The HIV status of cohabiting partners can be described as discordant if one partner is HIV-negative and the other is HIV-positive. This report assesses the prevalence of discordance and the level of HIV risk from discordance by locating discordant couples within a larger pattern of association among sex (male or female), having a cohabiting partner, and the HIV status of individual men and women. This three-way association varies by age, place of residence, educational level, and wealth quintile.
Data for the analysis come from DHS surveys in 10 countries in sub-Saharan Africa: Cameroon, Kenya, Lesotho, Malawi, Mozambique, Swaziland, Tanzania, Uganda, Zambia, and Zimbabwe. The countries were selected because their HIV prevalence exceeds 4%. Their most recent surveys were conducted between 2006 and 2012. Because DHS data are cross-sectional, and provide little or no information about date or source of HIV infection, duration of cohabitation, disruption due to death or dissolution of relationships, etc., the study is largely descriptive. Non-cohabiting partnerships are not represented in this analysis. It is found that in virtually all countries and subpopulations, there are far fewer discordant couples than would be expected under a random model. The two possible types of discordant couples are nearly always of equal size, regardless of differences between the HIV prevalences of men and women in the larger population. The percentage of HIV-negative individuals who have elevated risk of seroconversion because of discordance with a cohabiting partner ranges from 1% to 4%. The risks are similar for men and women.