Baseline prediction of two-year prevention and remission of anxiety, depression, and eating disorders in college students receiving a digital guided self-help intervention: derivation and external validation of a machine learning algorithm across 26 population-based cohorts.

INTRODUCTION: We examined whether machine learning identified baseline variables that predicted two-year prevention and remission from anxiety, depression, and eating disorders following digital cognitive-behavioral therapy guided self-help (D-CBTgsh). METHODS: Undergraduates at risk for or meeting criteria for anxiety, depression, and eating disorders who were randomized to population-based screening and D-CBTgsh were examined. A five-step hybrid feature selection strategy reduced 863 baseline questionnaire items to 37 spanning demographics, mental health problems, treatment motivation, and treatment targets. Five supervised machine learning algorithms were trained and tested using nested cross-validation among students from 13 U.S. colleges (n = 1472) to discriminate who did versus did not achieve two-year prevention and remission of all anxiety, depression, and eating disorders. External validation was conducted in a separate set of 13 U.S. colleges (n = 823). RESULTS: When externally validated, random forest outperformed all models in discriminative power, calibration, and clinical utility. Predictors of no disorders at two years included lower levels of behavioral avoidance, generalized anxiety interference, social fear, social anxiety interference, depressed mood, fear of observation, number of worry topics, perseverative thinking, and binge eating, along with higher feelings of calm/peacefulness, perceived need for treatment, and intention to seek help. Demographic predictors of no disorders were being male at birth, fewer financial difficulties, and a lower current body mass index. CONCLUSION: This predictive algorithm, using 37 single items, developed a scalable self-report measure that could be readily adopted by college campuses to help optimize and scale service delivery across campuses.