Time-Specific Differences in Stated Preferences for Health in the United States

Background: Changes over time in health state values from a societal perspective may be an important reason to consider updating societal value sets for preference-based measures of health. Objective: The aim was to examine whether stated health preferences are different between 2002 and 2017, controlling for demographic changes in the United States. Methods: Data from 2002 and 2017 US EQ-5D-3L valuation studies were combined. The primary analysis compared valuations of better-than-dead (BTD) states only, as both studies used the same time trade-off (TTO) method for these states. For worse-than-dead (WTD) states, the 2017 study used the lead-time TTO and the 2002 study used the conventional TTO, which necessitated transformation. Regression models were fitted to BTD values to estimate time-specific differences, adjusting for respondent characteristics. Secondary analyses examined models that fitted WTD values (using linear and nonlinear transformations of the 2002 data) and all values. Results: The adjusted BTD-only model showed mean values were higher for 2017 compared with 2002 (beta(Y2017)=0.05, P<0.001). WTD-only models showed negative changes over time but that were dependent on the transformation method (linear beta(Y2017)=-0.72; nonlinear beta(Y2017)=-0.35; both P<0.001). Using all values, 2017 mean valuations were lower using a linear transformation (beta(Y2017)=-0.11; P<0.001) but did not differ with the nonlinear transformation. Conclusions: Individuals in 2017 are generally less willing to trade quantity for quality of life compared with 2002. This study provides evidence of time-specific differences in a society's preferences, suggesting that the era in which values were elicited may be an important reason to consider updating societal value sets.