Hack Chat Transcript, Part 1

A event log for Hacking Diabetes Hack Chat

Diabetes met a hacker. Diabetes lost.

Dan MaloneyDan Maloney 10/16/2019 at 19:540 Comments

c448108011:48 AM
It's so annoying that you can't use the new medatronic pumps with closed loop projects :(

spaceneedle joined  the room.11:49 AM

IKR? I've got an old Animas pump sitting around doing nothing that I'd love to put to work

Casey Halverson11:51 AM

Hi, welcome!

Dana Lewis11:51 AM
hey @Casey Halverson!

Scott Leibrand11:51 AM
@c4481080 it is possible to use the 640G/670G as a CGM to upload to Nightscout and power a closed loop, but I agree it's a bummer that those aren't interoperable for insulin dosing.

Casey Halverson11:52 AM
hi @Dana Lewis ! Great to join you guys! I saw this cross my email just now. Perfect timing!

Dana Lewis11:52 AM

Pedrozamora11:53 AM

c448108011:53 AM
I've tried using the android app, with the 640G but it did ntot work

Scott Leibrand11:53 AM
Medtronic did announce they'll be making an interoperable ACE (alternate controller enabled) pump for projects like Tidepool Loop once they get FDA approval as an iController, but who knows what the timing will be on that.

c448108011:54 AM
@Scott Leibrand Do you have experience with that ? did not work for me, and I couldnt figure out why

Scott Leibrand11:55 AM
No, I know it's possible, and that it requires a Contour NextLink meter connected via USB to an Android phone, but I've never tried it myself.

c448108011:55 AM
Yeah, I connected it, but nothing happened (it was OTA)

c448108011:55 AM
I'm waiting for approval of the 670G here at the end of the year

Dana Lewis11:55 AM
there's a good gitter channel that might be worth joining to troubleshoot specifics:

Scott Leibrand11:55 AM
There are some good support channels linked in the docs if you want to get assistance to make it work.

Scott Leibrand11:55 AM
^^ that :)

Sean Hodgins joined  the room.11:56 AM

c448108011:56 AM
Thanks, I'll take a look into that (altough we are more of an iOS family :) )

mrb0t. joined  the room.11:56 AM

Andrea R. joined  the room.11:56 AM

Dana Lewis11:56 AM
us, too - although back in the day :) it was worth it to carry the extra Android phone as the original uploader to get CGM data into the cloud on the go!

c448108011:57 AM
Might be nice to publish this chat mabe in those glitter as well

Baruch Even joined  the room.11:59 AM

OK, let's get started. I want to welcome Dana Lewis to the Hack Chat today. Dana built her own artificial pancreas, which is about as awesome as it gets. We're going to talk all about that and hacking diabetes in general.

Welcome Dana! Can you start us off with a little about yourself?

irene.carrasco.shwartz joined  the room.12:00 PM

c448108012:00 PM
BTW "@c4481080 it is possible to use the 640G/670G as a CGM to upload to Nightscout and power a closed loop, but I agree it's a bummer that those aren't interoperable for insulin dosing." <-- power a closed loop, so you mean use 640G, and another pump ?

Dana Lewis12:01 PM
Sure thing. Hi everyone! I’m Dana Lewis, one of the people who created the open-source ‘artificial pancreas’ system, or closed-loop system. I’ve had type 1 diabetes for 17 years as of Monday, and although I’m not a developer or programmer by day job, I’ve spent the last ~6 years or so working to gain access to data from various diabetes devices; putting it together in algorithms to help me make better decisions, or even automate away a lot of the work I have to do. Looking forward to the chat today! Note you’ll see other OpenAPS/DIY co-developers like @Scott Leibrand @Jon Cluck, etc joining today as well - plus awesome open-source hardware partners like @Casey Halverson as well.

Kyle J. Rose joined  the room.12:01 PM

art joined  the room.12:01 PM

Franziska joined  the room.12:02 PM

Scott Leibrand12:02 PM
@c4481080 yes, you can use the 640G/670G as an oversized CGM receiver, and then another pump for actually dosing insulin. It probably makes more sense to instead use a Dexcom CGM if you can get one. AFAIK the only people looping with the 640G/670G as their CGM are doing so because that's all they can get funded by insurance.

c448108012:02 PM
Sorry for asking, is this a free-chat or not ? or Q&A later ?

Dana Lewis12:02 PM
There's a lot of great DIY projects in the diabetes community - everything from remote monitoring, to analytics/reporting, Autotune for adjusting pump setting, to open loop & closed loop software with a variety of hardware & software & phone options. If you have questions about any of those, ask away!

Jean-Christian de Rivaz12:03 PM
Hello, what CGM did you try and what was your results with them ?

