Evaluation of Opioid Prescribing Patterns and Drivers of Variation Following Endoscopic Procedures for Benign Prostatic Hyperplasia Across an Integrated Academic Health System

Abstract

To evaluate patient, provider, and facility factors associated with variation in opioid prescribing after endoscopic procedures for benign prostatic hyperplasia across a large academic health system to drive improvement efforts. Opioids prescribed at discharge for patients who underwent an endoscopic prostate procedure March 2018-November 2019 were analyzed. Multivariable logistic and linear regression were used to evaluate the relationship between patient, provider, and facility factors and the receipt of any opioid prescription and the quantity prescribed. We included 724 patients who had surgery with one of 26 urologists across five facilities. 222 (30.7%) received an opioid prescription, and the average morphine milligram equivalents (MMEs) prescribed was 97.9±33.5. We found wide variation in the proportion of patients who received an opioid prescription across surgeons (range 0%-88.9%) and facilities (range 19.9%-66.7%) and the average MMEs prescribed (range 25-188.5). Outpatient surgery (OR 2.32; 95% confidence interval [CI] 1.22-4.40, P = .010) and preoperative opioid use (OR 15.04; CI 9.65-23.45, P < .001) were associated with higher rates of opioid prescribing, while prescribing decreased with increasing patient age (OR 0.97; CI 0.95-0.99, P = 0.016). Multivariable linear regression analysis demonstrated an association between surgery at satellite facilities, having a surgeon in practice for at least 20 years, and higher surgeon volume with increased MMEs prescribed. Opioid prescribing following endoscopic prostate procedures varied widely. Targeted interventions tailored to younger patients, those taking opioids preoperatively, recipients of outpatient surgery and those undergoing surgery at satellite facilities may be particularly high yield given the association between these factors and increased postoperative prescribing.