Reducing uncertainties in global HIV prevalence estimates: the case of Zambia

Abstract

Background: The premise for using antenatal care (ANC) clinic data for estimating HIV prevalence in the general population is the finding from community studies in sub-Saharan Africa that total HIV prevalence in pregnant women attending ANC clinics closely approximate levels in the total general population of both women and men aged 15–49 years. In this study, the validity of national level HIV prevalence estimates for the total general population 15–49 years made from ANC clinic and population survey data was assessed. Methods: In 2001–2002, a national population HIV prevalence survey for women 15–49 years and men 15–59 years was conducted in Zambia. In the same period, a national HIV sentinel surveillance survey among pregnant women attending ANC clinics was carried out. Results: The ANC HIV prevalence estimates for age-group 15–49 years (rural: 11.5%; 95% CI, 11.2–11.8; urban: 25.4%; 95% CI, 24.8–26.0; adjusted national: 16.9%; 95% CI, 16.6–17.2) were similar to the population survey estimates (rural: 10.8%; 95% CI, 9.6–12.1; urban: 23.2%; 95% CI 20.7–25.6; national: 15.6%; 95% CI, 14.4–16.9). The HIV prevalence urban to rural ratio was 2.2 in ANC and 2.1 in population survey estimates. Conclusion: The HIV prevalence estimate for the total general population 15–49 years derived from testing both women and men in the population survey was similar to the estimate derived from testing women attending ANC clinics. It shows that national HIV prevalence estimates for adults aged 15–49 years can also be obtained from ANC HIV sentinel surveillance surveys with good coverage when ANC attendance and fertility are high.