Survival analysis

Origins of survival analysis lie in actuarial science of lifetimes and death rates of people. As a discipline, survival analysis has grown from many applications in biomedical fields. This chapter discusses life tables, Kaplan–Meier survival curves, and Cox’s proportional hazards regression model, which take into account censoring of the lifetimes. The Kaplan–Meier estimator gives a nonparametric estimate of the survival distribution. The chapter illustrates standard error calculation for Kaplan–Meier survival curves. Logrank test provides a test of the overall difference between two survival distributions by cumulating the differences between the survival distributions at the observed death times. Cox’s proportional hazards model postulates how the hazard rate depends on the covariates. It has a nonparametric component and a parametric component. Hence, the model is also called a semiparametric model.