We (now) have a gentle downward slope on the cgm and the Basal-IQ catches it and flatlines it almost exactly at 80.
The tail on the Fiasp seems quite short for us. I have been considering bringing it back further to 2-1/2 hrs.
Which just means we make more use of extended bolus then we did previous while on Humalog / Novolog. As well, we do not wait for 3 or 4 hours after eating to make additional bolus corrections but rather about 1-1/2 hrs after eating (depends on food obviously) if the BG looks a little higher than expected we will hit it with another unit or so.
Between the Fiasp and the Basal-IQ this is really letting us do much more fine tuning then we were able to do previously. The update is totally blowing me away. This is way more then I expected.
Plus to have the algorithm being so incredibly simple that anybody can understand how it operates. That is just awesome. I am a nuts-n-bolts guy. I like to know how the guts of things operate.
Tandem has a good page that completely explains this (you just have to scroll down on the page which at least for me was not immediately obvious)
How does Basal-IQ Technology work?
Basal-IQ Technology uses a simple linear regression algorithm that predicts glucose levels 30 minutes ahead based on 3 of the last 4 last consecutive CGM readings. If the glucose level is predicted to be less than 80 mg/dL, or if a CGM reading falls below 70 mg/dL, insulin delivery is suspended. Insulin delivery resumes as soon as sensor glucose values begin to rise. Insulin may be suspended for a minimum of 5 minutes and a maximum of 2 hours within a 2.5-hour rolling window.
That is it. The entire workings of the algorithm.