Comment on ‘Response –scale heterogeneity in the EQ-5D: Can we use vignettes to address response-scale heterogeneity in the EQ-5D? Not if but how’

Knott et al. (2016) focus on the issue of differential item functioning (DIF) in the EQ-5D. To address this issue, they recommend employing anchoring vignettes as ‘perhaps the most viable method worth pursuing in the immediate future, in terms of both time and cost…particularly if adjustments can be formulated using vignette responses from external samples’. I believe the Knott et al. idea of using vignettes for addressing DIF in the EQ-5D is very good. Therefore, in this commentary I do not question whether it is sensible to try using vignettes but instead I focus on how to do it. Indeed, I am somewhat sceptical about some of the conclusions put forward by the authors. Moreover, in the light of recent iterature, I will provide some methodological advice on how to implement anchoring vignettes at the survey stage.