The article, strangely, leaves out the name of the project it's reporting on. It's OpenAPS, http://openaps.readthedocs.io/en/latest/ . I use it and have worked on the code. It's a massive health and quality of life improvement over anything that exists which is FDA approved. The core problem is that the FDA-approved devices are optimized not for getting the best result, but for ensuring the manufacturer can't be blamed if something goes wrong. But that in practice means kicking most of the important things to the end user and their physician, who can't do nearly as good a job as software can.

Reading around, it's the controller software that is DIY, and it's grown to the point where a DIY device has been built to run the controller software. The sensor and pump/actuator are proprietary Medtronic devices, connected to the controller via wireless links.

The DIY controller came about because Medtronic didn't secure their wireless traffic and the protocol was reverse egnineered. New units now come with a secured wireless link, but the existing units remain unsecured. Consequently the project is constrained by the number of unsecured units remaining in the field?

Reading the project's website the FDA has been in touch with the project and so far the interaction has been constructive. The makers seem keen to have the safest system they can and the FDA seems to want to engage. I'd guess the project's future lies in being able to team up with someone who can provide ongoing access to sensors and actuators?

Wishing them much success!

I have type 2 diabetes and I've always been a bit envious of insulin pumps, not because I love having a machine plugged into my body (I don't), but because they offer better control of blood glucose peaks, since they use fast-acting insulin exclusively. The pancreas secretes insulin in pulses and pumps can mimic that. What has turned me off from pumps (apart from doctors generally not recommending them for type 2 diabetes) is the lack of a sensing mechanism. This DIY system sounds absolutely perfect. It could keep my blood glucose perfect at all times. At the moment, I'm injecting in the mornings and the evenings and have little control over my peaks apart from adhering to a strict diet (and we all know how hard that can be), and every peak does a little bit of damage to my eyes, nerves, etc. I've always wanted to have a more sophisticated treatment option, and this may be it.

Why exactly is a phone app involved? If I were designing this thing, DIY or otherwise, the CGM would talk to a dedicated micro controller, which would in turn talk to the pump. A phone, if used at all, would get an extremely limited API to read the glucose level, see the pump history, and perhaps change the control parameters within a narrow range. The point would be that no input from the phone should be able to kill or incapacitate the user.

And the controller would beep if communication were lost.

There would not be an Internet API to change any parameters at all.

One of my coworker has a DIY glucose monitor he uses to monitor his kids from work with a Raspberry Pi. He made a blog post about it. Here's the link if that can help anyone: https://www.robotshop.com/blog/en/xdrip-bridge-type-1-diabet...

Headline: "The $250 Biohack That’s Revolutionizing Life With Diabetes"

Article: "Sydney, now 15, is still using an updated version of that DIY system, which, because a fellow DIYer donated the pump, cost only $250 to make."

A continuous glucose monitor (CGM) uses a very thin gauge wire to sense interstitial fluid as a proxy for blood sugar.

My kids have tried Dexcom G4 CGMs a couple years ago. The G4 was never approved to dose insulin directly off the CGM reported value, we would confirm with a finger prick before dosing a correction. The G4 also required “calibration” where you would type in the reading from a finger stick and it would shift the curve a bit in that direction if it was off base.

We would routinely find the G4 reporting numbers up to 100 points off, or simply dropping out and reporting “???” for hours at a time. Bad or unreliable data being worse than no data, we discontinued using the G4.

The Dexcom G5 kept the same sensor but switched the transmitter (which connects to the sensor and sits on your body) to use standard Bluetooth to communicate the glucose readings. This is when Nightscout was really able to step in and intercept the glucose data and do neat things like put it in the cloud so you could watch your kids’ blood sugar throughout the day. And with the addition of a programmable pump, also try to build a “closed-loop” system which constantly adjusts basal insulin dosages to react to highs and lows.

The problem is the underlying sensor of the G5 was still not approved for direct dosing, a finger-stick still being the only method accurate enough. You don’t want to dose to correct a 250 blood sugar level and then discover you actually were 150 all along and now potentially heading for a dangerous low.

Early this year we started buying Freestyle Libre sensors which use a NFC reader to poll the latest readings on demand. This is a less ideal method because it can’t independently alert you of an impending high or low without you actively swiping the reader up against the sensor on your arm. Accuracy was significantly better (reading would often be within 15-30 points, and we didn’t see any that were more than 60 points off from a finger stick reading) and there were a couple times when it would stop reporting data but dropouts where much less common than what we experienced with the G4.

