Crosswalk algorithms for the conversion of mean EQ-5D indices calculated with different value sets

Abstract

Aims: The EQ-5D questionnaire of EuroQol is an important tool, not only for researchers, but also for quality registers. Until 2013, Sweden did not have a country-specific value set to convert the EQ-5D health states into a single index. Commonly, the UK time trade-off (UK TTO) value set has been used. The study reported here aimed to establish an easy to use tool for the bidirectional crosswalk of the mean EQ-5D values based on two different value sets: the UK TTO and the Swedish TTO value sets. Methods: Using an artificial data set encompassing all possible EQ-5D outcomes, we calculated the EQ-5D index using both the UK TTO and Swedish TTO value sets. Thereafter we modelled the relationship between the two indices using least-squares regression and major axis regression. A series of simulations was run to assess the feasibility of the obtained crosswalk algorithms. Results: Major axis regression was superior to ordinary least-squares regression. Converting the mean EQ-5D values from the UK TTO scale to the Swedish TTO scale was more accurate than the inverse conversion. Values close to the floor and ceiling of the EQ-5D index were more challenging to crosswalk. Conclusions: We established algorithms based on major axis regression to crosswalk EQ-5D values based on the UK TTO and the Swedish TTO value sets. The presented algorithm may facilitate comparisons of EQ-5D values when only mean values are available. The existence of a crosswalk algorithm will ease the transition from the UK TTO to the Swedish TTO value set.