Purpose To evaluate feasibility and acceptability, obtain preliminary efficacy data, and evaluate predictors of improvement with iDOVE, a brief in-person +longitudinal text-message violence and depression prevention intervention for high-risk adolescents seen in the emergency department (ED).

Methods A pilot randomised controlled trial was conducted with English-speaking adolescents (ages 13 to 17), presenting to the ED for any reason, who reported past-year physical peer violence and current depressive symptoms. The CBT- and MI-based intervention consisted of a brief, computer-guided in-ED intervention; and 8 weeks of daily, automated text-message mood queries and daily, automated text-message curriculum (tailored based on daily mood and baseline gender and violence). Follow-up surveys (including CTS-2, BDI-2, and qualitative questions) were administered at 8 and 16 weeks. Descriptive statistics, mixed effects longitudinal regression models, and latent class models were calculated.

Results iDOVE had high acceptability and feasibility: 86% of eligible youth consented, 95% completed 8 week follow-up, and 91% completed 16 week follow-up. High quantitative and qualitative satisfaction was reported by both intervention and control participants. The intervention had high engagement (96% of intervention group responded to at least 1 daily query, with a mean of 47/56 daily responses per participant [84% of days]). Comparing intervention to control, improved depressive symptoms (p=0.07) and physical peer violence (p=0.01) were observed among the more symptomatic youth in the intervention group (but no difference in symptoms between the full intervention and control groups) at 8 weeks. Latent class modelling showed that lower mood (measured through daily text-messages) at day 7 of the intervention correlated with no improvement in symptoms at 8 week follow-up.

Conclusion and Significance This pilot RCT of an innovative technology-augmented intervention shows high feasibility and acceptability, a promising signal of reduced violence and depressive symptoms among the highest-risk participants, and a potential marker of non-response for future adaptive interventions.