Survival Analysis

This chapter explains the basics of survival analysis, a statistical method for analysing time-to-event data. We begin with a brief overview, providing examples from different disciplines, including social sciences, where it is known as duration analysis. We then explain censoring, a concept that is central to survival analysis, and it occurs when the exact time of an event is unknown. We explain right censoring (event occurs after observation), left censoring (event occurs before observation) and interval censoring (event occurs within an interval) in detail. We then introduce the Kaplan-Meier estimator for undertaking survival analysis. Using R's built-in lung dataset, we demonstrate how to generate and interpret survival plots. We then address the limitations of the Kaplan-Meier method and show how the Cox proportional hazards model overcomes them by allowing for multiple covariates. The chapter concludes with a detailed, step-by-step explanation of the Cox model and interpretation of its output.