Please note: this event has passed
As health care data becomes more complex and multimodal, the structure of the data becomes increasingly complicated. There could be an increasing number of repeated measurements on the same individuals over time, longer duration of follow-up for time-to-event outcomes or nested hierarchies leading to non-independence of individuals.
The use of simpler statistical approaches to analyse these data is invalid because the key assumptions of the those approaches do not hold. In this module, we introduce the concept of multilevel and longitudinal modelling, including time-to-event or survival analysis.
The aim is for the student to understand the challenges of longitudinal and clustered data, and the concept and implementation of multilevel models. Students will also become familiar with the most common models for time-to-event data (inc. Cox proportional hazard models, additive hazard models), and finally link all these concepts together through joint modelling of survival and longitudinal data.
Students will discover Stata commands that can fit all of these models and become familiar with the resulting Stata output, whilst applying them to real data structures.
Institute of Psychiatry, Psychology & Neuroscience (IoPPN)
IoPPN, 16 De Crespigny Park, London SE5 8AB