Análise de inconsistências das preferências sociais por saúde no Brasil através da utilização do instrumento EQ-5D

Abstract

The elicitation of societal preferences through EQ-5D instrument has been implemented in several countries, in most cases based on the original MVH protocol developed by the EuroQol Group. This paper aims to compare two studies conducted in Brazil based on similar protocols that difered only in the study design. The first survey was conducted in the Minas Gerais state, and was based on a fixed balanced group of health states. The second study occured in 3 different cities: Porto Alegre, Recife and Rio de Janeiro and collected social preferences for all of the 243 health states from EQ-5D. The analysis will be based in the presence of inconsistent health state valuations. This phenomena can arise both from the description and nature of health states as well as the individual characteristics that may affect the valuations. Both studies ocurred between 2011 (Minas Gerais) and 2012 (Rio de Janeiro, Recife and Porto Alegre). The interviewers surveyed literate individuals between the ages of 18 and 64, living in urban areas. The logical inconsistencies were estimated through mathematical algoritms and were accounted in individual levels and in terms of distance of pairwise health state valuations. Three elicitation methods were analyzed: ranking, visual analog scale and Time Trade Off. The logical inconsistencies concentrate in health states that are close to each other, which can indicate that individuals fail to identify the diferences in the health state descriptions. Higher inconsistency levels were found in the second survey, which might indicate that the balancing of health states is a desirable feature in terms of consistente responses. Another result showed that the selective exclusion of individuals based on the distance between pairwise inconsistent valuations can improve the goodness of fit of the estimated social tariffs models, both on visual analog scale and TTO. These results could not be reproduced in the unbalanced study design. These paper findings might suggest that the balanced fixed block study design is a desirable feature when estimating societal preferences for health state, which might also reduce logistical costs of conducting the field research.