Psychometric properties of the Cannabis Abuse Screening Test (CAST) in samples of adolescents and young adults in Hungary.

Aim: The aim of our study was to analyze psychometric properties of the Cannabis Abuse Screening Test (CAST). Methods: Our sample comprised Hungarian high school (n = 476; male 56.3%; mean age 19.0 years, SD = 0.65 years) and college students (n = 439; male 65.1%; mean age 23.9 years, SD = 1.56 years) who reported cannabis use in the past year. The sample covered the five biggest universities of Hungary. Besides the CAST, participants responded to the Munich-Composite International Diagnostic Interview. Factor structure was analyzed by a confirmatory factor analysis. Receiver operating characteristic curve analysis was made to assess cut-off scores. Data collection took place in 2010. Results: CAST proved to be a reliable (Cronbach’s α 0.71 and 0.76) one-dimensional measure. Regarding both cannabis dependence and cannabis use disorders, a cut-off of 2 points proved to be ideal in both samples, resulting in optimal specificity, negative predictive values and accuracy, but less thanoptimal positive predictive values (dependence) and low sensitivity (cannabis use disorder). Discussion and Conclusions: In line with former results, the CAST proved to be an adequate measure for the screening of cannabis-related problems among adolescents and young adults in an Eastern European country where this scale has not been studied before