OpenAPS

[Paulbreen]

Has anyone had or has experience with the openAPS community? I’m very curious about what they are doing and would like peoples opinions on them and their ideas

 

[trophywench]

Never heard of it. Hence - no opinion.

 

[Inka]

No direct experience. I’ve googled and read info from them before and found it interesting. My pump is one that could be looped if I had a CGM so I’ve read a bit recently and over the years.

 

[helli]

I have a friend who has a diy loop.
She loves it and her DexComm graphs are amazing.
She eats a low carb diet.

I am less convinced as she still has to pre-bolus for food (mostly protein because of her low carb) and manually reduce her basal when exercising.
I believe the limitation is the speed of manufactured insulin compared to the speed of the stuff that a heathy pancreas can create.

 

[Paulbreen]

Well Just for Jenny they are a group of people who are engineers, programmers lay people who are trying to achieve an artificial pancreas in their spare time, google openAPS.org.
It’s very interesting to see that beyond the big organisations there are people who are doing something about this disease and I wondered if anyone had any interest in their community.

as I was poking around in the old inter web before sleeping last night (was waiting 2 hours for my CGM to warm up for no reason that I can find) and venting my frustration at the 70% viable pump/CGM system I like to use to keep me alive and kicking I found something new and very interesting for us pumpers, take a look at the video in this link
They say ready end of this year early 2021

YpsoPump with mylife Loop - mylife Diabetescare – International

Combine the small and lightweight mylife YpsoPump with the mylife CamAPS FX app and start looping.

www.mylife-diabetescare.com

 

[Paulbreen]

I have a friend who has a diy loop.
She loves it and her DexComm graphs are amazing.
She eats a low carb diet.

I am less convinced as she still has to pre-bolus for food (mostly protein because of her low carb) and manually reduce her basal when exercising.
I believe the limitation is the speed of manufactured insulin compared to the speed of the stuff that a heathy pancreas can create.

That can be overcome, AI will come into its own for this kind of application in the very near future, my company has been working with AI for the last 10 years, whilst not in a medical field if pushed we could develop something that learns an individual’s habits, how they use their medication and their lifestyle habits and make sure their insulin is balanced in their system.
Most of us now carry a very powerful computer around in our pocket these days so it’s just a matter of time. While the diabetic medical community have been responsible for researching the cause and effect of diabetes for the last XXX years they haven’t progressed very much in a realistic treatment or dare I say cure, that’s where engineers will take over and with the powerful tools available to them these days will at least get a realisable treatment

 

[mikeyB]

OpenAPS is just a community of developers aiming to produce a system mimicking an artificial pancreas, presumably by writing additional algorithms to sit on top of an existing loop system which sends alerts for highs, and switches off insulin for lows. It’s free, and open access allowing other nerds to fettle it.

The reasons why this is an exercise in futility are manifold. The companies that make pump loop systems, with their vast resources and experienced developers are all aiming for such a system. I don’t imagine it’s easy, which is why such a system doesn’t yet exist. It will soon, because it’s the next logical development in pump design - the complete artificial pancreas.

Secondly, such a system will have to be clinically tested to gain CE accreditation, so any system developed by OpenAPS will have to do the same. That costs money.

Finally, as Inka points out, they have been trying this for some years. It’s difficult because at the end of the day, it all depends on the accuracy of sensors, and as CGM users know, the current generation of sensors isn’t accurate enough with very high or very low scenarios, and that will persist whatever fancy system you add on top of that. And you can’t do that without venous sampling, so you need an implanted venous blood sensor. Such things do exist, in ICU units, but it does involve skilled medical input, and does have risks.

So all that considered, I don’t think we wIll ever have a hands off artificial pancreas, because in truth it isn’t worth all the bother. We’ll never escape BG testing.

We’ve had nearly a century of people with diabetes surviving a full life using insulin injections. It’s hardly an intolerable burden. CGM and pumps make life easier, but never free from some attention. If you think that is an intolerable burden, you need to change your mindset, because the alternative is a messy death, as it always has been.

