WIP: Microboluses #1205
WIP: Microboluses #1205
Conversation
* New strategy for asset overrides. Fixes popup warnings * Add DefaultAssets.xcassets back to WatchApp target
* Allow overrides to be enabled via APNs * Update to NS sent notifications * Cleanup prints * Query override history for NS upload * Bump submodule revs * Handle indefinite duration override treatment uploads * Use separate representation for indefinite duration treatments * Bump cartfile rev * Send any override changes when uploading loop status * Bump carthage revisions * Bump carthage revision * Fix typo, and log device token * Carthage dependency revisions
# Conflicts: # Cartfile.resolved
* draft UI * New Microboluses Screen
# Conflicts: # Loop.xcconfig # Loop/Views/MicrobolusView.swift # LoopCore/Microbolus.swift
|
@ivalkou thanks for this branch. We have been using it for the last few days with our 4 year old and working great. One thing I will throw out for your consideration is whether it would be possible to have separate partial application settings with and without cob. In the case with cob This would help us deal with a late re-rise that happens with slow GI foods like pasta several hours later. We set the pizza absorption time, but with high temping it’s not enough. Personally I would set a higher partial applicantion with cob to deal with this, but would like to have a lower % for instanced with no cob so loop doesn’t overdo it. |
|
@ddaniels1 Adjusting partial application is probably the wrong way to deal with carbs. If loop is mishandling carbs for you, then something about its carb model needs to change, or how you use the carb model. |
|
@ddaniels1 are you seeing better results with slow food using this branch? We have the same problem. I don't think I have ever had loop keep her in range in the last half of slow foods. Even accounting for fat /protein in the initial entry, trying longer and shorter absorption, I end up needing to double the initial entry halfway through. Best results seem to be a separate fat/protein entry 90 minutes to 2 hours after the first entry. But I can't quite figure out the calculation for it. I was hoping MB would be the solution if the problem was that the temp basal changes are just too slow. |
|
@ps2, loops ability to work well depends on a significant amount of insulin coming from basal. For kids under 6, often 3/4 of their insulin comes from bolus. I believe this Amy be because per lb, they eat significantly more carbs than an adult, but that lower body weight requires a proportionally Lowe basal. Olivia was diagnosed when she was 1, she is now 4 and we are slowly seeing a transition to more basal, and therefore more effectiveness from the loop model. This is very common both in taking to other parents and our Endo. It has been my feeling that because of this their are some changes needed to accommodate this. I can’t say we know the answer and in reply to @Nightfoxy, microboluses have helped, but we often add carbs 2-3 hours later, or recently i came up with a pasta override that runs for 8 hours after a pasta meal at 150% normal. This also seems to help. |
|
We have noticed I think since the update running microboluses we are using significantly more power and can’t make it through the day without charging. May be coincidence and battery is going bad on phone but wanted to see if others have seen similar issues? I tried rebooting and the metrics show the majority of power being used by loop. |
|
@ddaniels1 I have noticed the same with my phone battery since running microboluses . |
|
FWIW - we have not noticed and difference in phone battery life since starting Ivan’s MB branch. using an older iPhone 7 with a battery replaced about 8 months ago. |
# Conflicts: # Loop.xcodeproj/project.pbxproj # Loop/View Controllers/CarbAbsorptionViewController.swift # Loop/View Controllers/StatusTableViewController.swift
|
All further development of this option is underway at https://github.com/ivalkou/LoopWorkspace. News and discussions in Telegram https://t.me/LoopAPSEng |
Microboluses use internal Loop system to suggest boluses when basal change is not enough, but it forces a part of that dosage in the beginning of every loop. Be sure that you settings are correct because this system is more sensitive to mistakes than the original one. Also, microboluses are a lot more sensitive to BG sensor noise, so definitely keep that in mind
For microbolusing to work, be sure that the allowed basal increase is as minimal as possible (difference between basal delivery limit and current basal rate). The less it is, the harsher microboluses get
The part of the dosage is also configurable. By default it is set to 30%. That means that if Loops suggests a bolus of 1 unit of insulin at the beginning of the loop, 0.3 units will be implemented automatically. You can increase and decrease this setting to make microboluses more aggressive.
Maximal microbolus amount is set in basal minutes. This value determines the amount of basal insulin from your current profile for the given number of minutes (not TBR). For example, if the basal rate is 1 unit/hour, the 60 basal minutes amount is 1 unit. Basal minutes are also used in other artificial pancreas system (like OpenAPS or AndroidAPS), that's why they are used in Microbolus configuration. You can also choose not to limit the microbolus amount.
Microbolusing limits can be set separately to the situations when COB is 0 and when it's more that 0
Microboluses safe mode can be set to one of three options:
Microboluses limits
Current BG is lower than a suspend treshold
Microboluses are off
Current BG is below the target range
Maximal microbolus amount is limited to 30 basal minutes
Predicted BG at any moment is below the suspension treshold
Microboluses are off
Eventual predicted BG is below the suspension treshold
Microboluses are off
Eventual predicted BG is below the target range
Microboluses are off
In any other case microboluses are limited by your settings