Underage drinking and violent behaviors (i.e., physical aggression) are prevalent among adolescents and emerging adults, with enormous impact on morbidity and mortality, particularly among youth residing in socio- economically disadvantaged urban settings. An urban Emergency Department (ED) visit provides an opportunity for identifying youth involved in these comorbid risk behaviors and delivering interventions to alter risk trajectories. Despite promising findings from our prior work (the SafERteens Study), which demonstrated the efficacy of an ED-based brief intervention (BI) for alcohol and violence on reducing alcohol-related consequences and violent behaviors, the public health impact of single session BIs is limited by modest effect sizes. Further, there is a critical lack of data regarding how to optimize interventions to have maximal impact with parsimony of resources. Sequential, multiple assignment randomized trial (SMART) designs provide an innovative methodology to compare the efficacy of just-in-time adaptive interventions [(text messaging (TM) or remote health coach (HC)] for adolescents based on response/non-response. To date, such personalized adaptive interventions have not been applied to risky drinking and/or violent behaviors among youth. The proposed study harnesses technology which is particularly appealing to youth, for facilitation of just-in-time delivery of adaptive interventions (AIs). Thus, the specific aims are to: 1) Compare the efficacy of AIs that begin with BI+TM vs. BI+HC on reducing alcohol misuse and violent behaviors among youth while in the ED; and, 2) Identify the most efficacious second-stage strategy post-ED visit for those who initially respond and for those who do not. Specifically, 700 youth (ages 14-20) in the ED screening positive for alcohol use and violent behaviors will be randomly assigned to: BI+TM or BI+HC. After receiving the SafERteens BI in the ED, youth will complete weekly text assessments over an 8 week period to tailor intervention content and measure mechanisms of change, with one month determination of participant response (e.g., binge drinking, violent behaviors). Responders in each arm will be re-randomized to continued condition (e.g., maintenance), or reduced condition (e.g., stepped down). Non-responders will be re-randomized to continued condition (e.g., maintenance), or intensified condition (e.g., stepped up). Outcomes will be measured at 4, 8 and 12 months post-baseline. The AIs will be packaged to maximize future translation, with cost data provided (e.g., total implementation cost, cost per event averted). Secondary aims are to: 1) Compare and contrast pre-specified embedded AIs in terms of primary and secondary outcomes; and, 2) identify baseline and time-varying moderators of AI efficacy. Given the morbidly/mortality associated with alcohol use and violence, the proposed study will have significant public health impact by embedding a comparative efficacy study, testing state-of-the- art intervention delivery approaches (text messaging, remote therapy), within a SMART design to identify the optimal intervention strategy to produce and sustain outcomes among at-risk youth.
The proposed work will test the efficacy of state-of-the-art adaptive intervention delivery approaches (text messaging, remote therapy) for reducing alcohol use and violent behaviors among urban teens. Given the morbidly/mortality associated with alcohol use and violence, the proposed study will have significant impact by using a SMART design to identify the optimal intervention strategy to produce and sustain outcomes among at- risk youth.
