The Role of Explainable AI in Revolutionizing Human Health Monitoring: A Review
Authors:
Abdullah Alharthi,
Ahmed Alqurashi,
Turki Alharbi,
Mohammed Alammar,
Nasser Aldosari,
Houssem Bouchekara,
Yusuf Shaaban,
Mohammad Shoaib Shahriar,
Abdulrahman Al Ayidh
Abstract:
The complex nature of disease mechanisms and the variability of patient symptoms pose significant challenges in developing effective diagnostic tools. Although machine learning (ML) has made substantial advances in medical diagnosis, the decision-making processes of these models often lack transparency, potentially jeopardizing patient outcomes. This review aims to highlight the role of Explainabl…
▽ More
The complex nature of disease mechanisms and the variability of patient symptoms pose significant challenges in developing effective diagnostic tools. Although machine learning (ML) has made substantial advances in medical diagnosis, the decision-making processes of these models often lack transparency, potentially jeopardizing patient outcomes. This review aims to highlight the role of Explainable AI (XAI) in addressing the interpretability issues of ML models in healthcare, with a focus on chronic conditions such as Parkinson's, stroke, depression, cancer, heart disease, and Alzheimer's disease. A comprehensive literature search was conducted across multiple databases to identify studies that applied XAI techniques in healthcare. The search focused on XAI algorithms used in diagnosing and monitoring chronic diseases. The review identified the application of nine trending XAI algorithms, each evaluated for their advantages and limitations in various healthcare contexts. The findings underscore the importance of transparency in ML models, which is crucial for improving trust and outcomes in clinical practice. While XAI provides significant potential to bridge the gap between complex ML models and clinical practice, challenges such as scalability, validation, and clinician acceptance remain. The review also highlights areas requiring further research, particularly in integrating XAI into healthcare systems. The study concludes that XAI methods offer a promising path forward for enhancing human health monitoring and patient care, though significant challenges must be addressed to fully realize their potential in clinical settings.
△ Less
Submitted 26 February, 2025; v1 submitted 11 September, 2024;
originally announced September 2024.
