Daan Hulsmans

44 Chapter 2 used substance and compared at baseline and follow-up. If more than one substance was equally frequently or severely used at baseline, then the average baseline and follow-up scores for these substances were compared. For example, if a person used alcohol and cannabis daily at baseline and other drugs monthly, then the baseline score for frequency of alcohol and cannabis [(5), ‘almost every day’] was compared to the average frequency score for alcohol and cannabis at the follow-up measurement. Additionally, we assessed intervention effectiveness for each substance separately. All analyses were performed using R version 3.6.1 (R Core Team R, 2017). Mixed-effects regression models were used to test the effect of Take it Personal! on substance use frequency, substance use severity and binge drinking. Time, condition and the time × condition interaction were entered as fixed effects in the models. The intervention effect was estimated by the interaction effect on each dependent variable. Time was centered and sum-tozero contrasts were used. To correct for data clustering at baseline, each model included random intercepts for participant, gender and treatment centers. Random slopes were added to the models to control for a clustered effect of time (i.e. the change between baseline and follow-up) within gender and treatment centers. Graphical model diagnostics plots (Bates et al., 2015) were visually inspected to assess goodness of model fit. Attrition analysis by means of logistic regression and Little’s MCAR test indicated that values were missing completely at random, warranting the use of a multiple imputation strategy for intention-to-treat analyses. To obtain p-values, conditional F-tests were performed on both models using the Kenward–Roger approximation for degrees of freedom, a method that gives the most optimal type I error rates in linear mixed-effects models (Nakagawa & Schielzeth, 2013). 3. Results 3.1 Characteristics of the participants Participant characteristics are displayed in Table 1. Adolescents in each group did not significantly differ in age, IQ, all outcomes of substance use frequency or drug use severity at baseline. However, the groups differed significantly in gender, alcohol use severity and binge drinking. Overall, 24% of the adolescents were frequent alcohol users, reporting weekly or daily alcohol consumption at baseline, 41% used cannabis weekly or daily and 20% used illicit drugs weekly or daily. In total, 23% of the adolescents were weekly or daily poly-users of more than one substance.

RkJQdWJsaXNoZXIy MTk4NDMw