Kyle Rose is a former Team Type 1 cyclist and business development director for that group. He is now living in France, where he founded a consultancy called Delta PM Diabetes, focusing on chronic disease management in health care. Mainly, he advises companies in the diabetes management business on campaigns for products including insulin, drug delivery devices, and glucometers, as well as non-profit collaboration initiatives and medical/patient community outreach.
But recently, Kyle got the historical chance to be part of one of the first-ever trials of an artificial pancreas system outside of a clinical setting. He was was part of the “AP@home” European study, in conjunction with the JDRF Artificial Pancreas Project. And we — after bumping into Kyle at the recent ADA Conference — got the historical chance to learn all about it:
Like professionals in any industry, those of us working in the diabetes sector are often asked to step into the shoes of our customers. The design teams brainstorm for hours mapping out the desired user experience and how to get there. The marketers conduct days of research studying how potential customers react to different situations, and then their needs, wants, likes, and dislikes are all analyzed. Finally, a target set of specifications is created from which the scientists engineer a new product. The problem with this model is that our customers are people with diabetes, a chronic disease which they live with 24 hours per day, 7 days per week for the rest of their lives. How exactly can you step into those shoes and feel what people with diabetes are feeling, when you know you can step back out whenever you want?
Despite this, those of us who have diabetes, and work in the industry, rarely seem to be given the opportunity to test the products ourselves, even if we are part of the development team. That is why I was so excited when Montpellier University School of Medicine – CHU Lapeyronnie Hospital invited me to participate in the newest trial for the Artificial Pancreas (AP) Project. During the three-day trial, my time was split between two locations: the hospital and a hotel nearby. This trial was the first time the system was used outside of the hospital setting, the goal being to start working towards adaptation of the system for use in the home environment.
The System Components
The Artificial Pancreas System was made up of the following core components: Omnipod Insulin Pump (Insulet), Dexcom Continuous Glucose Monitoring (CGM) System with SEVEN® PLUS sensor, Sony Ericsson cell phone, and a PC computer tablet. The CGM number and trending arrow was displayed on the phone.
Both the insertion and calibration procedures used were the same as the normal procedure for the Dexcom CGM, however fingerstick blood glucose measurement results were entered directly into the phone instead of the Dexcom receiver. The Omnipod insertion was also the same as normal, however I was instructed to enter carbohydrate and blood glucose values at meal times into the phone rather than the PDM (Personal Diabetes Manager).
The phone display indicated a status of wireless connections for both the insulin pump and CGM, and most importantly, the real-time traffic light icons for hypo and hyperglycemia (shown in photo). Based on the threshold values entered, the traffic lights would caution for an impending low or high using the yellow light for warning, which changed into a red light as it became more evident that hypo or hyperglycemia was approaching. The algorithm then determined the best course of action based on a list of factors including: the time of both the last insulin bolus, and the last carbohydrates ingested. If I was in danger of hypoglycemia, it temporarily suspended insulin pump delivery. If I was in danger of becoming hyperglycemic, it administered the appropriate insulin dosage it had calculated. In both of these cases, it made me aware of the action it decided to take. This part of the user interface design was particularly important for me, and I will explain why later.
The Trial Experience
The trial started on a Wednesday afternoon. On the train ride to Montpellier, I began to feel a bit anxious about what I was about to experience. I wasn’t exactly sure why I was nervous , but going into a foreign environment for three days, which included being on a hospital bed for part of the time, suddenly didn’t seem so appealing anymore! This nervous excitement remained until I arrived at the hotel, where I was greeted by the team who would be taking care of me: two endocrinologists, two nurses, and two engineers. My first reaction was “Wow! This is the safest group I’ve ever had looking after me,” soon followed by “Man, I wonder what the heck they are going to do to me that needs six experts monitoring?” Team members were very warm and approachable, which helped a great deal.
I settled into my hotel room and the team began checking on the various components of the system. They put another Dexcom on me as a back-up and assisted me in inserting the Omnipod pump. The site locations used were all on my tummy which wasn’t typical for me (my normal pump/CGM sites are on my sides, lower back, or arms since I sleep on my stomach). It all went by very quickly, and before I knew it, I was several hours into the trial. The first day, the system is kept in ‘open loop’ mode and it is not until the following morning when the ‘closed-loop’ mode is activated. My blood glucose was measured on a fairly regular basis, with a separate stand-alone hospital glucometer. Mainly due to the weight of the PC tablet, I felt rather overloaded with gear, but was able to move around my room and work on my personal laptop computer which I’d brought from home. Surprisingly, after getting used to being hooked up to everything, the hardware portion of the experience began to feel like normal.
