BACKGROUND: High rates of alcohol and other drug use contribute to the ongoing generalized human immunodeficiency virus epidemic in South Africa. Despite the general link between alcohol use and condomless sex, findings from event-level studies of the relationship between alcohol use and condomless sex during the same encounter have been inconsistent.
METHODS: We conducted event-level analyses of the most recent sexual encounter reported by 636 women who use substances in Pretoria. Data were collected via a questionnaire that included questions about the number of rounds of vaginal and anal sex and condom use during each round. We used multiple logistic regression analyses to model the associations between alcohol use by both partners and having multiple rounds of sex, and alcohol use and condom use during all rounds of sex.
RESULTS: Over 50% of encounters involved multiple rounds of vaginal or anal sex. Encounters that involved multiple rounds of sex were associated with inconsistent condom use. Encounters in which both partners drank alcohol were more likely to involve condomless sex, as compared with encounters in which one or neither partner drank alcohol.
CONCLUSIONS: These findings raise the possibility that prior event-level studies, which do not ask about multiple rounds of sex, may underestimate the prevalence of condomless sex. The association between alcohol use by both partners and condomless sex may partially explain inconsistent associations between alcohol use and condom use in event-level studies that did not assess the number of partners drinking.