Effect of Peer Comparison Feedback, Individual Audit Feedback or Both to Clinicians on Opioid Prescribing in Acute Care Settings: A Cluster Randomized Clinical Trial
Abstract
Research Objective
Prescribing opioids, particularly the number of pills, is associated with greater likelihood of patients developing longer-term opioid dependence. Nudges targeted to clinicians are a low-cost strategy that could reduce unnecessary opioid prescribing. In particular, clinician-focused peer comparison feedback has been effective in influencing prescribing for other medications. However, peer comparison feedback has not been well-tested for opioid prescribing, alone or compared against other feedback approaches such as individual audit feedback.
Study Design
We conducted a pragmatic, four-arm factorial cluster randomized clinical trial among 48 emergency department (ED) and urgent care (UC) practice sites within Sutter Health System including a 6-month pre-intervention period and a 6-month intervention. 438 clinicians were cluster randomized by practice site. Interventions were delivered to clinicians monthly by email. Peer comparison feedback included the number of pills per opioid prescription and proportion of encounters with an opioid prescription in the prior 3 months, relative to that of their practice site peers. Individual audit feedback included the number of prescriptions with >30 pills in the prior 3 months, indicating potential outlier prescriptions. The primary outcome measure was the change in the number of pills per prescription. Secondary outcomes included changes in the proportion of pills and milligrams of morphine equivalents (MME) per prescription.
Population Studied
The sample included 263 ED and 175 UC clinicians comprised of 294,962 patient encounters, with a mean (SD) age of 49 years (19), 56% female, 9% Black, and 21% Hispanic.
Principal Findings
During the pre-intervention period, there was a mean (SD) of 15.1 pills (3.9) and 76.3 MME (23.0) per prescription, and 9.3% of encounters with an opioid prescription. In adjusted analyses compared to usual care during the intervention, there was a significant decrease in pills per prescription among clinicians receiving peer comparison feedback alone (−0.9 pills per prescription; 95% CI -1.5 to −0.3, P = 0.002) and receiving both peer comparison and individual audit feedback (−1.4 pills per prescription; 95% CI -2 to −0.8, P < 0.001), but not among clinicians receiving individual audit feedback alone (−0.4 pills per prescription; 95% CI -1.0 to 0.2, P = 0.24). There was also a significant decrease in MME per prescription for peer comparison feedback alone (−3.4 MME per prescription; 95% CI -6.6 to −0.3; P = 0.03) and combined with individual audit feedback (−4.3 MME per prescription; 95% CI -7.5 to −1.1; P = 0.009), but not for individual audit feedback alone (−3.1 MME per prescription; 95% CI -6.2 to 0.1; P = 0.06). There were no significant changes in the proportion of encounters with an opioid prescription.
Conclusions
Peer comparison feedback was effective, alone and together with individual audit feedback, for significantly reducing the number of pills and MME per opioid prescription.
Implications for Policy or Practice
This is one of the largest trials ever conducted testing the impact of nudges on opioid prescribing and provides promise for low-cost strategies to change clinician prescribing behavior more broadly.
Primary Funding Source
The Donaghue Foundation.
