Not that it matters much, Loop runs its algorithm every 5 minutes, timed by a clock on RileyLink and initiated by a Bluetooth wake-up call from RileyLink to iPhone.
okay that’s different from openAPS right? I remember openAPS running pretty much constantly, but usually every minute or two…
No, OpenAPS also runs at nominally 5min intervals. Both systems may do multiple things in between, such as attempts to read pump history, or update predictions based on new bg data, but dosing is updated at approx 5 min intervals, unless of course user initiates a bolus or anything else asynchronously. The 5-min interval is dictated by the 5-min interval at which we are getting bg readings - it does not make much use to make any more frequent dosing updates.
At home, he eats when he’s hungry ASSUMING his BG’s are conducive for it. We actually think the fact that he’ll be regimented at school is going to help us - if for nothing else, to “dial in” the correct settings for those regimented meals at school.
Not sure about you but I’m so nervous.
I’m nervous too. But because we’ve been in preschool and dealt with all that, I’m a little less nervous.
You know what I’m most nervous about though? It’s not his diabetes…it’s the attention that I’m sure no kid wants to have on them. It may not bother him now, but it may. I hate that he needs “special attention”. Other kids…perceptions…comments, bla bla bla.
Thanks for clarification. I had the remote control for my 522 (many years ago), which I thought could only do bolus. So was thinking that was the only external command it could process.
Thanks for this. Very helpful and a good reminder to review the carb screen!
Last night Samson started the night high (not sure why, that’s to be determined later). Loop gently brought him down to target range and he stayed there all night. So as far as I can tell the nighttime rates seem reasonably good. It’s possible the ISF could be very slightly changed, and it’s possible that one of the nighttime rates could be ever so slightly different. But to me this line looks pretty good.
I also upped his basal rate in the morning to 0.225, starting at 8a.m., because I had noticed a very slight rise even with no food on his MRI day. It’s hard to tell whether he was dropping because he ate his 55 g bagel shortly after (he also stayed fairly flat after that for a few hours, then started rising). So it’s possible 0.2 units/hr is slightly too little and 0.225 units/hr is slightly too much. I’ll have to check it at other times to see. But it seems like a decent enough setting.
He had been crashing around 1 to 2pm most days, so I lowered his noon basal to 0.375. I suspect his true basal need will probably be a lot closer to the 0.225 he has for most of the rest of the day, and that I was compensating for late-digesting food. Unfortunately, this is the absolute toughest time to do a basal test because it requires him to eat breakfast at 6am and then not eat lunch until 3pm. My goal for the summer is to get his settings dialed in really well for the period from about 12am to 3pm; nights are a safety issue and he gets off school at 1:50pm.
For now, my basal rate settings are:
Basal rates (units/hr):
12:00 a.m. | 0.125
3:00 a.m. | 0.175
4:00 a.m. | 0.2
6:00 a.m. | 0.225
12:00 p.m. | 0.375
2:00 p.m. | 0.225
7:00 p.m. | 0.275
9:00 p.m. | 0.4
10:00 p.m. | 0.425
11:00 p.m. | 0.3
I’m hoping that tomorrow I can feed him a bagel early, like at 6 a.m., and then keep the kids distracted long enough to test that 11 a.m. to 3 p.m. period. It’s tough though because bagels typically take several hours to digest for him.
It’s been a while since I updated this post. Basically, I’ve had a bunch of days where it was just not possible to do any kind of basal testing and certainly not for the periods between 11am and 11pm. I’m still struggling with getting numbers for the 11:30am to 4pm time frame, because it meansI’d have to feed Samson breakfast at 5am. And… I am just not willing to get up and wake him up at that time, LOL.
But this morning we had a day where we all slept late, then drove to a restaurant for breakfast, only to find they did not serve the one item either of my kids would eat. So we went to another one and Samson only started eating pretty late. Based on what I’m seeing for the past 18 hours, I’m thinking that his basal at night might be a little bit low at times, and that his ISF is too low as well (i.e. too strong). Why? Because it seems like he’s got those rolling hills overnight, where he’s rising high and then Loop cranks on for a while at a high basal rate, and then it drops him down low again and he has huge periods of 0 basal. He’s staying in range for most of the night which is good, and he’s not really going low, but it’s just a little less flat than is ideal. I suspect something like an alternation between 0.125 and 0.150 between 12am and 3am might work best, but I am not ready to do that level of fine tuning yet.
