Traditional Risk Factors and Logistic Regression Failed to Reliably Predict a “Bundle Buster” After Total Joint Arthroplasty

The purpose of this study was to determine if we could identify patient factors that were predictive of Medicare and privately insured patients being “high-cost.”

Ninety-day episode-of-care insurance company payments along with collected demographics, comorbidities, and readmissions were reviewed for a consecutive series of primary total joint arthroplasty patients from 2015 to 2016 at our institution. High-cost patients were identified by determining those patients above the cutoff, where the cost data became demonstrably nonparametric and both univariate analysis and logistical regressions were performed to identify risk factors that lead to increased costs. Receiver operator curves were created to determine the predictive nature of these risk factors.

Univariate analysis showed that high-cost privately insured patients were significantly older, more likely to be readmitted and less likely to be discharged to home ( P < .001) whereas high-cost Medicare total knee/total hip arthroplasty patients were more likely to have many of the comorbidities analyzed. Logistical regression did not find any predictive factors for privately insured patients and found that diabetes (OR 1.47 and 1.75, respectively), congestive heart failure (OR 1.94 and 3.46, respectively), cerebrovascular event (OR 2.20 and 2.20, respectively) and rheumatic disease (OR 1.78 and 1.78, respectively) were all predictive of being a high-cost Medicare patient.

Traditional risk factors for postoperative complications are not reliably associated with increased patient costs after total hip and total knee arthroplasty. Furthermore, the risk factors associated with increased costs vary greatly between privately insured and Medicare-insured patients. Further investigation is necessary to identify cost drivers in this patient subset to preventive higher costs.