Comparison of Responsiveness of Patient-Reported Outcome Measures After Total Knee Arthroplasty

Abstract

BackgroundThe aim of this study was to evaluate the responsiveness of different patient-reported outcome measures in patients with primary total knee arthroplasty (TKA). MethodsIn this prospective observational study, we assessed patients with TKA before the surgery, after 4 months, after 1 year, and after 2 years. Measures were the objective Knee Society Score (KSS) and the following patient-reported outcome measures: Knee injury and Osteoarthritis Outcome Score (KOOS), KOOS-12, Forgotten Joint Score (FJS-12), High-Activity Arthroplasty Score, and EQ-5D-3L. Responsiveness was determined by effect size (ES), standardized response mean (SRM), area under the receiver operating characteristics curve, floor and ceiling effects, and hypothesis testing. ResultsWe analyzed data from 309 TKAs (272 patients, 56% female). The ES and SRM for the change in KSS, KOOS, KOOS-12, FJS-12, and EQ-5D-3L from baseline to each follow-up were large (>0.8). The largest responsiveness from baseline to follow-up was found for the KSS, KOOS/KOOS-12 quality of life, KOOS-12 summary, KOOS-12 pain, and FJS-12 (2.0 > ES <3.9, 1.4 > SRM <2.4). The area under the curve from baseline to each follow-up was ≥0.7 for KOOS, KOOS-12, and FJS-12 (range 0.71 to 0.95) and <0.7 for KSS and EQ-5D-3L (range 0.65 to 0.74). We found floor or ceiling effects in the KOOS, KOOS-12, and EQ-5D-3L, but not in the KSS, FJS-12, and High-Activity Arthroplasty Score. ConclusionsOur study demonstrated that responsiveness differed between the various measures. The KOOS-12 and FJS-12 showed the greatest internal and external responsiveness, although ceiling effects occurred in the KOOS-12.