Objective: To identify subgroups of students with distinct profiles of mental health symptoms (MH) and substance use risk (SU) and the extent to which MH history and socio-demographics predict subgroup membership. Participants: University students (N = 10,935: 63% female). Methods: Repeated cross-sectional survey administered weekly to stratified random samples. Latent class analysis (LCA) was used to identify subgroups and multinomial regression was used to examine associations with variables of interest. Results: LCA identified an optimal 4-latent class solution: High MH-Low SU (47%), Low MH-Low SU (22%), High MH-High SU (19%), and Low MH-High SU (12%). MH history, gender, and ethnicity were associated with membership in the classes with high risk of MH, SU, or both. Conclusion: A substantial proportion of students presented with MH, SU, or both. Gender, ethnicity and MH history is associated with specific patterns of MH and SU, offering potentially useful information to tailor early interventions.
Keywords: Comorbidity; latent class analysis; mental health; substance use; university students.