Developmental Changes in Sexual Risk and Substance Use Among African American Females: an Integrated Data Analysis Approach Using Time-varying Effect Models

Abstract

The aim of this study was to describe age-related changes in sexual risk and substance use using existing data from three HIV/sexually transmitted infection (STI) prevention trials conducted in Atlanta, Georgia, that enrolled young African American women. We used two novel analysis methods: integrative data analysis (IDA) and time-varying effect models (TVEM). Each trial collected self-reported behavioral data and vaginal swab specimens assayed for chlamydia, gonorrhea, and trichomonas. Baseline data from all participants and follow-up data from participants not randomized to an active intervention arm were integrated in a pooled dataset using an IDA approach. The pooled dataset included observations for 1974 individuals, aged 14–25\xa0years for behavioral outcomes and 16–25\xa0years for STI outcomes. We used TVEM to model age-related changes in sexual risk and substance use behaviors and STI acquisition, adjusting for trial as a fixed effect. Coital frequency and condomless sex increased through the early 20s while multiple partnerships remained fairly steady. Alcohol use and cigarette smoking increased through about age 24\xa0years before declining. Marijuana use peaked at age 16–17\xa0years and thereafter generally declined. STI acquisition was highest at age 16\xa0years. This study demonstrates the feasibility and utility of innovative methodological techniques to address novel questions related to adolescent development using existing data from multiple trials. The results suggest that mid-adolescence and the early 20s may be periods of particular risk. The findings may be useful for timing culturally and developmentally relevant prevention interventions for young African American women.