11 ms·

Six months as a full time pancreas

121 points by grahar64 4y ago

grahar64 4y ago
This is a follow up post from “The Unreasonable Math of Type 1 Diabetes” (https://maori.geek.nz/the-unreasonable-math-of-type-1-diabetes-8c96bdf5b7fb https://maori.geek.nz/the-unreasonable-math-of-type-1-diabet...). It is more about the long term complications of T1D and the way we have adapted our strategies to minimize them.
- vertis 4y ago
  It's an absolutely amazing read. Not really a useful comment, because I lack anything useful to say. ¯\_(ツ)_/¯
  - grahar64 4y ago
    It’s better than the pro eugenics discussion above this comment. Cheers :)
- agileAlligator 4y ago
  - iLoveOncall 4y ago
    That's kinda the feeling I got from reading the post. Eugenics is a part of science deemed immoral and forever shunned because associated with the nazis, but maybe it would be good if we had made enough progress to avoid babies that come near death 4 times a day every day and require their two parents to be dedicated to their care 24/7.
    - consp 4y ago
      I've had this argument before: If you argue eugenics then you should really look into your family history and be 100% sure no-one has any "defects" anywhere. Most people who argue it have either needed some form of intervention somewhere in their lives to fix a "defect" and in other cases one of their predecessors and would they be unwanted in the eugenics-fanatics eyes. (Reflection usually works quite well with when someone is reasoning for eugenics)
      - iLoveOncall 4y ago
        Well that's a pretty week argument. Yeah sure if we killed every baby to be that had a defect I probably wouldn't exist. But I wouldn't care, because I wouldn't exist. Note I'm not saying to kill those unborn children, I'm saying that if this medical field had not been completely put to a stop following WW2, we wouldn't have to kill unborn children but we could do in utero interventions to prevent those conditions from developing. "Kill the babies" is the current state of eugenics because it has been at a full stop for more than half of a century. That's my point.
      - agileAlligator 4y ago
        Eugenics doesn't necessarily mean kill the baby. It meant that in the past only because science wasn't advanced enough. Techniques to modify the human genome exist; we could potentially "fix" the bad genes without killing the child.
    - borski 4y ago
      How many diabetic scientists and engineers have there been that have changed the world for the better?
      - iLoveOncall 4y ago
        Probably not as many as diabetic terrible people? I'm not saying to kill those unborn children, I'm saying that if this medical field had not been completely put to a stop following WW2, we wouldn't have to kill unborn children but we could do in utero interventions to prevent those conditions from developing.
        valarauko 4y ago
        How, though? It's not like we've had the ability to diagnose the majority of these diseases in utero until fairly recently, nor have we the means to intervene, even now. CRISPR as a tool has existed for just over a decade, and we're still very far from the ability to reliably modify fetuses, even in lab animals, and it's been only possible for, what, 5 years now? Lab research did not stop after WW2, and even then we've barely gotten to this point. The ability to sequence the whole genome and find mutations for the major disorders is still a far off goal. The majority of disorders are now known to be polygenic, affected by hundreds of different gene variants, with each population group seemly affected by different groups of genes. Nor have we settled on a definitive group of gene variants that can reliably be identified to be deterministic of any widely distributed disorder, outside a small handful like CF and DMD.
      - agileAlligator 4y ago
        I am not saying we should let diabetic babies die at birth. I am simply reflecting upon the fact that their genes would have been removed from circulation in the past; this is no longer the case. If we disallow nature from taking its course, it will have consequences on "genetic quality". Following that train of thought and assuming we don't want to kill genetically malformed babies, an alternate route must be chosen. Better in-utero screening, CRISPR to fix the deformed genes, etc.
        4y ago
  - datapolitical 4y ago
    Except type one diabetes is only partially affected by genetics. A woman with type one diabetes has less than a 4% chance of passing it to her child. The biggest risk right now to our genetic success is overwhelmingly the fact that the people most likely to reproduce are the ones with the least amount of education.
