No matter if on a high- or low-carb diet, fat and protein is tasty as well in a steak with green beans as in a pizza with extra thin crust and extra cheese on top. And as a type 1 diabetic I need insulin to cover it if I want my bg levels to stay in range. [1]
Traditionally I would give an extended bolus. Something that does not fit very well into the world of a closed loop as now the algorithm should take care of how much insulin comes out of the tip of my infusion set.
As the loop now increases the insulin when the bg rises and the rise from fat/protein is rather slow, the loop helps quite a bit. Still not 100% satisfactory as for larger amounts my CGM hovers above target for hours - and any unplanned BG increase won't be compensated by the loop if it is already at it's limits
Fake Carbs?
Hinting the loop that it can give insulin more agressively after a meal works quite well by entering "fake carbs". This is an equivalent of "carbs" you expect a meal to have impact on your blood glucose. I simply enter those as carbs into my system. Maybe "carbs equivalent" is a better name?With the dynamic carb absorption model from OpenAPS that AndroidAPS also adopted this works quite well. The algorithm will detect a rather slow carb absorption and give insulin accordingly.
As a safety feature, the algorithm will assume a minimum carb absorption. This means that after a couple of hours it will believe your meal to be fully absorbed. For smaller and medium meals this is no problem as the rest will be handled by the closed loop system even in "no COB" mode. With larger meals I still do see a small rise a couple of hours after the meal.
Future Carbs?
 |
| Setting Event Time in carbs in Nightscout/careportal |
A trick to handle the late rise after a protein-rich meal is to enter some carbs in the future. This way the closed loop knows that there is some cabs and it still can work a bit more agressively.
This also has the advantage that some advanced freatures like Autosense and Autotune will not consider this late rise as some form of insulin resistance but attribute it correctly to being meal-related.
Extended Carbs!
Starting this article with and extended bolus we are now at extended carbs:
The new 2.0 version of AndroidAPS (and NSClient as well if you use another system) has the possibility to not only schedule carbs in the future but also distribute it over an interval. Just like an extended bolus would distribute insulin over an interval.
Time for some food! :)
This is my test meal: 400g of pork strips in tomatoe sauce:
This was the situation before the meal:
With the new carbs dialog I entered what I thought could be the equivalent carbs (25g), scheduled it 60 minutes into the future and set an interval of 2 hours.
I didn't give any bolus. I do have SMB (super micro bolusing) activated. You can see them as the small blue triangles.
This is how it looked like after I confirmed, the carbs are scheduled as multiple events with 15 minutes in between:
This is how it looked like during the extended carb interval!
The COB curve (orange) has a saw tooth like shape. Due to the COB, there were a lot of SMB so the loop did its job perfectly :)
Enjoy your meal!The new 2.0 version of AndroidAPS (and NSClient as well if you use another system) has the possibility to not only schedule carbs in the future but also distribute it over an interval. Just like an extended bolus would distribute insulin over an interval.
Time for some food! :)
This is my test meal: 400g of pork strips in tomatoe sauce:
This was the situation before the meal:
With the new carbs dialog I entered what I thought could be the equivalent carbs (25g), scheduled it 60 minutes into the future and set an interval of 2 hours.
I didn't give any bolus. I do have SMB (super micro bolusing) activated. You can see them as the small blue triangles.
This is how it looked like after I confirmed, the carbs are scheduled as multiple events with 15 minutes in between:
This is how it looked like during the extended carb interval!
The COB curve (orange) has a saw tooth like shape. Due to the COB, there were a lot of SMB so the loop did its job perfectly :)
[1] HOLT, S. H.; MILLER, J. C.; PETOCZ, Peter. An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods. The American journal of clinical nutrition, 1997, 66. Jg., Nr. 5, S. 1264-1276.
