CHARLOTTESVILLE, Va. (CBS19 NEWS) — A new study finds that adding artificial intelligence software to an artificial pancreas improves its efficiency.
The first-of-its-kind study at the University of Virginia Center for Diabetes Technology found AI has the potential to increase the device’s effectiveness in treating Type 1 diabetes.
According to a release, researchers compared an advanced experimental artificial pancreas system with an artificial pancreas algorithm incorporating AI.
The system automatically monitors and regulates blood sugar for people with Type 1 diabetes, whose bodies may be incapable of producing the insulin needed to break down sugar.
The AI-supported system was able to keep blood sugar levels within a patient’s target range for an almost identical amount of time as the advanced system, but the AI system reduced the computational demands.
The release says the resulting increased efficiency could allow developers to implement the system in devices with low computational power, such as an insulin pump.
“So far, this is the first clinical trial of a data-driven artificial pancreas system, which used an extensively trained neural network to deliver insulin automatically," said Boris Kovatchev, PhD, director of the UVA Center for Diabetes Technology.
This was a small study involving only 15 adults with Type 1 diabetes.
In back-to-back sessions, the patients used both systems for 20 hours each and followed their normal daily routines as closely as possible.With the AI-supported system, blood sugar levels remained within the target range 86 percent of the time, just slightly under the 87 percent time with the advanced artificial pancreas system.However, the AI-supported system reduced the computational demands six-fold, which means it would be more suitable for devices with low processing power.While more research is needed, the UVA team hopes that by incorporating AI, the algorithm would be able to adapt and improve based on data from thousands of users instead of an individual.“Neural-net implementation allows the algorithm to learn from the data of the person wearing the system,” Kovatchev said. “This opens the door to real-time, AI-driven personalized insulin delivery.”The release says these efforts build on earlier groundbreaking work at the UVA Center for Diabetes Technology to create a separate artificial pancreas […]
Study finds AI improves efficiency of artificial pancreas system for Type 1 diabetics