Page 53 - AIH-2-2

Artificial Intelligence in Health

ORIGINAL RESEARCH ARTICLE
A stacked ensemble deep learning model for

predicting the intensive care unit patient mortality

Dimitrios Simopoulos * , Dimitrios Kosmidis 2 , George Anastassopoulos 1 ,
1
and Lazaros Iliadis 3
1 Department of Medicine, Faculty of Health Sciences, Democritus University of Thrace,
Alexandroupolis, Greece
2 Department of Nursing, Faculty of Health Sciences, Democritus University of Thrace, Didymoteicho,
Greece
3 Department of Civil Engineering, School of Engineering, Democritus University of Thrace, Xanthi,
Greece

Abstract

Accurate mortality prediction in intensive care units (ICUs) is essential for optimizing
patient treatment, nursing care, and resource allocation. Traditional models, such as
Acute Physiology and Chronic Health Evaluation and Simplified Acute Physiology
Score, have been very important in clinical practice, but they frequently have issues
with prediction accuracy and adaptability, especially when dealing with complex and
evolving patient data. These issues can be resolved, and the accuracy of mortality
prediction increased due to recent developments in machine learning, especially
deep learning. The present study introduces a new deep learning ensemble model
*Corresponding author:
Dimitrios Simopoulos that achieves a significant improvement over existing methods. Using stacked
(dsimopou@med.duth.gr) ensemble learning, our approach combines the advantages of one Random Forests
Citation: Simopoulos D, model and two CatBoost models. We achieved a notable performance in mortality
Kosmidis D, Anastassopoulos G, prediction by carefully training and optimizing this ensemble using the electronic
Iliadis L. A stacked ensemble deep ICU Collaborative Research Database. Our model boasts an accuracy of 94.19%,
learning model for predicting the
intensive care unit patient mortality. precision of 94.097%, recall of 94.29%, and F -score of 94.191%, demonstrating
1
Artif Intell Health. 2025;2(2):47-59. a substantial improvement over conventional approaches. The prediction of ICU
doi: 10.36922/aih.4981 mortality has been significantly improved using ensemble learning, which helps
Received: September 27, 2024 medical and nursing staff to better treat patients individually, allocate resources
efficiently, and enhance patient outcomes. This approach gives healthcare experts
Revised: November 21, 2024
the ability to make data-driven decisions, leading to more effective and efficient care
Accepted: December 2, 2024 within the ICU.
Published online: December 16,
2024
Keywords: Mortality prediction; Intensive Care Units; Healthcare; Machine learning; Deep
Copyright: © 2024 Author(s). learning; Stacked ensemble learning
This is an Open-Access article
distributed under the terms of the
Creative Commons Attribution
License, permitting distribution,
and reproduction in any medium, 1. Introduction
provided the original work is
properly cited. In recent years, predictive analytics in healthcare has focused on applying machine
Publisher’s Note: AccScience learning (ML) techniques to guide medical treatments, and the developments observed
Publishing remains neutral with seem to be in line with the rapid pace of artificial intelligence (AI) development. Especially
regard to jurisdictional claims in
published maps and institutional in intensive care units (ICUs), where timely prediction of various patient conditions
affiliations. and resource allocation is of paramount importance, ML is gaining momentum with

Volume 2 Issue 2 (2025) 47 doi: 10.36922/aih.4981