Statistical methods in epidemiology. IX. Survival (failure-time) models

Abstract

Purpose. This article introduces readers to survival (failure-time) models, with a focus on Kaplan–Meier curves, Cox regression and sample size estimation.

Methods. An example is used to show readers how to calculate a Kaplan–Meier curve from first principles.

Results. What makes survival data unique is censoring. Readers should understand censoring before undertaking an analysis of survival data.

Conclusion. The Cox model continues to set the standard for survival models, and will continue well into the future.

Keywords:

• Kaplan–Meier survival curves
• Cox regression
• sample size

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.