Association between cannabis use disorder symptom severity and probability of clinically-documented diagnosis and treatment in a primary care sample.

Brief cannabis screening followed by standardized assessment of symptoms may support diagnosis and treatment of cannabis use disorder (CUD). This study tested whether the probability of a medical provider diagnosing and treating CUD increased with the number of substance use disorder (SUD) symptoms documented in patients' EHRs. This observational study used EHR and claims data from an integrated healthcare system. Adult patients were included who reported daily cannabis use and completed the Substance Use Symptom Checklist, a scaled measure of DSM-5 SUD symptoms (0−11), during routine care 3/1/2015–3/1/2021. Logistic regression estimated associations between SUD symptom counts and: 1) CUD diagnosis; 2) CUD treatment initiation; and 3) CUD treatment engagement, defined based on Healthcare Effectiveness Data and Information Set (HEDIS) ICD-codes and timelines. We tested moderation across age, gender, race, and ethnicity. Patients (N=13,947) were predominantly middle-age, male, White, and non-Hispanic. Among patients reporting daily cannabis use without other drug use (N=12,568), the probability of CUD diagnosis, treatment initiation, and engagement increased with each 1-unit increase in Symptom Checklist score (p's<0.001). However, probabilities of diagnosis, treatment, and engagement were low, even among those reporting ≥2 symptoms consistent with SUD: 14.0% diagnosed (95% CI: 11.7–21.6), 16.6% initiated treatment among diagnosed (11.7–21.6), and 24.3% engaged in treatment among initiated (15.8–32.7). Only gender moderated associations between Symptom Checklist and diagnosis (p=0.047) and treatment initiation (p=0.012). Findings were similar for patients reporting daily cannabis use with other drug use (N=1379). Despite documented symptoms, CUD was underdiagnosed and undertreated in medical settings. • Documentation of cannabis use disorder (CUD) diagnosis and treatment was low. • A DSM-5 Symptom Checklist offered to patients could support diagnosis and treatment. • CUD diagnosis and treatment increased with report of symptoms on the checklist. • Gender moderated associations between reported symptoms and diagnosis and treatment. • There were missed opportunities to identify and treat CUD across all subgroups. [ABSTRACT FROM AUTHOR]