Data Sharing With Diabetes
You don't have to be a hacker to DIY diabetes care
People with insulin-treated diabetes and their loved ones know the fear all too well: that a nighttime low won’t be recognized, that a continuous glucose monitor (CGM) alarm will be slept through—both dangerous and even deadly risks. But how can you know when someone else is low?
Sharing Is Caring
The Challenge: Data are stuck on a single device.
A CGM works like this: A small sensor constantly reads glucose levels in the fluid just beneath the skin, sending results—via an attached transmitter—to a receiver that displays glucose readings. But there’s a catch. The receiver needs to be near the transmitter in order to get data, and until recently only the user could see real-time blood glucose levels.
That’s all well and good, but for parents of kids with diabetes, for people who live alone, for anyone who wants a little backup, that may not be enough. Data sharing is still generally done after the fact and requires cumbersome downloads. And caregivers and loved ones at a distance are not aware of highs and lows.
Thanks to intrepid engineers, developers, and people with diabetes, there are now many ways to share and access glucose data in real time. The Dexcom G5 Mobile CGM system made a splash when it came on the market last summer as the first CGM to allow people with diabetes to share real-time data with family, caregivers, and health care providers using a mobile device (for more on that, see “Code-free CGM Data in the Cloud,” p. 64). But thousands of people have been doing that DIY-style for years, with many more device options and at more affordable prices.
Ben West, 34, of San Francisco was one of a group of software engineers interested in maximizing diabetes devices. Living with type 1 diabetes, he had a stake in the results. While there were several teams working on pulling information from a CGM to the cloud—a secure online storage system—West saw how the different projects could aid one another in the pursuit of their shared goal. So he helped bring the best of each project into a single application: Nightscout.
An App Is Born
The Challenge: Loved ones don’t see warnings of highs and lows.
First released in April 2014 onGitHub.com, an online repositoryfor computing code that can be continuously shared and worked on by multiple users, Nightscout is an open-access, do-it-yourself guide to pulling data from a CGM transmitter and transferring to the Web or an app, where data can then be shared with whomever users allow.
The Nightscout app allows people with diabetes and their loved ones to set specific alerts on their phones or devices, such as for highs and lows. For families of young children or people living alone, this can offer a sense of security: If a person with diabetes doesn’t hear his or her low-glucose alarm, another person can check in.
James Wedding of McKinney, Texas, uses Nightscout along with his daughter Carson, 13, who has type 1 diabetes. Carson uses a Dexcom G4 Platinum and wears a Pebble smartwatch with Nightscout on it, programmed to alert her when her blood glucose drops to 85 mg/dl. Wedding’s own Pebble watch alerts him when Carson’s blood glucose drops to 72 mg/dl. It’s a safety net, he says, as she takes more control over her diabetes management.
“The thought process is, ‘I want you to be aware and have some time to deal before I get involved,’ ” he says. “Nine times out of 10, we see it. But we get a text that just says, ‘Treated,’ she carries on with her day, we carry on with ours.”
Here’s how it works: CGM transmitters broadcast their data to any device with Bluetooth or Internet capabilities. CGM receivers are programmed to read the signal of only one transmitter at a time. But they could theoretically read additional signals, and transmitters can be read by any computing device with the software to do so. Open source code, or freely available software, is available for pushing CGM data from the transmitter to the cloud in real time.
The first version of Nightscout allowed users to connect a cable from their Dexcom receiver to an Android phone with the Nightscout app and software. In doing so, the CGM’s glucose readings were uploaded to the app, which people with diabetes, as well as their loved ones and health care providers, could see on any device running the app—all in real time.
Since then, Nightscout has evolved to work with Apple products, the Pebble watch, and other devices. Any device that can connect to Bluetooth or Wi-Fi and house the Nightscout software can receive the data—no cable required. The devices cost money, but the software is free.
Anisha Dharshi, who has type 1 diabetes, also uses Nightscout. She lives alone, so her boyfriend gets an alert through the app if her blood glucose dips low. “Besides the day-to-day emotional support, someone knowing my numbers to help catch me when I’m falling—literally—I need that help,” Dharshi says. “I don’t hear alarms. So he will call me and wake me up.”
Another Use for Data
The Challenge: Manual entry is required for food and fitness data.
Some other people with diabetes just want to have all of their data in one place. For them, there are apps such as Glucosio, a blood glucose–tracking app that coordinates with other apps, such as MyFitnessPal, to pull in food logs, exercise data, and more to get a complete picture of how various factors affect blood glucose. It’s in the company of various other blood glucose trackers, such as Glucose Buddy, One Drop, and MySugr.
