Self-assessed performance-based function test versus patient-reported outcome measures for knee and hip osteoarthritis

Abstract

Background Physical function constitutes a key component of outcome assessment for almost all osteoarthritis interventions. The aim was to compare physical function measured using a self-assessed performance-based test versus self-reported function using questionnaires among individuals with knee or hip osteoarthritis (OA) participating in a digital exercise and education therapy. Methods We analysed data from individuals aged 40 + years participating in the digital program. We extracted data on the self-assessed 30-second chair stand test (30s CST) and the function subscales of Knee injury/Hip disability and Osteoarthritis Outcome Score 12 (KOOS-12/HOOS-12) at enrolment and 3- (n = 10884) and 12-month (n = 3554) follow-ups. Participants completed Numeric Rating Scale (NRS) pain, EQ-5D-5L, and an external anchor: global rating of change scale. Correlations were assessed using the Spearman correlation coefficient, responsiveness using standardized response mean (SRM) and receiver operating characteristic (ROC) curves, and agreement using weighted percent of agreement and weighted Gwet’s agreement coefficient. Results Correlations were weak between the 30s CST and KOOS-12/HOOS-12 function (r < 0.35 for raw and r < 0.20 for change scores). Correlations with NRS pain and EQ-5D-5L were stronger for the KOOS-12/HOOS-12 function subscale than for 30s CST. Greater internal (SRM > 1 vs. SRM < 0.5) and lower external responsiveness were observed for the 30s CST versus the KOOS-12/HOOS-12 function, even though external responsiveness was generally inadequate for both (the area under the ROC curves < 0.7). The direction of change was similar for the two function measures for about 70% of subjects with moderate agreement between them (weighted Gwet’s agreement coefficient range 0.45 to 0.50). Conclusion Weak correlations and moderate agreements between function measured using performance-based test and self-reported using KOOS-12/HOOS-12 in people with knee or hip OA suggest that they may capture different aspects of functional abilities in this population.