Aims: To date, evidence to support the construct validity of the EQ-5D-5L has primarily focused on cross-sectional data. The aims of this study were to examine the responsiveness of EQ-5D-5L in patients with stroke and to compare it with responsiveness of EQ-5D-3L and visual analogue scale (EQ VAS).

Methods: We performed an observational longitudinal cohort study of patients with stroke. At 1 week and 4 months post-stroke, patients were assessed with modified Rankin Scale (mRS) and Barthel Index (BI) and were administered the EQ-5D-5L and EQ-5D-3L, including the EQ VAS. The EQ-5D-5L index scores were derived using the crosswalk methodology developed by the EuroQol Group. We classified patients according to two external criteria, based on mRS or BI, into 3 categories: 'improvement,' 'stable' or 'deterioration'. We assessed the responsiveness of each measure in each patient subgroup using: effect size (ES), standardized response mean (SRM), F-statistic, relative efficiency and area under the receiver operating characteristic curve.

Results: A total of 112 patients (52% females; mean age 70.6 years; 93% ischemic stroke) completed all the instruments at both occasions. In subjects with clinical improvement, EQ-5D-5L was consistently responsive, showing moderate ES (0.51-0.71) and moderate to large SRM (0.69-0.86). In general, EQ-5D-3L index appeared to be more responsive (ES 0.63-0.82; SRM 0.77-1.06) and EQ VAS less responsive (ES 0.51-0.65; SRM 0.59-0.69) than EQ-5D-5L index.

Conclusions: The EQ-5D-5L index, based on the crosswalk value set, seems to be appropriately responsive in patients with stroke, 4 months after disease onset. As far as EQ-5D-5L index is scored according to crosswalk approach, the EQ-5D-3L index appears to be more responsive in stroke population.