Understanding the Effectiveness of Emotion Regulation Interventions for Sexual Risk Among Adolescent Mental Health Populations

Introduction: Emotion regulation (ER) promises to reduce engagement in sexual risk behaviors, particularly among youth living with mental health challenges. Results from individual controlled clinical trials have shown that ER interventions can reduce sexual risk among adolescents with mental health challenges. However, results among trials are difficult to compare because of different analytic approaches when considering outcomes. It is also not known which youth characteristics contribute to improved response to ER interventions.

Method: We harmonized individual participant data across 4 clinical trials of ER interventions (N = 1160, Mage 14.16, Female =43%, White=48%). Pooling across trials allows for consistency in analytic approach and for exploration of individual difference in treatment response. We examined the univariate impact of 6 baseline characteristics (gender, race, ethnicity, age, mental health diagnosis, and psychiatric impairment) on treatment differences in sexual risk behavior at 6-months and 12-months post intervention. Multiple imputation and inverse probability of treatment weights were used within each study to handle missing data and imbalance between treatment arms. Each baseline characteristic was examined separately using tests of moderation.

Results: Preliminary results indicate minimal differences between treatment and control conditions in terms of condomless sex acts, number of partners, or whether or not they engaged in any sexual risk behavior (any condomless sex and/or more than two sexual partners during the previous 6 months). Results among trials vary and the pattern of variability differed by outcome, with trials showing benefits on different outcomes. There were minimal differences in treatment effect by baseline characteristics.

Conclusions: Individual studies indicate support for ER, but effects are on different facets of sexual risk behavior. Consistency in the assessment and analysis of sexual risk outcomes is needed to understand the intervention effect across trials. Consideration of moderator variables in isolation using parametric models like those we used misses important information about baseline profiles. Next steps include non-parametric exploration of baseline profiles to better predict treatment outcome.