|
Abstract:
|
The current bedrock of precision prevention is based on selecting high-risk individuals, based on models of individualized disease risk, for screening or other preventive services. However, those at highest risk do not necessarily have highest benefit from preventive services, if many of those at highest risk tend to be elderly or have comorbidities that substantially reduce the life-years that they would gain from preventive services. We demonstrate how individualized gain in life-expectancy from preventive services can be estimated when the treatment effects of prevention services are heterogenous across the population and can vary over the course of screening. We develop a calculator for individualized life-years gained from undergoing lung cancer screening by combining component models for treatment effect, disease mortality risk, and competing mortality risks developed in separate data sources. We illustrate how use of this calculator to select US ever-smokers for lung cancer screening can identify individuals with fewer co-morbidities, better performance status, and favorable benefit-harm ratios compared to selections using an individualized lung cancer mortality risk model.
|
