Almost half of the 20 million reported sexually transmitted infections (STIs) each year occur in people aged 15 to 24. STIs increase risk for HIV infection and can lead to serious health complications such as pelvic inflammatory disease, infertility, epididymitis, and urethritis. Sexual risk behaviors, such as not using condoms and having multiple partners, place young people at risk for STIs, especially females and African Americans, who have higher rates of infection. Decades of research from longitudinal observational and experimental studies have identified numerous factors predictive of sexual risk behaviors, yet meta-analyses of interventions targeting these behaviors show modest effects. Before more potent interventions can be built, a comprehensive model of STI risk is needed, assessing factors at multiple levels of influence (i.e., individual, social, and environmental) and determining for whom each factor is most salient. Modern causal inference approaches provide a unique opportunity to quickly leverage existing resources to address these pressing research questions. For example, propensity score methods efficiently reduce the bias of many third variables that represent alternative causal explanations;thus, it becomes possible to assess factors that otherwise cannot be randomized for ethical or practical reasons. The proposed study will apply propensity score methods to data from the first three waves of the National Longitudinal Study of Adolescent Health (Add Health) to estimate the causal effects of individual (e.g., perceived benefits of sex), social (e.g., supportive friendships, parental attitues about sex, parent connectedness, school connectedness), and environmental (e.g., access to sexual education curriculum, access or referral to family planning services) factors on sexual risk behaviors in young adulthood. In addition, we will determine whether the pathways leading to STI health disparities vary by gender and race/ethnicity groups and explore how social ecological factors interact with one another. For example, we will examine whether high parental connectedness (social level) mitigates the effect of perceived benefits of sex (individual level) o sexual risk behaviors in young adulthood, and whether having access to sexual education curriculum (environmental level) offsets the effect of not having supportive friendships (social level) on sexual risk behaviors. Behavioral theories are virtually silent about interactions between factors and across levels, yet understanding their presence (or absence) could guide researchers in the construction of more effective and tailored interventions. The results of this study will inform the development of effective STI preventive interventions, thereby reducing the public health burden of STIs among young people.
Despite decades of research to identify and intervene on the social ecological factors for sexual risk behaviors, STI incidence rates remain stubbornly persistent, particularly among young people aged 15 to 24. By leveraging existing resources, we will use innovative causal inference methods to strengthen the understanding of which factors are most strongly associated with sexual risk behaviors that place young people at risk for STIs, how these factors mitigate or exacerbate one other, and for whom the factors are most salient. These findings will enable researchers to develop more potent STI preventive interventions, thereby reducing the public health burden of STIs among young people.