12nd year MBBS student, Department of Pharmacology, Rajarajeswari Medical College & Hospital, Bengaluru, Dr. M.G.R Educational and Research Institute (University). Email: priyankaviswanath007@gmail.com
2*Assistant Professor, Department of Pharmacology, Rajarajeswari Medical College & Hospital, Bengaluru, Dr. M.G.R Educational and Research Institute (University). Corresponding author. Email: prernatejaswi@yahoo.in
Introduction: Diabetes is a widespread lifestyle disease requiring early diagnosis and regular monitoring to prevent complications. Traditional methods may miss early warning signs or depend heavily on patient compliance. Artificial Intelligence (AI) offers tools that can predict risk, support diagnosis, and guide daily management. This review surveys current applications of AI in diabetes care.
Objectives: To assess the role of AI in the early detection of diabetes, treatment decisions and improves blood sugar control and manage diabetes in daily life, including apps and monitoring systems.
Methodology: This review conducted a systematic search of published studies on AI use in diabetes care. Databases including PubMed, Google Scholar, and Scopus were searched using terms such as "AI," "machine learning," and "diabetes." English-language articles focusing on diagnosis, prediction, or management were included. Titles and abstracts were screened, followed by full text evaluation. Duplicate and unrelated studies were excluded. Key findings were extracted and categorized into themes, early detection, treatment support, and long term management.
Results: Synthesis of 17 peer-reviewed studies identified three primary domains of impact. In Domain I (Diagnosis), machine learning models achieved Area Under the Curve (AUC) values >0.75 for early risk prediction, while deep learning (CNNs) reached specialist level diagnostic sensitivities of 90.5%–91% globally and up to 95.8% in Indian smartphone based cohorts (AUC 0.94–0.99). In Domain II (Treatment), Hybrid Closed-Loop systems significantly increased Time in Range (TIR) by an average of 11% (approx. 2.6 hours/day) and enhanced glycemic stability via predictive titration algorithms. In Domain III (Management), AI-enabled digital companions and decision support platforms improved monitoring consistency, provided real time emergency alerts for glycemic excursions, and reduced the cognitive burden of self-management through multi modal data fusion.
Conclusion: AI demonstrates strong potential to enhance early diagnosis, improve treatment decision making, and support ongoing self-management in diabetes care. Predictive algorithms, automated image based screening, and intelligent monitoring tools reduce diagnostic delays and support more stable glycaemic control. However, wider validation and real world implementation studies are needed before large scale clinical adoption.
Keywords: Artificial Intelligence; Diabetes mellitus; Machine learning; Diagnosis; Management; Digital health; Predictive modelling.
Conflict of interest: None
How to cite this article: Viswanath P, Tejaswi P. Artificial Intelligence in Diabetes: A Systematic Review of its Role in Early Diagnosis, Treatment Optimization, and Disease Management. Int J Drug Deliv Technol. 2026;16(4s): 111-116. DOI: 10.25258/ijddt.16.4s.14
Source of support: Nil.