Page 141 - EJMO-9-1

Eurasian Journal of Medicine
and Oncology

ORIGINAL RESEARCH ARTICLE
Clinical and demographic predictors of heart

failure outcomes: A machine learning perspective

Shivaprasad Chitta 1 , Supriya Chandu 2 , Krishna Chaitanya Katha 3 ,
5,6
7
Syam Sundar Junapudi 4 , Vinod Kumar Yata * , and Sunil Junapudi *
1 Department of Computer Science, Osmania University, Hyderabad, Telangana, India
2 Department of Public Health, College of Public Health, University of New Haven Brower Street,
West Haven, Connecticut, United States of America
3 Bioinformatics and Computational Biology, Morsani College of Medicine, University of South
Florida, Tampa, United States of America
4 Department of Community Medicine, Government Medical College, Mahabubabad, Telangana, India
5 Department of Molecular Biology, Central University of Andhra Pradesh, Anantapuramu,
Andhra Pradesh, India
6 Research Centre, KBK Multispecialty Hospitals, Hyderabad, Telangana, India
7 Department of Pharmaceutical Chemistry, Geethanjali College of Pharmacy, Hyderabad,
Telangana, India

Abstract

*Corresponding authors: Heart failure (HF) is a multifaceted clinical condition associated with high morbidity
Sunil Junapudi and mortality rates. It is an increasing public health concern, impacting millions
(suniljunapudi@gmail.com);
Vinod Kumar Yata globally and placing considerable strain on healthcare systems. In recent decades,
(vinod.yata@cuap.edu.in) there has been a growing interest in using machine learning techniques to predict
HF outcomes. Hence, this study aims to explore the clinical and demographic
Citation: Chitta S, Chandu S,
Katha KC, Junapudi SS, characteristics associated with HF outcomes using a comprehensive dataset obtained
Yata VK, Junapudi S. Clinical from Kaggle. The dataset, “Heart Failure Clinical Records.csv,” was preprocessed to
and demographic predictors of address missing values and prepared for analysis. Feature importance analysis and
heart failure outcomes: A machine
learning perspective. Eurasian J correlation matrix computations were conducted to identify significant predictors
Med Oncol. 2025;9(1):133-143. of death events among HF patients, including age, serum creatinine, and ejection
doi: 10.36922/ejmo.6583 fraction. Various machine learning models, such as logistic regression, random forest,
Received: November 27, 2024 support vector machine, and gradient boosting machine, were employed to predict
Revised: December 21, 2024 death events. The results revealed varying levels of performance among the models,
with some demonstrating promising accuracy and predictive power. However,
Accepted: December 30, 2024 further refinement of these predictive models is warranted to enhance clinical
Published online: January 21, decision-making and patient care in HF management. Overall, this study underscores
2025 the value of data-driven approaches in understanding HF outcomes and highlights
Copyright: © 2025 Author(s). the necessity for ongoing research in this field.
This is an Open-Access article
distributed under the terms of the
Creative Commons Attribution Keywords: Heart failure; Machine learning models; Logistic regression; Random forest;
License, permitting distribution, Support vector machine; Gradient boosting machine; Data-driven approaches
and reproduction in any medium,
provided the original work is
properly cited.
Publisher’s Note: AccScience
Publishing remains neutral with 1. Introduction
regard to jurisdictional claims in
published maps and institutional Heart failure (HF) is an increasing public health issue, impacting millions globally and
1
affiliations placing significant strain on healthcare systems. In recent decades, there has been a
Volume 9 Issue 1 (2025) 133 doi: 10.36922/ejmo.6583