Novel statistical methods for prognosis research

Abstract

This chapter introduces some advanced statistical methods that are growing in their application to address more complex data arising from prognosis research studies. Three major topics are covered: competing risks, multi-state models, and joint modelling of longitudinal and survival data. The advances in such statistical methods allow complex relationships and intricate prognosis pathways to be modelled, including multi-morbidities over time. They are needed to help identify prognostic factors at different parts of an individual’s time course, and to develop more dynamic prognostic models where outcome risk can be updated over time. Practical clinical examples are used throughout the chapter to illustrate the approaches.