Stephen Voida, PhD, Associate Professor of Information Science at University of Colorado Boulder talks about a new bionic pancreas for type 1 diabetes.
Interview conducted by Ivanhoe Broadcast News in 2024.
Explain what life is like for people with type 1 diabetes. Does it require a lot of constant day-to-day monitoring?
Voida: Having type 1 diabetes is challenging, especially for young adults who are diagnosed with it. And this is common because, frequently, people are diagnosed with type 1 when they’re either pre-adolescent or adolescent-aged. We’re working with a lot of the pediatric diabetes faculty members at the CU Anschutz Medical Center. These are clinicians who are helping teenagers figure out how to manage this new health challenge. This involves a huge amount of data work, in that they must understand what their blood sugar is doing at any given point in time. They must be thinking about whether they need to be responding to that with either a snack to bring their blood sugar up or insulin to bring their blood sugar down. It’s hard because they must be doing this at all times of the day, at school, at work, while they’re doing sports, at night, when they’re staying up late to do homework, and overnight when they’re sleeping. A lot of these kids have parents who have to wake up at two in the morning and check their blood sugar so that they’re not hitting a low and putting themselves at risk for medical conditions overnight.
Day-to-day life for type 1 diabetes is rapidly changing with the help of technology, correct?
Voida: There’s been a huge sea change in what it’s like to be an individual with diabetes in the last few years. We’ve gone from the traditional treatment and care, which is finger sticks to check your blood sugar and injections to add insulin when the numbers are too high, to continuous glucose monitoring—those sensors that you often see attached to the upper arm. These devices constantly check glucose levels and report it to a smartphone, smart device, or the cloud. Recently, there’s been the development of what’s called artificial pancreas devices, which are pumps that are attached to the abdomen that deliver insulin continuously at a low level throughout the day that helps the body respond and regulate. Then, when there is a mealtime or exercising or things that are going to change the blood sugar substantially, the pumps respond with additional insulin delivered directly into the bloodstream. There are no needles, there’s no necessity to constantly be carrying that equipment around. It’s all built into the device that’s attached to the body.
Is this where the bionic pancreas comes in? Explain what this is and how it works.
Voida: Our project, in collaboration with the clinicians at the CU Anschutz Medical Center, is to make these devices a little smarter. Right now, the devices monitor blood sugar, and they respond by delivering more or less insulin based on that one number, what the blood sugar happens to be at a moment in time. Blood sugar is one of these things that’s slow to respond. If you eat, it takes a little while for your blood sugar to ramp up. If you exercise, it takes a little while for your body to start metabolizing those sugars. What we’re trying to do is take other signals from electronic devices that teens are already carrying and make the device smarter—to make it know a little more about where they’re at, what they’re going to do, and how aware they might be about their blood sugar at a given moment to help the pump provide better insulin delivery. Another important benefit is in helping to reduce the interruptiveness and annoyingness of the alerts that these devices often give off, notifications that their blood sugar is getting too low or it’s getting too high. If we know the teenager is in a restaurant and they’re about to order food, there’s no reason to have the device alert them that their blood sugar is getting too low. They already know that!
How does the robotic part come in?
Voida: I wouldn’t say robotic. I think what we’re doing is we’re giving the pumps the ability to look at new kinds of data, and we’re adding algorithms that can pull those bits of data together. Instead of just looking at blood sugar and whether is it going up or down, we’re looking at what’s the blood sugar, what’s their location, what’s on their calendar, who else is around them, what has their routine been like on prior Thursdays if today is a Thursday? What are all these different factors that we can start to pull together? We think of it in the same way that a robot or an AI algorithm uses multiple sources of data to come to a richer understanding of what’s going on. That’s what we’re looking at: adding more sensing and intelligence to these devices to give them a better ability to predict blood sugar and sound alerts more appropriately.
Is this a function that an individual would use? Would it be prescribed, and would it need to be monitored by the person or does it work automatically?
