The Cannabis Use Disorders Identification Test (CUDIT) was used for the first time as part of a randomized controlled trial for brief interventions in mild to moderate alcohol-dependent out-patients. This sample may be seen as a population at increased risk of cannabis use disorder. The CUDIT was developed by modifying the Alcohol Use Disorders Identification Test (AUDIT). The ability of the CUDIT to accurately screen for cannabis abuse or dependence was examined in the portion of the sample who reported some cannabis use over the preceding 6 months (n=53), as was self-reported frequency of cannabis use in the preceding 6 months. The CUDIT was superior to the frequency measure, achieving positive predictive power of 84.6% and sensitivity of 73.3% at a cut-off of 8, compared to positive predictive power of 81.8% and sensitivity of 60.0% for 80 or more cannabis use-days. These results indicate the viability of a screening measure for identifying cannabis use disorder in at risk populations.