Time-to-event data are often subject to left-truncation. Lack of consideration of the sampling condition will introduce bias and loss in efficiency of the estimation. While auxiliary information from the same or similar cohorts may be available, challenges arise due to the practical issue of accessibility of individual-level data and taking account of various sampling conditions for different cohorts. In this paper, we introduce a likelihood-based method to incorporate information from auxiliary data to eliminate the left-truncation problem and improve efficiency. A one-step Monte-Carlo Expectation-Maximization algorithm is developed to calculate an augmented likelihood through creating pseudo-data sets which extend the form and conditions of the observed sample. The method is illustrated by both a real dataset and simulation studies.
Keywords: Augmented likelihood; Auxiliary information; Left-truncation; Time-to-event data.