A Randomized Trial of Behavioral Economic Approaches to Reduce Unnecessary Opioid Prescribing (REDUCE)

In this study, the investigators will evaluate the effect of a health system initiative aiming to change clinician opioid prescribing behaviors using two behavioral economic interventions - individual audit feedback and peer comparison feedback of clinicians.

Opioid-related abuse and overdose represent a growing national epidemic in the United States. Clinician practice patterns play an important role: opioid prescriptions impact the likelihood that patients will misuse or become dependent on these medications, with longer prescriptions leading to greater sustained use. In this study, we will evaluate a Sutter Health System quality improvement initiative using monthly individual audit feedback and/or monthly peer comparison feedback to clinicians to change opioid prescribing patterns. In partnership with Sutter Health System, this will be conducted using randomization to evaluate its effect. We will also conduct a process evaluation to understand factors associated with better or worse performance at the clinician level.

Clinicians will receive peer comparison feedback informing them of how their performance compares to their peers.

Clinicians will receive individual audit feedback informing them of their performance and peer comparison feedback informing them of how their performance compares to their peers.

Practice sites randomly assigned to have individual audit feedback will be sent information by email to each clinician at the site each month for the duration of the intervention period. The email will inform the clinician that the health system is doing monthly audits and provide them the number of patients in the last month for whom they prescribed 30 opioid pills per prescription or higher.

Practice sites randomly assigned to have peer comparison feedback will be sent information by email to each clinician at the site each month for the duration of the intervention period. Data on the mean number of opioid pills per prescription and the proportion of visits with an opioid prescription will be delivered using a 3-month rolling average as follows: a) If clinician is above median: informed how their data compares to the median; b) If clinician is below median but above 10th percentile: informed how their data compares to the 10th percentile; c) If clinician is 10th percentile or below: informed of their data and commended for being a "low prescriber."

The primary outcome is the change in the mean number of pills prescribed per opioid prescription from baseline to the intervention period.

The secondary outcome is the change in proportion of patient visits in which an opioid is prescribed from baseline to the intervention period.

The change in morphine milligram equivalents per opioid prescription from baseline to the intervention period.

The change in the mean number of opioid pills per patient-visit from baseline to the intervention period.

The change in proportion of patient-visits with non-opioid pain prescriptions (e.g., ibuprofen, acetaminophen, celecoxib, or muscle relaxants such as cyclobenzaprine, baclofen or tizanidine) from baseline to the intervention period.

Inclusion Criteria: Clinicians 1. Practiced primarily at participating emergency department or urgent care center Patients Presented to a participating emergency department or urgent care center during the study period Discharged to home from the visit Exclusion Criteria: Clinicians Saw less than 100 patients in the prior year Practiced primarily at another site that is not in the main trial Did not practice at Sutter Health in the prior 90 days Patients Currently pregnant