Use of a Behavioral Economic Intervention to Reduce Antibiotic Prescription for Upper Respiratory Infections

Overview

In an effort to implement Choosing Wisely guidelines and decrease patient harm, we will implement and evaluate a clinician audit-feedback and behavioral "nudge" initiative to reduce low-value antibiotics for URIs. Using a quasi-experiment (pre-post) design, antibiotic prescriptions for URI at LAC+USC Urgent Care Center (intervention site) vs. Olive View-UCLA Urgent Care Center (control site) will used to test the effects of behavioral "nudge" on antibiotic prescribing.

National prescription rates for low-value antibiotics for uncomplicated upper respiratory infections (URIs) remain unacceptably high, including at LAC+USC Medical Center-one of the largest safety net medical centers in the U.S. Using a quasi-experiment (pre-post) design, antibiotic prescriptions for URI at LAC+USC Urgent Care Center (intervention site) vs. Olive View-UCLA Urgent Care Center (control site) will be compared. A three-part intervention at LAC+USC consists of (1) the urgent care medical director emailing Choosing Wisely® guidelines and presented journal club to all 16 urgent care clinicians, and then (2) leveraging EHR performance data to provide individual clinicians with case-specific audit-feedback (both via emails and in-person while precepting nurse practitioners) on low-value antibiotic prescribing, and (3) using a behavioral "nudge", urgent care clinicians will sign a large poster committing to avoid prescribing low-value antibiotics for uncomplicated URIs, which will be displayed in the clinic. In contrast, the control site (Urgent Care Center at Olive View-Medical Center) will receive Centers for Disease Control prescription pads for non-antibiotic treatments (e.g., decongestants) that offer patients alternatives to antibiotics, in a broader health system effort to reduce antibiotic prescribing. Patient with URIs (e.g., acute bronchitis, bronchitis NOS, excluding guideline-based red flags such as COPD, HIV) will be identified using electronic health record clinical billable data and low-value antibiotic prescriptions rates per visit will be monitored at both sites. Differences in prescriptions rates will be determined using an interrupted time-series analysis comparing utilization between sites using a repeated measures logistic regression model. .