CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: On a general structure for hazard-based regression models Authors:  Laurent Remontet - Hospices Civils de Lyon (France)
Nicholas Jewell - University of Berkely (United States)
Aurelien Belot - London School of Hygiene and Tropical Medicine (United Kingdom)
Francisco Javier Rubio - King's College London (United Kingdom) [presenting]
Abstract: The proportional hazards model represents the most commonly assumed hazard structure when analysing time to event data using regression models. The context of excess hazard models, which is of great interest in cancer epidemiology, is also dominated by the proportional hazards assumption. We will give a brief introduction to excess hazard regression models, and we will present a general hazard structure which contains, as particular cases, proportional hazards, accelerated hazards, and accelerated failure time structures, as well as combinations of these. We combine these different hazard structures with a flexible parametric distribution (exponentiated Weibull) for modelling the baseline hazard. This distribution allows us to cover the basic hazard shapes of interest in practice: constant, bathtub, increasing, decreasing, and unimodal. An example with real data will be used to illustrate the usefulness of this model. We also illustrate the importance of studying flexible parametric distributions, with interpretable parameters and good inferential properties that control the shape of the hazard.