Open Source Glucose Monitoring On The Front Lines Of Innovation

Cloud-based CGM

[John] is the parent of a diabetic child, and his efforts to expand the communication options for his son’s CGM (continuous glucose monitor) have grown into a larger movement: #wearenotwaiting.

After receiving a new monitor—a Dexcom G4—[John] set about decoding its communication protocols. The first steps were relatively simple, using a laptop to snag the data from the CGM and storing it on a Google doc which he could access as the day went along. The next step involved connecting the monitor and a cellphone for around-the-clock data gathering. [John] managed to develop an Android app to accomplish just that, and shortly after people began to take notice. Both [Howard Look], the CEO of Tidepool, and [Lane Desborough], engineer and father of a child with diabetes, have thrown in their support, leading to further developments such as Nightscout, an open source solution for storing CGM data in the cloud.

This project is a victory not only for those with diabetes, but also for the open source community. [John] admits his initial hesitation for developing for the medical device platform: litigation from a corporation could cause devastation for him and his family despite his intentions to merely improve his son’s and others’ quality of life. Those fears have mostly subsided, however, because the project now belongs to both no one and to everyone. It’s community-owned through an open source repository. Check out the overview of [John’s] work for more pictures and links to different parts of the #wearenotwaiting community.

26 thoughts on “Open Source Glucose Monitoring On The Front Lines Of Innovation

  1. This is definitely cool! Open source health/medical projects are a huge untapped market.

    Now, if I could get a Ketone -monitor like this I’d be happy as a clam. I keep track of my ketone levels since I’m eating LCHF foods. Lost 39 pounds over 8 months – so I’ve gone from being a fat nerd to just a semi-fat nerd :-)

  2. I agree with matseng.
    I have been asked several times by diabetic family members about this. “there must be a better way.” Better methods of actually getting the test done (i.e. no pricking, no blood), cheaper (my god basic healthcare in the US is still way out of control and financially unavailable for more than half the population), and some way to both make sense of the data, log it, and report it automatically to your doctor.

    Like, why cant the hospital be aware of your situation BEFORE you are brought in? Or allow the meter to not only alert yourself (who, depending on how bad things get, are unable to help yourself) but also a loved one, your doctor, and to get an on call EMT on the phone?

    Id suggest that the next step here is to ditch the cell phone and embed a GSM device directly into it. Packet data only plans pay as you go are so cheap these days. Just buy disposable sims, load it up with several months and go. Perhaps some carrier can come up with a decent ‘Medical and Science device” plan in which they give you an extra SIM and it runs over your own data plan for a small additional fee on the bill. These things wont be using much data.

    Apps like this are where all the IOT buzz really shines and stands out as not just another bit of tech jargon BS (“hey man.. we just want to make the world a better place, with our product that does…”)

    1. Devices like this are currently undergoing clinical trial. My mother just this weekend completed one of (if not the first) trial of such a system in the U.S. The system runs on an android phone and collects data from the CGM *and* the pump and can make the data available for remote telemetry over a 3G/4G WWAN. The next step is to have the android phone *actually control* the glucose pump directly — effectively creating an artificial pancreas!

      I know the tag line is #wearenotwaiting … but just wait :) Things are about to get really awesome in the CGM/glucose pump/diabetes world.

      1. No doubt there are amazing products on the horizon, but why wait? Would you risk your 4 year old’s long term or immediate health, when you know you can improve life by simply reading and relaying data to any number or caregivers?
        This system by no means replaces the G4 Receiver in its purpose. Part of managing type 1 is contingency planning – if my pump fails, if my cgm fails, if my bgm fails… no solution or tool will ever be perfect, and you must know how to persist should it fail. Divine engineering isn’t perfect, how can we expect as much from people?

        The day products exist that exceed the outcomes of this one, I will trade up. Until then, this is just another way to use the data that we, as diabetics or parent’s of young diabetics, own.

  3. The interfacing between “anything” that is a *Direct Body Contact* device and the outside world demands sober awareness of what’s at risk. Even so – there’s a substantial risk>reward case to be weighted in favor of Not Waiting. IF this saves a life we’re heroic saviors. If this fails and Something Very Bad happens…

    Frankly- the bottom line is that remote watchdog monitoring of vital data can keep people literally alive. Or kill them in various ways if the data is munged/an exploit vector.

    I can see this concept safely serving as ASSIST but it scares me to consider it as a primary trusted system till we get substantial review data. Trust ,but set a level of trust?

