So, I know folks on here do not use openAPS or Loop for the most part, but I just wanted to give a shout-out to a new feature we’ve been using. It’s called super-microbolus and in the two weeks we’ve been using it, it has dramatically improved my quality of life, as well as improving my son’s BG.
For those who are not familiar with the original version of openAPS, it issues a temporary, 30-minute high or low basal rates based on the rise rate (and carbs eaten). Every few minutes it reevaluates its predictions and then reissues a temp basal if the predictions have changed.
It definitely helped us smooth out the highs and lows, and it gave us reassurance that our son wouldn’t go into DKA from a high assuming that his pump site was attached.
But there were persistent frustrations for us. For one, because openAPS was delivering insulin over a 30-minute window, that could mean the insulin is about 45 minutes behind when it’s needed. So for any rapid rise, you risked going high and then eventually going low, because of the additional time lag (on top of the insulin time lage)
Then too, we had tons of issues with WiFi at our son’s school. If he went out of range of the openAPS box and it issued a high temp basal, for instance, and then he stopped rising quite as fast, he would get a 30-minute temp basal, which could sometimes add up to an additional 0.3 or 0.4 units than he needed. Not a ton more, but enough that he would need a gummy.
And for very fast rises (more than +9 per five minutes) I just knew it wasn’t fast enough to diminish spikes so I would resort to both manual sugar-surfing and openAPS, which added additional layers of uncertainty. My son’s BG control was pretty good – better than most toddlers out there — but there were still many days that felt like a roller coaster. I felt like if I lost a reading or the box went offline, I needed to do something about it, sometimes going to school. And it also meant I had to really keep an eagle eye on his numbers at school, both when he was rising and dropping.
However, two weeks ago we implemented super-microbolus and our day-to-day management looks very different. Basically, instead of issuing high temp basals, it predicts how high our son is likely to go based on blood sugar rise, carbs eaten, and insulin on board. It calculates how much insulin he likely needs to stop that rise (maybe he needs 0.3 units) – then it looks at his basal rate (say, 0.2 units per hour). Then it makes a calculation: How much of the 0.3 units per hour can be given upfront, immediately, so that in the event he starts dropping or doesn’t need all of this insulin, the same amount can be subtracted from his basal rate during the course of the DIA? It then immediately issues a temp basal of 0 after the bolus for as many minutes as would be needed to subtract the extra insulin from future basal. It recalculates the same thing every few minutes. In other words, it is giving mini-superboluses all the time, always keeping in mind that any additional insulin it needs to give can be subtracted from future basal if needed.
This slight change in logic has made a huge difference for us. For one, he gets all the insulin faster, so he doesn’t spike as high – instead of capping out at 300 with fast changes he usually caps out at 190, even with no intervention on my part. And because it just needs to be connected to WiFi for the time it takes to deliver a single bolus, there are fewer situations where he gets a long high- or temp-basal and gets more than 0.1 units extra above what he needs. So he hasn’t been experiencing as many lows or preemptive low-treatments.
This has also been huge for his lunch. Every day at school, he eats his lunch and if I bolus for all of it upfront, he goes low. But if I don’t bolus at all he goes to 300. So I need to bolus for about 30% of his lunch upfront – and then an hour or two later, once he falls asleep, he starts to rise, and how much insulin he ultimately needed for the remainder of his lunch was always a huge guess. Now I just let the algorithm take care of the second rise and he rarely goes above 200 – and is not going low either. We have also been just bolusing some fraction of the carbs upfront at dinner (depending on the food, between 50 and 80%), which gives us a buffer when he eats less than anticipated, or if he’s just running low.
Also, I no longer sugar surf nearly as often; i trust that the system will be able to catch slight miscalculations of carbs or unexpected rises just about as well as I could, because it basically is doing sugar-surfing, just a little more precisely.
In the last two weeks our average standard deviation has gone done by 8, his time in range has gone up by 5%, his time spent low and high have both gone down.
It’s been great!