    - mschuster91 4y ago
      > The biggest risk right now to our genetic success is overwhelmingly the fact that the people most likely to reproduce are the ones with the least amount of education. Idiocracy wasn't a dystopia, it is a documentary... Sarcasm aside: A low amount of education is not really related to genetics (some forms of inheritable mental health issues aside). There are, however, extremely strong correlations with poverty: - poor families tend to live in areas with underfunded and thus lower-quality public schools, whereas rich people have better schools and can afford after-school private education - poor families tend to have issues with food security, which was brutally shown by covid as schools had to stay open simply so that the children could have at least one warm, somewhat healthy meal a day - poor families often tend to send their children off to work as soon as they can and are legally allowed to, simply to help make ends meet. That leaves them less time for homework and learning (leading to lower grades), and not being able to pursue higher education at all because having a job to survive is more important, which in turn locks them out of higher-paying careers and sets them up for being poor for life
      - mlyle 4y ago
        > A low amount of education is not really related to genetics Intelligence in adulthood is overwhelmingly heritable. Environmental and upbringing effects are small. Twin studies, unrelated sibling studies, etc, demonstrate very large effects. (Oddly enough, intelligence in childhood does seem to be very influenced by environment, but by adulthood your genetic destiny takes hold). Intelligence is also the biggest single predictor of your child's educational attainment-- even bigger than SES. Of course, it's not really clear what one can do with this information that is both ethical and useful. A confound: severe neglect and abuse is known to permanently reduce intelligence, and this is much less represented in adoption studies because of parental screening and social worker follow-up.
  - whostolemyhat 4y ago
    What a stupid thing to say
  - grahar64 4y ago
    Pro Eugenics comment to a post about keeping my kid alive. Wow!
    - sunmag 4y ago
      These comments and thoughts will come at you. You will quickly find reasons to dismiss them as noise. - father of T1D son.
    - agileAlligator 4y ago
  - piokoch 4y ago
    I think we no longer aim at purely physical strength. In the old days person who was not able to work on the field, in the garden, have 10 kids, hunt, fight was useless. This is no longer the case, maybe those kids when they grow up will be scientists or artists, maybe they will make our lives much better.
- robocat 4y ago
  Perhaps add “article author” or something similar, as it is not obvious on some of your comments. Thank you for your write up about such a personal topic.
- shagie 4y ago
  (hmm, visited link)... ahh. For other people passing through the comments, that post was on HN several months ago https://news.ycombinator.com/item?id=30376777 https://news.ycombinator.com/item?id=30376777 with 673 points and 304 comments. There is a lot of interesting material in there to read too.
bigmattystyles 4y ago
Amazing read, really well written - prior to reading, it seemed odd to me that a simple PID insulin or sugar pump armed with a glucose sensor could not mimic a pancreas. I didn't realize there were so many variables. It still seems from my engineering mindset that a PID device should be able to do this, but I'm starting to understand why it's not so simple now.
- bsder 4y ago
  This could be engineered around--mostly the limitations are liability. If you, the patient, take a glucose sensor, an insulin pump and tie them together with an app, you don't get to sue somebody if something fails and you die. If a company takes all three of those, packages them together, and sells them, then that company is now liable when one of the pieces fails and you wind up dead. The sad part is that AI/ML would be incredibly useful for all this. People could take pictures of their meals, upload their sleep and exercise data, and weigh themselves, feed all that into a cloud algorithm that could determine how you are likely to react and dial your insulin plan into your pump while feeding the glucose readings back to the cloud. Sadly, I suspect none of this is going to happen in the US due to the liability. Maybe China will pull this off.
