Big Data Takes On Type 1 Diabetes

The effect that food has on blood glucose levels is well established, but the role played by other factors, such as activity, stress and time of day, is less clear. That lack of knowledge is an obstacle to the advancement of care for Type 1 diabetes, a chronic disease that can cause serious health issues, from blindness to organ failure.

Now, a multidisciplinary research team at the University at Buffalo is looking to artificial intelligence (AI) to fill in those knowledge gaps and give diabetics more control over their disease.

Type I diabetes occurs when the pancreas doesn’t produce enough insulin to regulate blood glucose, requiring frequent monitoring. Until recently, diabetics had to perform a finger stick several times a day to test their blood sugar.

The development of artificial pancreas device systems in the past few years has made management of the disease easier. These systems mimic a healthy pancreas by continually monitoring glucose levels, typically through a tiny sensor inserted under the skin, and delivering insulin accordingly. But due to shortcomings with the modeling formulas that drive them, they can perform poorly.

“We’re developing new tools—combining data collected from diabetes-monitoring tools with AI systems as well as traditional time-series modeling approaches,” said the project’s leader, Tarunraj Singh, a professor of mechanical and aerospace engineering at UB.

This novel hybrid approach will allow for a more precise assessment of the many subtle but critical variables involved in the body’s management of blood glucose, as it reacts not just to food intake but to exercise, stress, time of day and other factors.

Ultimately, said Singh, “it could greatly improve how people manage their Type I diabetes.”