VIBASS9 - Invited Course

This course (12 hours) is provided by Denis Rustand (Research scientist at the National Institute of Health and Medical Research, Bordeaux Population Health Research Center, France). His research focuses on statistical models for the joint analysis of complex longitudinal and survival data in health. The content of the course is tailored to the framework presented in his book: Bayesian survival, longitudinal and joint models with INLA (Chapman & Hall/CRC, 2026).

Target audience

Statisticians and applied researchers with an interest in the analysis of longitudinal and time-to-event data. Basic experience with R is required. Participants are also expected to be familiar with the basics of mixed-effects models and survival analysis. No prior experience with Bayesian statistics or INLA is required.

Abstract

Analyzing longitudinal markers and time-to-event outcomes is a central task in modern biomedical research. While Bayesian methods provide a flexible framework for modeling these data, traditional MCMC approaches can be computationally demanding. This course focuses on a modern and efficient Bayesian framework specifically designed to overcome these challenges using the INLAjoint R package.

Recent developments in the Integrated Nested Laplace Approximation (INLA) methodology allow for the fast and accurate analysis of complex longitudinal and survival data. The course adopts a practical approach: participants will first learn to specify and fit flexible univariate models (mixed-effects and survival) before combining them into multivariate joint models.

Beyond standard modeling, we will explore diverse association structures and handle complex scenarios such as competing risks and multi-state models. Finally, the course bridges the gap to clinical application by illustrating how to perform dynamic predictions for personalized risk assessment.