Have African AIDS numbers been exaggerated by the UN?

New census reveals HIV numbers significantly lower than feared

By S H Cyr

The overwhelming numbers involved in the AIDS pandemic ravaging sub-Saharan Africa's population may have been dramatically overestimated, according to data from the latest Kenya population census.

The government of Kenya now estimates that about one million of the country's citizens are HIV-positive. Previous estimates ran as high as three million. Kenya is the third African country to include voluntary HIV testing in a randomly chosen subset of respondents as part of its national population census.

Mali and Zambia, the other countries to have done so, also found lower than expected rates of infection. Mali found the infection rate to be 1.7%, having previously estimated a figure of 4%. In Zambia, 21.5% of people had the disease, when prevalence had previously been put at 27%.

In Kenya, awareness has been growing that data-gathering methods used in the past have been flawed, and government figures for HIV prevalence have recently moved downwards. The Health Ministry's estimate in 2003 was 9.4%. The current UNAIDS estimate of HIV prevalence in Kenya is 15%. But the latest survey gives a figure of 6.7%.

Between them, Zambia, Mali and Kenya provide evidence from the south, the west and now the east of the continent that the spread of HIV is either slowing down, or more likely, has been overestimated. Cameroon and Tanzania, with estimated HIV rates of 12% and 8% respectively, are now planning to carry out similar population-based surveys to determine their true HIV infection rate.

In Kenya, occupants of 8,561 representative households around the country were offered HIV tests, and 70% of them agreed to them. The tests were conducted by officials from the US Centers for Disease Prevention and Control (CDPC).

Dr Kevin DeCock, the local director of the CDPC, called the new figures the best HIV statistics in Kenya to date. "The number of HIV-infected people in Kenya is lower than previously estimated. This is based on better, more accurate measurements," he said.