Valuing SF 6Dv2 Using a Discrete Choice Experiment in a General Population in Quebec, Canada

Abstract

Background: An updated version of the Short-Form 6-Dimension (SF-6D) Classification System has been developed. This new version (SF-6Dv2) with improved consistency and dimension descriptors is now requiring the development of new utility value sets. The aim of this study was to estimate an SF-6Dv2 value set from a general population in Quebec, Canada. Methods: A discrete choice experiment with time trade-off (DCE TTO) was conducted using two designs: binary choice sets (Design 1) and best-worst choice sets (Design 2). Design 1 consisted of binary choice sets along with an associated duration, and Design 2 included Design 1 and a third scenario describing "immediate death". Various logit model specifications were employed to estimate value sets separately for Design 1 and in combination with Design 2. Heterogeneity in preferences was assessed using a mixed logit model. Results: The survey was completed online by 1208 participants and 1153 were included for analysis. The model combining Design 1 and 2 data was considered as the best fitting model for estimating the final value set. It provided a value set with logical consistent coefficients and showed the lowest standard errors. Values ranged from -0.683 for the worst health state (555655) to 1 for full health (111111), with 13.01% of the values being negative. Preference values were the most affected by pain dimension and the least by vitality dimension. Preference heterogeneity existed for all the most severe levels of dimensions. Conclusion: This study provided the SF-6Dv2 value set for use in Quebec, Canada. The recommended value set is the anchored consistent model combining data from Design 1 and 2 using a conditional logit.