Things I've learned over the years

Terry4 started a topic awhile back about “brittle” diabetes & YDMV. And while I do feel there are some out there are some out there who do fall into the “brittle” category, like those that end up with transplants because the drugs are easier than the wild swings in blood sugar, I think the term is over used somewhat. I mean back in the early 70’s I was called brittle but who wasn’t, testing urine and taking one shot of Lente.
But many in that chat brought up some great points swinging both ways which made me think about some of my key learnings over the years.

  1. What works for me, might not work for the next person. And what works for me right now, might not work for me next week.
  2. Diabetes is mostly a mental disease. If you can’t wrap your head around the fact that this is forever and that you need to learn how to respect it, live with it and realize that it doesn’t always play fair, you will be in a battle with it forever which could lead one down that dark depression black hole.
  3. Most of those out of target numbers are not crazy and make no sense. Most of my out of target numbers are very explainable. When I finally started really tracking things & using a CGM, all those crazy numbers had reasons. A miscalculation on carb totals, a mistimed bolus, an extended bolus calculated wrong (% & length), not suspending, reducing or timing in basal for exercise. Most have a reason if I’m completely honest with myself.
  4. They are just numbers. They are not my failure to do this well. Diabetes feed into my personal belief that I wasn’t good enough and I was a failure. The disease can feed those beliefs. And I have learned that those numbers, high, low or in target are all learning tools. What works and what doesn’t work.
    5.With all the research & new tools, don’t be afraid to try new things. Just remember where we were.
    -starvation diet to injection of ground up pancreas from cows
    -urine testing to visual blood testing to meters
    -one injection a day, to 2, to upwards of 5 or 6
    -blood testing machines to ones that talk to phones or even your doctor
    -exchange diet to carb counting to low carb diets
    -sliding scales to insulin correction factors
    -old CGM’s, new CGM’s, different types of CGM’s
    -old pumps (backpack sized), DIY, AP
    -new drugs, Smylin, inhaled insulin, type 2 drugs off label
    clinical trials-new drugs, new devices (CGM, pumps, meter), stem cells, vaccinations
    Outside of starvation diet & very first insulin used in 1920’s, I’ve done it all and am so much better off with each new thing I tried. Did they all work well for me, no but I tried and learned and am not afraid to try the next new thing. There are many very dedicated, very passionate people trying to find a cure of a better way of life for all of us. And all these new things coming may help make my life and my families lives easier.
    I guess my point is to never give up, there is something out there that can help and you should try to keep learning new things. You can teach an old dog new tricks!

The advent of the CGM allowed us to rationally draw conclusions as to cause and effect with blood sugar lines. In times past we could easily just throw any untoward BG event into the “who knows?” category. That short-circuited many learning opportunities.


My blood sugar levels vary spontaneously by huge amounts all the time, and the reason is simple: The body’s level of glucose, in the absence of functioning beta cells, will vary not just depending on things the patient can calculate and control, such as weight, diet, insulin, and exercise, but also because of hormonal fluctuations, changing insulin resistance, subclinical infections and the thousands of other things that go on internally in the body without producing recognizable symptoms.

Injecting insulin does not equal the existence of functioning beta cells, so there is not even any scientific reason why we should be able to make injections match the performance of beta cells. Japanese scientists, trying to develop an algorithm for predicting blood sugars for one of the countless closed-loop, artificial pancreas devices being produced, find they had to resort to chaos theory for the necessary models, since the fluctuations were so unpredictable.

