Artificial Intelligence-Powered Glucose Prediction and Diabetes Management Investigated

A newly published systematic review examined the performance and clinical outcomes of artificial intelligence and machine learning models in glucose prediction and the improvement of glycemic management in diabetes patients.

Introduction and Background

Diabetes mellitus is a major chronic metabolic disorder requiring continuous glycemic management due to varying glucose responses among individuals.

Artificial intelligence has emerged as a potential approach to analyze complex healthcare data and support personalized glycemic management.

Glucose Monitoring Data

Continuous glucose monitoring systems provide a richer representation of daily glycemic patterns, and artificial intelligence models combine these data with features such as physical activity, dietary intake, and insulin use.

In the Maastricht Study, models based on CGM data predicted glucose at 15 and 60 minutes with root mean square error values of 0.19 and 0.59 mmol/L, respectively.

Model Performances

In machine learning models developed on 440 patients with type 2 diabetes, the LightGBM model achieved an area under the curve value of 0.869 for nocturnal hypoglycemia.

Evaluated in 118 adults with non-insulin-treated type 2 diabetes, an artificial intelligence-based digital therapeutic was associated with a 0.67 percentage point reduction in HbA1c after three months.

Review Methods

This systematic review was conducted in accordance with PRISMA 2020 guidelines, and adults with diabetes mellitus were included as the population.

A comprehensive systematic literature search was performed in the PubMed, Scopus, and Web of Science databases up to July 2026.