Learning to Detect Sepsis with a Multitask Gaussian Process RNN Classifier

Abstract

We present a scalable end-to-end classifier that uses streaming physiological and medication data to accurately predict the onset of sepsis, a life-threatening complication from infections that has high mortality and morbidity. Our proposed framework models the multivariate trajectories of continuous-valued physiological time series using multitask Gaussian processes, seamlessly accounting for the high uncertainty, frequent missingness, and irregular sampling rates typically associated with real clinical data. The Gaussian process is directly connected to a black-box classifier that predicts whether a patient will become septic, chosen in our case to be a recurrent neural network to account for the extreme variability in the length of patient encounters. We show how to scale the computations associated with the Gaussian process in a manner so that the entire system can be discriminatively trained end-to-end using backpropagation. In a large cohort of heterogeneous inpatient encounters at our university health system we find that it outperforms several baselines at predicting sepsis, and yields 19.4% and 55.5% improved areas under the Receiver Operating Characteristic and Precision Recall curves as compared to the NEWS score currently used by our hospital.

Presented at 34th International Conference on Machine Learning (ICML 2017), Sydney, Australia

References in corpus (1)

• A Framework for Individualizing Predictions of Disease Trajectories by Exploiting Multi-Resolution Structure

Cited by in corpus (28)

• Neural Ordinary Differential Equations
• A Review of Challenges and Opportunities in Machine Learning for Health
• GRU-ODE-Bayes: Continuous modeling of sporadically-observed time series
• Neural Controlled Differential Equations for Irregular Time Series
• MGP-AttTCN: An Interpretable Machine Learning Model for the Prediction of Sepsis
• Uncertainty-Aware Attention for Reliable Interpretation and Prediction
• Set Functions for Time Series
• Semiparametric Bayesian causal inference
• Multi-Time Attention Networks for Irregularly Sampled Time Series
• Cross-hospital Sepsis Early Detection via Semi-supervised Optimal Transport with Self-paced Ensemble
• Early detection of sepsis utilizing deep learning on electronic health record event sequences
• Predicting Acute Kidney Injury at Hospital Re-entry Using High-dimensional Electronic Health Record Data
• Counterfactual Normalization: Proactively Addressing Dataset Shift and Improving Reliability Using Causal Mechanisms
• Sequential Gaussian Processes for Online Learning of Nonstationary Functions
• Fast Matrix Square Roots with Applications to Gaussian Processes and Bayesian Optimization
• A Generalised Signature Method for Multivariate Time Series Feature Extraction
• Neural SDEs as Infinite-Dimensional GANs
• Self Attention with Temporal Prior: Can We Learn More from Arrow of Time?
• Mixture-based Multiple Imputation Model for Clinical Data with a Temporal Dimension
• Probabilistic Numeric Convolutional Neural Networks
• Using Contextual Information to Improve Blood Glucose Prediction
• Multi-Task Gaussian Processes and Dilated Convolutional Networks for Reconstruction of Reproductive Hormonal Dynamics
• Is In-hospital Meta-information Useful for Abstractive Discharge Summary Generation?
• Deep Mixed Effect Model using Gaussian Processes: A Personalized and Reliable Prediction for Healthcare
• The Consequences of the Framing of Machine Learning Risk Prediction Models: Evaluation of Sepsis in General Wards
• Real-time Prediction for Mechanical Ventilation in COVID-19 Patients using A Multi-task Gaussian Process Multi-objective Self-attention Network
• UNITE: Uncertainty-based Health Risk Prediction Leveraging Multi-sourced Data
• Deep Kernel Learning for Mortality Prediction in the Face of Temporal Shift