As cannabis use has increased across the U.S., so has cannabis use disorder (CUD), with estimates of 6.9% of adults meeting criteria for CUD. Patients with mild CUD may not be interested in treatment, but as the severity of their CUD increases, so too might their interest in care. Primary care providers are in a unique position to offer that care, but they need better tools to identify those patients.
This study, part of the NIDA Clinical Trials Network study CTN-0074 (PROUD), evaluated the use of a single-item screen for cannabis (SIS-C) as part of routine primary care, followed by an 11-item Substance Use Symptom Checklist (the “Checklist”) for those who report daily or “almost daily” cannabis use to evaluate for CUD.
While the Checklist has been evaluated psychometrically before, its discriminative validity had not been evaluated and is an important factor in determining its usefulness. (Discriminative validity measures whether two theoretically distinct measures are indeed unrelated, making sure the measure in question is capturing a specific construct.) In this study, the Checklist was compared to the Composite International Diagnostic Interview Substance Abuse Module (CIDI-SAM).
This study was done using electronic health record (EHR) data and a confidential cannabis survey given to Kaiser Permanente Washington primary care patients. Adult patients (equal to or over age 18) were eligible for the survey if they had EHR documentation of a cannabis screen completed as part of routine primary care between January – September 2019 and were randomly sampled to receive the survey (n=5000). The random sample was mailed survey invitation letters, followed by phone reminders, and received $20 for completing the survey, resulting in 1,688 respondents. Of those, 498 reported daily cannabis use on the SIS-C and were included in the final sample.
Of the 498 study participants, 17% met diagnostic criteria for moderate-severe CUD. The Checklist’s AUROC (“area under the receiver operating characteristic curve” – a metric used to evaluate performance) for moderate-severe CUD was 0.88, meaning the Checklist performed acceptably against the CIDI-SAM. Checklist scores of 1-2 balanced sensitivity and specificity, suggesting meaningful increases in the likelihood of identifying CUD for patients at average risk as Checklist scores increase. Checklist scores of 3 or higher increase the likelihood of moderate-severe CUD to 82% from an average pre-test probability of about 40% among those reporting daily cannabis use.
The Checklist performed best for patients whose prevalence of CUD was average and was less effective for patient subgroups with low or high prevalence of CUD. Given that prevalence of CUD varies by age, sex, and presence of mental health conditions or other SUD, these factors should be taken into account when interpreting Checklist results for individuals.
For example, based on this study’s results, for an 80 year-old woman (low prevalence) or a person with another mental health condition or SUD (high), the Checklist score wouldn’t add meaningful information about the probability of moderate-severe CUD to that provided by the SIS-C alone.
The DSM-5 Substance Use Symptom Checklist is a helpful tool for providers that’s easy to administer and increases the likelihood of correctly identifying moderate-to-severe CUD in patients at average risk. When providers can more accurately screen for CUD, they are better able to have meaningful and helpful conversations with patients about their needs and wants for additional support or care. Though this study reports on the use of the DSM-5 Substance Use Symptom Checklist in primary care settings, this tool is also used in substance use disorder treatment specialty settings as well. Find a copy of the Checklist here.
Need help getting a copy? Email, Meg Brunner, [email protected].