Background Reportable STIs have largely been increasing in the United States; yet, few national-level studies have examined potential reasons for the increase. We used data from the 2002 and 2011–15 National Survey of Family Growth to examine trends in high-risk sexual behavior among key subpopulations with disproportionate STI rates.

Methods The subpopulations we examined included sexually active: men who have sex with men (MSM) and adolescent, young adult, Hispanic, and non-Hispanic black women who had sex with men (WSM) and men who had sex with women (MSW). High-risk sexual behavior (HRSB) was a composite variable (yes, no) that included endorsement of any of the following: giving or receiving money or drugs for sex or having a partner who was bisexual, non-monogamous, HIV-positive, or who injected drugs. We used weighted data to examine the prevalence of HRSB over time for each subpopulation and conduct logistic regressions adjusting for race/ethnicity, age, marital status, poverty and education.

Results HRSB either did not change or declined over time. Overall, among WSM, reported HRSB significantly declined over time (15.1% in 2002 and 12.4% in 2011–15). Among men, we also found significant declines in reported HRSB for MSW (15.4% in 2002 and 12.2% in 2011–15) and MSM (25.9% in 2002 and 13.1% in 2011–15). In subgroup analyses, there were more declines among WSM than MSW or MSM. However, in adjusted analyses, we found significant declines in reported HRSB, comparing 2002 to 2011–15, among non-Hispanic black (aOR=0.78, 0.65–0.95) and adolescent (aOR=0.64, 0.50–0.95) WSM; Hispanic (aOR=0.49, 0.35–0.67) and adolescent (aOR=0.57, 0.41–0.79) MSW; and all MSM (aOR=0.42, 0.24–0.75).

Conclusion While STIs are increasing, HRSB was steady or declined among key subpopulations with disproportionate STI rates. Results from adjusted analyses suggest these findings are not the result of changing population demographics. Further research is needed to explain the STI increases.

Disclosure No significant relationships.