A Knowledge Distillation Ensemble Framework for Predicting Short- and Long-Term Hospitalization Outcomes From Electronic Health Records Data
| Autor: | Anthony Shek, James Galloway, Thomas Searle, Honghan Wu, Zeljko Kraljevic, Zina M. Ibrahim, Daniel Bean, James T. Teo, Sam Norton, Linglong Qian, Richard Dobson |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Computer science
Vital signs MEDLINE Machine learning computer.software_genre Data modeling Machine Learning Health Information Management Electronic Health Records Humans Electrical and Electronic Engineering Retrospective Studies Recall business.industry Length of Stay Ensemble learning Autoencoder Computer Science Applications Term (time) Hospitalization ROC Curve Gradient boosting Artificial intelligence business computer Biotechnology |
| Zdroj: | IEEE journal of biomedical and health informatics. 26(1) |
| ISSN: | 2168-2208 |
| Popis: | The ability to perform accurate prognosis of patients 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 from time-series vital signs and laboratory results obtained within the first 24 hours of hospital admission. The stacked 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 the prediction, incorporating static features of demographics, admission details and clinical summaries. The model is used to assess a patient's risk of adversity over time and provides visual justifications of its prediction, based on the patient's static features and dynamic signals. Results of three case studies for predicting mortality and ICU admission show that the model outperforms all existing outcome prediction models, achieving average Precision-Recall Areas Under the Curve (PR-AUCs) of 0.93 (95% CI: 0.878 - 0.969) in predicting mortality in ICU and general ward settings and 0.987 (95% CI: 0.985-0.995) in predicting ICU admission. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|