Cumulative Total Incidence for Estimating the Burden of Recurrent Events and Risk vs. Rate Concepts and Regression Models in Epidemiology

Abstract

In the first part of this thesis, a straightforward intuitive method for descriptive survival analysis, termed “cumulative total incidence”, is proposed to measure the total burden of recurrent events in a population by a given time. Using data from the Childhood Cancer Survivor Study, demonstrate the utility of this method contrasting this method to cumulative incidence. In the second part of this thesis, the concepts of risk and rate, and their relationship are discussed in the framework of survival analysis. Regression approaches for estimating the association between factors on event risk and event rate are discussed. Using data from the Childhood Cancer Survivor Study on two competing outcomes, we further demonstrate how competing-risk event affects the estimated association of covariates of interest with event risk and rate.