Using population-based, longitudinal data from 15- to 19-year olds in Rakai, Uganda, we examined temporal trends in the prevalence of sexual experience and potential risk factors for sexual experience (n = 31,517 person-round observations) using logistic regression. We then identified factors associated with initiation of sex between survey rounds, using Poisson regression to estimate incidence rate ratios (IRR; n = 5,126 person-year observations).
Sexual experience was more common among adolescent women than men. The prevalence of sexual experience rose for most age-gender groups after 1994 and then declined after 2002. Factors associated with higher prevalence of sexual experience (without adjustment for other factors) included age, not enrolled in school, orphanhood, lower socioeconomic status, and drinking alcohol in the past 30 days; similar factors were associated with initiation of sex. Factors independently associated with initiation of sex included older age, nonenrollment in school (IRR = 1.7 for women and 1.8 for men), alcohol use (IRR = 1.3 for women and men), and being a double orphan among men (IRR = 1.2). Sexual experience began to decline around 2000, whereas increases in school enrollment began as early as 1994 and declines in orphanhood occurred after 2004 (as antiretroviral therapy became available).
Sexual experience among youth in Rakai was associated with social factors particularly school enrollment. Changes in these social factors also appear to influence change over time in sexual experience.