Marital Partnership Selection and HIV Seroconcordance in a Generalized Epidemic Setting

Elizabeth A. Sully, Princeton University
Dominic Bukenya, MRC/UVRI Uganda Research Unit on AIDS
Janet Seeley, MRC/UVRI Uganda Research Unit on AIDS and London School of Hygiene and Tropical Medicine (LSHTM)

Serosorting, where individuals select partners of the same HIV status as themselves, is a potentially important HIV prevention tool. Using 13 years of sero-surveillance data from southwestern Uganda, this paper first documents the occurrence of serosorting in a generalized epidemic setting, and then uses a mixed-methods approach to explore the direct and indirect selection mechanisms that drive serosorting. We use simulation models to test whether mixing on non-HIV attributes can account for the observed level of serosorting. While disassortative age-mixing increases serosorting, it does not fully explain serosorting. We find no evidence that other factors associated with partnership choices (e.g., education, ethnicity, and marital histories), have a bearing on serosorting. Ruling out indirect selection, this paper invokes qualitative evidence of a direct preference for partners of the same HIV status. The paper concludes by discussing the implications of these findings for designing and evaluating HIV prevention programs.