Hierarchical dynamic time-to-event models forpost-treatment preventive care data on breast cancersurvivors

摘要

This paper considers modelling data arising in post-treatment preventive care settings, wherecancer patients who have undergone disease-directed treatment discontinue seeking preventive care ser-vices. Clinicians and public health researchers are interested in explaining such behavioural patternsby modelling the time-to-receiving care while accounting for several patient and treatment attributes.A key feature of such data is that a noticeable number of patients would never return for screening, aconcept subtly different from censoring, where an individual does not return for screening in the giventime frame of the study. Models distinguishing between these two concepts are known as cure ratemodels and are often preferred for data where a significant part of the population never experiencedthe endpoint. Building upon recent work on hierarchical cure model framework we propose modellinga sequence of latent events with a piecewise exponential distribution that remedies oversmoothingencountered in existing models with different latent distributions. We investigate simultaneous regres-sion on the cure fraction and the latent event distribution and derive a flexible class of semiparametriccure rate models.