Reasonable expectation for meal BG peak

77 views
Skip to first unread message

Luis Betancourt

unread,
Dec 16, 2016, 9:14:01 AM12/16/16
to OpenAPS Dev
Hello all, 

Considering Openaps is not originally meant to manage meal boluses, I have been wondering lately what I should actually expect to be a good meal BG peak, how high and how long, and if possible, how can one practically achieve best results?

Thanks a lot for your feedback.

Rgds
Luis

Dana Lewis

unread,
Dec 16, 2016, 12:39:32 PM12/16/16
to openaps-dev
There is a giant "it depends" on a lot of things - what each person's starting BG tends to be before meals, other confounding variables like exercise/stress/etc...so looking forward to hearing people's thoughts on what is "good" *for them* but keep in mind for anyone reading this thread that everyone has different goals, comfort levels, and situations.

That being said, here are some of my approaches that help me achieve my best post meal curves:
  1. Getting in the habit of being aware of netIOB going into a meal. Starting out with negative IOB is pretty much guaranteed to give me a spike, even if I bolus 100% perfectly for carbs and the loop kicks in, because without peak insulin activity at the moment the carbs go in my body, they're destined straight for the blood stream rather than the liver.
  2. So whenever possible, I try to trigger "eating soon" mode (a temporary target to 80 about an hour before the meal is optimal; or you can do the same target 80 with a manual bolus) up to around an hour before the meal. Some background posts on "eating soon" from my perspective: https://diyps.org/2016/07/11/picture-this-how-to-do-eating-soon-mode/ . Even half an hour helps, especially if you've been low temped for a long time heading into that meal.
  3. Make sure OpenAPS knows about any carbs. If your rig is hotspotted, it can pull carbs entered via NS, but if you're in offline mode (CGM plugged into a rig or ble-based), enter carbs on the pump through the bolus wizard. With knowledge of carbs, the system can react more quickly to carb-induced rises if they are "uncovered" by your bolus insulin, compared to if it just observed a rise with no carb information. And if you have slower than usual carb absorption (because of the meal composition, or you walk home from dinner, or for some reason the carbs seem to be disappearing for another unknown reason), this helps you be able to more easily (in my opinion) deal with later rises when the absorption is shifted back in time. If/when you feel comfortable with the basics of how oref0 works, you can enable "advanced meal assist", which is a more sophisticated feature that helps assist with meal rises more quickly, especially those that are shifted back from slow absorption. (This also enables the three "purple line" predictions you sometimes see in Nightscout, which also helps a human decide later if they want to do more manual action to influence future outcomes, or just let the rig take care of it.)
  4. I've also observed I have eliminated more post-meal crashes by looking at the prediction lines, and being conservative about any "more" bolus I might be inclined to do by observing a larger than usual spike. Usually, the loop has been adding enough insulin to bring me down to range, and I just have to accept that it will take a bit longer before I can get there. Also, instead of doing another .5u or 1u bolus at the top/near the top of a more extreme spike, I often achieve the same results with adding a post-meal "eating soon" (aka, again set a temp target of 80 mg/dL for half an hour or an hour). After the meal, this helps the loop do a little more insulin then (by targeting the lower than the normal post-meal target), but just for a small period of time; then it'll resume targeting the normal target - so this is almost like a tiny super-bolus post-meal. But, this also increases your chance of a post-meal low, so I'd only do this when you have time to pay attention and watch the BG coming down and can balance out with an activity mode or a 4g carb if needed if you end up overshooting.

With all that in mind...in optimal conditions using the above strategies, my ideal/most common meal scenario is one where I start 100-120 mg/dL and most of my meals then have a max "peak" at 150 mg/dL or so. Usually (based on eyeballs, not actually averaging the data) this means about a 20-30 mg/dL rise. Sometimes more, especially if I didn't get eating soon going and had negative IOB; sometimes less of a rise or no rise at all. Keep in mind for context of my personal situation, I don't do particularly low carb - I know there's some talk in the broader diabetes community about these results "only being possible with low carb" - but with my experience, getting the insulin timing right is the key to (my) success, even with higher carb meals.

-Dana


--
You received this message because you are subscribed to the Google Groups "OpenAPS Dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openaps-dev+unsubscribe@googlegroups.com.
To post to this group, send email to opena...@googlegroups.com.
Visit this group at https://groups.google.com/group/openaps-dev.
To view this discussion on the web visit https://groups.google.com/d/msgid/openaps-dev/f31cfa93-2dc8-451a-953b-54c0eea46b13%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Dana Lewis | http://OpenAPS.org | http://DIYPS.org | @DanaMLewis | http://www.linkedin.com/in/danalewis
"Doing something for someone else is more important than anything you would do for yourself."

Dana Lewis

unread,
Dec 16, 2016, 12:53:29 PM12/16/16
to openaps-dev
(One more note I forgot to add, in the flavor of "don't compare your behind the scenes to someone's finished footage" - for further context about some of what I shared from my perspective, I've been using a DIY closed loop now for 2+ years...so I have a *lot* of practice with meals on a hybrid closed loop.)

-Dana

Luis Betancourt

unread,
Dec 21, 2016, 11:31:02 AM12/21/16
to OpenAPS Dev
Danna,

thanks for this insight. I do appreciate it.

Two questions:

-Not sure I understood the relevance of entering the carbs to the pump. I understand that bolus wizard is good but how does this help openaps algorithm to improve its behavior.

-How do you trigger "Eating soon logic" besides doing it manually?

Cheers
Luis

Scott Leibrand

unread,
Dec 21, 2016, 1:34:18 PM12/21/16
to opena...@googlegroups.com

On Dec 21, 2016, at 10:31 AM, Luis Betancourt <bet...@gmail.com> wrote:

Danna,

thanks for this insight. I do appreciate it.

Two questions:

-Not sure I understood the relevance of entering the carbs to the pump. I understand that bolus wizard is good but how does this help openaps algorithm to improve its behavior.

OpenAPS meal assist uses carb entry (to the pump or NS) to know when to high-temp after meals. 


-How do you trigger "Eating soon logic" besides doing it manually?

By entering a temporary target into Nightscout. There is info in the docs on various ways to do that. 

-Scott


Cheers
Luis

--
You received this message because you are subscribed to the Google Groups "OpenAPS Dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openaps-dev...@googlegroups.com.

To post to this group, send email to opena...@googlegroups.com.
Visit this group at https://groups.google.com/group/openaps-dev.
Reply all
Reply to author
Forward
0 new messages