PIECE WISE KAPLAN-MEIER SURVIVAL FUNCTION FOR DEPENDENT DISEASES

Abstract

One of the assumptions of censoring in Kaplan-Meier Survival Function is independency i.e. disease under study must be independent but in real world, the situation is entirely different. In this case relative risk factors i.e. dependency plays an important role. To overcome this problem a new estimator called the Piece wise Kaplan-Meier Survival Function was developed and discussed in detail. Similarly, for measuring the variation, new variance estimator based on the Piece wise Kaplan-Meier Survival Function was introduced. Similarly, with the help of Greenwood Variance estimator and new variance estimator 95 Confidence Intervals are constructed. Results of the analysis showed that the Kaplan-Meier gives the overestimate results as compared to the new procedure. New procedures consider the competing risk factor and give more satisfactory results.