(PS12-76) Is There a Place for Cognitive Restructuring in Self-guided Digital Interventions for Adolescents? A Randomized Controlled Trial of Project Think

Background: Self-guided digital mental health interventions (DMHIs) teaching empirically supported skills and concepts (e.g., growth mindset and behavioral activation) have demonstrated efficacy for improving youth mental health, but we lack evidence for the complex skill of cognitive restructuring (CR). 

Method: In collaboration with public schools, we conducted the first-ever test of a CR DMHI (“Project Think”) against an active control (based on non-directive supportive therapy; “Project Share”) via randomized controlled trial (RCT). Pre-registered outcomes were DMHI acceptability and helpfulness post-intervention, as well as internalizing symptoms and CR skills use from baseline to seven-month follow-up, in the full sample and in the subsample of participants with elevated symptoms. 

Results: Participants (N = 597 [N_{Elevated Subsample} = 98]; M_Age = 11.99; 48% female; 68% White) rated both programs highly on acceptability and helpfulness. Both conditions were associated with significant internalizing symptom reduction across time in both samples, with no significant condition differences. CR skills use declined significantly across time for Project Share youths but held steady across time for Project Think youths in both samples; this pattern produced a significant condition difference favoring Project Think within the elevated sample at the final follow-up assessment (OR = 1.71, p = .031). Further, students with lower baseline CR skills use (relative to higher baseline CR skills use) evinced higher CR skills use at the final follow-up assessment if they had received Project Think than if they received Project Share (OR = 0.68, p < .001). Internalizing symptoms declined comparably for Think and Share participants; students in the Project Think condition with high baseline CR skills use coupled with high baseline internalizing scores demonstrated significantly reduced internalizing scores at 10-week follow-up relative to students with low baseline CR skills use and/or low baseline internalizing scores (OR = 0.80, p = .002). 

Discussion: Internalizing symptoms declined comparably for Think and Share participants. Consequently, future research should examine whether encouraging youths to share their feelings produces symptom improvements, and whether a single-session, self-guided CR DMHI produces beneficial effects relative to more inert control conditions. Further, the decline in CR skills use for Project Share youths versus sustained CR use by Project Think youths raises questions about the natural time course of youths’ CR use and the impact of these DMHIs on that course.