The statistical analysis of the information generated by medical follow-up is a very important challenge in the field of personalised medicine. As the evolutionary course of a patient's disease progresses, its medical follow-up generates more and more information that should be processed immediately in order to review and update its prognosis and treatment. Our objective in this thesis focuses on this update process through sequential inference methods for joint models of longitudinal and time-to-event data from a Bayesian perspective. More specifically, we propose the use of sequential Monte Carlo methods for static parameter joint models in order to update the posterior distribution of the parameters, hyperparameters, and random effects with the intention of reducing computation time in each update of the inferential process. Our proposal is very general and can be easily applied to most popular joint models approaches. We illustrate our research with two different studies: (i) a joint model for longitudinal data with informative dropout simulated through an own novel mechanism, and (ii) a joint model with competing risk events for a real problem about patients receiving mechanical ventilation in intensive care units.

中文翻译：

---

用于纵向和时间到事件数据的贝叶斯联合模型中的顺序蒙特卡罗方法

对医疗随访产生的信息进行统计分析是个性化医疗领域的一个非常重要的挑战。随着患者疾病的演变过程，其医疗随访会产生越来越多的信息，应立即处理这些信息，以便审查和更新其预后和治疗。我们在本论文中的目标侧重于从贝叶斯的角度通过对纵向和时间到事件数据的联合模型的顺序推理方法的更新过程。更具体地说，我们建议对静态参数联合模型使用顺序蒙特卡罗方法，以更新参数、超参数和随机效应的后验分布，目的是减少推理过程每次更新的计算时间。我们的建议非常通用，可以轻松应用于大多数流行的联合模型方法。我们用两项不同的研究来说明我们的研究：(i) 纵向数据的联合模型，通过自己的新机制模拟具有信息丢失的联合模型，以及 (ii) 具有竞争风险事件的联合模型，用于解决有关患者在重症监护下接受机械通气的实际问题。护理单位。