In a 7-year 11-wave study of low-SES adolescents (N = 856, age = 15.98), we compared multiple well-established transdiagnostic risk factors as predictors of first incidence of significant depressive, anxiety, and substance abuse symptoms across the transition from adolescence to adulthood. Risk factors included negative emotionality, emotion regulation ability, social support, gender, history of trauma, parental histories of substance abuse, parental mental health, and socioeconomic status. Machine learning models revealed that negative emotionality was the most important predictor of both depression and anxiety, and emotion regulation ability was the most important predictor of future significant substance abuse. These findings highlight the critical role that dysregulated emotion may play in the development of some of the most prevalent forms of mental illness.
Keywords: Anxiety; Depression; First incidence; Machine learning; Negative emotionality; Risk factor; Substance abuse; Transdiagnostic.
© 2024. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.