ABSTRACT
In this study, we investigated the robustness of the methods that account for independent left truncation when applied to competing risks settings with dependent left truncation. We specifically focused on the methods for the proportional cause-specific hazards model and the Fine–Gray model. Simulation experiments showed that these methods are not in general robust against dependent left truncation. The magnitude of the bias was analogous to the strength of the association between left truncation and failure times, the effect of the covariate on the competing cause of failure, and the baseline hazard of left truncation time.
MATHEMATICS SUBJECT CLASSIFICATION:
Funding
The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under EuroCoord grant agreement no. 260694.