I love it! Thanks for sharing.
Update: My husband’s got the meowlink working and is reading and talking to the Medtronic pump. We should have an open loop running in a couple of days (not on a human). Looking forward to it!!!
You techies might want to read this article if you haven’t seen it yet:
http://www.hanselman.com/blog/ThePromisingStateOfDiabetesTechnologyIn2016.aspx
3 Likes
So my son is now using the closed loop at home with us and at night. What a game changer. It’s still buggy and not portable and we haven’t quite figured out the right settings. At this point I’m not sure it will lower his A1C, but just having the ability to go to sleep knowing that the AP is gently nudging down a high BG, or that I don’t have to be a human calculator every 5 to 15 minutes when a bolus goes off the rails, has dramatically improved my peace of mind. It doesn’t seem able to completely forestall highs, especially if they’re caused by a carbguess bolus, but it seems like the wild crashes or uncontrollable spikes are reined in a bit too.
Any tips for how to make it portable the fastest?
I had to send the kiddo to daycare today and felt so nervous and sad that it wouldn’t be there, operating in the background. I was really just enjoying bolusing only for meals and relying on openAPS to do the fine-tuning…
I feel the same way about not having it on me now.
The advanced feature, advanced meal assist may help with that.
I am still dialing in settings, but mine runs on my phone with the help of a RileyLink… which is pretty small. I only put it in a case so I don’t lose it 
Results, so far, are beyond what I expected - I don’t think I’ve EVER had a day this good in all the time I’ve had T1… I never even hit 130 today. No lows either - a slight dip after bolusing for a late lunch, but I only went down to 78. (and I did eat breakfast, but didn’t bolus for it, it was mostly protein).
2 Likes
@Sarah28, everything looks awesome, congrats!
Sarah28, what’s the difference between the algorithm you’re using and the openAPS system? I know the hardware is different but am wondering if the AP itself is different too!
I’m not super clear on the differences, but I do know that the algorithm for what I’m using was modeled after the one used in OpenAPS.
One thing that has struck me though, as a “sugar surfer”, is really how closely this follows what I would be (trying) to do anyways… BG creeping up, push it down… BG heading low, try to bump it back up a bit. It’s just doing it in a slightly different way (I’d be bolusing/carbing, this uses short temp basal rates). But now I don’t have to think about it as much, and it can do it while I sleep - those are really the biggest advantages for me. Bolusing for meals and such is still 100% my responsibility, so I don’t want anyone to think that this is some sort of magical device that “automatically” makes their BG perfect 100% of the time, it’s not, and it still requires plenty of work/attention to be successful.
1 Like
Very well said @Sarah28 - thanks! Such statement should be added to any success story with any APS. One has to have good understanding of bg control to make best use of any D tech, including APS, One can indeed think of OpenAPS reference algorithm (and similar algorithms based on OpenAPS) as semi-automated “sugar surfing”. For me, pretty amazing nighttime performance is still the main benefit of APS.
1 Like
Thanks, that’s helpful.
Yeah, we just started our son on this too, and we noticed right away that a) wow, what a relief to know it’s basically doing what we were doing (crappily, at random times) persistently in the background but that b) it can’t make up for a bad meal bolus, it can just shave maybe 50-75 points off a spike or put a tiny crash pad beneath a screaming low, or recover better if we slightly miscounted carbs. I’m still trying to get the best meal boluses down, because it really does seem to work better if you give it a fighting chance with good meal coverage. It also seems to be good for figuring out if your parameters like ISF, DIA, etc. are right.
But it’s impossible to overstate just how much of a difference it has made so far to feel like I don’t have to be constantly on it!
Your setup looks so cool and compact – I admit to a bit of AP envy! 
1 Like
That’s one fantastic day! Is that the Loop app you’re running?
