Comparison between EQ-5D-5L and PROMIS-10 to evaluate health-related quality of life 3 months after stroke: a cross-sectional multicenter study

Abstract

BACKGROUND: Although the use of patient-reported outcome measures to assess Health-Related Quality of Life (HRQoL) has been advocated, it is still open to debate which patient-reported outcome measure should be preferred to evaluate HRQoL after stroke. AIM: To compare the measurement properties (including concurrent validity and discriminant ability) between the 5-dimensional 5-level EuroQol (EQ-5D-5L) and the Patient-Reported Outcomes Measurement Information System 10-Question Global Health Short Form (PROMIS-10) to evaluate HRQoL 3 months after stroke. DESIGN: Cross-sectional study. SETTING: Neurology outpatient clinics in 6 Dutch hospitals. POPULATION: The participants 360 consecutive individuals with stroke. Their median age was 71 years, 143 (39.7%) were female and 335 (93.0%) had suffered an ischemic stroke. METHODS: The EQ-5D-5L, PROMIS-10, modified Rankin Scale and two items on experienced decrease in health and activities post-stroke were administered by a stroke nurse or nurse practitioner through a telephone interview 3 months after stroke. The internal consistency, distribution, floor/ceiling effects, inter-correlations and discriminant ability (using the modified Rankin Scale and experienced decrease in health and in activities post-stroke as external anchors) were calculated for both the EQ-5D-5L and PROMIS-10. RESULTS: Ninety-six percent of the participants were living at home and 50.9% experienced minimal or no disabilities (modified Rankin Scale 0-1) 3 months after stroke. A ceiling effect and a non-normal left skewed distribution were observed in the EQ-5D-5L. The PROMIS-10 showed higher internal consistency (α=0.90) compared to the EQ-5D-5L (α=0.75). Both the EQ-5D-5L and the PROMIS-10 were strongly correlated with the modified Rankin Scale (r=0.62 and 0.60 respectively). The PROMIS-10 showed better discriminant ability in less affected individuals with stroke, whereas the EQ-5D-5L showed slightly better discriminant ability in more affected individuals with stroke. CONCLUSIONS: Both EQ-5D-5L and PROMIS-10 prove to be useful instruments to evaluate HRQoL in patients who are living at home 3 months after stroke. CLINICAL REHABILITATION IMPACT: The clinical rehabilitation impact depended on the setting and underlying goal which patient-reported outcome measure is preferred to evaluate HRQoL 3 months after stroke. The PROMIS-10 should be preferred to detect differences in less affected stroke patients, whereas the EQ-5D-5L provides slightly more information in more affected stroke patients.