 

[Bruce Stephens]

It’s difficult because at the end of the day, it all depends on the accuracy of sensors,

And their speed, and the speed of the insulin. If you're using an insulin where you'd normally prebolus 15 minutes before eating, an automated system isn't going to be able to match that without help since a CGM's not going to see much until half an hour (or more) after you'd have prebolused. Similarly for exercise.

(Maybe it could be OK: while the system's a bit late when you eat, maybe it can handle things so much better overall it works out better. And maybe for exercise you just know to have a quick snack right before you start exercising (or a few minutes after, after a bit of a warmup, say), to give the system time to notice.)

(Having said all that I have seen at least one report of an experimental system that works really well. I think it had pumps for both insulin and glucagon or something. I imagine calibrating the CGM regularly was also involved. I could also imagine a system working well with someone who mostly ate low-carb, so lessening the spikes.)

 

[Paulbreen]

OpenAPS is just a community of developers aiming to produce a system mimicking an artificial pancreas, presumably by writing additional algorithms to sit on top of an existing loop system which sends alerts for highs, and switches off insulin for lows. It’s free, and open access allowing other nerds to fettle it.

The reasons why this is an exercise in futility are manifold. The companies that make pump loop systems, with their vast resources and experienced developers are all aiming for such a system. I don’t imagine it’s easy, which is why such a system doesn’t yet exist. It will soon, because it’s the next logical development in pump design - the complete artificial pancreas.

Secondly, such a system will have to be clinically tested to gain CE accreditation, so any system developed by OpenAPS will have to do the same. That costs money.

Finally, as Inka points out, they have been trying this for some years. It’s difficult because at the end of the day, it all depends on the accuracy of sensors, and as CGM users know, the current generation of sensors isn’t accurate enough with very high or very low scenarios, and that will persist whatever fancy system you add on top of that. And you can’t do that without venous sampling, so you need an implanted venous blood sensor. Such things do exist, in ICU units, but it does involve skilled medical input, and does have risks.

So all that considered, I don’t think we wIll ever have a hands off artificial pancreas, because in truth it isn’t worth all the bother. We’ll never escape BG testing.

We’ve had nearly a century of people with diabetes surviving a full life using insulin injections. It’s hardly an intolerable burden. CGM and pumps make life easier, but never free from some attention. If you think that is an intolerable burden, you need to change your mindset, because the alternative is a messy death, as it always has been.

While I do respect your view point it seems a little pessimistic, without people pushing boundaries of what is possible we’d be still injecting pig insulin. The current Covid vaccines are a case in point, the common cold, a virus remains incurable yet within 9 months we have a vaccine on the market which used completely different technology to all the previous efforts at controlling viruses basically by injecting a small amount of the virus you are trying to control and then praying the persons own immune system takes over and protects them
The big players in the diabetic world are not the hard working medical researchers, doctors and professors they are the insulin producers where the only thing that matters is the bottom line, they have no interest in finding a treatment or a cure and by comparison their budgets are gigantic compared to the best of the best research departments or even the hardware manufacturers for pumps and CGM type technologies.

 

[helli]

@mikeyB I do not share your pessimism.
Just because something is developed by a community of developers does not mean it will not be adopted by vendors and approved for medical use.
Linux is a very popular example of software developed by a community and now in use in critical software. It passes rigorous quality standards and has been adopted by major vendors. Hey, it is the basis of iOS used by Apple.

I expect something similar with closed loop systems. Community developed code will be hardened, tested and adopted by pump and CGM vendors.

 

[everydayupsanddowns]

I don’t share the pessimism either!

In fact I think it can be argued that (because they are unhindered by formal FDA and other regulatory approval) the DIY open APS and ‘nightscout’ communities have really put a rocket under the device manufacturers because they have made so much progress so quickly.

I haven’t been at ‘Rise of the machines’ for a few years, but there was a company (maybe Bigfoot?) who were trying to get the FDA to approve the open source algorithm as part of a hybrid closed loop.

