ga-zelle
commented
2 years ago Here we go with an example: Let's say the ISF is 40 and we change profile to 80%. We would expect a change in sensitivity of 20%. The adjusted ISF now is 40/0.8 = 50 Say we want to have a bolus that brings us down by 40mg/dl. Originally we needed 40/40=1U f insulin. With the reduced profile we need 40/50=40/(40/0.8))=40/40 * 0.8=0.8U of insulin. That is 20% less than at full profile!
So, it is not the ISF changing by 20% but the sensitivity changes by 20%. ISF is somehow just the inverse of what mostzthink it is. It is like an engineer from Europe discussing a 20% improvement in fuel consumption with his colleague from the US. The one thinks in l/100km, the other in miles/gallon. Because one metric is more or less the inverse of the other they never find a common ground.
MilosKozak
The following issue was started by https://github.com/klalle in AAPS docs and should be delt by app team as the question text must be updated. https://github.com/openaps/AndroidAPSdocs/issues/1126
This is something that tricked me when going through the objective 3. If you set a 90 % profile, the example states that the ISF is devided by 0.9, which is not a 10 % increase last time I checked... And the example image in the docs is also wrong (2.2/0.9 != 2.4) its more like 2.44444. (yes you have 2 digits on some numbers). but an 10 % increase would be: 2.2+0.22=2.42.
Yes it's close to a 10 % increase, but lots of us are engineers, and we will have a hard time with this question since you'd have to answer wrong to get it right! Probably more than me that had to check the source-code to get this right ;) image
To prove the point, lets take something larger: if you set a profile switch of 60 % (not even sure its posible) and an ISF of 10. 10/0.6 = 16,6667, but a 40 % increase of 10 is 14.0!
Or am I misinterpreting this on some way?