Took me awhile to realize this. So now if I eat carbs for low, I bolus for them. Then I rise quicker, and this reduces basal suspend, and stay in range better.
1 Like
CIQ acts progressively based on your BG level as you are estimated to pass beyond its desired ranges.
You donât tell CIQ what you are eating, just the carb grams. Carbs combined with fat and protein in a meal cannot be relied on to effect BG predictably in a specific interval. Carb absorption rate varies with meal composition fullness and health. . CIQ doesnât know whether you ate pizza, oatmeal or a high fiber meatloaf, whether you have gastroparesis, diarrhea or constipation, or are dehydrated or how fast you eat. All those have an impact on how fast what you are about to eat reaches your bloodstream.
If you are low and/or expect your BG to drop to a suspension level, and you want to micromanage your therapy, then suspend Control IQ until your BG is rising.
But with higher fat, protein meals, you can use an extended bolus, to cover. I increase carb count to cover. Some refer to this as T-A-G, total available glucose). There are older posts about T-A-G if you search.
(Extended bolus also on Medtronic pumps).
Carbs information is only used by the bolus calculator, youâre not telling the CIQ algorithm anything about food, not even the count. The CIQ algorithm doesnât care whether you ate 1g of carb or 100g carb, it will react exactly the same way, based on BG thresholds and programmed ISF and basal rates. It gets away with ignoring carbs because it doesnât automatically correct within 1 hour of a bolus.
I mentioned suspending CIQ specifically because of my experience with BIQ . If you bolused far enough in advance of the food rise rise that BG would drop fast enough to trigger a cut off, Basal IQ would shut off all insulin delivery and forget the balance of the extended bolus.
Control lQ doesnât forget the balance of an extended bolus but can suspend delivery long enough to mismatch the delivery timing attempted with a 2 hour extended bolus. If this happens while youâre mid-range youâll get a larger rise, enough to rapidly rise beyond 160 and 180, still accelerating
The problem is the sensor and the 5-minute sampling interval and responding to speed instead of acceleration using only blood glucose⌠Within 10 minutes on an empty stomach fast carbs can make the BG rise accelerate much faster than reducing basal can compensate for as you actually pass 160 and overshoot the top of the range. You get an automatic correction that isnât needed to handle the total carbs and then you yoyo.
This is similar to what happens to people without diabetes. Somebody obsessively monitoring their blood glucose who can be upset by short-term effects is the difference.
A desire for instant gratification, obsession over details and micromanagement are incompatible with employing an assistant like CIQ or any limited insulin delivery system. No matter how much we want them to be they are not artificial pancreas.
Realistically, as long as you eat rationally, and bolus for the carbs, no matter how many grams of carbohydrates you eat and how high your blood glucose goes, once you stop eating eventually the level will drop.
Low blood glucose is immediately dangerous, high blood glucose is a threat
Control IQ doesnât control blood glucose, it provides moderation of lowering and rising blood glucose .
Itâs like having adaptive speed control on a car without brakes going down a long road with hills. It can step on the gas, lift itâs foot, and shut off the gas. The car will keep going until it collides or runs out of gas and stops because of friction losses.
The pump user is the one in control, the one who has to choose the road, watch whatâs going on and step in when the assistant isnât up to the job.
I realize this thread is from last year. Iâm wondering if this still applies to the latest Dexcom G7. Iâm comparing auto modes of Dexcom G7 and Medtronic 780 (Latest model orders supposedly available mid May). My body seems to be so sensitive, yet stubborn, Itâs difficult to predict which one would work best for me.
We donât have a lot of info available yet to the G7 (and Libre 2&3) integrations, since theyâre still pending FDA approval in the States. In the earnings call yesterday, Tandemâs CEO was talking about new Control-IQ algorithms in the works, two separate updates, independently of the new sensor integrations. He referred to them as Control-IQ 1.5, which with mage the system more agreeable to with low or high insulin needs outside the âaverageâ user model, and Control-IQ 2.0, which is the big update weâre all waiting on that will expand customization options and improve performance. I think itâs fair to say we wonât see any difference in how Control-IQ works when the G7 is added, though, and that algorithm changes will be implemented separately. Well, except for less system downtime, since you can skip the warmup window entirely by overlapping sensor sessions. (Referring to how one can start the 30 minute warmup of a new G7 sensor while the old sensor is in itâs grace period.)
