Cohort effects on sexual behavior and risk of acquisition of sexually transmitted infections and HIV in a South African HIV prevention trial

Introduction: As we approach the fourth decade with human immunodeficiency virus (HIV), a plethora of prevention interventions have been tried; however, prevention of new infections among women remain poorly understood and current behavioral change interventions are simply not working. Material and methods: We investigate whether women's sexual behavior risks vary by age cohort, and the effect of this on HIV and sexually transmitted infections (STI) incidence of women enrolled in the South African Carraguard HIV prevention trial, using Cox proportional hazard models. Results: Relative to women born in the 1960s or earlier, HIV incidence was approximately twice as high compared to women born in the 1970s (adjusted hazard rate [aHR]: 1.91; 95% confidence interval [CI]: 1.04-3.48; p < 0.001). Women in the youngest 1980s cohort had the highest risk of HIV infection (aHR: 3.77; 95% CI: 2.25-6.32; p < 0.001). Similarly, women in the 1980s relative to the oldest (<= 1960s) cohort had significantly higher hazards of acquiring STIs (aHR: 1.51; 95% CI: 1.16-1.96; p = 0.002), had a higher frequency of partner changes, higher coital frequency, higher proportion having had sex for money, to have an older partner, and less likely to have a circumcised partner. Conclusions: Our findings show HIV and STIs incidence remains highest amongst younger compared to older age cohorts, despite decades of STI/HIV risk-reduction interventions. Research is urgently needed to understand why after so many decades of socio-behavioral interventions little appears to have changed in the behavior of young women; instead, there appears to even be a glamorization of risky behaviors such as having older male sexual partners commonly called 'blessers'.