The fact that you are praising a 1920s technology, known to be inadequate almost as soon as it was developed (read the Nobel Prize acceptance speeches and hear the apologies for insulin’s failings), nearly a century later shows how utterly and absolutely inadequate scientific progress has been in the field of diabetology. I was told by a Harvard professor of endocrinology when I was first diagnosed with the disease that “a cure for diabetes is long overdue,” and that was in 1966.


yes, but that 1920 technology is what is keeping me and you alive! Isn’t that amazing?! In 1919. you and I would have died. And while it sounds like you have had a very tough and challenging life with diabetes, many of us have not. And while I would never say this is easy or fun or fair- I will say I have and will continue to lead a happy fulfilling life. Diabetes does get in the way sometimes, but that is my life and I choose to deal with it with as much grace and skill as I can and more on to the next thing in my life, fun, boring, necessary, scary. That is life whether you have diabetes or not.
And I can’t wait for a true closed loop pump to hit the market! Every one of these new tools or drugs that have hit the market, many have made my life so much easier. I can’t wait for a pump that I will no longer have to count, calculate, guess my carbs and hope the carb ratio will work right. Just think, a pump that will do all that thinking on it’s own.
And I don’t know anything about the chaos theory for a closed loop pump, I will say if the FDA is allowing a trial to use children in a summer camp and those children were able to be children and didn’t want to give it up after the trial was over, the algorithm must be better than a chaos theory.
And while each of us is different, I will say again one of my learning points over the years has been that all those times when the term “brittle” was used in my case, the reality was, I and my medical team didn’t have the tools to make sense of it. Now with all my tools, I can say,most times there is a reason for the wild numbers. And while the body is very unique and each of us is very different, for me it’s usually insulin, food or exercise and one of those was not handled as well as it should have been.
Who knows, maybe an insulin pump with a CGM could be very helpful. Who knows, getting into a clinical trial might help you find a tool that helps. But wow, to think where we are right now is no better than it was in 1920 is so very, very untrue. Right now is the best time to have diabetes. And all these tools will help me get to the end with my body functioning and ready to get in line for that cure.


Of course we all have our own individual experiences of the disease, but we should not rely on these to explain generally what the disease means in itself. A type 1 diabetic I know is 35 years old, blind, and on dialysis, and he screams his lungs out every time he is put on the dialysis machine and pencil-thick needles have to be inserted into the arterio-venous anastomosis which has been made in his arm to permit the blood cleaning to occur. This has to happen three times a week for the rest of his life, or until he gets a transplant, which in this area will be around a 10-year wait. But since his life expectancy is less than that, he will probably die on dialysis.

This is what diabetes can do, and while various people may have better luck with the disease or positive attitudes towards it, we should never pretend that the disease itself and the risks it creates is truly hideous.


Yes but as bad as it is for this young man, it is not what will happen to most of us. Yes, this is what my parents and I were told my life would be like. But that was 47 years ago. And things are sooo much better now. So much better. The technology you dislike so much has been what has helped me lead my happier and healthier life. My pump gives the flexibility to have different basal rates, different carb ratios at different times, ability to reduce insulin at any time for exercise or medical procedures, and I can turn it off any time vs injections once given are all in. Or the CGM and all the info it gives me. I can see what different type of food can do to my blood sugar in real time. I can see if maybe a mini dose of carb or insulin could help keep me in target vs waiting until the next blood test and than it’s a blind number-is that 150 heading up (maybe more insulin is needed) or heading down (maybe less insulin is needed). The reports really show how often I’m running low, not just my gut feeling or what the meter says. Numbers I can’t argue with. Insulins, how far we’ve come. I remember that one shot of Lente for years. No wonder why things didn’t go smoothly.
And I think you saying we will all live 12 years shorter than everyone else is so very wrong. I was told I’d be dead by 30. I am so far beyond that and still happy and healthy and working as hard or many days harder than the young “healthy” people I work with. I really believe I am healthier t han most the people in our society. You have such a negative outlook here and you have got to remember there are many new people out there to this disease and all this death and doom is so counter productive. I believed a lot of what you talk everyday but I have grown over the years and have realized my mental outlook is more important than my numbers. Please try and lighten up a bit. It is so much better than it was.


^^^ Yes!

Weirdly, the 670G was the first new advance since R/N that felt like a step backwards. Not to divert the conversation, but your discussion of the history of the treatments and your experience of them prodded me to respond because the same question has been very much on my mind in the context of this latest development.

