms over the course of the entire year. There don't seem to have been enough condoms in use to explain the decline of AIDS.