And here’s another night, earlier in the week, where he did seem pretty flat between 12am and 3am and as far as I can tell Loop wasn’t working, so I’m not sure if his basal is off, but his ISF might still be.
I also think the morning sometime between 8am and 11am might have a similar problem. For now, I’m just changing the basal rate at 10am because that’s when I have clear data on the settings. At that point, Loop was off and he was pretty clearly and consistently rising by 3 to 5 points every five minutes. I’m wondering if that may also be why he’s having huge spikes with meals the past few weeks; his basal is just insufficient so he is just skyrocketing the moment food touches his lips. I still haven’t had a chance to test Samson’s carbF at a time when he’s eating because his basals have not been dialed in for much of the day yet.
Here’s how I just changed his settings:
Basal rates (units/hr):
12:00 a.m. | 0.125
3:00 a.m. | 0.175
4:00 a.m. | 0.2
6:00 a.m. | 0.225
10:00 a.m. | 0.25
12:00 p.m. | 0.35
2:00 p.m. | 0.225
7:00 p.m. | 0.275
9:00 p.m. | 0.4
10:00 p.m. | 0.425
11:00 p.m. | 0.3
ISF (points/units of insulin):
05:00 a.m. | 275
16:00 p.m. | 250
23:00 p.m. | 345
I’ll keep folks posted on how these changes work. Do people see anything that jumps out at them? And does anyone have any tips for how to test his settings between 11am and 3pm, other than having him fast?
When I look at the July 3 graph, I see falling slopes. From 5:30am to 8:30, from 10 to 2, and from 3:30 to 4:45 or so. If I saw that in my graph for a few days in a row, I’d probably set a little less basal and see if that solved it. I very rarely do basal testing anymore, because I just get an impression of whether my BG is falling “all by itself” or rising “all by itself” day after day, which to me means basal. I adjust my carb ratio by seeing if I “always” end up with a stubborn high after a meal, or I “always” go low after eating. I know it would be better to do proper experiments, but I seem to get along well enough this way, and I find it much less burdensome.
When I look at the July 3 graph, I see falling slopes. From 5:30am to 8:30, from 10 to 2, and from 3:30 to 4:45 or so. If I saw that in my graph for a few days in a row, I’d probably set a little less basal and see if that solved it.
I think I would agree if I was seeing it all the time. But I am not seeing it on most of the other days in his week. I think that day he may have just been running low; for instance you can see I logged basically no carbs all day. He definitely ate more than 25 grams of carbs that day but probably what happened is I just fed him, waited to see if Loop could take care of things, and it did, so I didn’t log any carbs.
But I do think the 5:30 to 8:30 period is one where I go back and forth…sometimes it seems like he’s running high and I have to bump up his basal starting at 5am and sometimes it seems like he’s running low unless he eats in the morning. I think his “perfect” basal rate might be somewhere in between 0.2 and 0.225 for that time period maybe, on average. And then the next period I’m guessing is a step up.
I also know that Walsh says there should be just one “peak” and one “valley” in a sensible basal profile, and ours has two peaks (night and afternoon) and I’m not sure whether that’s a sign something is off, or if that’s outdated information.
I agree that your method of just looking at rough trends works well most of the time. That’s my preferred method too! More eyeballing and Fermian estimation than calculations with paper and pencil or spreadsheet. The thing is, for Samson that kind of rough sense has been not working so well for a while, and so I’m really trying to get a little tighter with my process. I mean, his numbers are okay. But his days feel so reactive and I know it’s not as good as i could be.
I think the added complexity of Loop means that you can get into a situation where things are working, then not working so well, you fix something, it seems to be a little better, then they’re not working again, you try another fix. You do a few things like that – and all of a sudden you’re in this weird place where you basically aren’t sure of any of your settings and just feel completely lost.
Yes, I agree. This graph to me looks like the basal is spot on for midnight - 3am, esp since it isn’t being influenced by Loop (not working), and still staying mostly flat. Then there is a falling trend 6-9am. If this is typical, then it looks like the basal is a bit too high, and you may need to decrease an hour or two before, at 4 or 5am.
Yes! I went through this, too! After so many fine-tunings, I finally reverted back to my original pre-loop values, and am having much better results, MOST days. On days where I do see issues, they require just temporary changes for that particular situation. For example, a change in food/eating out/exercise may cause a high. So to keep in range I may temporarily use very, very aggressive settings like reduce the ISF (stronger), or lower the Correction Range. But these are just temporary for a few hours. I almost never change my basal / IC. BUT, I’m an adult, and it is so much easier to manage one’s own self than someone else, much less a little one! 
