Mean survival by ordered fractions of population with censored data

TL;DR

This paper introduces a new method for estimating mean survival times across ordered population fractions with censored data, offering detailed insights and improved detection of differences in survival across groups.

Contribution

It presents a novel approach dividing populations into survival-ordered fractions to estimate mean survival, enhancing analysis of censored data.

Findings

Provides detailed survival distribution insights

Detects differences in mean survival across groups

Outperforms existing methods in certain applications

Abstract

We propose a novel approach for estimating mean survival time in the presence of censored data, in which we divide the population under study into survival-ordered fractions defined by a set of proportions, and compute the mean survival time for each fraction separately. Our approach provides a detailed picture of the distribution of the time variable while preserving the appealing interpretation of the mean. Our measure proves to be of great use in applications, particularly those where we are able to detect differences in mean survival across groups for certain fractions of the population that would have been overlooked using other available methods.