A Knowledge Distillation Ensemble Framework for Predicting Short- and Long-Term Hospitalization Outcomes From Electronic Health Records Data

Published in

IEEE journal of biomedical and health informatics(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

A highly-scalable and robust machine learning framework to automatically predict adversity represented by mortality and ICU admission and readmission from time-series of vital signs and laboratory results obtained within the first 24 hours of hospital admission is presented.

Abstract

The ability to perform accurate prognosis is crucial for proactive clinical decision making, informed resource management and personalised care. Existing outcome prediction models suffer from a low recall of infrequent positive outcomes. We present a highly-scalable and robust machine learning framework to automatically predict adversity represented by mortality and ICU admission and readmission from time-series of vital signs and laboratory results obtained within the first 24 hours of hospital admission. The stacked ensemble platform comprises two components: a) an unsupervised LSTM Autoencoder that learns an optimal representation of the time-series, using it to differentiate the less frequent patterns which conclude with an adverse event from the majority patterns that do not, and b) a gradient boosting model, which relies on the constructed representation to refine prediction by incorporating static features. The model is used to assess a patient's risk of adversity and provides visual justifications of its prediction. Results of three case studies show that the model outperforms existing platforms in ICU and general ward settings, achieving average Precision-Recall Areas Under the Curve (PR-AUCs) of 0.891 (95% CI: 0.878-0.939) for mortality and 0.908 (95% CI: 0.870-0.935) in predicting ICU admission and readmission.

Authors

A Knowledge Distillation Ensemble Framework for Predicting Short- and Long-Term Hospitalization Outcomes From Electronic Health Records Data

Published in

IEEE journal of biomedical and health informatics(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

Abstract

Authors

References73 items

An artificial intelligence system for predicting the deterioration of COVID-19 patients in the emergency department

Evaluation and improvement of the National Early Warning Score (NEWS2) for COVID-19: a multi-hospital study

Interpretable confidence measures for decision support systems

Using machine learning methods to predict in-hospital mortality of sepsis patients in the ICU

Clinical features of COVID-19 mortality: development and validation of a clinical prediction model

Clinical outcomes of COVID-19 in Wuhan, China: a large cohort study

Artificial neural networks improve early outcome prediction and risk classification in out-of-hospital cardiac arrest patients admitted to intensive care

Unsupervised anomaly detection for discrete sequence healthcare data

Machine Learning for Clinical Outcome Prediction

Hybrid approach for Anomaly Detection in Time Series Data

Assessment of the Accuracy of Using ICD-9 Diagnosis Codes to Identify Pneumonia Etiology in Patients Hospitalized With Pneumonia

COVID-19 and the elderly: insights into pathogenesis and clinical decision-making

A Machine Learning Early Warning System: Multicenter Validation in Brazilian Hospitals

Predicting Clinical Outcomes with Patient Stratification via Deep Mixture Neural Networks.

An interpretable mortality prediction model for COVID-19 patients

Comparison of Early Warning Scoring Systems for Hospitalized Patients With and Without Infection at Risk for In-Hospital Mortality and Transfer to the Intensive Care Unit

Supplementing the National Early Warning Score (NEWS2) for anticipating early deterioration among patients with COVID-19 infection

A clinical risk score to identify patients with COVID-19 at high risk of critical care admission or death: An observational cohort study

An Imbalanced Data Handling Framework for Industrial Big Data Using a Gaussian Process Regression-Based Generative Adversarial Network

Dynamic and explainable machine learning prediction of mortality in patients in the intensive care unit: a retrospective study of high-frequency data in electronic patient records.

Cardiovascular Implications of Fatal Outcomes of Patients With Coronavirus Disease 2019 (COVID-19)

Correlation Analysis Between Disease Severity and Inflammation-related Parameters in Patients with COVID-19 Pneumonia

Deep Interpretable Early Warning System for the Detection of Clinical Deterioration

Machine learning-based dynamic mortality prediction after traumatic brain injury

An Electronic CKD Phenotype: A Step Forward in Improving Kidney Care.

Improving Clinical Translation of Machine Learning Approaches Through Clinician-Tailored Visual Displays of Black Box Algorithms: Development and Validation

Outlier Detection for Time Series with Recurrent Autoencoder Ensembles

Prospective and External Evaluation of a Machine Learning Model to Predict In-Hospital Mortality

Exploring Interpretable LSTM Neural Networks over Multi-Variable Data

The incidence of cardiac arrest in the intensive care unit: A systematic review and meta-analysis

Fall Detection from Thermal Camera Using Convolutional LSTM Autoencoder

Prediction of sepsis patients using machine learning approach: A meta-analysis

Nurse staffing, nursing assistants and hospital mortality: retrospective longitudinal cohort study

The Number of Comorbidities Predicts Renal Outcomes in Patients with Stage 3–5 Chronic Kidney Disease

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

Artificial intelligence in retina

Learning from Imbalanced Data Sets

NEWS 2 – too little evidence to implement?

Interpretable Representation Learning for Healthcare via Capturing Disease Progression through Time

Predicting Intensive Care Unit Readmission with Machine Learning Using Electronic Health Record Data

Outlier Detection for Multidimensional Time Series Using Deep Neural Networks

Detection of Paroxysmal Atrial Fibrillation using Attention-based Bidirectional Recurrent Neural Networks

Rates and risk factors associated with hospitalization for pneumonia with ICU admission among adults

LightGBM: A Highly Efficient Gradient Boosting Decision Tree

Early hospital mortality prediction of intensive care unit patients using an ensemble learning approach

Risk factors for community-acquired pneumonia among adults in Kenya: a case–control study

Beat by beat: Classifying cardiac arrhythmias with recurrent neural networks

Patient Subtyping via Time-Aware LSTM Networks

Machine learning landscapes and predictions for patient outcomes

Real-Time Illegal Parking Detection System Based on Deep Learning

Dynamic Mortality Risk Predictions in Pediatric Critical Care Using Recurrent Neural Networks

Combining Static and Dynamic Features for Multivariate Sequence Classification

A mixed-ensemble model for hospital readmission

RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism

A Survey of Predictive Modeling on Imbalanced Domains

LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection

MIMIC-III, a freely accessible critical care database

Predicting ICU Mortality Risk by Grouping Temporal Trends from a Multivariate Panel of Physiologic Measurements

Why the C-statistic is not informative to evaluate early warning scores and what metrics to use

The precision--recall curve overcame the optimism of the receiver operating characteristic curve in rare diseases.

Chronic kidney disease in the elderly: evaluation and management.

Multicenter development and validation of a risk stratification tool for ward patients.

Machine Learning Methods for Mortality Prediction of Polytraumatized Patients in Intensive Care Units - Dealing with Imbalanced and High-Dimensional Data

Log-transformation and its implications for data analysis

The effect of comorbidities on risk of intensive care readmission during the same hospitalization: a linked data cohort study.

An experimental comparison of performance measures for classification

The relationship between Precision-Recall and ROC curves

Is Combining Classifiers with Stacking Better than Selecting the Best One?

Intensive Care National Audit and Research Centre (ICNARC)

National Early Warning Score (NEWS) 2: Standardising the assessment of acute-illness severity in the NHS

Mortality Prediction in the ICU Based on MIMIC-II Results from the Super ICU Learner Algorithm (SICULA) Project

Gaussian Processes for Personalized e-Health Monitoring With Wearable Sensors

Key statistics from the case mix programme april 2011-march 2012