Validation of the Korean Version of the Oxford Knee Score in Patients Undergoing Total Knee Arthroplasty

Abstract

Although translated versions of the Oxford Knee Score (OKS) in several languages are available, the absence of a Korean version precludes comparing data from Korea with that from other countries using the OKS. We therefore evaluated the reliability and validity of the adapted Korean version of the OKS. We first translated the English version of the OKS into Korean, then back into English, then held expert committee discussions to finalize the Korean version. We then mailed the Korean version of the VAS for pain, OKS, and the previously validated SF-36 to 142 patients who underwent TKAs for knee osteoarthritis. Factor analysis and reliability assessment using the kappa statistic of agreement for each item, the intraclass correlation coefficient, and Cronbach's alpha were conducted. To determine the subscales of the OKS, we used the factor analysis. We also evaluated concurrent and construct validity by comparing the responses to the OKS with the results of the VAS and SF-36 using Pearson's correlation coefficient. All items had a kappa statistic of agreement greater than 0.6. The OKS showed test and retest reliability as follows: OKS, 0.848; Factor 1, 0.867; and Factor 2, 0.819. Internal consistency of Cronbach's alpha was as follows: OKS, 0.932; Factor 1, 0.907; Factor 2, 0.867. The OKS correlated (r = 0.692) with the VAS. The Korean version of the OKS correlated with the SF-36 total and individual domain (physical functioning, role physical, bodily pain) scores. The adapted Korean version of the OKS was translated and showed acceptable measurement properties. The data suggest it is suitable for assessing outcomes in Korean-speaking patients having TKAs.