US valuation of the EQ-5D health states: development and testing of the D1 valuation model.

摘要: Purpose: The EQ-5D is a brief, multiattribute, preference-based health status measure. This article describes the development of statistical model for generating US population-based preference weights. Methods: A multistage probability sample was selected from adult civilian noninstitutional population. Respondents valued 13 243 states using time trade-off (TTO) method. Data 12 were used in econometric modeling. TTO valuations linearly transformed to lie on interval 1, 1. Methods investigated account interaction effects caused by having problems multiple dimensions. Several alternative specifications (eg, pooled least squares, random effects) also considered. modified split-sample approach evaluate predictive accuracy models. All analyses took into clustering and disproportionate selection probabilities inherent our sampling design. Results: Our D1 included ordinal terms capture effect departures perfect as well effects. specification yielded good fit observed data, with an overall R 2 0.38, mean absolute error 0.025, 7 prediction errors exceeding 0.05 magnitude. Conclusions: best predicts values states. resulting weight estimates represent significant enhancement EQ-5D’s utility assessment economic analysis US.