@c4481080 - Fire away! It's pretty much AMA-style

Dana Lewis12:04 PM
@Jean-Christian de Rivaz I've primarily used Dexcom G4 for the last 6 years. Most of our community has moved on to G5 and/or G6. I use a Libre for scuba diving/water stuff so I can get data underwater :).

c448108012:04 PM
@Scott Leibrand yep, insurence it is :)

Dana Lewis12:04 PM
Background reading - a little more about scuba diving with a Libre and the receiver in a waterproof bag: (and for general scuba diving/snorkeling tips with diabetes devices)

Colin McCrory joined  the room.12:04 PM

mrb0t.12:05 PM
I am currently wearing the Libre 2

Dexcom keeps trying to get us to switch to G6, but we stay with G4 - transmitters last more than a year, and we routinely get two or three weeks out of each sensor. MUCH cheaper this way.

c448108012:05 PM
I heard maybe that the G7 will finally talk with apple watch or so ?

Dana Lewis12:05 PM
And this is a little more about why I am still personally using G4 (multiple receivers connect to same transmitter, so I can hotswap and hotstart new sensor sessions):

Citr0n joined  the room.12:06 PM

Franziska12:06 PM
Hi, @mrb0t: me too; works great for me, but without looping right now ...

Alan Cohen12:06 PM
By chance, Is anyone working on an open-source insulin pump with community-driven features? I’m a medical device developer and understand that this is not a trivial undertaking, but I’ve thought for a while it would be an interesting project. Could be FDA cleared, although not if people build it at home!

Dana Lewis12:06 PM
@Dan Maloney yep. cost is a big issue globally for CGM access

c448108012:07 PM
I'm wondering how a jailbreaking scene world would look in the insulin pump world :/

Scott Leibrand12:07 PM
@Alan Cohen there are a lot of really serious concerns with trying to build an open-source insulin pump, and to date no one has presented a credible plan for how it could be done safely.

@Alan Cohen - I'd be highly skeptical that the FDA would ever go near open-source anything.

mrb0t.12:07 PM
@Franziska And with the encryption they added between the 10 and 14 day US sensors that will be a bit of a hack to get working again

Dana Lewis12:07 PM
For those considering looping with a FlashGM instead of a CGM, keep in mind that adds an extra layer of risk for processing that data and converting it into use for looping. Make sure you have extra alarms set, reasonable safety settings, etc. with that and understand what to do if your system is giving bad CGM data.

Alan Cohen12:08 PM
Open source would be fine with FDA as long as all of its regs are met.

Dana Lewis12:08 PM
@Dan Maloney given that we have full blown RCTs happening :) using open source software for DIYAPS (using already approved pump and CGM), I don't think that's as far off as you would think.

@Dana Lewis - I know, just sent $700 to Dexcom for a new transmitter and one sensor. Our insurance doesn't cover CGM, so it's all out of pocket.

Dana Lewis12:08 PM
^ yes. But likely to be open source algorithm/apps, not hardware, getting approved

Scott Leibrand12:08 PM
Insulin pumps have to be built to incredibly tight tolerances, and more importantly, have to be engineered and tested to an extremely high level of safety reliability. If you can't guarantee that the pump won't dose more insulin than instructed, you risk hospitalization or death.

Dana Lewis12:09 PM
@Dan Maloney ...ouch. I usually have to fight every year and appeal to get CGM covered, but this year (maybe because I broke my ankle?) they didn't deny it this year.

Sean Hodgins12:09 PM
Anyone have any thoughts on making an open source transmitter that connects to G4/G5/G6? Its not the most complicated thing. I always imagined most of the R&D went into making it last X amount of months.

Jon Cluck12:09 PM
@Dan Maloney there are some efforts to replace batteries in transmitters, and to extend their useful life

@Dana Lewis - Tough way to get coverage. But you took one for the team, I guess ;-)

c448108012:10 PM
BTW, is there any way to export the chat transcript to some sane text format ?

Dana Lewis12:10 PM
@Sean Hodgins I think most people are re-batterying old transmitters rather than DIYing those.

Dana Lewis12:10 PM
@Dan Maloney haha. no kidding.

Franziska12:10 PM
@mrb0t. I understand the reasons why (for me it works excellent). But ... let's say I'm interested in the problem they left with this coding stuff :-)

Andrea R.12:10 PM
@Dan Maloney I am sticking with the G5 for now. It's easy to extend the life of the transmitters (easier than the G6 from what I've seen) and I've had no issues extending the life of my transmitters using the Spike app.

