Machine learning approaches to identify profiles and predictors of psychosocial discomfort among Italian college students

Introduction. College students live a crucial period of transition from late adolescence to adulthood when they have to deal with important stressful tasks. Thus, university often represents a stressful environment, pushing students to cope with a high academic pressure. As a result, this period constitutes a sensitive age for the on-set of mental disorders. Typically, students are not aware of the early signs of their own compromised mental health until symptoms aggravate to an overt disorder. Therefore, it is important to timely detect subthreshold symptoms mostly related to generic mental distress. Objective. First, to assess psychophysical well-being and mental distress among college students in northern Italy, and to detect predictors, among socio-demo-graphic and academic characteristics, and risky drug use of these two outcomes. Method. The study involved 13,886 students who received an email explaining the purpose of the e-research. The questionnaires used were the General Health Questionnaire (GHQ-12), the University Stress Scale (USS), and a modified version of World Health Organization-ASSIST v3.0. Results. 3,754 students completed the web-survey. Students showed poor well-being and mental distress. The strongest predictor of mental distress and compromised well-being was physical health, followed by sex, study field, risky drug use, and academic performance con-cerns. Discussion and conclusion. This study shows that it is very important to promote in college students healthy behaviors in order to increase their physical exercise and reduce substance use. Moreover, it would be desirable to improve academic counselling facilities as an important front-line service to intercept mental health issues among young adults.