Evaluating Patient and Surgeon Characteristics Associated with Care Cost and Outcomes for Knee and Hip Replacement Procedures: A National Medicare Cohort Study.

Background: The role of physician credentialing has been widely considered in quality and outcome improvement studies. However, the association between surgeon characteristics and health-care costs remains unclear. Methods: Our objective was to determine the association of orthopaedic surgeon characteristics with health outcomes and costs, utilizing Medicare data. We used 100% Fee-for-Service Medicare data from 2015 to 2019 to identify all patients ≥65 years of age who underwent 2 common orthopaedic surgical procedures, total hip and knee replacement. After determining whether the patients had been readmitted after discharge from their initial admission for surgery, we computed 3 metrics of total medical expenditure: the costs of the initial surgery admission and 30-day and 180-day episode-based bundles of care. Hierarchical linear regression and logistic regression models were used to evaluate patient and surgeon characteristics associated with care costs and the likelihood of readmission. Results: We identified 2,269 surgeons who performed total knee replacements on 298,934 patients and 1,426 surgeons who performed total hip replacements on 204,721 patients. Patient characteristics associated with higher initial surgery costs included increasing age, female sex, racial minority status, and a higher Charlson Comorbidity Index. Surgeon characteristics associated with lower readmission rates included practice in the Northeast region and a higher patient volume; having malpractice claims was associated with higher readmission rates. Conclusions: A higher volume of patients treated by the orthopaedic surgeon was associated with lower overall costs and readmission rates. Information on surgeons' malpractice claims and annual volume should be made publicly available to assist patients, payer networks, and hospitals in surgeon selection and oversight. These results could also inform the guidelines of physician credentialing organizations. Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.