  - bayesian_horse 4y ago
    I don't want to call your comment ignorant, but you are lacking a lot of the information necessary to see why it's not that easy. For starters: Insulin is quite long acting. Maybe 4 hours for some injected "short-acting" types. Regular insulin (like the one from the pancreas) acts for maybe 6 hours. The ratio of how much insulin you need for a given weight of carbs doesn't stay constant. If you inject too much insulin, if you raise your carb intake, that leads to a higher ratio. Physical exercise modifies the ratio in a short-term fashion. Inflammation, like from a cold or other infection, can raise the ratio significantly. You could model the situation with a bunch of differential equations, and that's what a lot of scientific Most of these variables can't even be estimated precisely enough because there is too little data for too much variation. ML/AI is good at using information in the dataset that is hard to describe or see for Humans. But it can't use information that just isn't in the input data, it's not magic. For Type-2 diabetes there are relatively simple rules that work remarkably well. Type-1 is a lot harder, partly because a lot of the "buffering" of the system dynamics is missing.
bayesian_horse 4y ago
I only had a short period of injecting insulin and can empathize. I also had to do some of that Type-1 circus to cats and dogs in a clinical rotation. If you ever get Type 2 diabetes, go on a fast for maybe two or three weeks once your insulin requirements have stabilized somewhat... I'd like to say ask a doctor, but they mostly think fasting is bad for you, just without any evidence to that effect. Of course, BGL testing is mandatory, and insulin dosages have to be adapted (reduced) quickly. But I got over insulin dependence twice using that method and I blame the second crisis on my failure to follow through.
- bayesian_horse 4y ago
  I'd like to emphasize that I know Type-2 isn't very comparable to Type-1. I don't want to lessen or cheapen the struggle of Type-1 patients. I can only guess how hard that must be. I just wanted to pass on some knowledge because Type-2 is vastly more common.
Benjamin_Dobell 4y ago
Well written, great to see this getting attention. My daughter is 4, and was diagnosed at 2. Her blood sugar today has already been both 16 mmol/L (288 mg/dL) and 1.6 mmol/L (28.8 mg/dL). It's a constant struggle. 1.6 is not a normal occurrence. She was 6.7 and bolused 15 minutes before lunch (as per usual) and for whatever reason today she dropped incredibly fast. Treated before she reached that low, but metabolism can only work so fast. Here in lies the challenge of being someone else's pancreas. Just a note to anyone reading this who doesn't know much about Type 1 diabetes - and in all honestly, why would you? However, if someone says they or their child has Type 1, please don't start talking about your experience with Type 2 or gestational diabetes. Don't get me wrong, they're serious medical illnesses in their own right. However, it's sort of like someone saying they have debilitating rheumatoid arthritis and responding with how you broke your arm this one time. A broken arm can for sure cause life-long problems, but they're not the same thing. And just as someone with rheumatoid arthritis can also break their arm, someone with Type 1 can in fact still get Type 2, or gestational diabetes (if they're capable of childbearing).
- michaelgrafl 4y ago
  I think it was a mistake to call both those conditions diabetes.
  - grahar64 4y ago
    There are even more flavours of diabetes each with their own complications
- consp 4y ago
  Do note that all blood glucose meters get pretty inaccurate when it goes below about 4 mmol/l, depends a bit on the exact test used and varies wildly with environmental conditions. Consider it "low" from that point onward. The meter manufacturers will tell you it's always within the ISO bounds but that is under controlled conditions with new test strips accuracy goes out the window the lower you get and the less favorable the environment is. From my personal experience with bolus is that it can vary wildly with injection site especially for slower working insulin, I used to get massive hypoglycemia from time to time with actrapid simply to injection site changes.
godmode2019 4y ago
I'm not telling you how to parent, in saying that waking a growing boy in the night and limiting food during the day, may be more harmful in the long term. Childhood sleep and nutrition is probably the most important thing to a boys health. Forgive me if any offence is taken. Sometimes more data is not better and only leads to increased anxiety and over reactions.
- bayesian_horse 4y ago
  That you feel the need to give advice despite being monumentally unqualified to do so should give you pause. Your chosen name "godmode2019" speaks for itself. None of that comment is true except in the most general sense. You think you are giving advice, but the OP and other parents of Type-1 children have vastly more knowledge about diabetes, "childhood sleep and nutrition" and what it all means for their lives. Twice nightly blood sugar tests are unavoidable in many patients. No, you don't have to wake up the child necessarily. It's a simple finger prick at worst, so they might not even wake up or go right back to sleep. With modern sensors (which are less reliable) you only need to swipe the reader over the sensor.
