(December 2007) The Joint United Nations Programme on HIV/AIDS (UNAIDS) has released new estimates of the prevalence of the global HIV/AIDS epidemic.  In 2007, 33.2 million people are estimated to be infected with HIV (see figure), with the possible range of the estimate varying from 30.6 million to 36.1 million. This represents a significant decrease from the estimate made in 2006 of 39.5 million (a range of 34.1 million to 47.1 million).

Estimated Number of People Living With HIV Worldwide

Source: UNAIDS, 2007 AIDS Epidemic Update.

Much of the impetus for the lowered estimate stems from the results of HIV testing conducted as a part of the Demographic and Health Survey (MEASURE DHS) program, a series of nationally representative surveys conducted by Macro International Inc. with funding from USAID, international organizations, and private foundations. The key to the revised estimates is that DHS surveys are nationally representative while the earlier primary source of information on possible HIV prevalence, sentinel surveillance sites, was not. This issue was addressed earlier on the PRB website (see related article).

Before recent DHS surveys were conducted, developing countries did not have many sources for estimates. For example, because vital statistics such as births, deaths, and cause of death are incomplete at best in most developing countries, they can’t be used. Instead, these countries evaluated the extent of HIV by testing selected groups for infection, a process known as “sentinel site testing.” These include pregnant women at antenatal clinics and intravenous drug users. In this way, some information could be gleaned on the extent of HIV infection in the country as well as its geographic distribution. But difficulties arose when attempts were made to translate sentinel site prevalence levels to the population as a whole. Sentinel site groups are not representative of the national population.

The DHS survey samples, on the other hand, are specifically designed to be representative of the national population. DHS surveys have long provided needed data on a wide variety of demographic and health variables such as fertility rates, contraceptive use, immunization of young children, and infant mortality. The DHS has now conducted HIV testing in 30 countries, the majority in Africa, and the results show that past approaches to estimating HIV prevalence using sentinel site results overestimated infection rates. The new methods employed by UNAIDS that now take into account DHS survey results improve upon earlier estimates.

Unfortunately, changes in statistical estimates can create considerable confusion in the public mind and in media reports. They can even lead to the impression that HIV is not the problem it was once thought to be. UNAIDS now estimates that 2.5 million (a range of 1.8 million to 4.1 million) people became infected with HIV this year. It is also estimated that the number of new cases peaked in the late 1990s, thanks to HIV prevention programs. One can only hope that the new, lower estimates will not lead to a feeling of complacency, which itself could lead to a rise in new HIV case numbers yet again.

Visit PRB’s Graphics Bank to see the latest graphics on HIV/AIDS:

Carl Haub holds the Conrad Taeuber Chair of Population Information at PRB.

For More Information

MEASURE DHS, Demographic and Health Surveys

UNAIDS, 2007 AIDS Epidemic Update.