Modified Kaplan-Meier Estimator Based on Competing Risks for Heavy Censoring Data

Most follow-up studies are conducted to determine the survival rates of subjects affected by a specific risk. These subjects are also exposed to other risks. Every subject in a medical follow-up is exposed not only to the risk of dying, but also to the risk of being censored. In case of heavy censoring, the Kaplan-Meier estimates are biased and overestimate the survival distribution. A new methodology based on competing risks is proposed to estimate the survival function by using net and crude probabilities. These estimates reduce the bias and overestimation of the survival distribution noted in Kaplan-Meier estimators. In this study, the method of modified Kaplan-Meier (MKM) is compared with the Kaplan-Meier (KM), Huang’s method and also the two other methods namely Weighted Kaplan-Meier (WKM) and Modified Weighted Kaplan-Meier (MWKM). Either of the weighted methods depends heavily on the event times and censoring distributions. Due to this fact, the weighted methods can have misleading results when the censoring patterns are different in the individual samples. The results showed that the MKM estimator considers not only the problem of heavy censoring but also the problem of weighted methods and competing risks in complicated data. In this study “Stanford Heart Transplant Data” was used to investigate the effectiveness of the proposed methods.