MRC INTEGRATIVE EPIDEMIOLOGY UNIT (IEU)
SEMINAR SERIES

 Wednesday, 30^th May, 2018 : 15.00 – 16.00
Room OS6, Second Floor, Oakfield House

 Audinga-Dea Hazewinkel
University of Leiden

 A multi-state approach to modelling disease progression

 

Abstract

In the commonly used approach for the analysis of survival data from clinical trials a Cox regression model is applied to separate endpoints, such as relapse or death, or to composite endpoints such as disease-free survival. This, however, fails to capture the full history of other events a patient may experience. A multi-state model extends the simple Cox model by distinguishing various states a patient may occupy over the course of the study before the final event. This allows us to assess the impact of intermediate events on patient survival and also allows us to examine how the effect of prognostic factors differs over the stages of disease progression. In this presentation the theory behind the multi-state model will be discussed and illustrated by an application to osteosarcoma data from the EURAMOS-1 clinical trial

Biography

Ms Hazewinkel studied Biomedical Sciences (BSc), with a focus on oncology, at the University of Amsterdam. Currently, she is a master student in Statistical Science at Leiden University, where she is working on survival analysis in clinical trials with Dr Marta Fiocco and Prof dr Hein Putter. The emphasis of her thesis is on dynamic prediction, random survival forests and multi-state modelling. In November 2018 she will start a PhD in Mendelian randomization and survival analysis under the supervision of Dr Jack Bowden and Dr Kaitlin Wade.

  

ALL WELCOME