Glucosio developer Benjamin Kerensa, 32, of Portland, Oregon, started working on the app shortly after he was diagnosed with type 2 diabetes in the spring of 2014. Glucosio’s first version was available for Android users, but an Apple version is expected this spring.
As Glucosio grows and develops, it will have more sharing capabilities, including a data-sharing feature specifically designed for parents of kids with diabetes, Dexcom CGM support, and Google Vision, which allows those with standalone meters that lack Bluetooth capability to scan in their results (using the phone’s camera in the same way other barcode-scanning apps do), rather than add them manually.
Part of data sharing’s allure is that users can opt to share their results with diabetes researchers—part of a larger trend in using devices to help researchers gather data. Apple’s ResearchKit, an open-source software framework that helps developers create health-based apps, is probably the most notable version of this. Access to such data will allow researchers to get a clear picture of a typical day for people with diabetes, thousands of times over, which could lead to better understanding of the disease and improved tools for managing it.
Kerensa says a lack of breadth in diabetes research inspired the data sharing within the Glucosio app. “We want to be able to support research,” he says, adding that many long-term diabetes research projects are limited to maybe 100 people at a time. “Our hope is to get to roughly half a million users by the end of this year,” he explains. That would be a lot more data than most researchers have access to.
Another application, Tidepool, offers users multiple ways to see their diabetes data. Started by a group of self-professed “geek dads,” Tidepool is both the nonprofit organization overseeing the platform’s many aspects and the platform itself.
Like that of Nightscout and Glucosio, Tidepool’s code, the language that makes the app work, is open source and available on GitHub. Its products (three so far, with a fourth pending) are available on its website, tidepool.org. “We believe that there should be a freely available set of tools that make it easy for people to get access to their data from their devices and review it in intuitive and actionable ways,” says Howard Look, president, chief executive officer, and founder of Tidepool.
With so many diabetes data-sharing possibilities coming up, it might seem like there is some competition in the field. But Look says that variety is a good thing. “We don’t really see this as a competition and aren’t trying to set ourselves apart,” he says. “We think that people with diabetes should have a choice of what tools they use.” What’s right for the parent of a small child is different than the needs of teenagers and different from athletes or pregnant women.
Data Between Devices
The Challenge: Devices don’t “talk” to one another.
Diabetes data sharing can go beyond the cloud. Some tech-savvy people have figured out how to make their continuous glucose monitors communicate with their insulin pumps, reporting trends, predicting future blood glucose levels, and treating accordingly. They have effectively closed the loop, creating a partial artificial pancreas system through coding. It’s called open APS.
Dana Lewis, 27, of Seattle has built her own open APS with the help of data engineers. Building on the code for Nightscout and other open-source data-sharing projects, Lewis, who has type 1 diabetes, calls her closed-loop system a “do-it-yourself pancreas system,” in which trends are extrapolated from her CGM data. Those trends are then fed into an algorithm that recommends treatment based on the results, sending a signal to her pump to take action, whether by increasing or reducing basal insulin. (Safety note: This process overrides the user confirmation required for safe insulin dosing as indicated by the FDA.) As of March 22, Lewis reports, 44 people have created their own artificial pancreas systems.
The system is not without risks: One mistake in the code or a disconnect in communication between devices could result in missed highs and lows or wrong doses. There is no oversight from the government or pharmaceutical companies. But Lewis is excited to continue to work on herpersonal system.
“It’s always evolving and improving, so it’s always getting better,” Lewis says. “There are people who say, ‘Yeah it’s risky, but living with diabetes is risky.’ Letting your pump deliver while you’re low, especially while you’re sleeping, is really risky. The point of this is we are not waiting, because we have the tech. ... It’s incredible that we have this tool that we can choose to use.”
And while the technology for open APS and data sharing continues to evolve at a rapid pace, it’s not exactly new: Stephen Ponder, MD, FAAP, CDE, was part of the team that created the GlucoMON, a cradle for blood glucose meters that transmitted their most recent readings to a receiver for loved ones and caregivers, in 2002. Ponder, an endocrinologist, says several dozen of his patients use Nightscout or similar apps. “I have not seen any risks in using Nightscout,” he says. “I focus so much more not on how you get the information, but what you do with the information you get.”
