Compositional characteristics of neighbourhoods and of schools as correlates of youths' mental health: A cross-sectional study among Swedish adolescents.

Aims: The study aimed to analyse the association between neighbourhood- and school-level socio-demographic factors and self-reported mental health among adolescents.

Methods: The study population consisted of 3959 adolescents in southern and central Sweden (the KUPOL study), surveyed in the seventh grade (13 years old) and again in the ninth grade (15 years old). Cross-sectional associations were studied between socio-demographic indicators at neighbourhood level (proportion of adults with high education, without employment, of foreign-born residents) and at school level (proportion of students with at least one parent with high education, of students with a foreign background, mean qualification points in ninth grade) and mental health problems. These were assessed through the Strengths and Difficulties Questionnaire (SDQ). We derived odds ratios (OR) and corresponding 95% confidence intervals (CI) of scale scores above the threshold using multilevel logistic regression models.

Results: After adjustment for individual- and family-level socio-demographic factors, there was an indication of higher OR of mental health problems with increasing proportions of foreign-born and residents without employment, especially among girls and in public schools. Stronger associations in the expected direction were found with the internalising subscale. Socio-demographic factors at the school level were not associated with adolescents' mental health, except in private schools, where increasing proportions of students with highly educated parents were significantly associated with SDQ scores under the threshold (OR=0.91; 95% CI 0.85-0.97).

Conclusions: