Carolien Zeetsen

66 Analyses For descriptive purposes, differences in patient characteristics between subgroups and between the four addiction health care centres were explored using chi–square tests and univariate analyses of variance (ANOVA). Second, the prevalence of cognitive impairments was calculated for the total sample and per primary–problem substance. Third, MoCA total and domain scores (MoCA–TS and –DS) for the total sample and differences between primary–problem substances were analysed using univariate and multivariate ANOVAs, respectively. Finally, the effects of age, years of regular use, abstinence duration in days, severity of dependence and/or abuse (Section Alcohol & Drugs of the Composite International Diagnostic Interview), and depression, anxiety and stress (DASS–21) on MoCA total score were examined by Pearson correlations; abstinence (< 7 days/7–41 days/≥ 42 days, Walvoort et al., 2013) with a univariate ANOVA; and the effect of polysubstance use (yes/no) on the MoCA–TS was examined with an independent t –test. Hochberg’s GT2 (unequal sample sizes) or Games–Howell post–hoc tests (nonhomogeneous population variances) were used as post–hoc analyses in all ANOVAs. Alpha was set at 0.05 for all analyses and all data were computed and analysed using IBM SPSS version 25.0.