Preference Measurement for Multistate Health Profiles

点击翻译按钮获取中文摘要

DESCRIPTION (PROVIDED BY APPLICANT): This study aims at developing and testing a new method that can better capture preference for multistate health profiles. The motivation arose from the failure of the QALY (quality-adjusted life-year) model in capturing preferences for multistate health profiles. As past literature shows, the conventional QALY model violates one of its required assumptions, the additive independence. These findings imply that preferences between individual health states are not independent of each other. This study proposes a novel approach designed to measure preferences for multistate health profiles by looking at two consecutive health states at a time. It hypothesizes that the evaluation of a future health state is dependent or "conditioned" on the current health state. Characteristics of the current health state which are suspected to affect the "conditional preference scores" for future health state include duration of the current health state, direction of change, and the amplitude of change between the current and future health states. A full factorial design (three factors with two levels each) with three replications at three different levels of the future health state will be used to explore both main effects and interactions. Furthermore, this study will test whether the proposed technique, which assesses "conditional preference scores" for multistate health profiles, can better predict preference scores for an entire health profile than the unconditional health state assessments. Duration-weighted conditional preference scores and duration-weighted unconditional preference scores will be compared to directly elicited holistic scores for 10 hypothetical health profiles, each composed of carefully selected combinations of four health states. Two elicitation techniques will be employed for all tasks, visual analog scale (VAS) and time-tradeoff (TTO). A power analysis revealed that a sample size of 70 subjects for each technique will give at least 80% probability of detecting an effect size as small as 0.05. Subjects will be recruited from the student population at Georgia Tech. Human subjects approval has been obtained.