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{<a class="link-external link-https" href="https://github.com/chenzRG/BEHRMecom" rel="external noopener nofollow">this https URL</a>}

A Bayesian Dynamic Graphical Model for Recurrent Events in Public Health

Bayesian dynamic network modelling: an application to metabolic associations in cardiovascular diseases

Bayesian Modeling Longitudinal Dyadic Data with Nonignorable Dropout, with Application to a Breast Cancer Study

Bayesian Latent-Class Mixed-Effect Hybrid Models for Dyadic Longitudinal Data with Non-Ignorable Dropouts.

Bayesian Modeling of Dynamic Behavioral Change During an Epidemic

A Bayesian Approach to Modelling Longitudinal Data in Electronic Health Records

An N Time-Slice Dynamic Chain Event Graph

Bayesian graphical models for computational network biology

Joint Dynamic Models and Statistical Inference for Recurrent Competing Risks, Longitudinal Marker, and Health Status

GDBN: a Graph Neural Network Approach to Dynamic Bayesian Network

A Bayesian Framework for Causal Analysis of Recurrent Events with Timing Misalignment

Bayesian Robust Learning in Chain Graph Models for Integrative Pharmacogenomics

Dynamic Modeling and Statistical Analysis of Event Times

Bayesian semi-parametric inference for clustered recurrent events with zero-inflation and a terminal event/4163305

Prevalence estimation methods for time-dependent antibody kinetics of infected and vaccinated individuals: a graph-theoretic approach

Bayesian Nonparametric Modelling of Joint Gap Time Distributions for Recurrent Event Data

Bayesian Models for Heterogeneous Personalized Health Data

Dynamic Treatment Regimes Using Bayesian Additive Regression Trees for Censored Outcomes

Recurrent Events Analysis With Data Collected at Informative Clinical Visits in Electronic Health Records

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

Bayesian approach for flexible modeling of semicompeting risks data