Wealth or poverty? a multilevel analysis of socioeconomic status and HIV prevalence in Tanzania

Background: HIV/AIDS remains the leading cause of morbidity and mortality in Africa. Multiple factors, including socioeconomic status influence individual risk for HIV. However, not much is known about how HIV is influenced by socioeconomic factors beyond the individual level in Tanzania. Methods: Nationally representative, cross-sectional data from the 2003/04 Tanzanian AIDS Indicator Survey and the 2000/01 Tanzania Household Budget Survey were analyzed. 12,522 women and men aged 15 to 49 were interviewed and of these 10,747 also consented to anonymous HIV testing. Multilevel statistical modeling was used to analyze the data. Three level multilevel binary logistic regression models were calibrated to examine the relative contribution of regional, community and individual socioeconomic indicators to the variation in HIV seroprevalence. Results: HIV prevalence ranged from 1.6% to 15.5% across Tanzania's 21 regions. At the individual level wealth was found to be positively associated with being HIV positive (OR 2.04, 95% CI 1.36 to 3.08). A negative association was found between community (OR 0.24, 95% 0.09 to 0.61) and regional (OR 0.97, 95% CI 0.95 to 0.99) poverty levels and HIV prevalence. Conclusions: This is the first study to simultaneously examine the influence of context and composition on HIV prevalence in Tanzania. It provides evidence of the existence of significant community and regional variation in HIV prevalence above and beyond that which can be ascribed to variations in individual level socioeconomic status and other characteristics. These findings emphasize the importance of considering both context and composition in studying the social epidemiology of HIV.