The DSM Guided Cannabis Screen (DSM-G-CS): description, reliability, factor structure and empirical scoring with a clinical sample

Abstract

Clinicians need cannabis-specific diagnostic screens compatible with DSM-IV-TR and proposed DSM-5. A clinical sample (n=174) completed the DSM-Guided-Cannabis Screen (DSM-G-CS) 21 and 11 criteria versions and three drug comparison measures. DSM-G-CS descriptive statistics, reliabilities, three factor analyses, and eight ROC and discriminant analyses evaluated construct validity and empirical scoring. DSM-G-CS reliabilities are .88 (21-items) and .85 (11-criteria). Factor analyses (FA) with varimax rotation derived six and three factors explaining 62% to 60% of variances for the DSM-G-CS 21 and 11 respectively, with ≥.400 loadings supporting retention of all items. Cannabis withdrawal subscale reliability .952 (10-items) and FA supported one factor composite item. ROC and discriminant analyses supports DSM-G-CS 1.5 to 2.5 scoring cutoffs as empirically sound, based upon sensitivity-specificity maximums, accuracy probabilities, confidence levels and correctly classified percentages, optimal with Marijuana Screening Inventory (MSI) comparisons. Results support DSM-G-CS construct validity, empirical scoring and compatibility with DSM-IV-TR cannabis abuse or dependence and proposed DSM-5 cannabis use disorder diagnostic models. Clinically, DSM-G-CS scores of two to three (or more) suggest probable cannabis-use disorder, deserving assessment to determine diagnostic accuracy.