From Data to Decisions: The AI Revolution in Diabetes Care

Abstract

Diabetes Mellitus (DM) is a prevalent chronic disease that significantly increases the risk of developing other conditions such as ischemic heart disease, diabetic nephropathy, and atherosclerosis. This literature review investigates the application of artificial intelligence (AI) and machine learning (ML) in predicting and managing diabetes. The objective of the review is to explain how Artificial Intelligence and Machine Learning are currently employed in the provision of healthcare services with a specific focus on their application in diagnosing, predicting, countering diabetes through therapy. This paper presents a detailed stepwise systematic analysis based on PRISMA guidelines that sought to identify, choose, and assimilate select research works. For this review, 122 studies were reviewed out of 1235 articles first pulled from databases like PubMed, Google Scholar, Scopus, and IEEE Xplore. The results indicate that AI-driven predictive models significantly enhance risk assessment accuracy for diabetes management, achieving an area under the curve (AUC) of 0.85 for predicting the onset of type 2 diabetes. These models integrate genetic and environmental factors to improve prediction precision. Additionally, AI-based diagnostic tools, including image recognition for diabetic retinopathy, show high sensitivity (90.3%) and specificity (98.1%). The review highlights the need for thorough ethical and policy frameworks to ensure the safe and responsible implementation of AI in diabetes care. The findings suggest that audits of AI algorithms and the promotion of interoperability among AI systems are crucial for advancing AI-driven diabetes management. These insights contribute to policy development, clinical practice, and future research in the field. The conclusions emphasize the need for robust ethical frameworks and interdisciplinary collaboration to facilitate the effective implementation of AI and ML in healthcare systems.