Predictive Model of Septic Shock Staging Base on Continuing Invasive Hemodynamic Monitoring

Ruowen Liao

2022

Abstract

Septic shock is a major public health concern across the world, also is a typical cause for patients being admitted to the intensive care unit. It is easier to be misdiagnosed, yet the situation is getting worse. Septic shock can be classified into three stages: irreversible (early stage), compensated, and decompensated. Sepsis has long been misdiagnosed, but it develops and worsens at an alarming rate, often reaching irreversible levels within hours. This work has expanded the proportion of invasive hemodynamics to septic shock for the development of understanding of the phases of septic shock. This article aims to construct and develop a real-time prediction model of septic shock staging based on continuous invasive hemodynamic monitoring. The ultimate model of the article is a multi-classification prediction model. In this experiment, the eICU collaborative research database was employed, and four characteristics from the dataset were scored to indicate the stage of septic shock. Need to point out that deep active learning, a new approach that combines deep and active learning, was chosen as the research’s major learning approach. Margin sampling is the main query strategy used in the active learning approach, with the random selection strategy serving as a control strategy. There are two groups of query strategies, compare the two groups to see which one is more effective: random selection or active learning. As a result, the query strategy of active learning is considerably most stable than random selection in deep active learning. Although septic shock cannot be diagnosed purely based on hemodynamic characteristics, the model can nevertheless assist clinicians in making an early diagnosis or warning.

Paper Citation

in Harvard Style

Liao R. (2022). Predictive Model of Septic Shock Staging Base on Continuing Invasive Hemodynamic Monitoring. In Proceedings of the 1st International Conference on Health Big Data and Intelligent Healthcare - Volume 1: ICHIH, ISBN 978-989-758-596-8, pages 518-523. DOI: 10.5220/0011373200003438

in Bibtex Style

@conference{ichih22,
author={Ruowen Liao},
title={Predictive Model of Septic Shock Staging Base on Continuing Invasive Hemodynamic Monitoring},
booktitle={Proceedings of the 1st International Conference on Health Big Data and Intelligent Healthcare - Volume 1: ICHIH,},
year={2022},
pages={518-523},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011373200003438},
isbn={978-989-758-596-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Health Big Data and Intelligent Healthcare - Volume 1: ICHIH,
TI - Predictive Model of Septic Shock Staging Base on Continuing Invasive Hemodynamic Monitoring
SN - 978-989-758-596-8
AU - Liao R.
PY - 2022
SP - 518
EP - 523
DO - 10.5220/0011373200003438