Probability weighting and utility curvature in QALY-based decision making (Q1304532)

Consider a state of health \(x\) such that the \(U\) (1 year in state \(x\)) \(=1\) and a state of health \(z\) such that \(U\) (1 year in state \(z\)) \(=0\); in a standard gamble, one considers the risky situation where one spends \(t\) years in state \(x\) with probability \(p\) and \(t\) years in state \(z\) with probability \(1-p\) with the certainty of being in state of health \(y\) for \(t\) years and one may try to solve the following: \[ \begin{multlined} U\;(t\text{ years in health state }y)=\\ p'U\;(t\text{ years in health state }x)+(1- p')U\;(t\text{ years in state health }z)\end{multlined} \] now, \(p'\) denotes the quality weight of the health profile \(y\). This is the standard method of determining the Quality-Adjusted Life Years (QALY). The problem is that persons may not behave perfectly in accordance with expected utility theory and consequently, the standard gamble may not reveal accurate probabilities. The paper investigates whether rank dependent weighting can improve the descriptive validity of the standard gamble in QALY-based decision making.