Implementing Care for Cannabis and Other Drug Use in Adult Primary Care: Outcomes of a Cluster-Randomized Implementation Trial.

Abstract

Introduction: The US Preventive Services Task Force recommends substance use screening for adults when accurate diagnosis and follow-up treatment can be offered. This study evaluated whether implementation of population-based screening and symptom assessment increased treatment for cannabis and other drug use disorder (CUD/DUD) in a large primary care system.

Methods: Nineteen randomized regional healthcare system sites implemented population-based screening and symptom assessment for cannabis and other drug use as part of a stepped-wedge trial to integrate mental health in adult primary care from 1/2016-7/2018. Implementation strategies included: practice facilitation, electronic health record (EHR) decision support, and performance feedback. Outcomes included treatment initiation and engagement rates among patients with any new DUD, and secondarily, rates for specific DUDs (e.g., CUD), using EHR and insurance claims data. Mixed-effect logistic regression modeled binary outcomes before and after implementation (1/2015-1/2019), adjusted for randomization stratification and time, accounting for person-level repeated outcomes. Monthly outcome rates (per 10,000 patient-visits) were estimated using marginal standardization. Analyses were conducted 2023-2024.

Results: Pre-implementation, 244,542 patients had 942,400 visits; post-implementation, 287,696 patients had 1,087,565 visits. Implementation resulted in increased cannabis screening (9 to 153/10,000 patient-visits; p<0.001) and newly identified CUD (10 to 17/10,000 patient-visits, p<0.001). Treatment initiation for CUD increased 0.4 to 1/10,000 patient-visits (p=0.006), but treatment engagement did not change (p=0.147). Treatment initiation and engagement for any DUD did not change (p=0.777 and 0.584).

Conclusion: Findings underscore potential value of integrating population-based cannabis screening/assessment in primary care for treatment initiation and highlight the need for improved engagement for CUD.