Inter-rater reliability and responsiveness of patients to proxy responses of the EQ-5D-5L for older patients with musculoskeletal disorder in rehabilitation wards

Health-Related Quality of Life (HRQOL) is an essential domain for evaluating the true success of rehabilitation. The EuroQol 5-dimension 5-level (EQ-5D-5L) is a brief and widely used HRQOL measure. However, evidence is currently lacking regarding its responsiveness following rehabilitation for musculoskeletal disorder, particularly in Japan. Assessment of HRQOL is challenging for older patients with musculoskeletal disorders, particularly those with cognitive impairments or frailty, who may have difficulty self-reporting. Therefore, evaluating the reliability of proxy assessments is crucial in this population. Furthermore, the agreement between self-ratings and proxy-ratings is generally reported to be low. In contrast, a previous Japanese study reported high agreement when an Occupational Therapist (OT) acted as a proxy for stroke patients. This study therefore aimed to clarify the change (responsiveness) in HRQOL and investigate the inter-rater reliability between patient self-responses and OT proxy responses using the EQ-5D-5L in patients with musculoskeletal disorders in a Japanese convalescent rehabilitation ward. This was a multicenter, collaborative longitudinal study conducted across five hospitals in Japan. We enrolled 77 participants with musculoskeletal disorders who were capable of self-reporting. The mean age was 77.7 years. Evaluations were conducted at initial assessment (within one week of admission) and re-evaluation (one month later). HRQOL was measured using the EQ-5D-5L. Proxies were clinical OT who were blinded to patient responses during concurrent assessment. Responsiveness was assessed using the Standardized Response Mean (SRM). Inter-rater reliability for the QOL score was assessed using the Intraclass Correlation Coefficient (ICC), and the sub-items were assessed using weighted Kappa. Significant improvement was observed in all outcome measures post-rehabilitation (p < 0.01). The SRM for the QOL score indicated a large effect size for both self-reported (0.90) and proxy-reported (0.95) responses. Inter-rater reliability for the QOL score showed substantial agreement, with the ICC at re-evaluation was 0.72 (95%CI: 0.59; 0.81), which met the threshold for “good measurement properties” (≥ 0.70) as defined by the COSMIN criteria. The kappa coefficients for the five sub-items ranged from 0.31 (95%CI: 0.17; 0.45) to 0.56 (95%CI: 0.40; 0.72) (fair to moderate agreement). The EQ-5D-5L demonstrates high responsiveness and good inter-rater reliability when administered by an OT proxy at the one-month re-evaluation in this patient population. These findings support the use of OT proxy-responses to reliably estimate the HRQOL of patients with musculoskeletal disorders in the convalescent setting, particularly when self-reporting is difficult, providing evidence that proxy assessments can reliably complement patient reports in musculoskeletal rehabilitation settings.