Fast forward to last month when we got brand new Dexcom G6 CGMs. These are FDA approved to dose directly off the readout, and the accuracy is frankly unbelievable. We have regularly had finger sticks exactly match the CGM display, and even finger sticks have variation when you test multiple times back-to-back. The G6 does not require daily calibration (although you have the option if you find the numbers are off). It is not perfect, and once we did find it was 40 points off, but after 60 days of use my experience is that with the G6 we have finally arrived at a stable, trustworthy, reliable CGM which I would feel comfortable driving a closed loop system.

The only failing I have with the G6 is that we were promised it would work directly with an Apple Watch running the Dexcom app and syncing with the sensor over Bluetooth. In fact, as of now, a full fledged iPhone is required to pull data from the sensor, and then it can relay from the phone to the watch. I’m told that the FDA is pending approval of a stand-alone Watch operation, which would be exceedingly useful in my case. The goal is for the kids to wear LTE Watches which can read from their sensor directly and push glucose data to the cloud for us to watch on our phones. This would mean just the $5/mo cost to add a Watch to the plan versus having to add lines and gives full fledged iPhones to the kids in order to track their numbers remotely.

Both our kids use OmniPods as their insulin pumps, which AFAIK have not been hacked yet to allow 3rd party control of the pump. OmniPod is a self enclosed pump which sticks on your arm an auto-injects a cannula, and you control it wirelessly with a dedicated remote which is like a 1990s PDA. OmniPod has an ongoing study with their next-gen “Horizon” product which adds Bluetooth control and interfaces with the Dexcom G6 to automatically adjust basal rates.

Last thing I’ll mention is that this is not by any means an “artificial pancreas”. Even with a closed loop system you still must bolus (administer insulin) before meals in the correct ratio based on the number of carbs, and the glycemic index (speed of absorption). The closed loop is more rather a slight adjustment in the baseline insulin which is being automatically dosed every few minutes to cover your basal metabolic rate. If the CGM sees you are high or low, a closed loop system will temporarily increase or decrease the basal insulin rate to try to subtly move the needle so-to-speak and bring blood sugar back in range. This works best overnight when no new carbs are being eaten and blood sugar is usually more stable. So the idea is with a closed loop you are more likely to wake up with a blood sugar around 100 because it’s been working to gently bring you to that level all night long.

As the article states, this means a better nights sleep for both parent and child :-)

I can't help but wonder what it would take to develop a legitimate artificial pancreas that fits the standards precisely. Would animal testing even be that useful or required given precedent of other insulin pumps?

I suppose being able to adapt to a wide range of pancreas removed animal sizes and life-states could help to show some adaptability but I'm uncertain of how much that would help humans since diverse curve fitting may not be a good thing. Since for instance for humans protein is bad for the kidneys at too high of a percentage even if say cats wouldn't have problems with it.

I'm curious what the back of the envelope budget sizes, time tables, and liabilities would be for organic pancreases as opposed to an official approved smart insulin pump pancreases.

So some people are concerned that their cars have an internet connection and others connect their bodies to it (via the smartphone) and at the same time, both have good reasons to do so :-D

Sometimes it is a strange world we are living in.

While I admire open source and DIY solutions in general, they often lack the quality assurance a company specialized in medical equippment is used to. So I hope that soon there will be more devices from established manufacturers at reasonable prices. I would consider everything above $1k unreasonable if the hardware alone costs just $250 (for consumers).

Can you imagine how much money big pharma could make if they pulled their fingers out and actually created and supported a solution?

If a bunch of parents on the internet can hack it together, it's hardly impossible (I'm not trying to detract from the awesome work these people have done, but my point is that they have nowhere near the same kind of resources as pharmaceutical companies).

As a Type 1 Diabetic it's really frustrating seeing how stagnant treatment technology is.

Does anyone happen to have a Medtronic device on the compatibility list that they are looking to sell?

As a med student with a CS degree I wish we were further along with making this a reality for every child with DM. If anybody knows who in Aust is working on this - hit me up!

God bless Scott Leibrand and these folks.

How much are the continuous glucose monitors alone now? You could use it for dieting.

Anybody knows of other projects like this?

ketogenic diet is what is revolutionizing life with diabetes, by, you know, curing it (not type 1 though).

automating the process of releasing a chemical that is roughly speaking the cause of the disease is hardly revolutionary.

Or you could "hack" your body by eating keto-friendly foods:

https://blog.virtahealth.com/with-sustained-type-2-diabetes-...

https://medium.com/@drjasonfung/type-2-diabetes-reversal-the...