Classroom characteristics and implementation of a substance use prevention curriculum in European countries.

Partial implementation may explain modest effectiveness of many school-based preventive programmes against substance use. We studied whether specific characteristics of the class could predict the level of implementation of a curriculum delivered by class teachers in schools from some European countries.Secondary analysis of data from an evaluation trial. In seven European countries, 78 schools (173 classes) were randomly assigned to a 12-unit, interactive, standardized programme based on the comprehensive social influence model. Curriculum completeness, application fidelity, average unit duration and use of role-play were monitored using structured report forms. Predictors of implementation were measured by aggregating at class level information from the baseline student survey. Class size, gender composition, mean age, factors related to substance use and to affection to school were analysed, with associations estimated by multilevel regression models.Implementation was not significantly predicted by mean age, proportion of students with positive academic expectation or liking school. Proportion of boys was associated with a shorter time devoted to each unit [β = -0.19, 95% confidence intervals (CI) -0.32 to -0.06]. Class size was inversely related to application fidelity [Odds ratio (OR) 0.92, 95% CI 0.85 to 0.99]. Prevalence of substance use was associated with a decreased odds of implementing all the curriculum units (OR 0.81, 95% CI 0.65 to 0.99). Students' connectedness to their class was associated with increased odds of teachers using role-play (OR 1.52, 95% CI 1.03 to 2.29).Teachers' implementation of preventive programmes may be affected by structural and social characteristics of classes and therefore benefit from organizational strategies and teachers' training in class management techniques.