Life Expectancy Comparison between a Study Cohort and a Reference Population

Abstract

Questions of life expectancy are highly relevant in clinical practice, in particular if one is interested in the life expectancy loss due to a certain disease in comparison with the general population. Therefore, we propose a methodology which enables the researcher to compare more easily life expectancies between a study cohort and a reference population. Firstly, we take the formula commonly used in official statistics and adapt it to a survival analytic setting, thus establishing a link between official statistics and survival analysis. Further, we discuss two commonly encountered sources of potentially severe bias, and how to remedy them. We hope that the proposed approach facilitates the use of life tables, particularly in clinical studies with mortality as primary endpoint.