More info on Fridtjof will follow soon.

“Can we improve the prediction of health outcomes by taking into account the arising temporal dynamics?”

Joint models allow for dynamic prediction of time-to-event outcomes (such as disease onset or relapse), while taking into account time-varying covariates. As classical joint models are designed for only few longitudinal measurement occasions, we aim to model the covariates via vector autoregressive models. This extension can account for the temporal dependency arising in the intensive longitudinal data from the SiA project (e.g., ESM measure of affect/stress response). Additionally, the derived temporal dynamics can potentially improve the accuracy of time-to-event predictions.

PhD student,
University of Groningen