Scott Leibrand12:10 PM
@Sean Hodgins I haven't heard of any such efforts, outside of rebatterying (even with rechargable LiPos in some cases)

@Jon Cluck - I know, I've had transmitters apart destructively before, looking at ways to gently lap off the plastic to expose the batteries

c448108012:10 PM
@Dana Lewis Here, if you are trying to get pregenent, you can get CGM via basic country health insurence

c448108012:11 PM
for unlimited time

Roberto Marquez joined  the room.12:11 PM

F5OEO12:11 PM
Happy to say "hello" to Dana, as I contribute in OpenOmni. Just begin to open loop with FreeStyle , not a really good CGM but thanks to all presentation, french doctors are still not familiar with DIY looping.

Alan Cohen12:11 PM
Yes, of course. Need to be built by an ISO 13485-certified facility. Design needs to demonstrate safety via IEC 60601 etc. (I’ve developed roughly a dozen fda-cleared devices)

Dana Lewis12:11 PM
@c4481080 that's great. I know several countries are starting more pushes in that direction, but looking forward to the day where anyone who wants CGM can access it. but that's definitely a start.

Dana Lewis12:11 PM
@F5OEO hello and thanks for your work in the community!

Jeff Cacossa joined  the room.12:12 PM

Scott Leibrand12:12 PM
@Alan Cohen so if you meet those requirements, would that allow FDA approval?

BTW everyone, I will be posting a transcript for this chat right afterward. In case you miss any links, etc.

Dana Lewis12:12 PM
Thanks @Dan Maloney

Chris joined  the room.12:13 PM

c448108012:13 PM
@Dan Maloney Thanks, maybe also add is to the blog post as well

Dana Lewis12:14 PM
For anyone joining the chat who doesn't have diabetes, here's some background on what an APS (artificial pancreas system, or closed loop system) can do for those of us living with diabetes:

Alan Cohen12:14 PM
@Scott Leibrand That, plus submit other data and documentation that FDA wants. Performance testing, probably usability testing, requirements-specs-architecture... testing reports etc.

c448108012:14 PM
Is there a good site that maps all the loop possbilities ?

Dana Lewis12:14 PM

Dana Lewis12:14 PM
or cool, the gif loads too

leb904912:14 PM
How accurate are CGMs compared to regular test strips? My girlfriend works at a rehab clinic and she has seen them be way off.

Sean Hodgins12:15 PM
I've seen the rebattery, but I think OS trasmitter would open up possibilities for better data accessibility. I'm also out of pocket for my insurance and transmitter just died. I will probably end up just going to the G6.

Hochkbry joined  the room.12:15 PM

Jeff Cacossa12:15 PM
Anyone have links to efforts to swap out batteries on a g6? I saw one with lots of grinding and not much success as far as i could tell. Or rewiring to interrupt the circuit?

Dana Lewis12:15 PM
@leb9049 it's a little bit of apple and oranges - test strips are more likely to have issues due to lack of skin cleanliness, etc and are like a polaroid snapshot. CGM is like a moving picture of the BG. They can both be inaccurate, but with CGM the trend provides additional context, too

c448108012:16 PM
@leb9049 I think it depends on many factors, We had bad start with the first gen CGM of medtronic, but their new stuff works ok for us

leb904912:16 PM

Irene C. S.12:16 PM
Hi everyone! I just want to thank @Dana Lewis, you're such and inspiration for me. I'm currently (trying) developing a non-invasive glucose sensor for my Master disertation.

Dana Lewis12:16 PM
@leb9049 some CGM require calibration, and some of the new generations are factory calibrated

Bill Staffeldt joined  the room.12:17 PM

Dana Lewis12:17 PM
@Irene C. S. welcome! and cool!

c448108012:18 PM
@Irene C. S. cool in which part of the body ?

Andrea Limbourg joined  the room.12:18 PM

Scott Leibrand12:18 PM
@c4481080 has a good overview of the different closed loop options

Dana Lewis12:18 PM
@leb9049 this is a piece Chris Hannemann did a few years ago on meter accuracy:

Jon Cluck12:19 PM
@Jeff Cacossa most of the people who are doing it have a facebook presence, if you do some searching there

Andrea Limbourg12:19 PM
Hi all. A little late as I was updating my Loop to dev branch :)

Dana Lewis12:19 PM
@Andrea Limbourg good reason :)

Kyle J. Rose12:19 PM
Hi @Dana Lewis and @Scott Leibrand . Nice to "see you". Thank you for doing this chat and all of your continued work for the community. I realize this is a bit of a broad question how can PWD associations/non-profits and individual patient advocates get more involved? (understanding that many are already).

Andrea R.12:20 PM
@Andrea Limbourg I just did the same the other day. It's AWESOME.

c448108012:20 PM
I hate using facebook as a research platform :)

Scott Leibrand12:20 PM
It probably depends on their focus. Are you thinking of a PWD / non-profit that's focused on policy advocacy? Tech? General community support?

Scott Leibrand12:21 PM