UVA to test AI-powered insulin delivery to improve diabetes care

Author(s):

Austin Littrell

Key Takeaways

The trial explores AI's potential to enhance insulin delivery for Type 1 diabetes, aiming to improve blood sugar control and reduce management burdens.
The FDA-approved "Bolus Priming System with Reinforcement Learning" integrates with the existing AIDANET system, including a smartphone app, glucose monitor, and insulin pump.
Sixteen adults will participate in a three-week study to compare the AI-powered system's effectiveness with current technology in maintaining blood sugar control.
The trial seeks to make diabetes treatment more accessible and cost-effective by reducing the need for constant user input, easing the mental and financial burden.

UVA researchers test AI technology to enhance automated insulin delivery and improve blood sugar control for patients with T1D.

A new clinical trial at the University of Virginia (UVA) is exploring the potential of artificial intelligence (AI) to revolutionize insulin delivery for patients with Type 1 diabetes (T1D). The trial, led by UVA’s School of Data Science and the Center for Diabetes Technology, is evaluating an AI-powered enhancement designed to improve automated insulin delivery systems.

With recent approval from the U.S. Food and Drug Administration (FDA), the research team is set to test the “Bolus Priming System with Reinforcement Learning” (BPS_RL), which integrates with the existing Automated Insulin Delivery Adaptive NETwork (AIDANET), which includes a smartphone app, Dexcom glucose monitor and Tandem insulin pump.

The goal is to determine whether this new system can improve blood sugar control — particularly during meals and overnight — while reducing the burden of diabetes management for patients.

“This trial isn’t just about advancing technology — it’s a bold step toward transforming diabetes care and uplifting lives,” said Heman Shakeri, PhD, assistant professor of data science and co-lead of the study. “We are committed to creating a fully automated, intelligent insulin delivery system that redefines diabetes management, making treatment simpler, more reliable and entirely effortless for patients.”

The UVA trial aims to alleviate the many challenges faced by patients with T1D by testing a system that adapts to patients’ needs in real-time with minimal user intervention.

The trial

Sixteen adults who already use an Automated Insulin Delivery (AID) system will participate in the three-week study, which kicks off in March 2025.

Week 1: Participants use the standard AIDANET system at home to establish a baseline.
Week 2: Participants stay at a supervised testing location, alternating between the standard and AI-enhanced systems during 18-hour sessions.
Week 3: Participants return home to test the upgraded system under remote monitoring.

Researchers intend to analyze how well the AI-powered system maintains blood sugar control compared to existing technology. Half of the participants will start with the standard system before switching to the enhanced version, while the other half will follow the reverse order.

Moving forward with diabetes care

Beyond improving glycemic control, the trial looks to make diabetes treatment both more accessible and cost-effective. By reducing the need for constant user input, AI-driven systems could ease the mental and financial burden of living with diabetes.