Voida: Right now, the way that these devices work is that a doctor will prescribe a pump and a continuous glucose monitor—a CGM. An individual managing diabetes attaches both devices to their body and refills the pump with insulin every few days. The CGM collects their blood sugar data and wirelessly communicates it to the pump, which responds by sending a small amount of insulin directly into the body. A lot of these devices are getting a little smarter. They send the blood sugar and insulin data to the cloud so that when the teenager checks in with their clinician, the doctors, nurses, and dieticians can also see what has been happening and how the pump has been responding. In a lot of the clinical visits that we’ve observed, a lot of the discussion is about “How has your pump been working? Have you adjusted its settings? Is it delivering the right amount of insulin? Are you still getting highs and lows?” It’s really about fine-tuning what that pump will do. One of the interesting questions with this research, especially for young adults who have a fair amount of savviness with digital technology, is how much information are they interested in letting these pumps see about their lives? Is there agreement that accessing data about everyday life is appropriate? What happens when that data is recorded by the company that operates the pump as part of their cloud services? How much of those data do they want a clinician to see? How much access do they want to share with a parent? These are all really interesting issues, and we think there are some important trade-offs to be had between providing better insulin delivery and blood glucose management and leaning on all this extra data. Who gets to see it? Who gets access to it? How much information does the pump need to know about teenagers’ lives and what they’re doing? How comfortable are young adults in potentially having some of that information become part of their clinical medical record at some point?
It’s like you don’t want mom and dad you had two shots of something before class.
Voida: Or, you know, that you were just up until two in the morning working on a homework assignment when they thought you were done with it. It can be subtle things. And it’s an issue with all of these digital devices: they create a record of what we’re doing. Especially when you’re a teenager, and you’re figuring out who you are and what your identity is going to be, and how you’re going to live your life, you don’t necessarily want somebody peeking over your shoulder the whole time. You want to be able to do some of that identity formation on your own and make clear decisions about who gets to see that happen.
Because this is at a developmental stage, those decisions about how the system will work haven’t been determined yet, right?
Voida: Correct. A lot of what the research is about working with clinicians, parents, and patients to ask: What are the trade-offs? What are you comfortable with? What are you not comfortable with? Who do you want to be able to see this? When do you want us to ask you those questions? How do we set up the defaults so that it’s appropriate? The aim is that as these capabilities become part of commercial products in the future, we can hand the findings off to the companies that build the pumps so that they know where to set these boundaries and create algorithms that work well for the patients and the clinicians and the families.
Are there other artificial pancreases? How is this different from existing products?
Voida: My clinical partners would have a better answer to this, but I know that there are three or four artificial pancreas devices already on the market. I know that one is made by Medronic, one is made by a company called Tandem Diabetes, and one is made by a company called Beta Bionics. These are all devices that hold a small amount of insulin and attach to the body to deliver the insulin directly to the bloodstream. They are paired with one of these continuous glucose monitors, so they can tell what your blood sugar is doing at any point in time. Together, it creates what we call a hybrid closed-loop system. It’s a closed loop because it reads your blood sugar, and it responds with insulin that helps to nudge the blood sugar readings back into an appropriate range, closing the loop. Right now, the devices don’t know anything else unless the patient steps in and says, “I’m about to eat,” “I’m about to exercise,” “I’m about to sleep.” It’s the kind of system where the person using it has to stop and take action or respond to what’s going on any time a change is about to happen. Otherwise, these systems are not particularly “smart” because they just have this one kind of data that they’re working with, which is how much blood sugar is floating through the person’s body at a given point in time. At best, they have a little bit of history for how that number is changing.
Your approach goes a little bit deeper with the data than existing systems. Your team calls it a person-centered artificial pancreas. Explain that.
Voida: We call it a person-centered artificial pancreas device because we want to look at the whole person, what a person is doing, what a person knows, what a person is expecting to happen next. It’s more than just a physiological device, which is what patients have now. We’re looking more holistically at what are all the different facets of what a person is doing with their day, what they want to accomplish next, and who they’re around, as a richer context for doing the right thing with interpreting blood sugar measurements and delivering insulin.
How does this change patients’ lives, in your opinion?
Voida: What we’re hoping for is that this does two things. First, it reduces the amount of overhead in terms of the individual having to stop what they’re doing and say, “I’m going to do this next” to the pump. It helps the pump to anticipate things, so that they don’t have to be constantly declaring to it, “I’m about to eat,” “I’m about to work out,” “I’m about to sleep.” If we pay attention to these additional signals, the pump will be able to anticipate some of those changes more effectively. Second, we hear a lot from clinicians and patients both about alarm fatigue. These devices are regulated by the FDA, and so there are a bunch of rules about when and how they have to notify the wearer what they’re doing. If the wearer’s blood sugar is critically low, current devices are required by FDA regulation to beep audibly. There’s no way to turn that off. If the wearer’s blood sugar is higher than a particular level, current devices have to alert. If the wearer’s insulin is about to run out, they have to alert. What we hear from patients is they just get exhausted from constantly hearing these things go off. If we can help these devices to be smarter about understanding what’s going on, where the wearer is at, and what they’re doing, we can argue for reducing the intrusiveness of these alerts if the pump has a good idea that the wearer is already aware of the situation. But even better for teenagers, if there’s something that we can sense about the wearer’s social context, for example, that they’re in a crowded space, then the alert could automatically switch to vibrating that would be more appropriate than beeping out loud. Or if they’re located in a concert hall, perhaps delivering a silent alarm on a smartwatch would be better than an audible alarm. How do we deliver those alerts in a way that meets the FDA’s regulatory requirements but is better for the patients who have to have these things on their body 24 hours a day?
