A Bayesian Approach to Modelling Longitudinal Data in Electronic Health Records
This work addresses survival prediction in healthcare using EHR data, which is incremental as it builds on existing Bayesian and tree-based methods for handling sparse longitudinal data.
The paper tackles the challenge of integrating sparse, diverse longitudinal data from electronic health records to estimate survival distributions, proposing a nonparametric probabilistic model based on Bayesian trees that learns variable interactions over time, and demonstrates performance improvements on Primary Biliary Cirrhosis patient data.
Analyzing electronic health records (EHR) poses significant challenges because often few samples are available describing a patient's health and, when available, their information content is highly diverse. The problem we consider is how to integrate sparsely sampled longitudinal data, missing measurements informative of the underlying health status and fixed demographic information to produce estimated survival distributions updated through a patient's follow up. We propose a nonparametric probabilistic model that generates survival trajectories from an ensemble of Bayesian trees that learns variable interactions over time without specifying beforehand the longitudinal process. We show performance improvements on Primary Biliary Cirrhosis patient data.