A Question of Oversight
The Challenge: No regulatory oversight
Data-sharing tools are “secondary displays” of data, not put through the rigorous testing that other diabetes management devices get and not subject to FDA approval. But West says that Nightscout software has been built with protections in place. Users would have to know exactly how they are “breaking” the software, and confirm that they want to do so, in order to introduce errors—at which point the app won’t work anymore. And because the code is hosted on GitHub, engineers can see exactly who made what changes to the code, as well as which version of the code is being used. That provides protection for all users, he says.
There is another concern for some observers: Who can see this data, and is that a good thing? Alida Brill, social critic, advocate, and author of Nobody’s Business: The Paradoxes of Privacy, has written about both chronic illness and patient privacy concerns. She sees diabetes data sharing as a potential blessing with possible pitfalls. “We all want things that will make us better or safer,” she says. “We have a great desire to prevent and protect. [But] technology [is] racing so far ahead of where we are, ahead of us figuring out what we want to do with it.”
Brill notes that data sharing, particularly of medical information, may become a slippery slope in the future: “Once you let go … of your absolute rights to control your privacy, it could turn into the genie who jumps out of the bottle,” she says.
But even using a CGM without these data-sharing applications is not secure, says West. The transmitter for a CGM is constantly broadcasting data—as well as the serial number of the transmitter itself.
If users are already transmitting their blood glucose data out to the world, is it stealing or hacking if someone else reads it? West doesn’t think so. He likens seeing blood glucose data from Nightscout and similar apps to reading a blog: The information is out there if you have the Web address for it. “The express purpose of the Nightscout application is to share glucose and some treatment details with other people, easily,” he says. “Nightscout makes your data available for sharing. Since that’s the express purpose of the app, we don’t consider it a ‘hack’ or a security breach. We actually consider it a feature.”
Ponder says the risk that comes with data available to others is low. “Who the heck cares what your readings are ... unless you have a stalker out there [or] an estranged spouse or parent?” he says. “I could see someone then having the motivation to do that. But I don’t see that as a common problem.”
Dharshi says that it is important for people with diabetes to talk to their health care providers and decide for themselves how best to take advantage of all of the management tools available to them. “I think that the reality of type 1 diabetes is that you have to figure things out for yourself until all of the tools you need are FDA approved and are available at an affordable price,” she says.
“We understand that living with diabetes can be very difficult and [understand] the desire to develop tools to make ... lives easier,” says Alberto Gutierrez, PhD, director of the FDA’s Office of In Vitro Diagnostics and Radiological Health in the Center for Devices and Radiological Health. “We also understand that individuals may be modifying medical devices to address these issues but want to caution that doing so could lead to harm if the device fails, and we find it concerning when people modify devices used on other people.” Gutierrez says the FDA has been talking with diabetes advocacy groups about these issues, its concerns, and why medical regulations are in place.
The American Diabetes Association’s 2016 Standards of Medical Care in Diabetes guidelines do not address the issue of people with diabetes altering or tampering with their devices. “The Association is closely monitoring—especially from safety, security, clinical outcomes, and patient empowerment points of view—the topic of people modifying the technology of their diabetes management devices,” says Michelle Kirkwood, the ADA’s director of communications and media relations.
And while some may still question the safety or ethics of hacking one’s own diabetes management tools, Wedding believes the immediate benefits are worth it. “When I became involved with … the Nightscout project, every single [person asked], ‘How can we impact the lives of people with diabetes and the people that care for them, today?’ ” he says. “What can we do, today, to make things better?”
Data Sharing: What You Need to Know
You don’t have to be a software engineer to use Nightscout and other diabetes data-sharing apps. While the code is available for anyone to see and tinker with, users do not have to know anything about it when they set up their apps and devices. Nightscout, Tidepool, and others have step-by-step setup guides on their websites. And for those worried about “breaking the code,” says Ben West, a software engineer who helped develop Nightscout, there are safeguards put in place to keep that from happening. “To prevent people from inserting bad or malicious data, or misappropriating the computing resources, the database, and the equipment itself, whenever any potential changes are made, we require either physical presence or a password that is set by the user while configuring Nightscout software,” he says.
Code-free CGM Data in the Cloud
If you like the idea of diabetes data sharing but the setup makes you cringe, there is a product on the market for you: the Dexcom G5 Mobile system. This FDA-approved CGM system works like others on the market but also sends readings to your phone in real time. Friends, family, and health care providers can access your data using the Dexcom Follow app. The G5 works specifically with Apple products, including the iPhone and Apple Watch. It does not connect with other devices.