Version 1
: Received: 27 July 2022 / Approved: 4 August 2022 / Online: 4 August 2022 (04:00:20 CEST)
How to cite:
Chiu, I.; Cheng, C.; Chang, P.; Li, C.; Cheng, F.; Lin, C.R. One-Class Machine-Learning Model to Screen for Dysglycemia Using Single Lead ECG in ICU, toward Noninvasive Blood Glucose Monitoring. Preprints2022, 2022080098. https://doi.org/10.20944/preprints202208.0098.v1
Chiu, I.; Cheng, C.; Chang, P.; Li, C.; Cheng, F.; Lin, C.R. One-Class Machine-Learning Model to Screen for Dysglycemia Using Single Lead ECG in ICU, toward Noninvasive Blood Glucose Monitoring. Preprints 2022, 2022080098. https://doi.org/10.20944/preprints202208.0098.v1
Chiu, I.; Cheng, C.; Chang, P.; Li, C.; Cheng, F.; Lin, C.R. One-Class Machine-Learning Model to Screen for Dysglycemia Using Single Lead ECG in ICU, toward Noninvasive Blood Glucose Monitoring. Preprints2022, 2022080098. https://doi.org/10.20944/preprints202208.0098.v1
APA Style
Chiu, I., Cheng, C., Chang, P., Li, C., Cheng, F., & Lin, C.R. (2022). One-Class Machine-Learning Model to Screen for Dysglycemia Using Single Lead ECG in ICU, toward Noninvasive Blood Glucose Monitoring. Preprints. https://doi.org/10.20944/preprints202208.0098.v1
Chicago/Turabian Style
Chiu, I., Fu-Jen Cheng and Chun-Hung Richard Lin. 2022 "One-Class Machine-Learning Model to Screen for Dysglycemia Using Single Lead ECG in ICU, toward Noninvasive Blood Glucose Monitoring" Preprints. https://doi.org/10.20944/preprints202208.0098.v1
Abstract
Blood glucose (BG) monitoring is an important issue for critically ill patients. Previous studies reported that poor sugar control was associated with increased mortality in admitted patients. However, repeated blood glucose monitoring can be resource-consuming and cause a healthcare burden in clinical practice. In this study, we aimed to develop a personalized machine-learning model to predict dysglycemia based on electrocardiogram (ECG) findings. The study included patients with more than 20 ECG records during single hospital admission in the Medical Information Mart for Intensive Care III database, focusing on the lead II recordings, along with the corresponding blood sugar data. We processed the data and used ECG features from each heartbeat as inputs to develop a one-class support vector machine (SVM) algorithm to predict dysglycemia. The model prediction for dysglycemia using a single heartbeat had an AUC level of 0.92 ± 0.09, with a sensitivity of 0.92 ± 0.10 and specificity of 0.84 ± 0.04. Based on 10 s majority voting, the model prediction for dysglycemia improved to an AUC of 0.97 ± 0.06. In this study, we found that a personalized machine-learning algorithm could accurately detect dysglycemia using a single-lead ECG.
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
