Optimal scaling of the CAST and of SDS scale in a national sample of adolescent

Purpose: Psychometric and screening properties of the Cannabis Abuse Screening Test (CAST) and of the Severity Dependence Scale (SDS) were investigated using DSM-IV diagnoses of cannabis dependence (CD) as external criteria. Performance of the SDS and of the CAST were compared. Methods: Cross-sectional European School Survey Project on Alcohol and Other Drugs (ESPAD) was carried out in Italy in 2009. The sample consisted of 5787 Italian adolescents aged 15-19 who reported cannabis last year use. Uni-dimensionality, internal reliability, external validity, and optimal scaling of the 6 items for CAST and 5 items for SDS were performed. The Munich Composite International Diagnostic Interview (M-CIDI) was used as a gold standard for DSM-IV diagnoses, and all outputs were assessed by 10-fold cross validation procedure. Results: Both scales were uni-dimensional and Cronbach's α was 0.74 for SDS and 0.78 for CAST. High and comparable area under curve (AUC) values indicate a good ability of both scales to discriminate between individuals with and without dependence diagnosis. Based on balanced sensitivity and specificity, the optimal cut-off scores for problematic use disorders were 7 for CAST MCA and 4 for SDS MCA. Both CAST and SDS overestimated CD prevalence. Conclusions: The CAST and SDS are equally useful for screening for problematic cannabis use disorders. Both clinical and research applications of the scales are possible