A groundbreaking AI-based solution could transform the way young adults with Type 1 diabetes (T1D) manage their condition, relieving some of the daily burdens and cognitive strain often associated with the chronic disease.
In 2015, Sam King’s 9-year-old son was diagnosed with Type 1 diabetes, a moment that reshaped his perspective on both parenthood and diabetes care. At the time, King was working in the tech industry, but he quickly realized how complex and exhausting diabetes management could be.
“The daily burden of diabetes care is really amazing and extensive,” says Dr. Stephanie Crossen, a pediatric endocrinologist at UC Davis Health. “It’s a constant mental drain if you want to do it well and achieve good results.”
Around 2 million Americans live with T1D, including more than 300,000 children and adolescents. The autoimmune disease impairs the pancreas’s ability to produce insulin, a hormone necessary for glucose metabolism. As a result, individuals must monitor their blood glucose levels continuously, adjusting for meals, exercise, stress, and other factors to maintain balance.
King’s family had managed his son’s diabetes effectively for years, but he still encountered difficulties, especially as his son grew into his teenage years. There were moments when insulin doses were forgotten or check-ins neglected.
Then, in a twist of fate, King himself was diagnosed with Type 1 diabetes. As he adapted to managing the condition in his own life, he identified a significant gap in the technology available for diabetes care, particularly with continuous glucose monitors (CGMs). These devices, which use sensors placed on the body to track glucose levels every five minutes, alert users when their blood glucose is out of range. However, King noticed that by the time he received an alert, it was often too late to prevent a dangerous imbalance.
“I wanted software that would give me a little nudge before things went badly so I could take action early,” King says.
This realization led to the creation of BeaGL, a “metabolic watchdog” designed to help patients stay ahead of potential glucose imbalances. The AI-driven software works alongside CGMs but incorporates predictive machine learning algorithms to forecast glucose fluctuations before they occur. The alerts are delivered to users via a smartwatch, offering a more proactive solution to managing diabetes.
The goal of BeaGL is to automate diabetes management as much as possible, reducing the cognitive load on patients—especially adolescents, who often struggle to balance diabetes care with other life priorities. Dr. Crossen, a co-principal investigator on the project, believes that while technologies like CGMs and insulin pumps have made significant advancements in diabetes care, they are not yet fully optimized for young adults.
“CGMs, insulin pumps, and smart pumps have been amazing breakthroughs in diabetes technology over the last decade, but we haven’t yet optimized them for this age group,” she explains. “They have competing priorities, immature decision-making skills, and a lot of other life management tasks to juggle on top of diabetes care.”
In June 2024, six UC Davis students with T1D began using BeaGL in a pilot study. Early results have been promising, with participants reporting a reduced cognitive load despite receiving frequent alerts. One student, for example, was able to complete an experiment in a lab setting without interruption, thanks to BeaGL’s predictive alerts. These notifications allowed the student to finish their work before managing his glucose levels.
“He was able to manage his time better, knowing he had a few extra minutes before needing to address his blood sugar,” King says. “Once participants trusted the system, they spent less time worrying about their glucose levels because BeaGL was constantly monitoring them.”
Another feature students appreciate is the ability to customize alert frequencies. Initially, King’s team had trouble creating an algorithm that struck the right balance between too many and too few notifications. After receiving feedback from users, they made adjustments to give patients more control over their alert intervals.
Additionally, user feedback has influenced practical adjustments, like the ability to analyze sensor data locally on a phone instead of relying on cloud-based servers. One student, for instance, was able to use the system during a canoeing trip in an area with no cellular service, demonstrating BeaGL’s adaptability to real-world situations.
As BeaGL moves toward the market, King and Crossen have a larger vision for its future: a closed-loop insulin delivery system powered by AI. In this system, BeaGL would not only monitor glucose levels but also administer insulin doses autonomously, based on the data it collects. This would significantly reduce the burden on patients, allowing them to live more typical lives without the constant need for manual blood sugar checks or insulin adjustments.
“This is the ultimate goal the international diabetes community is working toward,” Crossen says. “If we can make a fully automated insulin delivery system a reality, it would be a game-changer, particularly for young adults. They wouldn’t have to be their own pancreas anymore.”
While the technology is still in development, King and Crossen are optimistic. Early-stage testing with an insulin-dosing model has shown promising results. However, for a fully functional closed-loop system to reach market, it would need to undergo extensive trials, FDA approval, and careful regulatory processes.
King believes that BeaGL represents just the beginning of what could become a new era of AI-driven healthcare. He envisions a future where AI systems monitor and predict health conditions in real-time, across multiple areas of medicine.
“What we’re looking at is an extreme version of distributed healthcare driven by AI,” King says. “We’re providing a glimpse into the future of healthcare, and the potential applications are limitless.”
Related topics:
New Data Resource to Support Diabetes Research in Arkansas
Diabetes Cases Reach Record High in UK Amid ‘Hidden Health Crisis’ Warning
