Thanks Robyn and Helmut. I just lowered my CF from 1:50 to 1:40 and will see what happens.
My goal is to find a way to keep below 100mg/dl with my t-slim Control IQ, so if you have more insights please share them.
Type 1 (LADA) less than 30g carbs/day.
Safety first has ever been the governing philosophy of insulin therapy. I remember back in the R/NPH days I’d get yelled at if my A1C dipped below 6.5 b/c it meant too many lows. “First do no harm,” I get it. But the problem those of us with tight control goals have with these commercial looping systems is that this attitude gets hardwired into them, and they have the ability to lock you out of parameters that you could hurt yourself with if you don’t know what you’re doing, which also locks you out even if you do. Keeping your fumbling little paws off the scary stuff is a feature, not a bug, from the designers’ p.o.v. But this hyper-protectiveness then results in people like us struggling to find ways to circumvent the blackbox with crazy stuff like deliberately mis-calibrating our CGMs, entering erroneous data, running “sleep mode” 24/7 etc etc. Which is probably more dangerous than if they just gave us better access to the controls, and kinda defeats the whole IDEA of a blackbox-that-does-it-all-for-you.
This is why I’m so stuck about doing a pump upgrade, which I’ve been eligible for since back in March. I know you can run these things in manual, but I’ve already got a stodgy-but-reliable manual pump (Paradigm). And trying to use the automated features just seems like an endless game of cat-and-mouse trying to get around the built-in obstacles that are there precisely because the manufacturers don’t want you to and have deliberately designed these things around having you run at higher averages because FDNH. If you’re continually fighting with the thing to circumvent what the automation is designed to do, then the automation is the problem, not the solution, isn’t it?
Seems like for those of us at this end of the T1-control spectrum it’s as much a hindrance as a help. Open source looping seems to be the real solution if truly customized automation is what you want, but there’s a lot more start-up effort, including locating a pump that has the open communications you need. So infuriating that to achieve genuinely customizable automation, you need an obsolete pump!
@Derek_Nunez, even with DIY Loop and target 80 mg/dL I run CF 1:40 during the day and 1:20 during the night. My true CF is 1:50. It is all trial and error. I like @Laddie’s regimen of manual control during the day. I am much better than Loop anyway when I pay attention. I pay attention most of the time. When BG target cannot be lower than 110 mg/dL, I expect to use closed-loop only during the night and the rare occasions when I don’t want to pay attention.
It seems like all the commercial efforts to design an automated insulin dosing system have been unable to give the user the freedom s/he needs to choose normal glucose settings. Their hand is held back by the fear that such freedom will hurt someone.
There’s a relatively small percentage of us glucose managers who know how to shoot for more normal glucose target without increasing personal danger. I propose that the designers allow users to earn more freedom by demonstrating their ability to hold glucose variability at levels that truly mitigate exposure to severe hypos. Once the user can achieve a standard deviation of say, < 25 mg/dL, then more aggressive algorithm goals may be selected.
I’m sure things are more complicated than I perceive but I do think the designers could embed earned freedoms within their code. Up until now, all the commercial designers have locked out the higher achievers since they are only a small slice of users.
I am really curious how Tidepool will handle these issues with Tidepool Loop, a not-yet-approved FDA app that purports to keep almost all of the freedoms of DIY Loop. This is as much a test of the FDA as it is of Tidepool.
I feel your frustration, @DrBB. I can’t help but think that if you were able to use DIY Loop with an older MiniMed pump, you would be successful. I’m hoping that the Tidepool project will help people like you.
If you were trying to DECREASE your insulin sensitivity, you went the wrong direction. The value of the setting and the degree of sensitivity are inverse. Making the number smaller makes you MORE sensitive to insulin. And that was a really big adjustment, too. I’m afraid you might be in for a wild roller coaster ride of highs and lows if your previous settings were close to Tandem’s goal.
Granted, things get all confusing when you talk about correction factor, so did I misunderstand the question?
@Laddie – This resonated with me.
I have used Control IQ for over a year and a half. I have accepted an average BG between 112 and 120. What I can’t deal with is Control IQ suspending my insulin when my BG is flatlined at 100. I am amazed watching my BG tracings throughout the day without Control IQ. I can flatline for hours at a time with minor up-and-down waves. But add Control IQ to the mix and I have insulin suspensions as my BG drops below 110. The pump screams that I will drop below 70 but without Control IQ it usually stabilizes in the 80’s or 90’s. Then future highs from those suspensions are another unknown as I navigate my next meal or my next couple hours of D-existence. Ups and downs and more ups and downs.
