A spline-based framework for the flexible modelling of continuously observed multistate survival processes

Multistate modelling is becoming increasingly popular due to the availability of richer longitudinal health data. When times at which events characterising disease progression are known, multistate process greatly simplified as it can be broken down in a number traditional survival models. We propose flexibly model them through existing general link-based additive framework implemented R package GJRM. The associated transition probabilities then obtained simulation-based approach mstate, appealing its generality. integration between two seamless and efficient since we transformation function, rather than hazard commonly found. This achieved use shape constrained P-splines elegantly embed monotonicity required for functions within construction themselves. proposed allows inclusion virtually any type covariate effects, including time-dependent ones, while imposing no restriction on assumed. exemplify usage this case study breast cancer patients.