Modelling the EuroQol data: a comparison of discrete choice conjoint and conditional preference modelling

Abstract

This article compares two measurement strategies for measuring EuroQol health state preferences: (a) conditional preference modelling, implemented using rating scale and standard gamble scaling methods and (b) discrete choice conjoint modelling. The nature of the model form of the EuroQol health status preference function and the predictive ability of each measurement strategy formed the basis of the comparison. Data were collected via personal interviews with 140 US patients, 139 of whom provided usable responses. Both strategies supported a multiplicative model form as representative of the EuroQol health status preference function and were acceptable in terms of predictive ability. The agreement of the two measurement strategies on the nature of the model form provides evidence of the convergent validity of the multiplicative nature of the EuroQol health status preference function in this patient population. Since both strategies were found to be acceptable in terms of predictive ability, further research comparing preference scores and measuring respondent evaluations of the methodologies is necessary to illustrate the relative strengths and weaknesses of different health state preference measurement methodologies.