OpenAPS new feature

You may have heard that OpenAPS got a couple new features in development…super microbolus (SMB) and Unannounced meals (UAM).  I’m currently testing out SMB features this week…seeing how they work, trying to wrap my head around some new ways of thinking, finding out if we have to change anything in our patterns.  I thought I’d report in on how that’s going.

We started SMB just before going out of town for a wedding.  Not a big deal except…lots of “perfect storm” events conspired against us in the usual diabetes way.  Anna wore a dress (hard to access pump and we had some unbolused food as a result), the music was loud and I misheard her food choice (and therefore gave her a bad carb estimate to bolus for dinner…she ate a large brioche bun and pasta.  I heard chicken.  yeah…you know where that went.), we happened to be in the only cell dead spot near LAX for several hours where I wasn’t paying any attention (and therefore I can’t say how the system would’ve behaved because it was offline).  We even had a couple random pump resets that made us need to reset the pump time (which proved to be important learning lesson on smb…pump and rig time must be within 60 seconds of each other).

So, a rough start just for my mental preparedness (although BGs on Sunday after the wedding were great).  We went back to normal non-SMB use for Monday while I read some more details about SMB and UAM.  I needed to understand more about the decisions and actions because I was not so pleased at first glance (there’s a lot more “suspend” after meals than I was mentally prepared for…I needed to understand how those played a part).  After a lot more reading, we restarted with SMB.  We are not yet trying UAM features…one step at a time.  Plus, we don’t usually have many unannounced carbs (thankfully Anna is a pretty diligent bolus unless she is wearing a dress, at a wedding, in a cell dead spot, with large brioche buns around.)

We restarted SMB yesterday and have been running it since.  I am very pleased with it now.  The basic idea is that SMBs are given in little increments earlier to help control a rising BG, as opposed to waiting the whole 30 min temp basal to deliver a needed insulin amount.  The program only gives as much as it can safely temp suspend for later…like the idea of a traditional super bolus…just on a micro scale and with the aid of a computer doing all the math better than my brain or time can permit.

Want to see some examples?  How about last night…



Here’s her 15g dinner (yes, low carb for this one).  But, it nicely suspended and prevented a low early after the bolus…and then SMB picked up on the tail end of meal when it needed to come back online.  See those little tiny boluses of 0.1 and 0.4 units?  Those are SMBs…automatic little boluses of insulin.  The only time she touched her pump in the above graph was the 15g meal bolus.

You can also see an earlier SMB help with controlling a rebound from a low treatment at school.  Why did she go low?  She gave a correction after cheetos-involved lunch without noticing that SMB already had done so.  So, note to kids…no corrections with SMB.  She’s pretty happy with that instruction.

How about her school mornings?  Here’s this morning…


Before DIY closed looping, school mornings were such a cluster @#$%.  We had huge basal set from 5am-8am plus an extra (always a guess) bolus when she left the house to help control the school morning nerves.  And then, this year, she has PE second period…oh but wait…not always second period because it might be a block day schedule.  You parents know what I’m talking about…always having to stay on my toes about what day of the week it is and make sure the bolus/basal program is the right one.  We regularly battled 220s in the morning using that system.  Plus interrupted her school day with about 5 text messages to try to get things right.

Then we got on Loop and OpenAPS…so much better.  We had to keep the basal patterns different for school days vs weekends when we used Loop…but that was ok.  When we went to OpenAPS, we’ve only used one basal pattern and autosens has taken care of our basal changes (magic?!).  But, we have still had about 75% of the mornings where she slowly climbs to 120s and even 130.  When her BG hits 130, Anna’s pebble will buzz and she mindlessly gives herself a set correction of 0.5 unit. (Parent note: Teens will not do math to check how much a loop has already given.  We finally told her to not give more than 0.5 unit for correction anymore, which equates to about a 15 point BG drop for her ISF.  That amount seemed to keep her from overcorrecting but also helped out on those days where she needed some help.)  We were pretty happy with this system because we no longer needed to text her, she wasn’t overdosing, and our basal switching around was no longer necessary.  Win-win-win.

And now two mornings on SMB have been flat-lines.  Like above.  It’s not a big sample set yet to draw conclusions, but I’m loving it.  I know SMB is playing a part (I can see the SMBs on my graph after all…none of those were administered by her or I.  She has been told to give no manual corrections while we test this SMB stuff.)

But also look…that 8g snack snuck in around 10am.  What is that?  I have no idea.  Fruit cup?  Likely.  Well she bolused for it but I have no idea what it is.  Here it comes…


I don’t interrupt with a text message asking what it was.  I don’t intervene.  But I do get to watch and that was fun.  So what was OpenAPS doing when that climb started?  It lowered her BG target. (This is a new feature, too.)  We normally have her BG target set for 90-90 mg/dl.  We have allowed OpenAPS now to temporarily lower her target when her BG rises like this to help get insulin in a bit faster.  Basically, it’s like having OpenAPS automatically set “eating soon” mode for me in these situations. That temp target adjustment also will be automatically stopped when not needed anymore.  Very nice.  (I will still need to watch how it behaves in other situations like exercise.)


OpenAPS also increased her basal reasonably and gave her a starting microbolus of 0.4 units based on this unexpected deviation up.

IMG_2796And then more BG data came in…and I could verify the 0.4 unit SMB had been given.


At that point, OpenAPS wanted to do a 0.1 unit SMB and lower that temp basal.  Things were slowing down, but not done yet.


So then we seemed to peak out about here…

IMG_2802Before starting a smooth decline…thanks to about 0.7 units of help from SMB, temp basals, AND suspended temp basals afterwards…because that’s how SMB works.  Able to give some reasonable part of the needed help up-front early and then suspend later.  Super nice.

IMG_2809But wait, that teen wasn’t done.  It’s lunch time now that I’ve spent all this time writing up the morning.  And so it begins for lunch…

IMG_2812OpenAPS will be there, helping with some SMBs and temp basals.







Note: I realize the BG control represented here is tight.  BUT, did you also notice that it was completely hands off?  I haven’t texted Anna.  She hasn’t done a correction.  I didn’t adjust a target, nor did I adjust a basal or ISF in over four weeks.  It’s been automated.  She has only done her usual bolus for carbs.

We will keep testing things out for awhile.  Keeping a close eye on things…looking if we need to change our habits, change our settings, or wear fewer dresses at weddings.  I’ll be back with a higher carb day (Panda Express anyone?) and more details on that (foreshadow though…she ate two Randy’s donuts for breakfast and in-n-out for lunch on the way home from wedding. And it was nice…I just didn’t grab screenshots.)

2 thoughts on “OpenAPS new feature”

Leave a Reply

Your email address will not be published.