I joined the T1 club in 1983, in the post-Lente but pre-meter R/NPH days, and before now every advance—from R/N to Lantus-Novolog, then to a pump, then to a pump-plus-CGM—has felt like it was giving me more and more control, empowering me to adjust and fine-tune my treatment to fit what I was doing, as opposed to that original experience of having to adjust every aspect of my life to fit my treatment, not the other way around. Which is why the 670 experience was so disorienting. I did make some progress with it, and it’s not like it didn’t live up to some of its promise even for me. I had had a pretty rough ride adjusting to a pump in the first place, so I knew it would take some work. I was in the FB groups and saw how many people were having a great experience, and I really wanted to succeed with it. For those reasons it took me a long time–I gave it better than three months–to realize what it was that bothered me so much, and it was that this was the first time in all my major treatment transitions that I felt like all the controls were being taken back out of my hands. This was the first time since R/N that I felt like I had less control over what was being done to me, and my only tools for getting it to work the way I felt like I needed it to ultimately came down to adjusting my life to it, rather than the other way around. Not quite like the bad old days, but definitely not what I was looking for.

A lot of people do love the thing, and I’m glad for them–and not a little jealous. There’s still an atavistic little voice saying it’s my fault for not making it work. But I think it’s to some degree endemic to the whole automation enterprise: in trying to do everything for you, it’s also going to have to limit what you can do for yourself. I think for a lot of people a black-box that takes more of the fear and responsibility off their shoulders is exactly what they do want. I also think that MT has drawn tighter limits around what you can do than necessary. But the whole system has a kind of “Don’t touch that!” Nanny-knows-best attitude built into it that became more palpable and frustrating the longer I used it.

1 Like

It has been five years off ever since I was dx’d in 1983 (and before that of course).

1 Like

For some recent data on the loss of life expectancy in type 1 diabetics, consider the following study:

"JAMA. 2015 Jan 6;313(1):37-44. doi: 10.1001/jama.2014.16425.
Estimated life expectancy in a Scottish cohort with type 1 diabetes, 2008-2010.
Livingstone SJ1, Levin D1, Looker HC1, Lindsay RS2, Wild SH3, Joss N4, Leese G5, Leslie P6, McCrimmon RJ5, Metcalfe W7, McKnight JA8, Morris AD3, Pearson DW9, Petrie JR2, Philip S9, Sattar NA2, Traynor JP7, Colhoun HM1; Scottish Diabetes Research Network epidemiology group; Scottish Renal Registry.

Type 1 diabetes has historically been associated with a significant reduction in life expectancy. Major advances in treatment of type 1 diabetes have occurred in the past 3 decades. Contemporary estimates of the effect of type 1 diabetes on life expectancy are needed.
To examine current life expectancy in people with and without type 1 diabetes in Scotland. We also examined whether any loss of life expectancy in patients with type 1 diabetes is confined to those who develop kidney disease.
Prospective cohort of all individuals alive in Scotland with type 1 diabetes who were aged 20 years or older from 2008 through 2010 and were in a nationwide register (n=24,691 contributing 67,712 person-years and 1043 deaths).
Differences in life expectancy between those with and those without type 1 diabetes and the percentage of the difference due to various causes.
Life expectancy at an attained age of 20 years was an additional 46.2 years among men with type 1 diabetes and 57.3 years among men without it, an estimated loss in life expectancy with diabetes of 11.1 years (95% CI, 10.1-12.1). Life expectancy from age 20 years was an additional 48.1 years among women with type 1 diabetes and 61.0 years among women without it, an estimated loss with diabetes of 12.9 years (95% CI, 11.7-14.1). Even among those with type 1 diabetes with an estimated glomerular filtration rate of 90 mL/min/1.73 m2 or higher, life expectancy was reduced (49.0 years in men, 53.1 years in women) giving an estimated loss from age 20 years of 8.3 years (95% CI, 6.5-10.1) for men and 7.9 years (95% CI, 5.5-10.3) for women. Overall, the largest percentage of the estimated loss in life expectancy was related to ischemic heart disease (36% in men, 31% in women) but death from diabetic coma or ketoacidosis was associated with the largest percentage of the estimated loss occurring before age 50 years (29.4% in men, 21.7% in women).
Estimated life expectancy for patients with type 1 diabetes in Scotland based on data from 2008 through 2010 indicated an estimated loss of life expectancy at age 20 years of approximately 11 years for men and 13 years for women compared with the general population without type 1 diabetes."