It just doesn’t seem right that an automated insulin dosing system particularly penalizes active glucose managers who should enjoy the best chance of success. It’s like some perverse handicap scoring system is being employed.
Your situation reminds me of the troubles that @DrBB has witnessed. It’s well past time that the algorithm designers recognize high achievers and allow them to use the full power of the algorithm.
The fact remains: insulin is as dangerous as it is life-saving. Any good tool can also be used as a weapon. Any system that completely protects us from poor choices also loses is usefulness. It’s time for the algorithm designers and regulatory gatekeepers to get creative and to take some chances!
@Robyn_H, my DIY Loop displays the following explanation:
“Insulin sensitivity describes how your blood glucose should respond to a 1 Unit dose of insulin. Smaller values mean more insulin will be given when above target. Values that are too small can cause dangerously low blood glucose.”
My values are 20 mg/dL between 12 am and 4 am and 40 mg/dL otherwise. I can tell from my experiments that above explanation is correct.
@Terry4, I was afraid that I would be denied coverage for CGM and pump because my A1C was 5.8 with MDI. Once, when I had to get a new prescription for my Dexcom I overheard another doctor saying: Why does he need a CGM?
@Helmut We’re saying the same thing, but from different sides of the equation. A CF that is too low CAN cause most definitely cause dangerous lows, because it gives too much insulin when you’re advice target, driving you straight down past the target. Just like it can cause massive highs when rebounding from a low, because it withholds too much insulin… Resulting in a vicious rollercoaster of continuing over-corrections at both extremes. Setting the CF too low scares the 
out of me! (Thus my wig out when @Derek_Nunez said they dropped it from 1:50 to 1:40)
But too high of a CF can be equally dangerous. Especially in this example where we’re talking about hacking the automation to maintain a lower target than Control-IQ allows… since that requires programming an excessive basal too, otherwise you’ll just bounce all around the default 110 target. In this example, we need to hinder the automation from raising BG up to the forced110 target, by raising the CF and slowing the correction down. If someone increases their basal rate TOO aggressively, though, and didn’t strike the right balance and set the CF too high to balance the basal, then they’ve crippled the ability of the system to prevent hypos. It won’t be able to withhold enough insulin, and the heavy basal will keep driving them down. Hence why I think this strategy only works at night or when you’re fasting. I don’t think it’s responsive enough to protect us from inevitable bolus mishaps and usual day-to-day variations.
Tandem is desperately trying, since it’s one of the biggest criticisms holding people back. The FDA has halted their proposed software update that was going to allow more customization. They asked for more clinical trials… Which puts things behind another year or so. In fact, they’ve asked for more clinical trials on nearly everything Tandem has brought to them recently. I’m beginning to think they’re using it as a stall tactic while they’re overwhelmed in the wake of Covid. There’s still an update coming soon, but it’s going to be much more disappointing than most of us were hoping for. We have to be patient for the bigger changes. I’m not so good with the patience thing, though.
As a medical tech user, I’ve often observed both the tech company as well as the FDA remaining quiet about the facts of regulatory review until the announcement of their reality. The regulatory dance is between the company and FDA. Investors and users are deliberately kept ignorant until a ruling has been announced.
As a result, I’ve witnessed tech companies blaming the regulatory process to save face with investors. In these cases the FDA stays quiet and does not confirm or deny the device manufacturer’s position.
As users, we don’t get a clear view of what’s happening. In this case, where the FDA may request additional clinical trials, it might be true that the med-tech company submitted an incomplete application when it should have known better. The company can find it convenient to lay blame with the FDA in order to protect them from further stock price erosion.
In the last several years, the FDA has tried to be as expeditious as possible. The Dexcom G6, for example, was approved so quickly that it surprised the company and caught them flat-footed in their ability to ramp up production and meet expected demand.
It’s entirely possible, but I don’t think so. I’ve always found Tandem to be pretty transparent, at least when it comes to talking to the investors (which I am not, I just like the insider info), if not in all arenas. For example, they’ve been saying for a while that they hoped all the existing compiled data they had of Type 2s using the pump off-label would be enough to get an easy, speedy Type 2 indication. But they recently admitted that it’s not good enough, that trials are indeed needed. It was never a blame game, though. It was a “hey, we swung and missed… So we’ll get it via this more detailed route.”
