Using a discrete choice experiment to elicit time trade-off and willingness-to-pay amounts for influenza health-related quality of life at different ages

Abstract

Background: Recent research suggests that values for health-related quality of life may vary with the age of the patient. Traditional health state valuation questions and discrete choice experiments are two approaches that could be used to value health. Objective: To measure whether public values for health vary with the age of the affected individual. Methods: A discrete choice experiment was administered via the Internet in December 2007 to measure preferences for different attributes of influenza-related health-related quality of life: age of hypothetical affected individual (range 1-85 years), length of episode (days of illness), severity of illness (workdays lost) and time trade-off or willingness-to-pay amounts. Each respondent answered identical choice questions for a hypothetical family member and for himself/herself. Data on sociodemographic characteristics and influenza illness experience were also collected. Respondents were US adults randomly sampled from an Internet survey panel (n = 1,012). The relative value of attributes was estimated using generalized estimating equations and controlling for sociodemographic characteristics and illness experience. Marginal time traded and marginal willingness to pay using discrete choice and traditional time trade-off or willingness-to-pay questions were compared. Results: Respondents preferred shorter influenza episodes but did not significantly prefer fewer workdays lost if episode length was held constant. Respondents were more likely to choose to avert uncomplicated illness in children and less likely to choose to avert uncomplicated illness in working-age adults. Marginal time trade-off and willingness-to-pay amounts elicited using discrete choice questions were larger than those elicited using direct valuation questions. Conclusions: Approaches that assume values for health-related quality of life do not vary with the age of a patient may bias economic analyses that use these values. If patient age could affect valuations, then age should be included in the valuation exercise. Additional research should evaluate the effect of patient age on values for other conditions. © 2013 Springer International Publishing Switzerland.