(February 2010) In the late 1990s and early 2000s, the global AIDS epidemic overwhelmed health care systems, especially in sub-Saharan Africa. The epidemic appeared unstoppable as infection and prevalence rates continued to increase dramatically. However, data now show that since its peak in the late 1990s, new HIV infection rates and prevalence have begun stabilizing in much of the world, including in sub-Saharan Africa, the world's most affected region. John Bongaarts, vice president and distinguished scholar of the Population Council, discussed recent AIDS data, the reasons for the decline, and policy implications at PRB's Policy Seminar on Jan. 20, 2010.

HIV prevalence rates vary dramatically across the world. The rate is more than 15 percent throughout southern Africa, compared with less than 2 percent in much of the rest of the world. Four factors account for variation in epidemic size: sexual behavior (multiple and concurrent partners), male circumcision, condom use, and the prevalence of other sexually transmitted infections. Bongaarts focused on the first two factors to explain the variation. There is correlation between premarital sex and prevalence because of frequent partner changes, concurrent partners, and highly infectious partners (people are most infectious in the first weeks after getting HIV). He noted that four countries with older ages at marriage (South Africa, Swaziland, Botswana, and Namibia) have higher prevalence rates. In terms of male circumcision, clinical trials in South Africa, Kenya, and Uganda measured HIV infection rates and found a correlation between male circumcision and lower prevalence. In addition, male circumcision rates are lower in southern Africa than in other areas of sub-Saharan Africa, while this region has the highest HIV prevalence rates.

The AIDS epidemic has largely leveled off in sub-Saharan Africa. Prevalence appears to have declined and AIDS deaths have stabilized. The number of new infections has declined by 30 percent since the peak in the 1990s. Increased access to antiretrovirals (ARTs) and behavior change—including abstinence or postponement of first sexual intercourse, increased condom use, and fewer sexual partners—has lowered HIV incidence. According to Bongaarts, another key point is that high-risk groups are already saturated: HIV prevalence cannot get much higher than it is in these groups. Also, those who have been HIV-positive for longer are becoming less infectious because HIV is less contagious among people who have lived with HIV for a long time compared with newly infected people.

So what does the future hold? There is a difference of 5 million annual deaths between the projections of AIDS deaths from UNAIDS and WHO. Why the divergence? In 2006-2007, UNAIDS reduced their estimate of the size of the AIDS epidemic. Models showed that countries had based their AIDS estimates on antenatal clinics (extrapolating prevalence from pregnant women who attended clinics) whereas DHS data based on representative population samples, especially from India, had lower rates. The gaps between the two data sources were connected, resulting in a lower projection.

Finally, Bongaarts discussed the policy implications, especially regarding funding of AIDS treatment and prevention programs. According to Bongaarts, "high allocation is justified when infection is spreading rapidly and is cost-effective." The problem is that the nature of the AIDS epidemic is changing. Prevention is much more cost-effective than treatment but there is a large divergence between funding for treatment (through ARTs) and prevention. Bongaarts argued that the UN goal of universal access to treatment is unsustainable and that too little funding is being allocated to basic health, immunizations, and family planning. To illustrate, he quoted David Halperin, senior research scientist at the Harvard AIDS Prevention Research Project, "...the best health practitioners have abandoned lower-paying positions in family planning, immunization, and other basic health areas in order to work for donor-financed HIV programs."

Bongaarts ended with two messages on the current status of the global AIDS epidemic: The epidemic has peaked but incidence and prevalence will remain substantial in many countries; and the current allocation of scarce health sector resources to AIDS treatment as opposed to HIV prevention and other health problems needs to change.

Eric Zuehlke is an editor at the Population Reference Bureau.