So, exactly as I said, the loss is around 12 years.

The theme of all your comments is that genetics determine outcomes and complications are inevitable. I agree luck plays a role and that genetics will determine how easy or hard of a time we each have managing the disease and outcomes to an extent, no matter how good control is. However, from what I understand, most diabetics struggle to get their A1Cs under 7 and many of us on this forum are outliers. I dont think most of the research out there, sampling a broad pool of diabetics, applies to people, who through hard work, luck or otherwise, maintain tight control. I would like to see some research showing outcomes for people who maintain near-normal to normal A1Cs. I would suspect the rates of serious complications would be very, very low and life expectancies would be similar to the general population.


But I think it’s to some degree endemic to the whole automation enterprise: in trying to do everything for you, it’s also going to have to limit what you can do for yourself.

It doesn’t have to be that way. Think of it like GPS navigation. It tells you where to turn and so on, but if you decide to go a different way, they’ll just say “Recalculating.” and figure out their next step based on what you did. For me, the open-source LOOP software ( cooperates with me like that.

The way the Medtronic 670 implements hybrid closed loop may have been necessary for them to get such a quick FDA approval, but, to go back to the GPS navigation example, it would be like the GPS says “Turn left here,” but you go straight, and then the GPS says “Well, if you’re not going to follow my directions I can’t help you. You’re on your own.” And then the GPS screen turns completely dark and won’t turn back on. Yeah, I’d find that outrageous and completely intolerable.

Hybrid closed loop software doesn’t have to be like that. It can be respectful and helpful. But simply turning off; dropping out of closed loop, is Medtronic’s approach. It is similar to how their CGM responds to a missed calibration. It completely turns itself off and refuses to work. That’s a Medtronic decision, nothing inherent to the process of automation—consider Dexcom CGM, which after a missed calibration indicates that their estimate of BG may be less reliable, but it continues to work just fine.

I also think that MT has drawn tighter limits around what you can do than necessary. But the whole system has a kind of “Don’t touch that!” Nanny-knows-best attitude built into it that became more palpable and frustrating the longer I used it.

I’m hopeful that this is just the current state of affairs. If you want to be “Deliberately Non-Compliant” you could investigate one of the open-source hybrid closed-loop solutions (LOOP, OpenAPS, and AndroidAPS are the three salient ones.) Within a year or two, perhaps we will have more agreeable systems from Tandem, BigFoot Medical, or Medtronic.

1 Like

Another problem is that for those of us who were diagnosed before the development of the home glucometer in the 1980s, we may have had anything from 10 to 30 years of the disease before we had any idea of our glucose levels, other than that provided by the useless ritual of urine sugar testing. The typical A1c levels being recorded in patients from this era when the transition to home glucometers began were in the 10 to 12 range.

The problem with this now is that because of hyperglycemic memory, the effect of this initial period of hyperglycemia is fixed in the form of damage to the genes so that the complications now unfold on their own, regardless of current sugar levels. In my case, for example, I have had A1c values in the four-range for a decade now, but I have required four laser photocoagulation treatments because of worsening retinopathy during that time, while in my entire previous life as a diabetic I required only two. This can only be the effect of hyperglycemic memory from what was going on in the 1960s and 1970s, and without a time machine, there is no fixing that.