I know several PWD including respected diabetes consultants who have adopted open source options for their own use.

I am trying to remember the name of a new member who I think was using AAPS.

Good luck with your investigations @Paulbreen

 

[everydayupsanddowns]

Found the thread!

It was new member @Shaolin

https://forum.diabetes.org.uk/boards/threads/not-so-newbie.90104/

 

[Paulbreen]

I really do feel that we are on the cusp of important breakthroughs, I work in industrial robotics and we have been harnessing AI (artificial intelligence) for at least 10 years,
We no longer need to fully program a machine to follow a step by step routine, we are able to let the machine make decisions within defined parameters. It’s not just the hardware that I see morphing into this new realm but if the drug suppliers want to keep up they are going to have to evolve their products or some younger, sharper supplier will take their businesses.
We have all heard “it will be ready in 5 years” before but actually now it’s the first time I have started to believe it in forty odd years of diabetes.

 

[mikeyB]

It’s not pessimism about closed loop systems that my post is really aimed at. It is the glucose sensing. To have a true artificial pancreas, you need a sensor that can cope with all readings - a pancreas does. A real pancreas responds to blood glucose levels. Show me a sensor that reads blood glucose levels. There aren’t any. They all work with interstitial fluid, and until you can contrive a system that samples blood glucose that can continually operate without bleeding you to death, you will never have an external artificial pancreas. Or, indeed, an internal one unless you want an insulin supply tank and batteries secreted about your body.

Until you get the sensors right, it doesn’t matter how small or clever the technology gets. And not a single respondent has addressed this issue, so if anyone would like to throw the sunlight of their optimism on that particular issue, I’d be more than pleased to hear it.

 

[Docb]

I would be interested in that as well @mikeyB. Most technology development has a really hard bit in it somewhere which the proponents sort of gloss over. On board hydrogen storage in fuel cell powered transport and battery power density in electric cars are two others that come to mind.

 

[Paulbreen]

I would say look at the engineering challenges someone like Elon Musk and his TEAMS have over come to make space ships reusable, those booster rockets returning are amazing sight, the sensor technology is out there already it just needs the right set of minds to develop it for the diabetic application.

 

[Docb]

Engineering problems are generally soluble with effort, money, brains and application. I'm convinced that the next real advance in medicine will be in application of non invasive sensors, in particular monitoring neurological activity.

It's when you come up against thermodynamics and Newton that the fun really starts.

 

[Inka]

The DANA RS pump, which I have, was developed in consultation with the #wearenotwaiting people. I feel OpenAPS is a great idea and has solid ideas and people behind it. They’ve done so much.

There’s nothing like having Type 1 yourself or having a family member with it to spur people on. People without Type 1 just don’t get the cognitive load of it all. Looping reduces that cognitive load, and is an amazing idea.

@helli I’ve seen a looping friend’s graphs too and they’re incredible. I could see how the algorithm pushed down their blood sugar after a pizza and how it constantly adjusted itself. Absolutely brilliant!

 

[Bruce Stephens]

@helli I’ve seen a looping friend’s graphs too and they’re incredible. I could see how the algorithm pushed down their blood sugar after a pizza and how it constantly adjusted itself. Absolutely brilliant!

Yes, it feels like even if the CGM delay (and slowish insulins) can't be improved much, we could plausibly end up with systems with two buttons: one to press before eating something significant and one to press before exercising significantly, with the system handling the details.

 

[Inka]

The two main weak points in my opinion are the difficulty in getting a proper CGM prescribed (or should that be “impossibility”?) and the issue of pump cannulas and bad sites. Most of my unexpected highs are due to poor sites.

Much as I respect and like the idea of looping, only a real cure will be reliable. As said above, insulin companies like to make money and invent ever newer ‘improved’ insulins, but if only all the money wasted on them had gone towards research towards a cure. And yes, I do mean “wasted”.