A comparison of some two-sample tests with interval censored data

Interval censored data arise naturally in large scale panel studies where subjects can only be followed periodically and the event of interest can only be observed between some time intervals. Peto and Peto's rank invariant tests, including the logrank test and their generalized Wilcoxon test, of comparing two survival curves for right censored data can be applied to interval censored data. Simulation studies are conducted to assess and compare their performance, along with that of Mantel's test and Petroni and Wolfe's IWD test. It seems that the logrank test is most sensitive to detect late differences of two distributions while the generalized Wilcoxon test works best in distinguishing early differences. Mantel's test seems to lie between them. The methods are illustrated by reanalyzing the Breast Cosmesis Study dataset.