Simulating failure times when the event of interest is unobservable with emphasis on animal carcinogenicity studies

This paper discusses and summarizes the methods which can be used to simulate failure times when an intermediate event of interest is unobservable. Data from animal carcinogenicity experiments are typically of this form. The simulation of time-to-event data where cause of failure is known and the case where cause of failure is unknown are discussed. The methods proposed for simulating failure times use identifiable hazard functions and allow for random sampling of functioning experimental units at fixed times. Some useful design and programming tools are also given.