    1. How can monitoring the data remotely kill them?

      If the data is corrupted and shows things are bad, then the patient will be whisked in to see the quack. If the data is corrupted and shows things are fine (by masking data that shows things are bad) then the true situation will be detected at the next scheduled checkup. In the meantime, if the patient uses their built-in monitoring system and weakly moans “I don’t feel well” then once again they’ll be whisked off to the quack pronto.

      Of course, someone will say “But you have remote monitoring! How could you have missed it?” the correct answer is “Them’s the breaks.” In this world, things are not, and can not, be perfect. We should all do our best, but not beat ourselves up over something not working out. The worst thing we can do is not to try.

      1. If a device falsely advises that the patient is fine, and the patient trusts the device, they may not seek care, and they may die.

        The worst thing you can do is try something that has not been properly tested.

        Also those “quacks” help keep you and others living healthy lives, show a little respect for your physicians.

        1. There are a lot of “may”s in your comment, which means that the doomsday scenario you envision is unlikely. Anyone who has some special condition that requires monitoring is going to have regular checkups *despite* what their personal instruments say.

          And “quack” is a term of endearment as far as I am concerned.

  4. “Open Source Glucose Monitoring an the Front Lines of Innovation”



    C’mon, you can put someone into space, and yet, you can’t write.

    Oh, and great project BTW.

  5. Ah, the G4.. had it in/on me couple times. Nice meter and all but the transmitter and cannulas are pricey! Also get’s in the way sometimes, too bulky.
    Artificial pancreas implant would be nice and non-contact BG meter too. Both have been in development quite some time now but as all things medical the progress is slow as molasses. T1 for life :P

  6. The problem with technology like this is HIPPA law… There is too much concern of information not being secure… Honestly, WHO CARES !! If My meter can call or text someone when I am low, well, hell yeah go for it !

  7. Interesting project, but this guy probably needs to be more afraid of the FDA rather than the company.
    Unfortunately medical devices are expensive to produce, the FDA regulates every aspect. I write firmware for glucose meters, so I have to deal with this every day. There is an huge amount of code that goes into these little things and its incredibly complicated. Every line and branch is verified and tested. Years of R&D and millions of dollars go into even the simplest meters. Yes, there can be harm to the user. What happens if your ‘hacked’ interface produces incorrect data and you administer too much or too little insulin? CGM is even scarier when its controlling a pump. Your basically putting your life in the hands of the people who wrote the code.

    Priority 0: Safety of the customer
    Priority 1: Cost and schedule

  8. The problem this project will not get very far in the US. Having worked in a FDA regulated, electronic medical firm I know about the FDA process having worked as a technician with engineers that work with regulatory for return units including FDA and death investigations for defibrillators. Glucose monitors com under the FDA as a Class II device There regulatory issue that need to be followed:

    Establishment registration,
    Medical Device Listing,
    Premarket Notification 510(k), unless exempt, or Premarket Approval (PMA),
    Investigational Device Exemption (IDE) for clinical studies
    Quality System (QS) regulation,
    Labeling requirements, and
    Medical Device Reporting (MDR)

    One cannot simply whip up or legally modify the device. Then there liability. This make Open Source verer impracticable. It is not just company the guy needs to worry about but also the FDA.

  9. I am a type one diabetic. I am looking into getting the OmniPod and a G4 within the coming weeks / months, and am currently researching how I will be able to combine the data flow from the G4 and the OmniPod PDM UST400 along with an (or several) android device(s), and my PC to create a system that will [at first] give real time suggestions as to insulin injections / reductions. I think it is obvious where such a system could go (and I can only hint at it, as it would quickly become a medical device if I were to flip the switch and tell it to actually interact with the PDM and input the glucose readings as current levels that should then be corrected for, but I digress). I know there are many systems / software similar to this “available” – even those that are functioning in clinical settings right now, but I am tired of waiting, and want something that is truly available to be happening now. I am a college student, currently taking some time off of school, and have 12 – 14 hours a day to give to this project. If you are able to send me information that may help with this cause (I am reading into Nightscout, and have 34 other tabs open with information that I am attempting to digest on the subject), or if you have time to or knowledge to give to the cause, please do so. Everything I will be doing will be open source, cloud based, and completely non profit (of course). I do not know tons about mobile apps (frankly I am willing to learn if no one volunteers to help move things forward) – so that is currently my weak point, and I would appreciate any possible support there.

  10. With the event of Dexcom releasing the Share 2 Remote monitoring device, it’s time to update this story. Yes #Wearenotwaiting is changing the way diabetes data is seen by parents,school nurses and adults.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.