A New Method for Regression Analysis of Case–Cohort Interval‐Censored Failure Time Data

ABSTRACT The case–cohort design is widely used as a way of reducing costs when the disease incidence is low and covariates are expensive to measure. In this paper, we consider regression analysis of interval‐censored failure time data arising from the case–cohort design under a class of semiparametric transformation models. Although some methods have been proposed for the situation, they only make use of the information on the subjects in the subcohort and thus can be less efficient. To address this, we propose a class of linear inverse probability weighted estimators that make use of the information on all subjects. Furthermore, a method is provided to determine the optimal weights, and the proposed estimator is shown to be consistent and asymptotically normal. A simulation study is performed and indicates that the proposed method works well for practical situations, and it is illustrated by using the data arising from an Alzheimer's disease study.