I look at that as a sensible starting point, not a commandment. The body needs whatever insulin dose it tells you that it needs, and rules like “one peak per day” don’t supersede that. It’s likely easier to build upon a simple basal profile than one that wanders all over the place.
It almost sounds like you’re suggesting that science and engineering are more reliable than “by the seat of the pants” and “Kentucky windage.” I’m certainly not going to take the other side of that argument. I’m just saying that proper experiments to determine the correct parameters won’t get you to a flat line in real life, because there are so many factors that push the BG around.
From those graphs, his numbers are quite a bit better than just okay. You’ve set Samson up for a healthy long life, and to say otherwise would be an inaccurate assessment. As high achievers we’re prone to severe self-criticism as a motivational technique, but that can be a path to burnout. Striving for continuous improvement is a good thing, but don’t undervalue your achievements.
The pancreas is reactive. LOOP is reactive. Surfing is reactive. We use basal schedules and carb counting and pre-boluses and exercise basals to deal with some of the major factors that we can anticipate, then we simply must react to the outcomes. LOOP can take some of that burden, but for me the constant reactivity is kind of like dusting the furniture. It’s just something to keep up with. From time to time we can neglect it, but the need to keep after it never goes away. (Maybe when we get the dual-hormone pumps…)
Just keep dusting the furniture for Samson until he’s ready to do it for himself. Minor nuisance; no big deal.
As an aside, they’ve started making zoodles as an option at Olive Garden. It’s not on the menu but you can sub it for pasta. I’m guessing @TiaG isn’t a fan (if I recall from a previous thread) of zoodles. But I adore them and being able to order them out is kinda fun.
/thread hijacking off
I agree with @Katers87! Thanks for sharing with this thread. It’s been really informative and hopefully helpful for Samson too!
So I know kids can be mean. There. I said it. BUT I am frequently surprised by the ability that many kids have to help one another - especially to be aware of other kid’s dietary needs. I know a lot of kids who are looking out for their friends. Not out of pity. Out of kindness. They know and understand what gluten containing ingredients are, for instance, and they help spot them. And it doesn’t feel like special attention, it is just kids looking out for other kids. I’ve seen it with gluten intolerance and more noticeable physical differences (a friend of mine who’s 10 years old has a skin condition called EB for which he wears bandages over much of his exposed skin from the neck down - it’s noticeable). And yet he has little to no problems at school in relation to his different appearance. Actually the parents are the biggest problem, there is one or two every year that tend to think he’s contagious. I really think that kids can have a great capacity to care for their friends. Even in school! So I’m sending good thoughts for those kiddos preparing for school or in school with T1. Because I see the pride that pancreas normal kids take in knowing about diabetes and being able to help their T1 friends and it is seriously heartwarming.
Update on settings.
Yesterday I had a chance to do a very rough basal test from like 2pmish to 5pm, and this morning I had a chance to (very roughly) test an elusive half-hour period I have been trying to test for weeks: 12pm to 12:30 pm.
Here’s the first:
Now this is a “rough basal” test from about 1pm onward or so because he technically hadn’t gone 6 hours since his last bolus. What’s not shown on here is that because Loop was running, he actually had negative IOB starting at about 1pm and was still dropping. That told me his basal at 12pm was probably too high.
Then, due to camp pickups, he wasn’t able to eat a full meal till 5:30pm. Around 2pm he flatlined at around 70 and then was slowly rising, but he was hungry and I didn’t feel good about just letting him rise to see how long it took, so I gave him a small snack that sent him shooting up. And then Loop kicked in, giving him too much insulin – which tells me my ISF is still too low there, or that his basal rate is too high. I gave him a low treatment and after that, put Loop in “open loop” mode, which means it wouldn’t ramp up insulin if he started rising, but would instead give him his scheduled basal. He flatlined after that last low treatment and then started rising a bit before he ate pizza. I guess to me that suggests his scheduled basal at that time is probably close to fine.
