Predicting Adolescent Response to School-Based Mindfulness: A Machine Learning Analysis in the MYRIAD trial

Importance: Depression most commonly first emerges during adolescence, making early prevention critical. While school-based mindfulness training (SBMT) offers a scalable prevention approach with broad reach, evidence of its effectiveness is mixed, and there is a compelling case for a more personalized approach to prevention.Objective: To develop a data-driven algorithm from baseline characteristics to predict which adolescents are most likely to benefit from SBMT.Design: Secondary analysis of the My Resilience in Adolescence (MYRIAD) cluster randomized controlled trial conducted between 2016-2018. Machine learning models were trained on 70% of schools and validated in the remaining 30% from different schools.Setting: Eighty-four broadly representative UK secondary schools across England, Scotland, Wales, and Northern Ireland.Participants: A total of 8,376 adolescents aged 11-13 years at baseline were included. Students had to be enrolled in participating schools with parental consent and student assent. The sample was 55% female, 76% White ethnicity, with mean age 12.2 years.Interventions: SBMT teaching core mindfulness skills through psychoeducation, class discussion, and practices, compared to standard social-emotional learning (teaching-as-usual; TAU) covering topics such as relationships and mental health education.Main Outcomes and Measures: Change in depressive symptoms from baseline to post-intervention measured by the Center for Epidemiologic Studies Depression Scale. Causal forest and elastic net regression models computed Personalized Advantage Index (PAI) scores quantifying individual expected benefit from SBMT vs. TAU.Results: While the overall RCT found no average benefit of SBMT over TAU, the PAI model identified a subset of adolescents (19%) with significantly better depression outcomes when assigned to SBMT (d = 0.22, p = .037), and a larger subgroup (37%) who benefited more from TAU (d = 0.22, p = .004). Top predictive features were primarily student-level variables, including baseline emotional symptoms, executive functioning difficulties, mindfulness skills, and anxiety. Prediction was strongest for depressive symptoms, with mixed findings for two other trial outcomes (social–emotional–behavioural functioning and well-being).Conclusions and Relevance: Machine learning can identify adolescents differentially likely to benefit from SBMT versus TAU, offering a path toward more personalized school-based prevention. Future prospective trials should test whether algorithm-guided intervention assignment improves clinical outcomes.