Inconsistencies in TTO and VAS values for EQ-5D health states

Abstract

Background: Most EQ-5D health states can be ordered logically. When people assign values to different health states, they may violate this logical order, which can be seen as inconsistencies.

Objective: The aim of this study was to assess the effect of inconsistently valued EQ-5D health states on mean visual analog scale (VAS) and time trade-off (TTO) values. The authors also examined the effect of removing respondents with high numbers of inconsistent states on the estimation of tariffs, which is used to interpolate values to all possible EQ-5D states from the direct valuation of a subset of states.

Method: Data from the Dutch EQ-5D valuation study were used. A representative sample of 309 Dutch adults valued 17 EQ-5D health states by VAS and TTO. A state was valued inconsistently when it had a higher value than at least 1 logically better state. Mean values of groups with various numbers of inconsistently valued states were compared.

Results: Of the respondents, 65% had inconsistencies for VAS and 89% for TTO. The mean VAS values of consistent respondents tend to be lower. For TTO, those with inconsistencies gave lower values. Removing data of respondents with the highest number of inconsistently valued states for VAS (13%) and TTO (9%) did not result in statistically significant different coefficients in the estimated tariff using all data.

Conclusion: The majority of respondents valued at least 1 state inconsistently. For both VAS and TTO, the presence of these inconsistencies did not affect the estimated tariffs.