Machine Learning of Physiologic Waveforms and Electronic Health Record Data: A Large Perioperative Data Set of High-Fidelity Physiologic Waveforms

Sungsoo Kim; Sohee Kwon; Akos Rudas; Ravi Pal; Mia K Markey; Alan C Bovik; Maxime Cannesson

doi:10.1016/j.ccc.2023.03.003

Machine Learning of Physiologic Waveforms and Electronic Health Record Data: A Large Perioperative Data Set of High-Fidelity Physiologic Waveforms

Crit Care Clin. 2023 Oct;39(4):675-687. doi: 10.1016/j.ccc.2023.03.003. Epub 2023 May 18.

Authors

Sungsoo Kim¹, Sohee Kwon², Akos Rudas³, Ravi Pal², Mia K Markey⁴, Alan C Bovik⁵, Maxime Cannesson⁶

Affiliations

¹ Department of Anesthesiology and Perioperative Medicine, University of California Los Angeles, Los Angeles, CA 90095, USA; Department of Electrical & Computer Engineering, The University of Texas at Austin, Austin, TX, USA.
² Department of Anesthesiology and Perioperative Medicine, University of California Los Angeles, Los Angeles, CA 90095, USA.
³ Department of Computational Medicine, University of California Los Angeles, Los Angeles, CA 90095, USA.
⁴ Department of Biomedical Engineering, The University of Texas at Austin, Austin, TX, USA.
⁵ Department of Electrical & Computer Engineering, The University of Texas at Austin, Austin, TX, USA.
⁶ Department of Anesthesiology and Perioperative Medicine, University of California Los Angeles, Los Angeles, CA 90095, USA. Electronic address: mcannesson@mednet.ucla.edu.

PMID: 37704333
DOI: 10.1016/j.ccc.2023.03.003

Abstract

Perioperative morbidity and mortality are significantly associated with both static and dynamic perioperative factors. The studies investigating static perioperative factors have been reported; however, there are a limited number of previous studies and data sets analyzing dynamic perioperative factors, including physiologic waveforms, despite its clinical importance. To fill the gap, the authors introduce a novel large size perioperative data set: Machine Learning Of physiologic waveforms and electronic health Record Data (MLORD) data set. They also provide a concise tutorial on machine learning to illustrate predictive models trained on complex and diverse structures in the MLORD data set.

Keywords: Deep neuronal networks; Machine learning; Perioperative medicine; Physiologic waveforms; Prediction; Surgery.

Publication types

Review

MeSH terms

Clinical Relevance
Electronic Health Records*
Humans
Machine Learning*