This morning, he ate breakfast early because I had to take his brother to an appointment in the South Bay. I got back around 12:30, and he was going low. Loop was apparently off the whole time, but the first 6 hours don’t count really because he still had insulin in his system. To me, it seems like his basal at 12pm is definitely too high, and that his basal is probably a little too high from around 10am on – I had just changed that to 0.25 because a previous basal test suggested that his previous setting was a wee bit too low. Anyways, my guess is he needs some alternation between 0.25 and 0.225 during that time. Also, last night he ate pizza, which takes forever to digest. But with the weakened ISF I changed, he was flat – but stuck at a high number. So I’m going to strengthen the ISF (lower the number a bit) to see if that helps. Another possibility is that since I also weakened the basal from 3am to 5am, that may have been overkill and his original settings are fine. But I want to wait a few days to change that as he also had pizza, and that can take hours and hours to leave his system. Here’s the graph of last night and this morning – which is from NightScout because I can’t see the data yet on Tidepol.
Here are the new settings:
Basal rates (units/hr):
12:00 a.m. | 0.125
3:00 a.m. | 0.150
4:00 a.m. | 0.175
5:00 a.m. | 0.200
6:00 a.m. | 0.225
10:00 a.m. | 0.25
**11:00 a.m. | 0.225
12:00 p.m. | 0.25
2:00 p.m. | 0.225
7:00 p.m. | 0.275
9:00 p.m. | 0.4
10:00 p.m. | 0.425
11:00 p.m. | 0.3
ISF (points/units of insulin):
05:00 a.m. | 280
16:00 p.m. | 260
23:00 p.m. | 343
I don’t like to open the loop because I forget to close it again. Instead for this purpose I use the exercise mode to set the higher target for 45 minutes (if glucose) or 2 hours (if some other snack). That gives the temporary spike time to level off before loop starts trying to overcorrect because of it’s prediction of a continuing fast rise. I’m thinking I might like to try the jojo branch of loop and use a temp ISF to make loop not react strongly for 45 minutes or something.
Could be. I notice that the flatline at 2:30 came at just about the time the action of the 10:30am bolus(es) wore off, and I notice some basal resumption at 1pm and 2pm. So I’m not sure if the noon basal was too high or not. In any case it’s always acceptable to tweak it a little to see if it makes the next few days better, even in the absence of clear evidence. If you don’t like what the tweak does, it’s easy enough to go back to where it had been.
It looks like the 60g carb at 9:30am got a post-bolus rather than a pre-bolus. Hence the spike. Similarly, maybe the 55g at 5:45pm could have been handled more gracefully by some combination of pre-bolus and maybe larger bolus amount, rather than the chase and catch at 7pm and 8pm. But if circumstances make pre-bolusing inconvenient, you’re surfing it reasonably.
It looks like the 60g carb at 9:30am got a post-bolus rather than a pre-bolus. Hence the spike. Similarly, maybe the 55g at 5:45pm could have been handled more gracefully by some combination of pre-bolus and maybe larger bolus amount, rather than the chase and catch at 7pm and 8pm. But if circumstances make pre-bolusing inconvenient, you’re surfing it reasonably.
Yeah we almost never do a prebolus mainly because it’s so unreliable what Samson eats. But also because I suspect our basal was a bit higher and that essentially allowed us to be sloppy with bolusing. Downside is you sometimes have to feed the basal but it wound up working out. I think I’m not yet at the point where I feel confident in all my basal rates to figure out the carb ratios yet, and whether I need to figure out a better bolus strategy. We may have to prebolus for some X amount, then give the remainder after he eats, for instance. I’ve also been manually overriding Loops suggested boluses a lot because it always suggests giving more upfront than I know Samson usually handles. For instance, for pizza I give 1 unit upfront, but Loop would suggest something like 2 units. That kind of thing. So it may be that with basals appropriate, Loops predictions are closer to the mark. Still lots of testing to be done!!
I highly recommend the JoJo branch with its overrides. There is a new JoJo, JoJo-beeps. It is the same as JoJo but includes a new feature (coded in the RileyLink branch) which generates two-beeps on bolus start, and one-beep on bolus finish. Very helpful to confirm that the bolus was actually performed.
If you haven’t already, make sure you use Autotune
It’s a very handy tool for reading your NS loop profile and providing feedback on tweaks the tool thinks you should make. I’ve found it EXTREMELY useful and has saved me so much manual testing (which can be very hard for someone Liam’s age).
Hi @ClaudnDaye, I have tried using Autotune but honestly I never trust the results enough to use them, partly because until recently we’ve been pretty bad about logging appropriate carbs. For instance, at one timepoint it tells me that Samson needs a 0.475 unit basal. I think it sees late food rises and thinks it’s a basal issue. But now that we’ve gotten better, maybe I should give it a go again.
