Potential Biases in Post-Stroke Health Utility Estimates by Modified Rankin Scale Scores

I read with great interest the recent article by Pinckaers et al1 published in Value in Health, I read with great interest the recent article by Pinckaers et al1 published in Value in Health. This study is an excellent contribution to the field of health economics, addressing one of the most pressing unmet needs: the estimation of costs and health utility values per modified Rankin Scale (mRS) score, which are essential for health economic evaluations related to stroke. Although the research directly contributes to the context of health economic evaluations in the Netherlands, the robustness of its cost and health utility estimates will also significantly contribute to subsequent evidence. Although the study is valuable, I would like to point out 2 potential biases that are not mentioned in the article. The first bias is the proxy response bias. In this study, EQ-5D-5L data from patients with a diagnosis of stroke were collected through telephone interviews. The EQ-5D-5L was administered not only to the patients themselves but also to their representatives. In principle, the EQ-5D-5L should be administered to the patients themselves, and administration to proxies raises concerns about the proxy response bias, which may overestimate the health utility status.2 I understand that patients diagnosed with stroke may not be able to respond because of neurological deficits. Considering the potential proxy response bias, estimating and presenting health utility values separately for each respondent (patient/representative), in addition to pooled estimates, may help improve future research designs.