Selective Mixing Based on HIV Status in Sub-Saharan Populations? Some Indirect Evidence from Demographic and Health Surveys and Implications for HIV Prevention

Stephane Helleringer, Columbia University
Georges Reniers, Princeton University

While most studies of behavior change in Sub-Saharan countries have focused on abstinence, partner reduction or condom use, partner choices and selective mixing may also represent important strategies to reduce exposure to HIV in generalized epidemics. In this paper, we investigate whether individuals are more likely to form new relationships with sero-concordant partners. While this practice has been documented in high-risk groups in developed countries (e.g., serosorting among men who have sex with men), it has garnered little attention in Sub-Saharan Africa. It may however become more and more relevant as HIV testing becomes widespread. We use data from the Demographic and Health Survey to compare the proportion of concordant relationships among recent relations (i.e., formed within a year of the survey) to what would be expected under a baseline model of random mixing. Initial results suggest selective mixing among both HIV-positive and HIV-negative individuals. These results have strong implications for prevention and treatment programs.