If you’re in a location where it’s been set up that you don’t want it to be audible, it wouldn’t beep, right?
Voida: Location is an important cue. We’re excited thinking about what sensing location could enable these pumps to do.
I wonder if Google will want a piece of that financially.
Voida: If you have an Android phone, or if you run a Google app on an Apple phone, these companies already do know where you’re at. They know where you’re spending your time. They know where you call home and where you work. They know that you’re about to leave and they use this information to tell you the traffic times. Since location information is already being collected by a lot of these companies, how do we put it to use in a way that is more person-centered? Our idea is to use location data in a way that has value for people, especially for managing a chronic illness, which is a hugely inconvenient, time-consuming, attention-grabbing thing to have to deal with. How do we use data about people’s physical and social context in a way that’s going to be of benefit to the patients in the end?
I think you’ve already covered this, but instead of stopping and doing a blood prick, you just look at your smartwatch, right?
Voida: That’s exactly right. Existing devices are making it so that someone can glance at their smartwatch and see their current blood sugar levels. But that’s just one bit of information. There are a bunch of other things that the smartphone, the smartwatch, and other digital devices know. How do we make that information equally accessible to individuals managing diabetes, and to their insulin pump?
For people who have loved ones, your children, or adults – I want to hear your thoughts on type 1 diabetes.
Voida: The story for parents is important, especially when we’re talking about families with younger kids. That is, for families that receive a diagnosis for a child that’s seven or 10 years old. Having access to information about what’s raising and lowering blood sugar and having quick access to that information to be able to deal with blood sugar changes is really important, especially when the parent and the child aren’t in the same place, for example, when the child is at school or at camp. That’s such a valuable tool for helping parents to feel comfortable and confident in how the family is going to be able to treat this condition. One of the interesting places where we’re seeing this play out is when teenagers get a little bit older and they’re now off doing their own thing more: they’re driving themselves to sports practice, they’re taking on more independence, and they’ve got more distance from their family. Having these technologies that are both more robust and that the parents can trust to say, “we know that this is going to be helping to take care of my kids, to provide that information when I need it.” But also, to just give them a sense of confidence that diabetes management is working out well is important. The other interesting pivot point that we’re seeing is when kids go off to college, they must take responsibility and independence for managing their diabetes for the first time. It’s a whole new environment, it’s a whole new eating schedule. It’s exposure to new places and new people. Thinking about how a device could be smart about helping young adults to make that transition, to be independent on their own, about managing their chronic condition—that’s super interesting, and it’s one of the things that we’re really excited to be looking at this in a university context. We have students who are in that very situation that we can be working with here to ask questions about what makes the most sense for them.
What do we expect in the future?
Voida: It’s interesting how quickly this technology landscape is changing. These artificial pancreas devices have only appeared on the market in the last five to 10 years, so this is a brand-new way to treat diabetes. It’s changing clinical practice guidelines for our medical colleagues, and the device capabilities are evolving incredibly quickly, as well. Even since we began this research project, we’ve seen the device capabilities expand. A couple of years ago, a person managing diabetes had to push buttons on the pump device to request additional insulin before a meal. Now, they can do that through a smartphone app or using a smartwatch. The companies building these devices are innovating constantly. We’re really privileged to have good working relationships with insulin pump device manufacturers. As we discover ways to augment the capabilities of these devices, we can get those new ideas into patients’ hands quickly. The companies are always on the lookout for innovations that will help them to deliver better devices. One of our partners in this research, Tandem diabetes, is excited to start to integrate our algorithms and our data management practices into their apps and their pumps. Our goal is to help get these products into the doctor’s office and as part of diabetes clinical visits as soon as possible.
END OF INTERVIEW
This information is intended for additional research purposes only. It is not to be used as a prescription or advice from Ivanhoe Broadcast News, Inc. or any medical professional interviewed. Ivanhoe Broadcast News, Inc. assumes no responsibility for the depth or accuracy of physician statements. Procedures or medicines apply to different people and medical factors; always consult your physician on medical matters.
If you would like more information, please contact:
Daniel Strain
Sign up for a free weekly e-mail on Medical Breakthroughs called First to Know by clicking here
