The open-source movement to improve continuous glucose monitoring (CGM) technology
by Adam Burrack, PhD
As more and more folks become fluent in the programming languages, and as diabetes technology continues to inch into the 21st century, groups like the open-source CGM project are popping up. This group is part of a recent interesting trend to ‘democratize’ the technology and science of how type 1 diabetes is managed through hacks of the hardware and software developed by companies including Medtronic and Dexcom. My own local paper, the Star Tribune, ran a recent Sunday article about this trend about a father who got fed up with his inability to track blood sugar levels, so wrote his own software. These approaches raise internal legal and ethical questions about data security and privacy concerns, as well as what I would call the moral imperative to optimize health care delivery.
So-called ‘big data’ and cloud data storage raise the specter of both unlimited data generation to track health care outcomes – a very good thing – as well as the prospect of individuals with bad intent hacking these systems and running amok. Without knowing the data – how our bodies respond to various insulin doses in various exercise/diet/stress situations – it can be very challenging to develop our own appropriate insulin dosing guidelines. Taking this concept one step further, including data from multiple people in a learning algorithm would potentially facilitate development of appropriate situation-specific guidelines across a larger range of people and situations. This is essentially one goal of the open artificial pancreas group – to fast-track development of broadly applicable insulin dosing algorithms based on CGM sensor data from multiple users.
Along these lines, recent news shows Medtronic is partnering with Samsung to improve data tracking/archiving. As noted above, many issues related to the definition of a medical device, who owns the data, and data security will need to be worked out at the level of federal laws for this work to proceed in the long-term.
In addition, I can’t speak to anyone else’s experience wearing both a CGM and an insulin pump, but I would rather have one device in my side than two, especially when the second insertion site takes away half of my placement options – in terms of real estate – for the first (and more important) insertion site. Along this important pinch-point for diabetes management (pun intended), Google is partnering with Dexcom to miniaturize CGM technology.
So, welcome to the brave new world of miniaturized CGM devices, big data, and algorithm-based insulin dosing? It could be great, it could be all hype, or it could be a little bit of both.