Strategies to Deimplement Opioid Prescribing in Primary Care: A Cluster Randomized Clinical Trial.

Importance: Centers for Disease Control and Prevention guidelines advocate reduced opioid prescribing for chronic pain, yet research on their implementation remains limited. Objective: To compare 4 deimplementation strategies to promote guideline-concordant opioid prescribing. Design, Setting, and Participants: This cluster randomized clinical trial was performed at 32 primary care clinics from 2 US health care systems from February 2020 to March 2022, using a hybrid type 3 sequential multiple-assignment design focused on patient outcomes. Clinics were recruited through volunteer sampling, including 268 clinicians and 8978 patients. Data were analyzed from September 2020 to March 2022. Intervention: Deimplementation strategies were targeted at the system, clinic, and prescriber levels. All clinics received a system-level strategy consisting of quarterly educational meetings with monthly audit and feedback (EMAF) reports. At month 3, half the clinics were randomized to receive practice facilitation (PF), a clinic-level strategy that targets clinic workflows. At month 9, half the clinics were again randomized to add prescriber peer consulting (PPC), a prescriber-level strategy focused on challenging patient cases. Main Outcomes and Measures: The primary outcome was change in mean morphine milligram equivalent (MME) dose in clinics receiving the least intensive bundle of deimplementation strategies (EMAF) vs the most intensive (EMAF plus PF plus PPC). Secondary outcomes included adherence to guideline metrics aimed at mitigating opioid risk. Results: Among the 8978 patients included in the analysis, 5142 (57.3%) were female; 42 (0.5%), American Indian or Alaska Native; 74 (0.8%), Asian or Pacific Islander; 411 (4.6%), Black; 187 (2.1%), Hispanic or Latino; 8127 (90.5%), White; and 137 (1.5%), other or unknown. Mean (SD) age was 58.3 (14.3) years. Eight clinics (including 66 prescibers and 2044 patients) assigned the most intensive strategy (EMAF plus PF plus PPC) had statistically significant effects on the primary outcome compared with 7 clinics (including 60 clinicians and 2427 patients) receiving the least intensive strategy (EMAF); clinics in the high-intensity group decreased the mean MME dose by 2.4 (95% CI, -4.3 to -0.5) mg/d more than the EMAF group (P = .02), representing a 6% reduction, and increased screening for pain severity, enjoyment of life, and general activity by 5.4% (95% CI, 0.4%-10.4% [P = .04]) more. Compared with EMAF, the most intensive strategy resulted in statistically significant decreases in urine drug screening (difference, -7.3% [95% CI, -11.5% to -3.0%]; P < .001) and use of treatment agreements (difference, -6.7% [95% CI, -11.1 to -2.3%]; P = .003), in the opposite direction of the hypothesis. There were no significant differences between groups in benzodiazepine coprescribing, mental health screening, or patients receiving an MME dose greater than or equal to 90.0 mg/d. Conclusions and Relevance: In this cluster randomized clinical trial, a high-intensity deimplementation strategy targeted at prescribers significantly decreased the MME dose and increased screening for pain intensity and pain-related interference while reducing use of treatment agreements and urine drug screening. Providing clinic- and prescriber-level deimplementation strategies may help health systems take positive steps toward reducing reliance on opioid medications for chronic pain management in primary care settings. Trial Registration: ClinicalTrials.gov Identifier: NCT04044521.