A two-sample test with interval censored data via multiple imputation

Interval censored data arise naturally in large scale panel studies where s ubjects can only be followed periodically and the event of interest can onl y be recorded as having occurred between two examination times. In this pap er we consider the problem of comparing two interval-censored samples. We p ropose to impute exact failure times from interval-censored observations to obtain right censored data, then apply existing techniques, such as Harrin gton and Fleming's G(p) tests to imputed right censored data. To appropriat ely account for variability, a multiple imputation algorithm based on the a pproximate Bayesian bootstrap (ABB) is discussed. Through simulation studie s we find that it performs well. The advantage of our proposal is its simpl icity to implement and adaptability to incorporate many existing two-sample comparison techniques for right censored data. The method is illustrated b y reanalysing the Breast Cosmesis Study data set. Copyright (C) 2000 John W iley & Sons, Ltd.