But in addition to that, there is also evidence that some of the complications may proceed on their own from genetic priming, while others may come from the damage done by the continuing autoimmunity of diabetes, which after attacking the beta cells goes on to damage other parts of the body. If you look at the data for diabetic nephropathy, for example, there is a large spike in new onset cases at 17.5 years after onset of diabetes, but assuming that all these people haven’t had exactly the same level of blood sugar control, why is there this coincidence in time of onset? This looks more like a genetic process than a side effect of hyperglycemia.

I think GPS is a good analogy, though the stakes aren’t quite as high. I think there are a lot of factors, including many that have more to do with regulatory and marketing than technical concerns, influencing why the 670G is as locked down as it is. But I’ve also worked in medical technology and UI design, and part of that process is that you always have to circumscribe user choices to a greater or lesser degree or the development process becomes unmanageable. My own expectation with the 670, going in, was that you were going to have to accept some limitations as part of the deal with an automated system but the other part of the deal was that it was going to do a better job than I could. Turned out that it couldn’t, and I totally agree that that was because they locked it down too hard. Just being able to adjust the target within some reasonable range would have made a huge difference. Especially since that target, 120mg/dl, which sounds all very precise and concrete, is in this context more notional than absolute, much the same as the “Active Insulin Time” parameter, which virtually everyone has to change from whatever setting they used on their traditional pump.

But also consider the soon-to-be-released G6, which doesn’t need calibration but also has a hard-wired sensor life limit of ten days. My view is that the reason for this is that they are looking toward integrating with looping systems, and the FDA requirements are much steeper because the sensor is actually controlling insulin delivery. I’ve gotten far more than ten days out of my G5, but they do start to get erratic eventually, some sooner than others, and that ten day limit is a conservative estimate of the line beyond which they can’t certify it’s safe for it to be telling your pump how much insulin to deliver. MedT has made a different design call with the Guardians: “calibrate or get kicked out,” but they’re both driven by the same considerations.

I did! I still have my pager-style Minimed, which I’ve actually gone back to because as a manual pump it has far fewer annoying habits than the 670. But unfortunately its firmware is a version or so too recent to work with OpenAPS. Is LOOP different? One of those terms it’s hard to search for on Google. Tried stuff like “insulin pump loop systems” and whatnot but all I get is lots of general references and links to the major pump manufacturers.

Sally7, what an excellent article, “What I’ve Learned Over All These Years!” So true on so many accounts. I too am (an old dog) a Type 1 diagnosed 60 years ago. I too went through so many of the same things you have #5, and am so very grateful to even be alive today. I remember my mom crying when the doctor told her I had juvenile diabetes. She thought at the time if I had a life expediency of 40 years without any complications it would be a miracle. Well here I am so I guess we are miracles! :wink:
So many changes and more yet to come. Thanks for the upbeat message!


Is LOOP different?

No. The issue is that the older Medtronic pumps could use a clicker to remotely command the pump to dose insulin without pulling out the pump to press buttons. This radio interface is what the looping software uses to control the pump. But then there was a paper given at one of the hacker conferences like BlackHat or DefCon that explained how it was possible to build a cheap transmitter that could command someone’s insulin pump to give a fatal overdose from 30 feet away. So Medtronic quickly updated their firmware to completely remove the ability to give dosing commands by radio. The LOOP documents specifically describe the pumps and firmware versions that can accept dosing commands by radio, at

The canonical LOOP documentation is at and it gives plain-talk explanations of everything.

But also consider the soon-to-be-released G6, which doesn’t need calibration but also has a hard-wired sensor life limit of ten days. My view is that the reason for this is that they are looking toward integrating with looping systems, and the FDA requirements are much steeper because the sensor is actually controlling insulin delivery.

