Objective: Research has demonstrated school effects on health, over and above the effects of students' individual characteristics. This approach has however been uncommon in mental health research. The aim of the study was to assess whether there are any school-contextual effects related to socioeconomic characteristics and academic performance, on the risk of hospitalization from non-fatal suicidal behaviour (NFSB).
Methods: A Swedish national cohort of 447,929 subjects was followed prospectively in the National Patient Discharge Register from the completion of compulsory school in 1989-93 (≈16 years) until 2001. Multilevel logistic regression was used to assess the association between school-level characteristics and NFSB.
Results: A small but significant share of variation in NFSB was accounted for by the school context (variance partition coefficient <1%, median odds ratio = 1.26). The risk of NFSB was positively associated with the school's proportion of students from low socioeconomic status (SES), single parent household, and the school's average academic performance. School effects varied, in part, by school location.
Conclusion: NFSB seems to be explained mainly by individual-level characteristics. Nevertheless, a concentration of children from disadvantaged backgrounds in schools appears to negatively affect mental health, regardless of whether or not they are exposed to such problems themselves. Thus, school SES should be considered when planning prevention of mental health problems in children and adolescents.