Adaptive intervention to prevent adolescent suicidal behavior following psychiatric hospitalization: A Sequential Multiple Assignment Randomized Trial

ABSTRACT The increasing rates of suicide among adolescents is an urgent public health concern. In parallel, emergency department (ED) visits and psychiatric hospitalizations due to youth suicide risk have also been on the rise. Although psychiatric hospitalization provides critical stabilization, discharged adolescents remain at elevated risk for recurrent suicidal crises (e.g., suicide attempts, rehospitalizations). New approaches are urgently needed to alter risk trajectories and prevent suicidal behavior among adolescents transitioning from inpatient care. Bridging an adaptive intervention (AI) strategy—wherein type, intensity, and timing of treatment can be individualized to address suicidal adolescents’ changing and heterogenous treatment needs—together with mobile technology to reach adolescents during the high-risk post-discharge period, our goal is to identify an optimal AI for reducing youth suicidal behavior after inpatient care. We build on promising evidence from our recently completed NIMH-funded pilot sequential multiple assignment randomized trial (SMART) of a multi- component intervention comprised of: a Motivational Interview (MI)-enhanced safety plan delivered to adolescents and parents during hospitalization (MI-SP); post-discharge booster calls provided to youth and parents; and text-based support (Texts) delivered to adolescents daily for a month after discharge. In addition to feasibility and acceptability, the pilot SMART results suggested that certain intervention sequences improved key mechanisms of change (self-efficacy, safety plan use, coping) and were, preliminarily, associated with lower suicidal behavior 3 months after discharge. In this application, we propose to conduct a full-scale SMART to address the following specific aims: (1) Compare the AIs that begin with MI-SP alone or MI-SP plus Texts on the primary outcome of suicidal behavior (actual, interrupted, or aborted attempts) within 3 months post discharge and the two secondary outcomes of time-to-suicidal behavior and severity of suicidal ideation over 6 months; and (2) Determine the optimal sequence of intervention components by comparing four AIs embedded in the SMART on primary and secondary outcomes. Adolescents (N=300), ages 13-17, hospitalized due to suicidal ideation and/or attempt will be initially randomized to MI-SP alone or MI-SP with Texts. Based on adolescents’ responses to daily surveys, those classified as non-responders within 2 weeks post discharge will be re-randomized to added low-intensity booster calls (one call with adolescent and one call with parent) or high-intensity boosters (six calls with each). Secondary aims are to: (1) Identify moderators of initial intervention options and of augmentation strategies for non-responders; and (2) Examine whether mechanisms of change (self-efficacy, safety plan use, coping) mediate the impact of MI-SP plus Texts and that of high- intensity boosters. The proposed research will have significant public health impact by testing different sequences of continuity of care strategies within an innovative SMART design to identify an optimal technology-augmented AI targeting suicidal behavior reduction among youth transitioning from inpatient care.