Beyond Kaplan-Meier: A comprehensive R Package for Interval-Censored Survival Analysis using Turnbull’s Approach

Abstract

Interval-censored data frequently arise in survival analysis when the exact time of an event is unknown but is known to occur within a specific time interval. Traditional methods like the Kaplan-Meier estimator are inadequate for such data, necessitating specialized approaches. This paper presents an R library designed to handle interval-censored data, emphasizing the use of Turnbull’s estimator for nonparametric survival estimation. The package offers flexible functionalities, including the calculation of survival estimates, the generation of both static and interactive plots, and the construction of bootstrap-based confidence bands. Additionally, the library provides users with detailed outputs such as Turnbull intervals and their corresponding weights, which are instrumental in understanding the survival distribution and serve as an analogue to Kaplan-Meier weights in right-censored contexts. These weights enable the extension of survival analysis methods to more complex models, including multi-state frameworks. The practical utility of the library is demonstrated using real-world datasets, highlighting its potential to support advanced survival analysis and foster the development of new estimators beyond traditional survival probabilities.