Approved Research

Valuing Medical Innovations

Lay summary

In this project, we seek to estimate the value of potential medical advances in terms of quality-adjusted life years (QALYs). Say that medical researchers eliminate a specific cause of death. How many QALYs are saved?

At first, one may simply subtract the expected age of death with that disease from the population average life expectancy. However, causes of deaths are "competing risks" - if someone does not die from one cause of death, he will die from another. These competing risks could be strongly correlated. For example, a person who lives through one potentially fatal disease may be more likely to soon die from another. Therefore, an estimate based upon average life expectancy could be severely biased. Using biased estimates could result in the inefficient allocation of limited resources for medical research.

To properly estimate QALYs saved, then, we require a model of competing risks. This model can only be estimated using valid instruments - otherwise exogenous factors that only correlate with the cause of death and not with potential confounders.

In this project, we will use genetic instruments (Mendellian randomization) in a competing risks framework to estimate the value of potential medical advances. Specifically, we will use de-identified data on genetic predisposition for certain diseases, causes of death, and lifespan from the UK Biobank to estimate the value of curing a single disease.

We expect this project will take 36 months. Ultimately, we hope that this project will provide better estimates of the value of medical innovation. These results can be used to more efficiently allocate funds to medical research. Further, these results can be used to better understand the disease targets most likely to reduce demographic health disparities.