How smart watches and mobile apps could transform life for Type 1 diabetics
Associate Professors Steven Voida and Casey Fiesler, from the .
By Lisa Marshall (Jour, PolSci'94, MJour’22)
Photos byKimberly Coffin (CritMedia, StratComm’18)
In the not-so-distant future of Casey Fiesler’s imagination, the complex and burdensome task of managing Type 1 diabetes will be easier.
Instead of counting carbs, doing math and fiddling with her pump multiple times per day, she could essentially set it and forget it, only occasionally receiving nudges via her phone to take action.
When the geolocation device on her phone detects she’s at a restaurant, it might nudge her to snap a picture of what she’s ordering, then signal the pump attached to her abdomen to deliver just the right amount of insulin to metabolize her meal.
When her digital calendar shows she is at a Pilates class and her smartwatch senses she’s exercising, her pump might back off on delivery of the blood-sugar-lowering hormone.
Throughout each day, a sophisticated algorithm will hum silently along, combining data from her wearable sensors and mobile apps with survey data about her habits and real-time glucose monitoring to help keep her blood-sugar in check.
“Anything that helps people be able to think about their diabetes less would be an improvement in their quality of life,” said Fiesler, an associate professor in the Department of Information Science who was diagnosed three years ago. “It’s something you have to think about constantly.”
With a new $1.2 million grant from the National Institutes of Health, Fiesler, Information Science Associate Professor Stephen Voida and collaborators from computer science and CU Anschutz hope to develop a “person-centered artificial pancreas” that uses real-time cues from our devices to move us closer to replicating the real thing—all while taking into account privacy concerns patients might have.
“What can a smartwatch or phone sense about your physical location or your social context or your activity history to help the system be smarter?” asked Voida, who studies personal informatics. “We’re trying to make the system more aware of all the different things going on in the real world so that it can be a better partner in delivering insulin at appropriate times.”
A unique burden for young patients
Unlike with Type 2 diabetes, in which the body makes less insulin and grows resistant to it, with Type 1, the pancreas makes little to no insulin at all.
That’s problematic for the 1.3 million people who have it because insulin ushers sugar from the bloodstream into the muscles for energy. If we have too little insulin, glucose can linger in the bloodstream damaging tissues. If we have too much insulin or blood glucose levels dip too low, we can end up feeling weak, or worse.
The disease is typically diagnosed in childhood, tasking young people with a daunting obligation.
“They are suddenly responsible for thinking like a pancreas: They have to think about what their blood sugar is doing and whether or not they need insulin, and there’s not much room for error,” said Laurel Messer, an assistant clinical professor in the Department of Pediatrics at CU Anschutz and an original collaborator on the project.
Each year, a fair amount of patients end up in the hospital with a condition called diabetic ketoacidosis. Others have seizures from hypoglycemia. Tragically, some occasionally die.
“The amount of stress diabetes puts on top of already-stressful adolescence is tremendous,” said Messer, who recently took a position with Tandem Diabetes Care, a supporter of the project, to develop technological solutions to make their lives better.
The emerging artificial pancreas
Sriram Sankaranarayanan, a professor of computer science and project investigator, notes that diabetes care has already come a long way.
“Type 1 diabetes is one of the few diseases where the treatment could be technological,” he said. “It sits at the convergence of health, human behavior, computer science and mathematics.”
half of Type 1 diabetics, including Fiesler, already use continuous glucose monitors , which attach to the body to track blood sugar levels, minimizing the need for painful finger pricks.
Roughly two-thirds of adults with the disease use a pump, rather than self-injecting insulin multiple times per day.
Algorithmic advancements now enable the monitors to talk to the pumps in an automated system known as an artificial or “bionic” pancreas. The newest systems even allow users to check their levels and self-administer insulin from their cell phone. But what’s still missing is information from the outside world.
“The only thing the system knows without me telling it is what my blood sugar is,” Fiesler said.
That’s where the gadgets come in.
“Anything that helps people be able to think about their diabetes less would be an improvement in their quality of life. It’s something you have to think about constantly.
Casey Fiesler
Associate Professor
How technology could make life better
While existing artificial pancreases are helpful for those who can afford them, there’s still a significant lag time between what the glucose monitor senses and what the pump does.
As a result, individuals must calculate how many carbohydrates they plan to eat about a half-hour before each meal and manually plug it into the system.
They’re also encouraged to let their pump know before they exercise, as it tends to lower blood sugar. A host of other things, from alcohol consumption to stress to menstruation, can also influence the insulin-blood sugar equation.
Some youth interact with their pumps 20 or more times per day. Others forget or are too embarrassed to mess with it, prompting disruptive alarms to go off when blood sugar levels begin to drift outside predictable ranges.
“If you’re getting beeped all the time, it’s very easy to get alarm fatigue,” Messer said. “We need to find better ways to get this information to people and not make them guess about what to do.”
The team has already begun to survey young patients from the Barbara Davis Center for Childhood Diabetes to inform development of algorithms that can predict where blood sugar is headed, based on user habits and device data, and subtly nudge the user when needed.
For instance, a 20-year-old may be, according to her devices, feeling stressed and exhausted after studying all night and headed toward a blood sugar crash. The system might text her a reminder to eat a snack.
Sankaranarayanan doubts that a true set-it-and forget-it “bionic pancreas” will come any time soon: The real one can sense blood glucose level changes almost instantaneously and secrete fast-acting insulin in real time.
Replicating that will require not just technological fixes, but also faster-acting insulin and other medical advances.
But ultimately, Voida and Fiesler believe the person-centered artificial pancreas can paint a more holistic picture of what’s going on in a user’s life, incorporating things they might not even notice.
“This would enable the system to sort of get out of your way when you live your life without you having to continually do this dialogue back and forth,” Voida said.
Fiesler, who studies privacy and ethical issues surrounding digital technologies, is particularly interested in what young people will be willing to share, or not, in exchange for such a luxury. She’ll explore these questions through youth surveys.
For now, she’s incredibly grateful that she was diagnosed amid such a renaissance in diabetes innovation.
“I have trouble imagining what it was like to manage this before all of this technology,” she said.
Now she’s imagining an even brighter future, and through her work, helping to make it happen.