Project Description
The joint modelling of longitudinal and time-to-event data has become a highly-active field of research over the last twenty years with recent focus on developing models to handle multiple longitudinal profiles along with several, potentially competing, time-to-event processes. Several situations remain in their infancy when it comes to methodological advances. Examples are the treatment of asynchronous longitudinal profiles, data encompassing multiple levels (i.e. patients within a hospital within a region), sample size and power calculations. There is also a need to develop models to handle longitudinal data in the form of counts, or percentages and improving flexibility in time-to-event models by incorporating barycentric rational interpolation in place of splines. Software to implement methods is also key in order to allow routine implementation of the most appropriate model for the data to hand. The focus of this project will be two-fold: developing novel models for situations not adequately dealt with in the current literature, encompassing ‘pre’ and ‘post’ data accumulation in the form of study design and methodological advances, and implementing these models in user-friendly R software to enhance dissemination.