Dependency and health utilities in stroke: Data to inform cost-effectiveness analyses.

INTRODUCTION

Health utilities (HU) assign preference weights to specific health states and are required for cost-effectiveness analyses. Existing HU for stroke inadequately reflect the spectrum of post-stroke disability. Using international stroke trial data, we calculated HU stratified by disability to improve precision in future cost-effectiveness analyses.

MATERIALS AND METHODS

We used European Quality of Life Score (EQ-5D-3L) data from the Virtual International Stroke Trials Archive (VISTA) to calculate HU, stratified by modified Rankin Scale scores (mRS) at 3 months. We applied published value sets to generate HU, and validated these using ordinary least squares regression, adjusting for age and baseline National Institutes of Health Stroke Scale (NIHSS) scores.

RESULTS

We included 3858 patients with acute ischemic stroke in our analysis (mean age: 67.5 ± 12.5, baseline NIHSS: 12 ± 5). We derived HU using value sets from 13 countries and observed significant international variation in HU distributions (Wilcoxon signed-rank test  < 0.0001, compared with UK values). For mRS = 0, mean HU ranged from 0.88 to 0.95; for mRS = 5, mean HU ranged from -0.48 to 0.22. OLS regression generated comparable HU (for mRS = 0, HU ranged from 0.9 to 0.95; for mRS = 5, HU ranged from -0.33 to 0.15). Patients' mRS scores at 3 months accounted for 65-71% of variation in the generated HU.

CONCLUSION

We have generated HU stratified by dependency level, using a common trial endpoint, and describing expected variability when applying diverse value sets to an international population. These will improve future cost-effectiveness analyses. However, care should be taken to select appropriate value sets.