Abstract:
|
It has been well recognized that in the presence of competing risks, the effects of a variable on the time to the occurrence of a particular type of failure, say type 1 failure, is not completely characterized by the type 1 cause-specific hazard (CSH1) alone. Additional quantities such as the cumulative incidence function for type 1 failure, the overall hazard (OH) due to any cause, and the cause specific hazard due other causes other than 1, need to be considered jointly. In this talk, we consider sample size determination for jointly testing CSH1 and OH based on the joint tests recently developed by Li and Yang [1]. These pair of quantities correspond to important study endpoints such as the disease specific survival and overall survival, which are frequently used as co-primary endpoints in clinical trials. An R package has been developed to implement our methods. We illustrate our methods and the potential sample size saving of the joint tests over the Bonferroni method through simulations and the 4-D (Die Deutsche Diabetes Dialyse Studie).
|