The Impact of Surgeon Variability on Patient-Reported Outcome Measures, Length of Stay, Discharge Disposition, and 90-Day Readmission in TKA.

Abstract

Studies involving total knee arthroplasty (TKA) have focused on patient-related factors as drivers of outcomes. Although some studies have investigated surgeon and/or surgery-level factors (i.e., approach, volume), the measure of variance in postoperative outcomes across surgeons following TKA has not been determined. The aim of the current study was to evaluate the relationship between the surgeon and 1-year patient-reported outcome measures, the length of stay, the discharge disposition, and 90-day readmission after TKA, as well as the differences in these variables among surgeons. Data were prospectively collected for 5,429 patients who underwent TKA at a large health-care system with 36 surgeons between 2016 and 2018. Likelihood ratio tests were performed to investigate the relationship between surgeon and the 1-year Knee injury and Osteoarthritis Outcome Score (KOOS)-Pain subscale, KOOS-Physical Function Shortform (KOOS-PS), KOOS for Joint Replacement (KOOS-JR), patient acceptable symptom state (PASS), length of stay, discharge disposition, and 90-day readmission. The minimal clinically important difference (MCID) was used to assess meaningful outcomes. Variable importance was determined by the Akaike information criterion (AIC) increase, using ordinal and binary-response mixed-effect models. There was a significant association between surgeon and KOOS-Pain (p < 0.001), KOOS-PS (p = 0.001), KOOS-JR (p < 0.001), PASS (p = 0.024), length of stay (p < 0.001), discharge disposition (p < 0.001), and 90-day readmission (p < 0.001). When modeling 1-year KOOS-Pain (AIC increase, 15.6), KOOS-PS (AIC increase, 6.8), KOOS-JR (AIC increase, 13.5), PASS (AIC increase, 1.9), length of stay, and discharge disposition, the surgeon variable contributed more to the 1-year outcome than some patient-level factors (e.g., body mass index, Charlson Comorbidity Index). The difference between the highest and lowest median probabilities of attaining the same value for the KOOS-Pain (11.2%), KOOS-PS (9.4%), KOOS-JR (11.8%), PASS (5.9%), length of stay (46.6%), discharge disposition (22.8%), and readmission (13.1%) indicated surgeon-level variability. Surgeon-related factors may be stronger contributors to the 1-year patient-reported outcome measures and length of stay than patient characteristics emphasized in the literature. Current findings have suggested variability in patient-reported outcome measures, length of stay, discharge location, and 90-day readmission among surgeons. Surgeon variability should be considered when model-fitting in the setting of TKA. Prognostic Level II . See Instructions for Authors for a complete description of levels of evidence.