On Comparison of Survival Curves with Interval Censored Data

Abstract

Survival comparison is one of the main goals and interesting problems in most survival studies such as clinical studies. In this paper, we compare through a Monte Carlo study, several tests for comparison of survival functions for interval censored failure time data. In particular, three nonparametric generalized log-rank tests, a parametric score test and an imputation-based test have been considered. It is observed that the parametric score test and the imputation test outperform the nonparametric generalized log-rank tests in most cases. Finally, a real dataset is studied for illustrative purposes.

Key words and phrases. Exponential distribution, Generalized log-rank test, Interval censoring, Monte Carlo simulation, Survival comparison.

1991 Mathematics Subject Classification. 62N01, 92B15