Which groups are most affected by sexual violence? A disaggregated analysis by sex, age, and HIV-status of adolescents living in South Africa.

Background: Estimates on sexual violence (SV) among adolescents are rarely disaggregated by key indicators.

Objective: We examine the independent effects of sex, age, HIV status and mode of HIV acquisition on the exposure to SV using generalised estimating equations models and proportional hazard model.

Participant And Setting: Data from a 4-year prospective cohort study of 1447 adolescents (ages 10-19, 57% girls, 76% living with HIV) in South Africa.

Methods: We describe the lifetime prevalence of three forms of SV: non-contact (unwanted showing of private parts), contact (coerced sexual debut, attempted and completed forced penetrative or oral sex,) and exploitative (sex in exchange for goods/money).

Results: Overall, 23.9% (95%CI = 21.7%-26.1%) reported exposure to SV at some point in their lives: non-contact = 5.7%, contact = 9.0% and exploitative = 15.8%. While girls reported higher rates of exploitative (18.4% vs 12.3%; p < 0.001) and contact SV (12.3% vs 4.6%; p < 0.001), there were no differences in levels of non-contact SV by sex (6.5% vs 4.6%; p = 0.086). Exposure to any SV doubled in late adolescence (10-14 years = 7.0% vs ≥15 years = 31.7%; aIRR = 2.07; 95%CI = 1.82-2.37). Though level of SV were comparable between those living with HIV and those HIV-uninfected (22.9% vs 26.2%; p = 0.182), adolescents who recently acquired HIV were twice as likely to experience SV compared to adolescents who perinatally acquired HIV (42.3% vs 15.7%; aIRR = 2.03; 95%CI = 1.73-2.39). This association persisted when analysis was restricted to incidence SV during follow-up (aIRR = 1.53; 95%CI = 1.23-2.10).

Conclusion: Exposure to SV was high for both sexes, increased with age and more prevalent among adolescents who recently acquired HIV. SV prevention and response services must also be offered to boys and strengthened in HIV care services.