Variationally Regularized Graph-based Representation Learning for Electronic Health Records

TL;DR

This paper introduces a variationally regularized graph neural network for EHR data that adaptively learns medical concept connections, improving predictive performance and interpretability over existing methods.

Contribution

It proposes a novel variational regularization approach for graph neural networks to enhance robustness and interpretability in EHR representation learning.

Findings

Outperforms existing graph and non-graph methods in EHR prediction tasks

Provides interpretability through singular value analysis of regularization effects

Demonstrates robustness in learning medical concept connections

Abstract

Electronic Health Records (EHR) are high-dimensional data with implicit connections among thousands of medical concepts. These connections, for instance, the co-occurrence of diseases and lab-disease correlations can be informative when only a subset of these variables is documented by the clinician. A feasible approach to improving the representation learning of EHR data is to associate relevant medical concepts and utilize these connections. Existing medical ontologies can be the reference for EHR structures, but they place numerous constraints on the data source. Recent progress on graph neural networks (GNN) enables end-to-end learning of topological structures for non-grid or non-sequential data. However, there are problems to be addressed on how to learn the medical graph adaptively and how to understand the effect of the medical graph on representation learning. In this paper, we propose a variationally regularized encoder-decoder graph network that achieves more robustness in graph structure learning by regularizing node representations. Our model outperforms the existing graph and non-graph based methods in various EHR predictive tasks based on both public data and real-world clinical data. Besides the improvements in empirical experiment performances, we provide an interpretation of the effect of variational regularization compared to standard graph neural network, using singular value analysis.