Learning Representations of Missing Data for Predicting Patient Outcomes
This addresses the challenge of handling complex, incomplete EHR data for healthcare providers, but it is incremental as it builds on existing representation learning techniques.
The paper tackled the problem of predicting patient outcomes from Electronic Health Records (EHRs) with missing data, varied modalities, and hidden relationships, and showed that their representation learning approach based on message passing is competitive or outperforms state-of-the-art methods for tasks like in-hospital mortality, length of stay, and discharge destination.
Extracting actionable insight from Electronic Health Records (EHRs) poses several challenges for traditional machine learning approaches. Patients are often missing data relative to each other; the data comes in a variety of modalities, such as multivariate time series, free text, and categorical demographic information; important relationships among patients can be difficult to detect; and many others. In this work, we propose a novel approach to address these first three challenges using a representation learning scheme based on message passing. We show that our proposed approach is competitive with or outperforms the state of the art for predicting in-hospital mortality (binary classification), the length of hospital visits (regression) and the discharge destination (multiclass classification).