Yes it is… for me as a G5 user this was the best solution for offline looping, any OpenAPS rig would have had a few significant downsides, primarily needing to carry around my Dexcom receiver, which I didn’t really want to do… I have a job (I teach figure skating) that would make it almost impossible to carry a rig like that around conveniently
Did you run the openaps loop at all? I wonder if one algo is more aggressive than the other. Have it all in the iphone and the riley link is pretty cool. Riley link communicates with the dexcom, right? What is doing the pump comms?
Briefly, but I wasn’t running a fully closed loop yet, and I switched over when my RileyLink arrived so I haven’t continued to work on my OpenAPS. I don’t think either algorithm is necessarily “more aggressive” because it all depends on user settings so that’s kind of a tough question. Either could probably be “set” to be more aggressive based on the settings that the user defines (max IOB, max Basal, ISF, DIA, targets, etc). Loop doesn’t actually have the ability to limit IOB the way OpenAPS does, so as a result I haven’t tightened up my target ranges (and they are not set as low as I’d be aiming for “off” either system) OR given Loop the ability to set very high temp basals yet - I’m taking it slow-ish for safety, and watching what it does - as I get more comfortable using it, I’ll lower/tighten my ranges and give it the ability to give me more insulin
really awesome. Thanks. That’s interesting that loop doesn’t have a way to limit basals. I think I’d take tightening my targest and max IOB slowly as well! Keep us posted on how it goes.
It does give you the ability to limit how much basal it can use, you just can’t set a max IOB like you can with OpenAPS. The net effect is likely the same when just restricting your max basal though… OpenAPS just allows (and recommends) restricting both when starting out for safety.
That said, I’ve not seen my IOB go very high at all on just basal insulin, even as I’ve allowed Loop access to higher basal rates… that may change as I increase the max though. However, I also haven’t really had anything resembling a low - I did dip down to 78 after bolusing for a meal, but I came right back up since I was eating so I don’t really count that. I’ve had several long stretches at 0 basal while I was trending downward which stopped that.
Hi,
Do the hardware requirements of the OpenAPS apply for the Loop + RileyLink as well? I.e. the pump model. It’s becoming difficult to get hold of the old pumps… I actually did a good thing by donating my old pumps to less fortunate people and now I’m in trouble because I don’t have one for this project. Irony of life.
OpenAPS is not an algorythm. It is a set of tools that allows you to communicate with a few different medtronic pumps (so far, more coming) as well as medtronic and dexcom cgm’s.
Openaps allows those communications with the pump to happen using several different devices including things like a rileylink, carelink stick, TI stick and others. Openaps can be run several different ways on different hardware including raspberrypi, intel edison, ios, laptops, etc.
There are a few different algorythms that compute temp basals. One of them is oref0, which is what many people use on linux based devices. Another is openaps-predict which also runs on linux setups and has also been adapted as part of Loop for the iPhone. There are a couple other algos that a few people used but are more privately built.
The two algos are different in some ways. Both are open to user edit to make their own and add bells/whistles and add/edit constraints. Both encourage people to make them your own.
I have used both for approx 3 months each. Both are good. Both required me to learn and adapt and fine tune my settings (insulin action duration, carb ratios, target rates, etc). Both are aggressive if you want them to be and both are conservative if you want that too.
My personal take is that oref0 worked slightly better for me. I think it was mostly due to advanced meal assist which has a means to detect when you are resistant for some reason and will become more aggressive.
I have used Loop/openaps-predict for the last 3mo and it has been working well too. It has some great features like exercise mode built in. It is very straightforward in how it calculates the basal adjustments. It displays what it predicts your bg to be over the next 3 hours which is very useful.
I think if you are considering an OpenAPS based solution you should look at both. Both algos and what platform makes the most sense for you. Hopefully soon we will have multiple algo choices for the iOS platform like we do with Linux.
CrushingT1D
2 Likes
Yes, same pump restrictions. I know people are getting close to having an omnipod solution. Animas ping has been cracked but does not allow temp basal so it’s not a true openaps option.