Discriminative validity of a substance use symptom checklist for moderate-severe DSM-5 cannabis use disorder (CUD) in primary care settings

The prevalence of cannabis use disorder (CUD) is increasing in the US and primary care providers need tools to identify patients with moderate-severe CUD to facilitate treatment. A single-item screen for cannabis (SIS-C) has outstanding discriminative validity for CUD. However, because the prevalence of moderate-severe CUD is typically low, the probability that an average patient who screens positive for daily cannabis has moderate-severe cannabis use disorder is low, making follow-up assessment important.

This study reports the discriminative validity of a DSM-5 Substance Use Symptom Checklist ("Checklist") for moderate-severe CUD among 498 primary care patients who reported daily cannabis use on the SIS-C. We evaluated the performance of the Checklist (score 0-11) completed during routine care, compared to ≥4 DSM-5 CUD symptoms (moderate-severe CUD) on the Composite International Diagnostic Interview Substance Abuse Module from a confidential survey (reference standard). We estimated areas under receiver operating curve (AUROC), sensitivities, specificities, and post-test probabilities.

Of 498 eligible patients, 17 % met diagnostic criteria for moderate-severe CUD. The Checklist's AUROC for moderate-severe CUD was 0.77 (95 % CI: 0.71-0.83), and Checklist scores of 1-2 balanced sensitivity and specificity. Among patients from a population with average prevalence of CUD before screening (~6 % prevalence) and daily use on the SIS-C, a Checklist score of 3 indicated a post-test probability of 82.1 %.

Overall performance of the Checklist was good and the high specificity made it useful for identifying patients likely to have moderate-severe CUD among those at average risk.