Total Ankle Arthroplasty in the Alternative Payment Model Era. Issue 4 (16th October 2019)

• Record Type: Journal Article
• Authors: Jones, Stephanie M.; McDowell, Arthur R.; Altieri-Dunn, Stephanie C.; Bilderback, Andrew; Conti, Stephen; Hasselman, Carl; Kline, Alex; Saar, William; Yan, Alan Y.; Hogan, MaCalus V.
• Is Part Of: Foot & ankle orthopaedics. Volume 4:Issue 4(2019)
• Publication Date: 2019-10-16
• Subjects: Total ankle arthroplasty -- Alternative payment model -- Bundled-payment -- Cost -- Quality; Foot -- Surgery -- Periodicals; Ankle -- Surgery -- Periodicals; Orthopedics -- Periodicals; 617.584
• DOI: 10.1177/2473011419S00236
• Languages: English
• ISSNs: 2473-0114

Abstract

Category: Ankle, Ankle Arthritis Introduction/Purpose: Ankle arthritis often results in functional decline and decreased quality of life. Total ankle arthroplasty (TAA) is an acceptable option for end-stage disease. With a growing active and aging population, there's increasing demand for total joint arthroplasty (TJA). In response, the Centers for Medicare and Medicaid Services (CMS) implemented the Comprehensive Care for Joint Replacement (CJR) program to curb costs and optimize outcomes associated with TJA. TAA is included in CJR; however, little is known about episode of care costs and quality of care delivered to patients within a bundled- payment model. This study provides an analysis of quality and cost across an episode of care for primary TAA performed under the CJR model and a commercial bundled-payment health plan. Methods: At a large hybrid academic institution, CJR and bundled-payment health plan data was queried for TAA discharges at 7 affiliate hospitals within Metropolitan Service Areas (MSAs). Patient-reported outcomes (PROs) between April 2016 and August 2018 were assessed pre-operatively, 0-3 months post-operatively and >3 months post-operatively. Foot and Ankle Ability Measure (FAAM), Global Rate of Change (GRC), and Patient Acceptable Symptom State (PASS) were analyzed. Episode of care cost and associated cost domains were assessed between April 2016 and March 2018. Statistical analysis was performed using Stata® 15.1. Pre-operative and post-operative FAAM scores were analyzed using two-tailed t-tests. Pre-operative and post-operative GRC and PASS scores were analyzed using likelihood ratio tests. Associations between cost and FAAM scores were analyzed using Spearman correlations. Comparisons between cost and GRC or PASS were analyzed by Kruskal-Wallis tests and Wilcoxon rank- sum tests, respectively. Statistical significance was defined as a p-value <0.05. Results: 155 (55% male, mean 66.1 years) patients were identified. There was no difference between pre-operative and 0-3 months post-operative function (p=0.4943). However, there was significant improvement in functional outcomes at >3 months post-operatively (p=0.0113). Patients reported significant improvement in perceived change in ankle status (p<0.001, p<0.001) and satisfaction of ankle status (p<0.001, p<0.001) at both 0-3 months and >3 months post-operatively when compared to pre- operative assessments. The average TAA episode of care cost was $24, 398. Average contracted diagnosis-related group (DRG) costs accounted for the majority (67.58%) of the average total cost. The average readmission cost was $9, 440. Costs for skilled nursing facilities (SNF) and inpatient rehabilitation were on average $7, 462 and $19, 635, respectively. There were no significant correlations observed between outcomes and cost (p>0.05). Conclusion: Primary TAA within a bundled-payment model demonstrates that alternative payment models can reasonably maintain quality of care standards while concurrently promoting cost-efficient utilization of medical resources. Significant functional improvement and increases in patient satisfaction were observed following primary TAA in the bundled-payment system. With the progressive shift from fee-for-service to value-driven reimbursement, continued evaluation of the relationship between quality and cost for TJA within an alternative payment model is necessary.