The analysis of case cohort design in the presence of competing risks with application to estimate the risk of delayed cardiac toxicity among Hodgkin Lymphoma survivors

Abstract

The case-cohort design is an economical solution to studying the association between an exposure and a rare disease. When the disease of interest has a delayed occurrence, then other types of event may preclude observation of the disease of interest giving rise to a competing risk situation. In this paper, we introduce a modification of the pseudolikelihood proposed by Prentice (Biometrika 1986; 73:1-11) for the analysis of case-cohort design, to accommodate the existence of competing risks. The modification is based on the Fine and Gray (J. Amer. Statist. Assoc. 1999; 94:496-509) approach to enable the modeling of the hazard of subdistribution. We show through simulations that the estimate that maximizes this modified pseudolikelihood is almost unbiased. The predictive probabilities based on the model are close to the theoretical probabilities. The variance for the estimates can be calculated using the jackknife approach. An application of this method on the analysis of late cardiac morbidity among Hodgkin Lymphoma survivors is presented.