1 Like
Thanks for the info.
1 Like
What the what what?? I didnât know you could overlap the sensors. Holy cow!
Which kinda defeats, or at least inhibits, the effect of pre-bolusing doesnât it? This is one of the things that irks me, when I look at the Connect app and I see that itâs lowering my basal because Iâve done a meal bolus. I wish it wouldnât do that! I guess the idea is that it also predicts the post-prandial rise on the other side but getting ahead of that when itâs already trending up is a struggle. I know it does its own corrections, but IIRC theyâre limited to I wanna say 80% of what your numbers actually dictate. So Iâm ending up doing a lot more proactive corrections myself if I see it rising steeply, or adding a bit if I see itâs done one on its own.
Maybe this is why some people go to 24/7 Sleep Mode? No automated corrections but you get a lower target, and if youâre having to do a lot of corrections yourself anywayâŚ
1 Like
You just nailed my biggest complaint about algorithm-based insulin pumping. I am an avid Sugar-surfer. Iâve learned to judge how a complex mix of fats/carbs/proteins will impact my blood glucose level. Even the time of day is considered. So Iâll bolus up to an hour before I start to eat and watch as my blood glucose trends down and, depending on the food, I may not start eating until it hits 70. The only way Iâve found to do this on either the Tandem X2 pump or the OP5 is by running it in manual mode with almost all alarms turn off. I get it - the pump isnât designed for my use. Itâs designed to make life a whole lot easier and safer for most diabetics, especially kids.
2 Likes
I actually find the automated corrections somewhat dangerous especially when it does it after eating fast acting carbs for a low or slightly low BS. It not smart enough to know that your BS will rise quickly and then level off and go back down. That is one of the reasons i only use sleep mode. I also find automated corrections in exercise mode dangerous. The system isnât smart enough and doesnât have enough inputs to be able to safely do the corrections.
1 Like
I have two pumps, the tandem tslim:x2 and the MiniMed 780G. Comparing the algorithms is extremely interesting, however, studies show you get essentially the same results from both algorithms even though Smartguard has a 100 mg/dl target.
The Smartguard is quite interesting in that in learns your insulin needs, and doesnât consider or need your basal profile at all. It only needs carb ratio and active insulin time. Active insulin time can be adjusted down to 2 hours for a quite aggressive experience. The algorithm starts off quite horribly in that it is being so safe that you will spend a lot of time floating at ~130-140 mg/dl. Overtime it begins to learn and then starts to work as advertised, but is still extremely safe in working you down to the target.
My last A1C using control IQ for 3 months solid with extremely well built basal profiles showed 5.6. Iâm willing to give Medtronic another shot with the 780G (I hated the 770G, the CGM, and the target), and see if I can get lower. However, I have doubts that will be achieved as the Smartguard algorithm is too safe. They care more about time in range than about achieving the actual target. This makes the target more of an advertising gimmick when compared to a 110 target, rather than an actual improvement versus other algorithms.
I would think Control IQ 2.0 will arrive with adjustable targets and removing basal profiles from the algorithm.
Interestingly enough, I float between 5.6 and 5.9 A1C using MDI. I use pumps for convienence rather than improved A1C. The Medtronic extended infusion sets need to be a strong focus of Tandemâs.
Overall I find the algorithms are essentially equivalent, however, Smartguard is probably easier for those less experienced with basal profiles or just down right donât care. Practically, the tslim mobile bolusing feature is great and you get more basal information with their app. You also move straight into the algorithm without a 48 hour warm up, or if only off for two weeks and 5 hour warmup, 2 two hour sensor warmup, and a 45 minute transmitter charge.