For dinners, I had preselected my favorite meals off of a list of ‘ Picard’ grocery store items (equivalent of US frozen Lean Cuisine Microwave ‘TV Dinner’), in order to ensure precise carbohydrate counting, since the nutritional value information of the serving plate was written on the box (not always the case in Europe!). All meals were eaten at the daily times specified by the protocol, which also allowed for an optional late-night snack.
As I prepared for bed, the reality of what was about to happen the next morning began to sink in. For 16 years, I had almost thought about the following nonstop: my last blood sugar value, when I was going to eat, where my meter was, what my pump was saying, the activities I had planned for the day, how much insulin I had in my pump reservoir, was it time to change my infusion set?, temporary basal rates, fast-acting carbohydrates, and the list goes on. With the support of the team around me, the Artificial Pancreas System I was wearing would now take over.
I suppose there are two ways one could react to this: a huge sigh of a relief or an enormous gasp of terror. Unfortunately, my reaction was the latter. I’m not sure if it was due to my Type A perfectionist personality or the fact that I had maintained a 5.7% A1C over the course of my diabetic life, but I could not imagine allowing anyone else to manage my diabetes for me, let alone a machine! It is perhaps ironic that an engineering graduate would have such a reaction.
The team noticed my apprehension and asked about my concerns. They let me peer into the hotel room next door to mine, where I observed the “Misson Control” Center. Computers were laid across the room with several large LCD monitors displaying graphs and charts that were all showing my data real-time! “Don’t Worry! We are keeping a close eye on you,” they said. It was a strange feeling to have so many eyes on my data, but it reassured me enough to go back to my room and go to sleep.
The next morning, I woke up with a BG a little high (~150 mg/dl). I went over to the hospital as planned and the closed-loop system was activated. I ate breakfast and entered my BG value and the carbohydrate total into the phone based on what the hospital dietician told me. It delivered the bolus like my pump would normally. The next day and a half went by very quickly. I experienced some of the diabetes rollercoaster that we all know too well with ups and down in my glucose values, but nothing extraordinary.
While I became used to the device setup I was wearing from a hardware perspective, the concern about giving up control never went away. I did become less anxious over time. A major contributing factor to this was that the phone would let me know when it was taking action. So even if I didn’t have a choice, it at least made me aware of the actions it was taking. That made a big difference and on occasion, I notified the team if I thought one of its actions was questionable, and they would discuss it with me. The doctors and the rest of the team were interested in my concerns and available to answer my questions. An extra plus for me was that I learned that being a bit more patient with my insulin correction boluses could be a better strategy for me to avoid hypoglycemia when I returned to my normal routine back at home.
Taken as a whole, the experience was incredibly exciting. It helped me realize that all of us diabetics have routines, and even if we don’t think about it in these terms, we have defined glucose value ranges in which we aim to operate for different times/activities/meals etc… As such, if someone or something else steps in to manage our diabetes for us, we react with much anxiety over the change. That is precisely what happened to me, but in fact, by sitting back and observing, I learned a bit about my own insulin therapy regiment including a need for new a insulin-to-carbohydrate ratio in the evenings.
Overall, today’s version of the Artificial Pancreas System is an intelligent system that grew to know me better every hour and every meal. However, the pattern recognition piece of the algorithm was difficult to observe in the short three-day session, and I wish I would have had more time, because both it and I needed to become fully compatible with one another. I believe that my expectations were realistic going into the trial and I was not disappointed. There is more work to be done, but I look forward to advances in sensor technology and AP algorithm development which will hopefully allow for this to become a commercially available system in the future.
It’s important to recognize that my experience was incredibly safe, and I wish that governing authorities like the FDA would take a more open approach to studies of this kind. For example, I have been wearing my Medtronic Veo pump for several years, since I have had access to it in Europe.
While I respect the enormity of the task the FDA undertakes, I find it incredibly unfair that the U.S. government is preventing children with diabetes in the United States having access to the hypoglycemia preventing “low-glucose suspend” technology. They claim that they need more clinical data, even though the device has been approved and used for three years already in many other countries. Parents of these children find this difficult to comprehend. I wonder if the relevant FDA officials have visited their local diabetes camps? Have they spent 24 hours with a child with type 1 diabetes? If they have, then surely they understand the danger these kids face now (without that powerful technology) and why it is so vital to approve this new technology.
If they have not, I encourage them to do so.
Too right, Kyle: there’s nothing like experiencing the realities of diabetes treatment yourself. Thanks for being an AP “guinea pig” and sharing your insights with us!