Psychosocial determinants of smoking in Swiss adolescents with special reference to school stress and social capital in schools.

OBJECTIVES: To identify the major psychosocial determinants of smoking in adolescents and the school influence on these determinants. METHODS: Cross-sectional questionnaire survey in 8th grade (age 14.8 years, n = 459) of 14 schools. Logistic regression with smoking as the dependent, and psychosocial indicators as independent variables. Total climate score for each school computed as the sum of scores of five school-related indicators. Linear regression analysis on aggregate data (school level), controlling for gender, ethnicity, and social class. RESULTS: Five out of 15 tested psychosocial indicators were identified as independent protective factors. The prevalence of smoking decreased steeply with an increasing number of protective factors. In the regression analysis on the aggregate level the mean number of protective factors per school and the prevalence of smoking were significantly related to the school climate score (R2 = 0.650, p < 0.001, and R2 = 0.456, p < 0.001). CONCLUSIONS: Provided a causal interpretation of the cross-sectional statistical associations is correct, efforts to improve the general climate in schools appear as a promising strategy to enhance individual protective factors. Longitudinal evaluative studies are needed to prove the effectiveness of such a strategy.