Screening for cannabis use disorder among young adults: Sensitivity, specificity, and item-level performance of the...

The Cannabis Use Disorders Identification Test Revised (CUDIT-R) is a widely used screening tool. However, its utility in screening for cannabis use disorder (CUD) among young adults requires further investigation. The current study evaluated the accuracy of the CUDIT-R in distinguishing between young adults with and without CUD. We also examined the individual item properties of the CUDIT-R in this sample. A community sample of young adults (N = 153, ages 19-26 years) who reported at least one lifetime period of regular cannabis use (i.e., twice or more per month for at least six months) completed the CUDIT-R and the Structured Clinical Interview for DSM-5 (SCID-5) Substance Use Module. The CUDIT-R showed good accuracy in distinguishing between young adults with and without CUD as determined by the SCID-5 (area under the receiver operating characteristic curve = 0.84). A cut-point of nine yielded 87.80 % sensitivity and 70.42 % specificity, with positive and negative predictive values of 77.42 % and 83.33 %, respectively. Item response theory analyses revealed that CUDIT-R items assessing impaired control, failure to meet expectations, time devoted to use, and memory and concentration problems displayed high discrimination and were relatively difficult, whereas items assessing frequency of use and attempts to reduce use showed moderate discrimination and relatively low difficulty. Items assessing hours high and use with potential for physical hazard displayed poor discrimination. The CUDIT-R appears to validly distinguish between young adults with and without CUD and may be useful in screening for CUD in this high-risk population. Further research considering individual item properties is needed to refine the CUDIT-R for use among young adults.