A Proposed Alternative Sampling Strategy for Centers for Medicare & Medicaid Services Arthroplasty Patient-Reported Outcome Performance Measure Compliance.

Abstract

The Centers for Medicare & Medicaid Services (CMS) mandate reporting of Hip disability and Osteoarthritis Outcome Score for Joint Replacement (HOOS, JR.) and Knee injury and Osteoarthritis Outcome Score for Joint Replacement (KOOS, JR.) to support a hospital-level patient-reported outcome performance measure (PRO-PM). The current CMS requires matched pre- and postoperative PROMs for ≥ 50.0% of all eligible patients, regardless of hospital volume. This may yield unreliable performance estimates for low-volume centers and potentially biased and unnecessarily large estimates for large hospitals. The purpose of this study was to determine the smallest sample required to produce statistically valid PRO-PM estimates. We calculated the minimum number of completed PROMs needed in order to achieve a 95% confidence interval (CI) width 10.0% (± 5.0%) for hospitals with five, 50, 500, and 5,000 eligible total hip arthroplasty (THA) and total knee arthroplasty (TKA) cases annually, assuming a conservative substantial clinical benefit (SCB) achievement of 50.0%. These values were compared with current CMS requirements. Sensitivity analyses explored the effects of confidence interval width, hospital volume, and frequency of SCB attainment on required sample sizes. The CMS's volume-proportional rule requires three, 25, 250, and 2,500 patients for hospitals with five, 50, 500, and 5,000 annual cases, respectively. In contrast, a statistically grounded random sampling strategy would require five, 45, 218, and 357 patients, respectively. Sensitivity analyses confirmed that random sampling reduces required sample sizes for high-volume hospitals. The current CMS reporting rule may produce unreliable performance estimates for low-volume hospitals while imposing excessive reporting burdens and potentially biased results for high-volume hospitals. A random sampling strategy may provide more precise and valid estimates while substantially reducing hospital burden, especially among the hospitals performing 500 or more THA and TKAs per year, which comprise approximately 35.0% of all THAs and TKAs annually.