It’s not even really possible to egregiously shorthand an FDA submission. It’s not like handing in a final paper in school, where you get one chance to get it right. It’s a discussion. They tell you what info they need, many times they tell you exactly how to do it, and what the next step is throughout the procedure, before it ever goes off to committee for a final decision. And sometimes annoyingly, when you think the work is all done and you’re in the midst of preparing to launch the new device in anticipation of the new approval… They come back and ask for more work instead. It’s an unfortunate part of the process. Doesn’t always happen that way, but sometimes it does. Just seems like it’s happening more often than expected with all the approvals Tandem is seeking.
The designers are held back on this by needing to gain FDA approval. Do not forget this. It is the FDA who mandates certain “safety” standards and not the pump manufacturers.
NOTE: Wrote this before reading the subsequent thread exchanges.
I came to the conclusion that APS is for people with pretty high A1c’s…compared to most of us. I don’t think that we are the target patient population for this tech to operate successfully. Some people do OK, but some people are fighting with the automation all the time.
Would you all agree with that?
Like, if you wanna tune someone from an A1c = 9 to an a1c = 7.5, maybe that’s in the realm of very possible. But, for ‘high performers’ in the A1c = 5.something to 6.smething, the fine tuning gets awful speculative and ‘wonky’ with APS systems. I’m not saying its impossible, just that its speculative and people start balancing a house of cards. Then, when unknown updates get pushed, the house of cards can tip over and we may see some dangerous circumstances arise.
Mis-calibrating the sensor might be one example of “house of cards” type stuff. (Not that the sensor is even accurate to 10 points most of the time…maybe 20 points. We could prob be a little flexible on 10 pnts here or 10 pnts there.) When your mis-calibrating the sensor, you are starting to build a house of cards. But there other examples.
Big picture, its not so important that someone run a a1c of 5.6 versus 6.5. That’s a difficult point to raise to this community and we should always try to improve, but we’re running critical systems and we need to make sure no one gets put into dangerous situations. I don’t know where that line is. That’s something we will need to figure out as a community.
Really interesting perspectives that you all raise here.
@Helmut and @Robyn_H Changed the Correction Factor (CF) from 1:50 to 1:40 on 8/21/21. Experiencing very good results between 80-100 now. Before doing this the range was roughly between 90-120.
After changing the CF, I’m thinking it would be good to lower my overnight basal rate from 0.4U/hr to 0.2U/hr because my Tandem report showed 0.0U/hr for most of the night.
Thought concerns much appreciated.
Yes, @mohe0001, and I can only speak to the Tandem IQ. The system is the most help to those whose A1c is elevated. CIQ brings their A1c down into the 6 percentiles. I was well under 6.5% for years but with CIQ I actually went up into the high 6s. But… my nighttime lows disappeared entirely. And that was good, very good since I live alone. However, my post-prandial peaks became higher and persist longer. Not so good. I may not be in a very good place constantly battling the highs, but others are very happy with their lower A1cs in the 6s. Trouble is that I don’t think there are many with high A1cs who opt for a pump. Some do, most don’t. And with the narrowed pump options today in the U.S., there is not much out there to choose from, is there? I came off 13 years on an Animas Ping and do realize that I could again have a “dumb” pump by simply turning off CIQ. But would I give up my nighttime flat line? Not! I also spend more time messing with my pump than I used to and I really don’t want to but feel I have to. And there it is…
@Derek_Nunez, when I reduced my CF number it worked for several days and then the pump algorithm seemed to re-assess what it was doing and my numbers went back up at night. Still a flat line with gentle curves up and down, but a higher average.
My a1c was running in the low 7s before automation.
I’m on a tandem for almost a year. I’m running 5.5 now.
The tweaks that I do are only adjustments to insulin sensitivity and carb ratio and my weight.
It brings my target down on sleep mode to right about 100 mg/dL.
I don’t really need it to be lower, but I would like the ability to set my own targets.
I’m hoping we can get new insulin’s going that are more responsive so that we can keep it as tight as we want and not go low.
I’m happiest with my new tech as I’ve been on any other.
It’s good that there are lots of choices. Because not one solution works for everyone.
That’s a good point.
Maybe you could turn on the automation during the night time and turn it off during the day? Best of both worlds?
I could do that. I could also go back to MDI and have full control. The idea was that the pump would require less attention and management. I find the pump requires MORE attention. I love the low-less nights. Not so happy with the beeping days going up and down on the rollercoaster. I know that is probably caused by me overcorrecting. But I am neither used to nor willing to go up into the 200s and staying there for hours until the pump brings it down by manipulating the basal.