In this article, we develop a flexible model for left truncated right censored (LTRC) data - which is a commonly observed structure of data obtained from survival studies. The model involves piecewise linear approximation of the cumulative baseline hazard function of the proportional hazards model; it is flexible, data-driven, and is shown to be robust in fitting LTRC data generated from different processes. In addition to results of extensive Monte Carlo simulation studies, analyses of two real LTRC datasets by using the model are also provided as illustrative examples. Applications of the model in some practical prediction issues are discussed.
CO Author: A. Ganguly, N. Balakrishnan, and D. Kundu
Journal: Statistics in Medicine
Click here
