DOS 2019

DOS Kongressen 2019 · 177 Translation and classical test theory validation of the Danish version of the Oxford Knee Score Anne Mørup-Petersen, Michael Krogsgaard, Rasmus Nielsen, Aksel Paulsen, Anders Odgaard Dept. of Orthopaedics, Copenhagen University Hospital Herlev-Gentofte; Dept. of Orthopaedics, Copenhagen University Hospital Bispebjerg; Dept. of Orthopaedics, Stavanger University Hospital Background: The Oxford Knee Score (OKS) a is a joint specific questionnaire, developed for the assessment of knee osteoarthritis patients. A Danish version was developed in 2007 but has not undergone formal validation. Purpose / Aim of Study: To translate and validate the Danish version using the classical test theory validation techniques. Materials and Methods: The original version was translated into Danish using a forward/backward protocol. Electronic patient-reported outcomes (PROMs) of 351 patients undergoing unicompartmental or total knee arthroplasty from Sept. 2016 to May 2018 were studied. Preoperative OKS was repeated after 2-7 days along with EQ-5D and a global knee anchor question. 1 year after surgery (± 2 mo.), PROMs were reassessed and patients were asked about sat- isfaction and willingness to repeat. Findings / Results: Mean OKS difference from test to retest was 0.29, SD 3.85, P=0.16). A Bland-Altman plot revealed no systematic deviation and 95 % of patients were less than 7 points from their first answer. None of the preop- erative measurements reached floor (0) or ceiling (48), while 7 postoperative measurements (2.0%) reached the ceiling. Internal consistency was appropri- ately high (Cronbachs alpha 0.83 preop., and 0.90 postop.). Construct validity measured by Spearman rank correlation between OKS and the anchor ques- tion was strong/moderate (rho=0.79/0.66, pre- and postop., respectively). Between OKS and EQ-5D dimensions mobility, activity and pain, correlations were fair/strong/strong preop. (r= -0.47/-0.72/-0.75) and moderate post- op. (r=-0.70/-0.67/-0.71) while, as expected, correlations to self-care and anxiety/depression were fair/poor (rho=-0.51/-0.23 and -0.36/-0.20). OKS changed from 20.6 (SD 6.2) preoperatively to 38.0 (6.4) postoperatively, with an individual improvement of 17.4 (7.2). Postoperative OKS was moderately correlated to satisfaction (rho=-0.71) and willingness to repeat (rho=0.62), as was OKS improvement to global improvement (rho=0.63). Conclusions: The Danish OKS was comparable to other versions regarding con- struct validity, reliability (test-retest and floor/ceiling effects) and responsive- ness. We suggest further validation based on item response theory, e.g. Rasch analysis, for different knee pathologies. 133.

RkJQdWJsaXNoZXIy NjEzNTY=