Empirical transition matrix of multistate models: the etm package

When dealing with complex event history data in which individuals may experience more than one single event type, multistate models provide a relevant modelling framework. Well known examples include the competing risks model in which subjects may die from several possible causes, and the illness-death model that permits to study the impact of an intermediate event on a terminal event. Quantities of interest in this framework are the transition probabilities that can be estimated by the empirical transition matrix, that is also referred to as the Aalen-Johansen estimator. In this talk we present the R-package etm that computes and displays these transition probabilities. etm also features a Greenwood-type estimator of the covariance matrix, which has recently been found to be the preferable estimator in the competing risks situation. The use of the package is illustrated through a prominent example in bone marrow transplant for leukaemi