Background: The association between surgeon prescribing practices and new persistent postoperative opioid use is not well understood. We examined the association between surgeon prescribing and new persistent use among total hip arthroplasty (THA) patients.Methods: A retrospective analysis of Medicare claims in Michigan was performed. The study cohort consisted of orthopedic surgeons performing THAs from 2013 to 2016 and their opioid-naive patients, aged >65 years. High-risk prescribing included high daily doses, overlapping benzodiazepine prescriptions, concurrent opioid prescriptions, prescriptions from multiple providers, or long-acting opioid prescriptions. The occurrence of a preoperative prescription, initial prescription size, and 30-day prescription dosage were examined as individual exposures. Surgeons were categorized into quartiles by prescribing practices, and multilevel hierarchical logistic regression was used to examine associations with postoperative new persistent opioid use.Results: Surgeons exhibited high-risk prescribing for 66% of encounters. Patients of surgeons with the highest rates of high-risk prescribing were more likely to develop persistent use compared with patients of surgeons with the lowest rates (adjusted rates: 9.7% vs 4.6%, P = .011). Patients of surgeons with initial prescription sizes in the "high" (third) quartile (adjusted odds ratio, 2.91; 95% confidence interval, 1.53-5.51), and of surgeons in the "highest" (fourth) quartile of 30-day prescription dosage (adjusted odds ratio, 1.93; 95% confidence interval, 1.03-3.61), were more likely to develop persistent opioid use compared with patients of surgeons with low initial and 30-day prescription sizes, respectively.Conclusion: The development of persistent opioid use after surgery is multifactorial, and surgeon prescribing patterns play an important role. Reducing prescribing and encouraging opioid alternatives could minimize postoperative persistent opioid use. (C) 2020 Elsevier Inc. All rights reserved.