MedML: Fusing medical knowledge and machine learning models for early pediatric COVID-19 hospitalization and severity prediction
Access full-text PDFOpen Access
Check access options
Check access options
UMass Chan AffiliationsCenter for Clinical and Translational Science
Document TypeJournal Article
Artificial intelligence applications
MetadataShow full item record
AbstractThe COVID-19 pandemic has caused devastating economic and social disruption. This has led to a nationwide call for models to predict hospitalization and severe illness in patients with COVID-19 to inform the distribution of limited healthcare resources. To address this challenge, we propose a machine learning model, MedML, to conduct the hospitalization and severity prediction for the pediatric population using electronic health records. MedML extracts the most predictive features based on medical knowledge and propensity scores from over 6 million medical concepts and incorporates the inter-feature relationships in medical knowledge graphs via graph neural networks. We evaluate MedML on the National Cohort Collaborative (N3C) dataset. MedML achieves up to a 7% higher AUROC and 14% higher AUPRC compared to the best baseline machine learning models. MedML is a new machine learnig framework to incorporate clinical domain knowledge and is more predictive and explainable than current data-driven methods.
SourceGao J, Yang C, Heintz J, Barrows S, Albers E, Stapel M, Warfield S, Cross A, Sun J; N3C consortium. MedML: Fusing medical knowledge and machine learning models for early pediatric COVID-19 hospitalization and severity prediction. iScience. 2022 Sep 16;25(9):104970. doi: 10.1016/j.isci.2022.104970. Epub 2022 Aug 17. PMID: 35992304; PMCID: PMC9384332.
Permanent Link to this Itemhttp://hdl.handle.net/20.500.14038/51753
Funding and AcknowledgementsThe UMass Center for Clinical and Translational Science (UMCCTS), UL1TR001453, provided data for this study.
RightsCopyright 2022 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).; Attribution 4.0 International
The following license files are associated with this item:
- Creative Commons
Except where otherwise noted, this item's license is described as Copyright 2022 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).