Volume 11, Issue 2 (summer 2025)
JMIS 2025, 11(2): 0-0 |
Back to browse issues page
Ethics code: IR.US.PSYEDU.REC.1404.007
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Ronaghi M H, Malekpoor A R. Applications of Data Mining in Healthcare: A Systematic Review. JMIS 2025; 11 (2)
URL: http://jmis.hums.ac.ir/article-1-606-en.html
URL: http://jmis.hums.ac.ir/article-1-606-en.html
Department of Management, School of Economics Management and Social Sciences, Shiraz University, Shiraz, Iran.
Abstract: (797 Views)
Introduction: This study aimed to systematically review the applications of data mining in disease prediction and diagnosis within the healthcare domain.
Information sources or data: A systematic search was conducted in PubMed, Scopus, ScienceDirect, and Magiran databases for studies published between 2010 and 2025.
Selection methods for study: This study was conducted based on the PRISMA 2020 guidelines. After removing duplicate studies, the titles, abstracts, and full texts of articles were screened according to the inclusion and exclusion criteria, and finally, 44 studies were included in the analysis. Study quality was assessed using the Joanna Briggs Institute checklist, and data synthesis was performed using a qualitative narrative approach.
Combine content and results: The findings showed that data mining was most frequently applied to the prediction and diagnosis of cardiovascular diseases, diabetes, kidney diseases, cancers, and other chronic conditions. The data used mainly consisted of structured clinical data and electronic health records. Commonly used algorithms included decision trees, Naïve Bayes, support vector machines, neural networks, and random forests. More recent studies, in addition to predictive accuracy, focused on model interpretability, feature selection, and the clinical applicability of results using methods such as SHAP and LIME. Furthermore, no single algorithm demonstrated absolute superiority across all problems, and model performance depended on disease type, data quality, and preprocessing methods.
Conclusion: The results of this systematic review showed that data mining can play an effective role in early disease diagnosis and clinical decision-making support. However, challenges such as data heterogeneity, information quality, and limitations in clinical implementation still remain. The development of data infrastructures, standardization of health information, and greater attention to model interpretability are essential for the effective use of data mining in healthcare systems.
Information sources or data: A systematic search was conducted in PubMed, Scopus, ScienceDirect, and Magiran databases for studies published between 2010 and 2025.
Selection methods for study: This study was conducted based on the PRISMA 2020 guidelines. After removing duplicate studies, the titles, abstracts, and full texts of articles were screened according to the inclusion and exclusion criteria, and finally, 44 studies were included in the analysis. Study quality was assessed using the Joanna Briggs Institute checklist, and data synthesis was performed using a qualitative narrative approach.
Combine content and results: The findings showed that data mining was most frequently applied to the prediction and diagnosis of cardiovascular diseases, diabetes, kidney diseases, cancers, and other chronic conditions. The data used mainly consisted of structured clinical data and electronic health records. Commonly used algorithms included decision trees, Naïve Bayes, support vector machines, neural networks, and random forests. More recent studies, in addition to predictive accuracy, focused on model interpretability, feature selection, and the clinical applicability of results using methods such as SHAP and LIME. Furthermore, no single algorithm demonstrated absolute superiority across all problems, and model performance depended on disease type, data quality, and preprocessing methods.
Conclusion: The results of this systematic review showed that data mining can play an effective role in early disease diagnosis and clinical decision-making support. However, challenges such as data heterogeneity, information quality, and limitations in clinical implementation still remain. The development of data infrastructures, standardization of health information, and greater attention to model interpretability are essential for the effective use of data mining in healthcare systems.
Type of Study: Research |
Subject:
Special
Received: 2025/11/29 | Accepted: 2026/02/1 | Published: 2025/06/22
Received: 2025/11/29 | Accepted: 2026/02/1 | Published: 2025/06/22
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |