lol - Not a test.
Our pump does not have micros. So not something we would be experienced with.
But your explanation makes sense.
2 Likes
Here’s the thing… when I do a food bolus, it often knocks out my micros. It can knock them out for 15 minutes or for an hour— you never know. It even will continue micros from time to time— again, you never know. I remember asking you all about how the superbolus works, and learning that it takes the amount in your food bolus as well as some upcoming basal for the bolus but then knocks out that basal for the predetermined amount of time. My grammar is falling apart. Sorry. Anyway, that made a lot of sense, but this doesn’t. This doesn’t calculate any of your basal… it just knocks it out. So it’s one bolus for two purposes. Am I wrong for thinking this way?? If my food bolus knocks out my basal, aren’t I short some insulin??
Sorry if this is a silly question, but is a micro-bolus in place of what another pump would just consider basal? (I.e. is it just another word for basal?) Or is it something your pump does in addition to basal when it senses a BG rise?
2 Likes
It’s your actual basal… or is supposed to replace your basal. It’s a “dynamic insulin delivery” and is delivered every 5 minutes based on an algorithm that is kept secret from everyone.
2 Likes
Thanks for explaining! I have no useful insight for you, lol, but thanks for explaining 
From my absolutely inexperienced perspective, I’d agree with you that it seems counterintuitive for it to skip the micro boluses while you’re high and climbing, which if I’m understanding your graph right, is exactly what it did, no? It shouldn’t care if you have food bolus on board, because in my mind, that should be handling the carbs you are eating and doing nothing more. Unless you took more than it recommended for the meal?
Sorry, I haven’t the foggiest idea how this all works so I’d best not complicate matters 
1 Like
All I can contribute is that you have more courage than I do. I would flip my lid if I had the slide that you experienced from what appears to be 4PM to 6PM. Holy cow. I’m not saying that in a judgey judgey way. But holy cow that would freak me out for it to keep going down without dampening.
Side question: is auto mode serving you well in a fasting state (like overnight) currently? I remember it having trouble for waking up and having coffee, I think. I was just thinking that if it works out while fasting, maybe you’re onto something with it over-adjusting micro boluses after food boluses.
1 Like
I admire that. And I strive for knowing the “why” in all things bg in my life, too.
I hope you do nail down why Auto Mode is less than stellar. But I’ll also point out that this technology is in its infancy, IMO. There are things that I have to imagine will be improved with each new system or update that will hopefully make it work better each time. This may not be something for you to figure out at this juncture. It may end up that it is a thumbs up or a thumbs down bc it did or didn’t do something well enough for you. That might be good enough.
If you are the person to determine why it’s not great at what it does, I’m sure as hell glad I already know you so you can split your millions with me! You’ll be rich!! 
2 Likes
Nicky, you are not wrong.
-
Those microboluses ARE your basal. A pump delivers basal in the form of microboluses. The 670G shows you everyone it delivers.
-
Your food boluses cannot take the place if your basal. As you said, they are for your carbs, they can’t be used twice.
-
Your food boluses should not knock out your microboluses, and I am sure they don’t.
-
But something else may: the forward predictions of the algorithm may do that.
*Although I cannot for the life of me understand how the 670G would suspend basal while you are going up to 350 fast.
For my understanding: the blue correction is yours, or auto-generated by the 670? I assume it is generated by the 670G since you are in auto.
The forward predictions are a part of it, yes, but the algorithm also includes your current IOB. Unfortunately, it does mean that if you do a big bolus (and that can mean 1.5 - 2 units!), it’ll knock out your micros. So it’s the super bolus… minus the super part.
Thank you for responding. I’m wrapping up with this. I was surprised to see how strong auto started off this time around. My guess is it did the last time, too; I just wasn’t looking. So it evolves, but in my case, that’s just meant a steady retreat from a steady delivery. This is where I think people start changing their settings in order to buy themselves the extra insulin (or doing the fake carbs or manual injections), but it just seems like such rough management for something that can be streamlined.
That’s the food bolus. I’m seeing it a couple of times a day at this point.
It’s my correction in that I’ve prompted one, but I have no control over what it gives me, and it often is a fraction of what I would get in manual… again, because of the algorithm. I think it gave me around 2 units for that yesterday, but that was generous for an auto mode correction. It’s part of its design, to bring down a high BG slowly. In fact, it’s called the 200 minute excursion. It’s factoring in what it can deliver to you in a bolus as well as what it can deliver through the micros in order to bring you back to a 120. It’s not that it never works, but often it’s just not enough. And all along, it’s hitting the breaks on the micros after your correction… enabling that rise even more.
Just thinking out loud. I appreciate your responding.
2 Likes
It still does not seem to make sense though. As you discussed, insulin for one (microbolus/food bolus) cannot be used twice–unless the 670G team discovered a new insulin principle where effect(bolus+basal )= effect(bolus), which is possible, but not a general belief.
1 Like
You’ve shown two graphs that I think I now understand more or less, and I have a different take on how well the 670G was handling things. I really hope you won’t take my comments as criticism; because they really aren’t intended that way. You had the guts to show your actual data on a public forum, and I have never had the guts to do that (I don’t have a CGM so my data probably wouldn’t be of interest to anyone, but anyway you know what I’m saying).
This first graph is an overnight run:
First I will assume that your CGM is reading accurate BG, and I know you have complained that it does not always do that. But the 670G algorithm does the best it has with the data it gets, so I’ll ignore that problem and assume that your graphs accurately reflect your actual BG.
To me it looks like your carb/insulin ratio is wrong, and set to deliver too much insulin. Your food bolus at 10PM caused your BG to tank, so that the microboluses were disabled from 10pm to 1am. Arguably this was the correct thing for the 670G to do, since as it is your BG bottomed out in the 60’s and your BG would have gone even lower if it had continued to deliver basal insulin via microboluses. Once your BG turned the corner at 1am the 670G restarted your microboluses, and you ended the night with your microboluses active. You were therefore pretty much in the 670G target range all night. When you woke up at 6am it looks like you over-bolused again for a meal, and the 670G turned off microboluses again as your BG was dropping fast.
All told it looks like you averaged a BG of about 110 overnight, with a minimum of 65 and a maximum of 160. Which isn’t bad in my opinion. Especially when you consider that you would have likely gone quite a bit below 65 if not for automode.
Now your latest graph is from the daytime, and I’ll remind you that I don’t think the 670G can possibly do as well during the day, with variable exercise, eating and food bolusing as it can do overnight. Which is why I think it is better to turn it off during the day, possibly forever or at least until you are comfortable with it overnight. But in any case here is your latest graph:
Here it looks like your boluses at 1pm and 2pm were too small, and set you up for the dramatic rise in BG you experienced. You were already at 300 when the 670G turned off your microboluses. Clearly it would have made more sense for the 670G to maintain your microboluses during the 2:30 to 4pm time period, as that contributed to your rise to 350. But notice that the 670G turned microboluses off at 2:30 when your BG seemed to be leveling off, and I assume that with all the active bolus insulin you had, that it was forseeing a BG fall if it kept microbolusing additional insulin after you had turned the corner and at 3pm actually start heading down (though still above 250). Then your correction at 3:45pm, followed by food/insulin at 4 and 4:30 seemed to instead be too much insulin this time, so that your BG by now was dropping dramatically into the low range. Note your microboluses were off this whole time as they should have been (and you would have crashed harder if they had not been turned off), until you turned your BG around at 6pm by eating again.
The only part of this I find hard to explain is your microbolus being off from 3pm to 4pm when clearly your BG was high and rising. The only possible explanation I can think of is that once it turns microboluses off when there is lots of active insulin (at that point your boluses were only 1 and 2 hrs old), it delays turning them back on for some amount of time. Clearly this didn’t work well in this case.
So I think your overnight result was good, though your daytime result clearly left a lot to be desired. But I don’t know how much of that is due to eating and bolusing several times in a small time period, and how much is due to the algorithm. Because of my programming job I need to think like an algoritm, and if you try to do that, I think you can see why the pump responded as it did. Which unfortunately gets me back to my first statement that I’ve made several times - that overnight is when the 670G will do its best, so that is when you should use it, and maybe go without it during the day when you are awake and able to manage your BG better yourself.
Finally I’ll mention one thing that helps me a lot using just a syringe and BG meter. But it looks like it would also help the 670G during the day in automode, is to wait until your BG is more or less normal before starting a meal. You were at 175 and rising fast at 1pm when you first ate, and later in the afternoon ate at 2pm, 4pm and 4:30pm when you were at 225 or above. This is always difficult for a human to handle (at least it is for me), so I’m not surprised it is hard for a pump too.
2 Likes
If this is based on a new discovery of theirs, I can only hope they don’t waste too much time on it. It’s not the most impressive design. It’s not just me either. My Facebook group is at almost 700, and it’s a pervasive theme.
It’s all very interesting though, and I will finish up with changing my ICR and AIT settings. The myth is that in doing so, one can unlock bigger micros. I will say that if the micros I got were bigger or steadier, this would be a different experience. In the first couple of days, they were. Auto would deliver 160% of my regular manual basal in response to a rise. With that kind of delivery, I was holding well even with the minimum delivery. More than 2 weeks in now, I’ve seen those rates collapse and am finding myself at 60 or 70% in the same scenario.
It’s interesting anyway, but I’m about ready to get back to better control.
@T1Allison, @kpanda01, and @jag1, I have read your responses and really appreciate them. I’m too tired to write back right now, but I will get to it this weekend. You’ve given me some interesting ideas!