UNAIDS/WHO estimate (2002 figures) that around 42 million people worldwide are infected with HIV-1, with sub-Saharan Africa accounting for just under 30 million infections. Conventional estimates of HIV-1 prevalence are based on the HIV-1 status of pregnant women attending antenatal clinics. Several countries have recently done national population-based surveys that include HIV-1 testing, and more countries are planning to do so.
J Ties Boerma from WHO, Geneva, Switzerland, and colleagues reviewed estimates of HIV-1 prevalence in adults from three recent national surveys and compared the prevalence data with that obtained by conventional estimates for three countries: Zambia, Mali, and South Africa.
HIV-1 prevalence was found to be around 25% higher according to conventional UNAIDS/WHO estimates than data obtained from population surveys; however the authors concede that data from population surveys are at risk of possible underestimation, mainly due to an incomplete response-rate from some households.
J Ties Boerma comments: "Population-based surveys probably underestimate true prevalence. However, they do provide better quality data for rural populations than do most surveillance systems. The value of antenatal clinic-based surveillance for HIV-1 estimates is mainly in the assessment of trends over time. Surveys, mostly done at 3-5 year intervals, should be regarded as part of such surveillance systems, and a means to improve estimates of HIV-1 prevalence and associated trends. Since the coverage of health programmes for prevention of mother-to-child transmission, antiretroviral treatment, and testing and counselling for HIV-1 is increasing, future HIV-1 estimates will have to be assessed from an even greater number of data sources, all with different drawbacks and biases."
Contact : Dr. J Ties Boerma
Coordinator Surveillance
Research & Monitoring & Evaluation
HIV/AIDS Department
World Health Organization
CH1211 Geneva, Switzerland
T) 41-22-791-1481
F) 41-22-791-4834
E) boermat@who.int
Issue 6 December 2003
Lancet 2003; 362: 1929-31
The Lancet