Addititive hazards regression with missing censoring information

In this article we study estimation in the additive hazards regression model with missing censoring indicators. We develop simple procedures to obtain consistent and efficient estimators for the regression parameters as well as the cumulative baseline hazard function, and derive their asymptotic properties. The
estimator of the regression parameters is shown to be asymptotically normally distributed, while the estimator of the cumulative baseline hazard function converges to a Gaussian process. We address both the situations where the mechanism for
missingness of the censoring indicators is independent of any other factors, and those in which the missingness may depend on the covariates. Monte Carlo studies are also conducted to evaluate the performance of the estimators.