Comparison of the responsiveness of the SF-36, the Oxford Knee Score, and the Knee Society Clinical Rating System in patients undergoing total knee replacement

Objectives: To compare the responsiveness of the Knee Society (KS) Clinical Rating System, the general health status measure Short Form 36 (SF-36), and both the raw and Rasch-based scores of the condition-specific Oxford Knee Score (OKS) in patients undergoing total knee replacement (TKR) Methods: Data were prospectively collected as part of routine care from adult patients who underwent TKR between 2001 and 2006. OKS data fit the Rasch partial credit model after removing items regarding limping and kneeling. Responsiveness was assessed using effect size (ES), standardised response mean (SRM), and relative validity (RV). Results: Among 702 patients with complete data at baseline and two follow-ups, the pain subscale of the KS (KS-P), raw-OKS, and Rasch-OKS consistently had higher levels of responsiveness than all eight SF-36 and the other KS subscales. At 6-month follow-up, Rasch-OKS had the largest ES and KS-P had the largest SRM (2.7 and 2.0, respectively). When compared to raw-OKS, the RVs of KS-P, Rasch-OKS, SF-36 bodily pain, and SF-36 physical functioning were 1.1, 0.66, 0.49, and 0.36, respectively. A similar ordering of responsiveness was observed at 24-month follow-up. Conclusion: The OKS and KS-P are more responsive than most SF-36 subscales in TKR patients. Raw-OKS and Rasch-OKS have comparable responsiveness. Different responsiveness indices may give different results.