Quality-of-Life–Adjusted Hazard of Death: A Formulation of the Quality-Adjusted Life-Years Model of Use in Benefit-Risk Assessment

Abstract Background Although the quality-adjusted life-years (QALY) model is standard in health technology assessment, quantitative methods are less frequent but increasingly used for benefit-risk assessment (BRA) at earlier stages of drug development. A frequent challenge when implementing metrics for BRA is to weigh the importance of effects on a chronic condition against the risk of severe events during the trial. The lifetime component of the QALY model has a counterpart in the BRA context, namely, the risk of dying during the study. Methods A new concept is presented, the hazard of death function that a subject is willing to accept instead of the baseline hazard to improve his or her chronic health status, which we have called the quality-of-life–adjusted hazard of death. Results It has been proven that if assumptions of the linear QALY model hold, the excess mortality rate tolerated by a subject for a chronic health improvement is inversely proportional to the mean residual life. Conclusions This result leads to a new representation of the linear QALY model in terms of hazard rate functions and allows utilities obtained by using standard methods involving trade-offs of life duration to be translated into thresholds of tolerated mortality risk during a short period of time, thereby avoiding direct trade-offs using small probabilities of events during the study, which is known to lead to bias and variability.