Validity of the EQ-5D-5L for routine assessment in residential aged care - results from a feasibility study.

Objective: To examine the psychometric performance of the EQ-5D-5L in residential aged care and to explore the measurement properties of the EQ-HWB(-S), ASCOT and QOL-ACC in a smaller subsample.

Methods: Residents were eligible if consent was provided for one of the following: (i) self-reported QoL; (ii) staff proxy reported QoL; or (iii) family proxy reported QoL. The EQ-5D-5L was administered first, with a second measure randomly assigned from either the ASCOT, QOL-ACC, or the EQ-HWB. Feasibility, floor-and ceiling effects, convergent validity, known-group validity and inter-rater reliability were examined.

Results: We gathered cross-sectional QoL data from 103 consenting participants through self-report (n=90), staff proxy-report (n=101) and family proxy-report (n=49). Missing values on the EQ-5D-5L ranged from 1% for staff to 14% for family members. No floor/ceiling effects were detected and the EQ-5D-5L showed good convergent validity with the QOL-ACC (r=0.62-0.82). Inter-rater reliability was good between residents and family (ICC=0.84) and excellent between residents and staff (ICC=0.96). Only staff-reported EQ-5D-5L scores could discriminate residents by cognitive impairment; but all raters could discriminate by dependency levels. Other measures showed minimal missing data, no floor/ceiling effects, and moderate to good inter-rater reliability. While most measures could discriminate by dependency level, some did not reach statistical significance.

Conclusions: The EQ-5D-5L showed promising performance and could serve as a potential candidate for implementation in residential aged care. While this study also provided additional evidence for the EQ-HWB(-S), QOL-ACC and ASCOT, further validation in larger samples are needed.