Intelligent Technology for Predicting the Risk of Patient`s Death from Coronavirus Based on PRINCIPLE-Methodology for Selecting Indicators Collected from Medical Devices

Title:Intelligent Technology for Predicting the Risk of Patient`s Death from Coronavirus Based on PRINCIPLE-Methodology for Selecting Indicators Collected from Medical Devices

Authors:Vitalii B. Mokin, Olena V. Kovalchuk and Nadiia O. Muzyka

Conference:ELNANO-2022

Tags:Bayesian network, COVID-19, medical devices and prediction of death

Abstract:

In Ukraine, COVID-19 has contributed over 104,106 deaths. Multiple risk factors for COVID-19 almost have been identified. The new PRINCIPLE methodology for selecting indicators of patient screening using medical equipment is developed. The name of this methodology is an acronym of the criteria for selecting indicators: "Provability", "Reproducibility", "Informativeness", "Numerical", "Clinical", "Importance", "Prevalence", "Lungs", "Electrocardiography". Tools based on the Bayesian network to predict the high risk of patient mortality based on these indicators are developed. An example of application of the proposed methodology, construction of the model, and formation of conclusions by them are given for anonymized data consisting of 22 features from adults 280 alive and 140 dead patients, diagnosed with COVID-19 at the hospital in Vinnytsia. The work offers an improved method for processing and analyzing the biomedical indicators and medical diagnostic data for the clinical decision-making tool for COVID-19 inpatients construction.