Ali Cinar, director of Illinois Tech’s Engineering Center for Diabetes Research and Education, praises the first United States Food and Drug Administration-approved commercial artificial pancreas (AP) system, which was offered to the public in 2017. But he says the technology behind it can be compared to “cruise control.”
“What we are after is the equivalent of the driverless car,” says Cinar, who has been developing a fault-tolerant, next-generation multivariable artificial pancreas system since 2009 that automatically monitors and infuses insulin according to metabolic changes that occur in response to food intake and various types of exercise, from solo fitness efforts to organized sports. With a grant from the JDRF, he and his team of collaborative researchers are now further refining the algorithms that control their AP system by factoring in additional physiological signals from wearable devices that indicate future variations in blood-glucose concentration.
“To get to the whole picture, we have to check the physical activity, not exercise as a voluntary and well-known activity, but all kinds of physical activity,” he says. “Say you’re sitting in a room and hear a child shout in the next room. There is an element of surprise or stress and then an element of physical activity where you would run into the room to figure out what is going on.” He explains that “the body has an increased sensitivity to insulin when people are exercising, so for a certain amount of insulin, the amount of glucose that would be taken from the bloodstream into the cells would be increased as compared to the non-exercise state. But stress has the opposite effect on insulin sensitivity. If we were using just a single physiological signal such as heart rate and had made the assumption that all heart rate increases indicate exercise, then in the case where there was stress, we would be giving an insulin dose in the opposite direction of what the dose should have been.”
For his projects Cinar has been collecting data from people with type 1 diabetes through collaborators at the University of Chicago (UofC), the University of Illinois at Chicago (UIC), and York University (Toronto). During summer 2016 he also obtained real-world exercise data from kids participating in a basketball camp coordinated by the organization Slam Dunk for Diabetes. In addition to wearing armbands, the young players also wore chest bands with built-in sensors. Cinar and Nicole Frantz (BME ’16), one of his biomedical engineering Ph.D. students, analyzed the data to help determine the accurate dispensing of insulin via Illinois Tech’s AP system. In the current study, Illinois Tech students will be utilizing a variety of new exercise equipment in the controlled environment of Cinar’s lab to streamline the positioning of the sensors and to conduct more experiments utilizing different types of exercise. As collaborators, the UofC and UIC teams will conduct additional experiments with people under free living conditions.
“The FDA-approved AP system relies only on glucose information and requires manually entered meal and information and exercise. When it senses that it cannot handle things automatically, it turns control over to the patient,” says Cinar. “Our system does not expect any manual entries from the user. By using the additional information provided by physiological data collected from a wristband, we will be in a position to make more accurate decisions on what glucose values will be in the near future and how we should adjust insulin to make sure that while glucose is changing, insulin would be there to moderate those changes and keep glucose within the desirable range.”