The Hospital Frailty Risk Score is Not an Accurate Predictor of Treatment Costs for Total Joint Replacement Patients in a Medicare Bundled Payment Population

Background: Medically complex patients require more resources and experience higher costs within total joint arthroplasty (TJA) bundled payment models. While risk adjustment would be beneficial for such patients, no tool currently exists which can reliably identify these patients preoperatively. The purpose of this study is to determine if the Hospital Frailty Risk Score (HFRS) is a valid predictor of high-TJA treatment costs. Methods: Retrospective analysis was performed on patients who underwent primary TJA between 2015 and 2020 from a single large orthopedic practice. ICD-10 codes from an institutional database were used to calculate HFRS. Cost data including inpatient, postacute, and episode of care (EOC) costs were collected. Charlson comorbidity index, demographics, readmissions, and complications were analyzed. Results: 4936 patients had a calculable HFRS and those with intermediate and high scores experienced more frequent readmissions/complications after TJA, as well as higher EOC costs. However, HFRS did not reliably predict EOC costs, yielding a sensitivity of 49% and specificity of 66%. Multivariate analysis revealed that both patient age and sex are superior individual cost predictors when compared with HFRS. Secondary analyses indicated that HFRS more effectively predicts TJA complications and readmissions but is still nonideal for clinical applications. Conclusion: HFRS has poor sensitivity as a predictor of high-EOC costs for TJA patients but has adequate specificity for predicting postoperative readmissions and complications. Further research is needed to develop a scale that can appropriately predict orthopedic cost outcomes. (C) 2021 Elsevier Inc. All rights reserved.

Automated summary

The study indicates that the Hospital Frailty Risk Score (HFRS) has low sensitivity in predicting high treatment costs for total joint arthroplasty (TJA) patients, but has sufficient specificity in predicting postoperative readmissions and complications. The results show that patient age and sex are superior predictors of individual costs compared to HFRS, and HFRS is still not reliably predictive of treatment costs for TJA patients.