Hierarchical dependency models for multivariate survival data with censoring.

Abstract

A family of partial likelihood logistic models is proposed for clustered survival data that are reported in discrete time and that may be censored. The possible dependence of individual survival times within clusters is modeled, while distinct clusters are assumed to be independent. Two types of clusters are considered. First, all clusters have the same size and are identically distributed. Second, the clusters may vary in size. In both cases our asymptotic results apply to a large number of small independent clusters.