FUDiabetes

Tidepool: FDA-approved Loop project will access OMNIPOD!

closed-loop
loop
tidepool

#1

As many know, Tidepool recently engaged into a project to produce an FDA-approved version of Loop. Loop is a DIY closed-loop project, similar to OpenAPS, focused on running the control loop app on an iPhone platform. Many in the diabetes community run Loop, including a number of us here at FUD, and @dm61 wrote a great wiki to get people started. @TiaG reported the original Tidepool project kickoff:

I don’t know how I missed this—but, earlier this month (11/2), Tidepool announced that they will support Omnipod DASH in the Tidepool Loop app. That is big! Until then, I was doubtful of the initiative. But this gives them access to a powerful, cutting-edge pump, with the agreement of the manufacturer!

https://tidepool.org/author/howard/

Update #1 (11/2/2018): We’re quite thrilled to be partnering with our friends at Insulet to support Omnipod DASH in Tidepool Loop.

To me, support from Omnipod is a game changer. That said, there are many hurdles on the way to FDA certification. I wish them the best of luck, and truly hope to see them successful!

You can get direct updates from Tidepool Loop.


#2

It would be interesting to read the other half of the press announcement.

ie - From Insulet.


#3

Here is Diatribe on the topic on Nov 20:

I still can’t find an Insulet announcement :frowning:


#4

It was in the Nov 2 earnings call!

Courtesy of YDMV.net:

As a first step in our support of interoperability, we are pleased to share that Insulet is the first pump partner for Tidepool’s Loop Program. For those who aren’t familiar, Tidepool is a non-profit organization that is working to get an open source, iOS-based Looped app and algorithm, approved by the FDA. Together, we are developing an interoperable, automated insulin delivery system that allows Omnipod DASH to be controlled by the Tidepool Loop algorithm from an iPhone. This partnership is additive to our internal Omnipod Horizon program and offers a potentially faster avenue to market, an automated insulin delivery system with iPhone control for our Podders, and a terrific way for us to support the DIY diabetes community.” (About 23 minutes into earnings call starts getting more interesting for PWD aka less businessy at about 20:00)


#5

Interesting.

Thanks for the pointer. I just read through the Insulet 2018 Q3 earnings transcript.

Slight mention of iPump from an analyst which was basically swatted down by Insulet. Per my understanding, without iPump designation, there will be required a clinical trial and regular PMA submission to allow the combination of the Tidepool Algorithm to run on the Omnipod Dash system using the Dexcom G5 or G6 system.

Point being, the progress should be able to be tracked from the Clinical Study which would reasonably be expected to be required prior to FDA submission.


#6

Could not agree more. Very smart move by Insulet - they will have their own closed-loop system on Android platform. In parallel, making pods operable with Tidepool Loop, which runs on iOS platform, makes perfect sense - they are simply expanding the market for pods without any further development costs. Lets hope Tidepool will be able to navigate successfully and quickly through FDA approval process. They have a great team, and they’ve hired the key people behind DIY Loop, which makes me optimistic.


#7

If the FDA needs to sanction it, that means it probably won’t allow complete user control, right? That sucks.


#8

That is the $100,000 dollar question.

The FDA is clearly promoting “interoperability”. How far they are willing to go with that in terms of giving control over to the end user is IMHO the big question.


#9

Not sure though. I attended a pres by the two FDA women who piloted the 670G through certification. They were able to make many changes to the process, and blazed some new ground. I was impressed by their gung-ho spirit. I think the FDA is much more open than it used to be. I guess we’ll have to see.

At the same time, there is some need to be careful in how many mistakes you allow users to make! This evening, my son made a mistake and let a +30% basal run way past its need. It was sobering for the two of us to realize that, had he made that mistake before going to bed, if he lived on his own he would have been dead tomorrow.


#10

Does he not use a cgm ?


#11

Yes, but he does not wake up to alarms, like 70% of teenagers according to his endo.


#12

Ok. I missed the part about living on his own.

I just thought you meant living some place other than with you. Such as in a dorm or an apartment with other college students.

I would very much prefer for my daughter NOT to be living alone. I like backup plans and I like safety nets. When the time comes, I will very much encourage her to live with friends and not alone. There are so many situations even aside from T1D where having another person in an apartment can turn a critical situation into a mild nuisance.


#13

I SO agree with this statement.

In fact, I think it was @glitzabetes who mentioned how much she had been helped by her roommates in college!


#14

A safer way to set basal increases is to never set them for long extended times.

If you think you will need a basal increase that goes 4 hours, you can make one that goes 2 hours, check again, and if needed, make another 2 hour increased basal.

Especially if you are going to sleep!

A smaller increase, like 10%, is no big deal if it goes longer. But for a bigger one, do several that are shorter duration.

Also, I have a pump suggestion for him when he is off on his own.


#15

Perhaps. But, DIY Loop will continue to exist and will evolve further. Importantly, Tidepool has made commitments to keeping their version of Loop open source to the extent legally possible, which means that fully user-customizable DIY Loop will (hopefully) also be able to work with Dash pods. The pod driver will likely have to remain a closed-source software module, but that should not be a problem. DIY Loop may never be FDA approved, but that’s fine with me.


#17

Why? Only the official Tidepool App would be FDA approved, not the DIY self-compile.


#19

But that is only true as a commercial device though. If Tidepool, like OpenAPS or loop, put its code in an open depository, I don’t think it breaches anything, because those who use it compile it themselves and run it on their own, like OpenAPS. I could be wrong though.

But, if I am wrong, I think that Loop and OpenAPS would be targettable too. Without first hand knowledge, I imagine they have probably had a good look at the legal issues of their position?


#20

DIY people (self not included) have had conversations with FDA. The understanding is that as long as apps are not distributed, and you yourself must assemble the system together, there is really nothing FDA or anyone else can do about that.


#22

There are no two ways. Tidepool will never distribute DIY Loop. DIY Loop is not and will not be developed by Tidepool. They do not have and will never have any liability or legal issues related to DIY Loop. DIY Loop has been developed under MIT License, which allows Tidepool to legally incorporate the already developed and any future code into whatever way they like, and eventually (and hopefully) have their Tidepool Loop FDA approved. They also have every right to legally keep their software open source under the same or similar open-source licensing terms.


#23

The thing we don’t know is the extent to which the FDA would put restrictions on the Tidepool fork that would ‘prevent’ unapproved software from using the DASH command interface. For instance there’s probably some way to arrange it so that DASH could refuse to listen to commands from software whose binary was not cryptographically signed by Tidepool. Tidepool could publish the bulk of its source code without publishing the ability to authenticate to DASH.