I agree with your explanation of the 10-day limit, but I think there’s more to it. They approved the G6 in a new category of device called “integrated CGM.” The FDA will not require pre-approval clinical trials of the combination of a specific CGM with a specific pump to get an approval for that combination to do (hybrid) closed loop. If a pump and its algorithm gets approval for closed looping, it can use any of the approved “integrated CGM” devices interchangeably. So the requirements for approval as an iCGM are quite a bit stricter than before. It’s not just a “within 15% except for 15% of the measurements” kind of requirement, it has several added requirements. In particular, there’s a pair that basically say the device is absolutely forbidden to give an error in Clarke Error Grid region E. (I.e., it is absolutely 100% guaranteed that the CGM never says you are over 180 when you’re actually under 70, and the reverse.) In the past, the standards never required perfection like that, they always allowed a small number of really bad measurements as long as most of the measurements were reasonably good.

MedT has made a different design call with the Guardians: “calibrate or get kicked out,” but they’re both driven by the same considerations.

But the draconian user-hostile approach that Medtronic has taken doesn’t make me safer. The Dex’s estimated CGM readings when a calibration is overdue still give me the information I need to keep my BG in a safe range. Perhaps the Medtronic approach is shaped by their past experience with sensors, where their CGM would sometimes be dangerously off, with no warning. The ancient Medtronic Sof-sensors have indicated that I need lots of insulin when I actually needed glucose, and have reassured me that everything is fine when I actually was around 40. Thankfully the new Medtronic sensors are much better.

No argument here. And “user-hostile” is a good way to put it*. Re LOOP, same roadblock I hit on OpenAPS: “Firmware 2.4A or lower is required for using Closed Loop and Bolus features.” Mine’s a version 3-something. But I appreciate the additional info.

ETA: *I suspect long-term that this is leaving a huge opening for competitors, though having been orphaned by one such competitor, the Asante Snap, a few years ago I’m impressed at how unfriendly the market is to new entrants. The Snap wasn’t perfect but it was much more of a user-centered design, which is STILL not a trend in UI that MdT seems to know anything (or care) about. Hopefully Bigfoot will be the evolution Snap should have been allowed to undergo.

Bigfoot certainly has potential. As does Tandem with their hybrid AP software (look for it in '19). I’m also intrigued by the Dana-RS pump by SOOIL (Korea) which was designed for compatibility with open source AP systems. There is some hope that it can be approved in the US in '18 or '19.

Re LOOP, same roadblock I hit on OpenAPS: “Firmware 2.4A or lower is required for using Closed Loop and Bolus features.

I had a Revel 723 with v2.4A that failed and was replaced under warranty with a Revel that has v2.5 so I was also out of luck. I watched craigslist and was fortunate to find a good 722 within driving distance for $500. That may be one possible approach for ambitious folks inclined to join the Deliberately Non-Compliant crowd. I did this after discussing it with my endo and CDE. The endo knew that I was generally competent in managing my diabetes, and said he would be interested to see how it works out. The result for me was more-or-less an identical BG profile (95% between 70 and 180, A1C 5.9) with much much less manual effort. Sleeping uninterrupted through the night is a good thing and I wouldn’t voluntarily go back to manual sugar surfing.

1 Like

Thanks! And while I would so prefer to not be part of this diabetes club, I do feel that we are in such a good time for outcomes of this disease. Many can manage their blood sugars to near normal levels and still lead a normal, happy, healthy life. It is the best time to have diabetes vs 20, 30, 40, 50 years ago!

1 Like


thanks for perfect sharing of wisdom.
! We are so connected , those of us with T1 for decades! I could not have added or deleted any of your words of wisdom.
For me, it took 40 years to come to the place of peace and acceptance. It is a great freedom to finally look at the Bgs as teachings/ not diminished sense of self… Bring on the next 50 years!

Gratitude ~!


Very interesting information! The question that I have is if a radio interface is necessary for activating a pump with LOOP, why is the Animas Ping with a radio remote not appropriate?

I think I just answered my own question… because the Ping remote does not control basal release nor bolus without a person’s input. Am I correct?