Analyze baseline concurrent opioid prescribing metrics at the individual prescriber level in the Duke Health System on the identified three main outcome measures.

Test the impact of reports on opioid prescriber behaviors with the following primary measures: number of prescriptions with concurrent benzo within reporting period, number of prescriptions with concurrent muscle relaxants within reporting period, and number of encounters with naloxone prescriptions for patients with any opioid-related diagnosis within reporting period.

Create a blueprint to implement the concurrent opioid prescribing nudge intervention in other settings.

The Concurrent Opioid Prescribing Nudge project intends to address multiple points within the opioid-use cycle through the development of standardized and scalable reporting mechanisms to provide social comparisons and feedback to physicians across the Duke Health System regarding their concurrent and co- opioid, benzodiazepines, and muscle relaxant prescribing practices. Concurrent are hereby defined as:

• writing a new opioid prescription for a patient with a benzodiazepine prescription within the last 3 months,
• writing a new benzodiazepine prescription for a patient with a opioid prescription within the last 3 months,
• writing a new muscle relaxant prescription for a patient with an opioid prescription within the last 3 months,
• writing a new opioid prescription for a patient with muscle relaxant prescription within the last 3 months, or
• writing a new prescription for both an opioid and benzodiazepine or opioid and muscle relaxant.

The proposed project will leverage insights from behavioral economics to design informational and social incentives to reduce concurrent practices and mitigate opioid harm. Opioid prescribers (attending physicians, residents, and advanced practice providers) at participating departments and clinics in the Duke Health System will be randomized to a control or intervention arm. Over six month reporting periods beginning fall 2019, providers in the intervention arms will receive monthly reports with their individual prescribing patterns and comparison to peer prescribing patterns for the following measures: number of prescriptions with concurrent opioid active prescriptions of opioid/ benzodiazepines, number of prescriptions with concurrent opioid active prescriptions of opioid/muscle relaxants, and number of missed opportunities to prescribe naloxone to patients with any opioid-related diagnosis. The control arm will receive usual clinical education and feedback. Interventions will be implemented at participating departments and clinics utilizing a stepped-wedge timeline.

Inclusion Criteria:

The primary population of focus for this study is:

• attending physicians
• residents
• advanced practice providers

hereby referred to as opioid prescribers in Duke University Health System and may include these departments and clinics:

• Emergency Department
• Neurology, Pain Management
• Primary Care
• Psychiatry, Sleep Disorder Clinic
• Spine

All opioid prescribers in these settings will be identified in partnership with Duke University Health System.

Exclusion Criteria:

• Providers not identified above