A novel estimator of survival without assumption of censored survival time

Abstract The Kaplan-Meier estimator is commonly used to analyze time-to-event data, but it may over estimate the survival function when censored events occur. Several methods have been developed to adjust the traditional Kaplan-Meier estimator. Most of these adjusted methods are based on their assumptions for survival time of censored events. We here propose a novel estimator of survival without assumption of censored survival time, and it gives sensible estimates of survival probabilities in various censoring conditions.