BernGraph: Probabilistic Graph Neural Networks for EHR-based Medication Recommendations

TL;DR

This paper introduces BernGraph, a probabilistic graph neural network that models binary EHR data as continuous probabilities to improve medication recommendation accuracy using only binary records.

Contribution

It presents the first binary EHR-based drug recommendation system that models event relationships via Bernoulli probabilities and leverages graph neural networks for improved performance.

Findings

Achieves state-of-the-art results on large-scale EHR datasets.

Outperforms methods using secondary information by a large margin.

Effectively models event-event relationships with probabilistic transformations.

Abstract

The medical community believes binary medical event outcomes in EHR data contain sufficient information for making a sensible recommendation. However, there are two challenges to effectively utilizing such data: (1) modeling the relationship between massive 0,1 event outcomes is difficult, even with expert knowledge; (2) in practice, learning can be stalled by the binary values since the equally important 0 entries propagate no learning signals. Currently, there is a large gap between the assumed sufficient information and the reality that no promising results have been shown by utilizing solely the binary data: visiting or secondary information is often necessary to reach acceptable performance. In this paper, we attempt to build the first successful binary EHR data-oriented drug recommendation system by tackling the two difficulties, making sensible drug recommendations solely using the binary EHR medical records. To this end, we take a statistical perspective to view the EHR data as a sample from its cohorts and transform them into continuous Bernoulli probabilities. The transformed entries not only model a deterministic binary event with a distribution but also allow reflecting \emph{event-event} relationship by conditional probability. A graph neural network is learned on top of the transformation. It captures event-event correlations while emphasizing \emph{event-to-patient} features. Extensive results demonstrate that the proposed method achieves state-of-the-art performance on large-scale databases, outperforming baseline methods that use secondary information by a large margin. The source code is available at \url{https://github.com/chenzRG/BEHRMecom}