“I’m feeling funky right now,” Aimee Jose says, glancing at her “bionic pancreas” for a status check on her blood sugar levels. “The number here says that I’m fine, but I don’t feel fine, so I need to confirm it.”
Jose, a certified diabetes educator and nurse at the Palo Alto Medical Foundation, has had Type I diabetes, a disease affecting as many as 3 million people in the United States alone, since she was 12.
While her “bionic pancreas” suggests that she is, in fact, fine, her body is telling her otherwise, so she pricks her finger to get a tiny amount of blood for her glucose meter, a device that has been available to diabetics for over 20 years. “The meter is the gold standard,” Jose says.
While this gold standard is a device two decades old, now emerging are medical devices providing life-saving diagnostics, monitoring and treatment that are becoming smaller, smarter, integrated, talkative and even extendable through clever software work.
The Bionic Pancreas
Jose has been participating in a 40-person study with Boston University and Massachusetts General Hospital testing out an apparatus called the “bionic pancreas.” This device is a connected, collection of devices actually, consisting of a Dexcom continuous glucose monitor (CGM), an iPhone 4S with a specialized app, and two Bluetooth-connected automated pumps (one with insulin and the other with a hormone called glucagon). Together, the bionic pancreas automatically monitors and regulates Jose’s blood sugar levels.
Type I diabetes is an autoimmune disease where the body attacks insulin-producing beta cells within the pancreas. Without insulin, the body doesn’t metabolize certain nutrients like carbohydrates properly, leaving them to sit in the system. This sugar floating in the blood is what a diabetic measures, explains Jose. The body needs insulin to take that sugar out of the bloodstream to get it to muscles and cells to use as fuel.
Not only does the bionic pancreas monitor Jose’s glucose levels, it automatically regulates those levels using infusions of either insulin or glucagon into her bloodstream based on her sugar levels as reported by the CGM, which communicates via RF (radio frequency) with a platinum sensor injected under Jose’s skin. The sensor is replaced weekly. The iPhone is physically connected to the monitor, which sends readings every five minutes. To manage sugar levels, the iPhone is paired via Bluetooth to two Tandem t:slim pumps to automatically infuse insulin and glucagon into the bloodstream.
Traditionally, diabetics had to manage their sugar levels manually. Since 1995, Jose has been on an insulin pump that infuses insulin directly into her bloodstream. Before the pump, Jose had to manually determine how much insulin she would need to cover the food she was going to eat and then inject herself with the appropriate dosage. The insulin pump made things a bit “easier.”
“The beauty of the insulin pump was that you pre-programmed it with the appropriate ratios, and it did the math for you,” says Jose. But there are still a considerable amount of manual steps required according to Jose: use the finger prick, go to the insulin pump, count carbohydrates prior to eating, enter in her blood sugar level into the pump, receive the insulin infusion, and then wait 15 minutes prior to eating.
The bionic pancreas aims to fully automate this process. “Now, I don’t do anything,” Jose says.
Challenges still face the bionic pancreas, which is in clinical trials and pending FDA approval. The hormone glucagon, which raises sugar levels, is not shelf-stable as a liquid. Type I diabetics carry an emergency kit with them as a “rescue medicine” when they hit a “really bad low,” explains Jose. The liquid must be mixed with a powder in order for it to be used. This means the glucagon pump must be refreshed with new liquid daily. A shelf-stable version of glucagon is being developed.
“The invention of the continuous glucose monitor, in my opinion, is probably the most helpful tool in diabetes management outside the development of insulin,” says Jose.
The traditional finger prick meter is “just a moment in time,” explains Jose. The CGM provides a more continuous reading, taken every five minutes. According to Jose, all blood sugar meter are required by the FDA to have an accuracy of plus or minus 20 percent as compared to a lab blood draw. The latest Dexcom CGM now has an accuracy of 9 percent, according to the company.
“That’s the best technology that we have right now to manage medicine decisions – life threatening dosages of medicine,” says Jose.
Parents are Hacking the Continuous Glucose Monitor
Others recognize the importance of the CGM in diabetes management. In fact, parents who desire remote monitoring of their kids’ sugar levels are developing means to extend the capabilities of such devices. A group of parent engineers, developers and hackers, created a DIY project called Nightscout and have essentially open-sourced a means to allow real-time, remote access to a CGM data via browsers, smartphones, tablets and Pebble watches.
John Costik, a software and systems engineer, whose 6-year-old son, Evan, was diagnosed with Type I diabetes in 2012, felt that the tools he and his wife were given to care for their son were “inadequate for such a data-heavy disease.”
“We realized that if we collected the right information and analyzed it correctly, we could do a lot more for him and keep him healthier than we were expected to,” says Costik on the treatment of his son. “When we were looking at technology options, the Dexcom G4 was and still currently is the best glucose monitor available.”
Costik and his wife received the CGM for their son in February 2013, and after one week of his son in daycare, they realized they didn’t want to be disconnected from the data coming from their son’s monitoring unit. Out of the box, it can only transmit data from the arm sensor to the unit.
“I quickly realized that I didn’t want to be away from that information,” says Costik. “I wanted to see his blood sugars every five minutes.”
Costik began analyzing the CGM’s Windows software and discovered a way to communicate with the unit directly. He wrote a simple Windows application that polled the CGM receiver every five minutes for data and uploaded the data to a Google Docs spreadsheet automatically. He also wrote an iOS app for him and his wife to use.
Next, to make the process more convenient and mobile, reducing the need to lug a laptop with the monitoring unit, Costik created an application that worked on an Android phone and was connected via USB. As the cellphone was attached to the CGM receiver, the data could be then uploaded to the cloud. Costik shared his code so others could develop with it.
At that point, Costik was contacted by Lane Desborough, whose son, Hayden, was diagnosed with Type I diabetes in 2009. Desborough wanted to develop a system similar to Costik’s and through their collaboration formed Nightscout, a term Desborough concocted.
“What Nightscout does is liberate that data [from the CGM],” says Desborough. “Nightscout takes the data out of the CGM, pumps it up into the cloud, from which it can then be viewed on any web browser around the world and even things like a smart watch.”
Desborough developed a software program using Costik’s shared code and various open source web services and frameworks. In a few weeks, Nightscout became an “effective and reliable” system to remotely monitor Hayden’s diabetes. The code was then shared on a GitHub repository, which now includes code for the Pebble Watch, iOS, Windows Phone, Android and other devices.
Jason Adams, who had been in contact with Costik throughout the process and whose daughter also suffers from Type 1 diabetes, hired a software engineer and created a Facebook group called CGM in the Cloud. Awareness of the project skyrocketed, says Costik. The Facebook group currently has over 10,000 members. And, the Nightscout Foundation was formed in 2014.
“The story is very simple,” says Costik. “I want my son to have as normal a childhood and life as possible.
“We want to validate that these systems really do improve outcomes. And I think we have enough data and we have enough volunteers to share their data to show that people are healthier with remote monitoring,” says Costik. “Moving forward we want to look at how do we take this data and nutrition data and carb data and insulin data and develop better tools to help folks manage that aspect.
“If we can treat highs and lows without as much intervention from users that would be wonderful. We will continue to see it evolve into that,” says Costik on the future of Nightscout and other systems similar to the bionic pancreas.
This content was originally published on the Intel Free Press website.