  - godmode2019 4y ago
    39min of rem sleep is not enough for any human. My Grandad has type 1 diabetes he treats it with a strict routine of eating at the same time every day and injecting at the same time, and testing only if he feels low. The article I read was a parent in panic, and my only advice could possibly be to try and chill out, which is almost impossible as its your son and you would want to do as much as you can. Which is why I felt empathy to the poster, I could feel his anxiety.
    - grahar64 4y ago
      Your grandad was following advice given before the 1980s when typically using only long lasting insulin (which was all that was available) and strict routines were the only way to survive and be healthy with T1D. You should read that DCCT citation I link in my post, it is a good read, about the impact of following the treatments of nearly 1,500 diabetics and their long term complications. DCCT lead to a lower HbA1c being the goal as it is the largest predictor of long term issues. There are many books like “think like a pancreas” and “sugar surfing” which explain and use the same methods I explain in the post. If you are interested. Also, I would love to meet a person who can write a multi page blog post in a panic. Like “help my car is on fire: a history of car fires”
braingenious 4y ago
Out of curiosity, does the G6 require no calibration over 10 days? The way it’s worded is kinds of confusing because it kind of sounds that way. From my T1D partner of many, many years: “Great job! You’ve learned so much! All I would say is you’ve learned a lot, but as he ages, you’ll have to toss that out and learn that again, and again, and again.” Solid A1C! Solid care! Kudos!
- Benjamin_Dobell 4y ago
  No calibration required for the entire 10 days. The one significant inconsistency with BGL between my daughter's regular meter vs. G6 CGM was when my daughter had paracetamol in her system. Dexcom do claim (in Australia at least) that the G6 is not affected by paracetamol - which is the opposite claim they made about the G5. It was actually incredibly dangerous since it was 3am and the sensor was indicating she was very high and she was in fact slightly below target. Glad I thought it seemed sus and chose to do a finger prick, rather than telling her pump to deliver a bolus. That one (scary) incident aside, we've had no other issues with the G6. In general I've found it to be much more reliable than even a regularly calibrated G5. Perhaps the paracetamol incident was a coincidence and the paracetamol didn't play a part, but at least personally, I'm careful with medicines that have been known to interfere with CGMs - otherwise I tend to trust the G6.
wjnc 4y ago
This is an amazingly good read. The author pretty much describes in full all the requirements for a system / model that could help his son. (And which they have developed on an n=1 basis.) And it’s a complex system. Food, type of food, time, insuline (two types), body response times and curves / ratios, a measurement model. And then after admiring his technical writing I come away with a deep respect and empathy for what they are going through.
PetahNZ 4y ago
Its almost uncanny the similarity I have to Graham. NZ, programmer, daughter T1 diagnosis about 2 years ago. A CGM and insulin pump were absolute game changers for us. We went with the Dexcom CGM (not funded) and t:slim pump (funded). The control IQ software in the pump is great and we have got her time in range to > 95%. One scary story to share though, as we started trusting her school/teacher aids to administer/supervise bolus and corrections (before she was on a pump) it only takes one small mistake to potentially kill her. She was had high BGL and needed a correction to take her back down to 7 mmol. There is a bit of math to do, but they made a mistake and corrected her to 0 mmol. Luckily because we had the Dexcom we were alerted. We tried contacting the school when we saw her dropping, but got no answer. We rushed to the school to find he delirious lying on the field alone. It could have been worse but we are grateful for the technology that we have that basically saved her. Needless to say, we don't trust anyone else with those calculations any more.
- Benjamin_Dobell 4y ago
  Holy hell! This is literally my nightmare. I'm so sorry to hear you went through that and so glad she is okay!
- grahar64 4y ago
  Hey, We do have a lot in common. We have not trusted anyone with Sam’s management yet. That would be my nightmare situation though, and more reason to put technology with safeguards into the loop.
  - GiorgioG 4y ago
    It's been 8 years since our now 10 year old son was diagnosed with T1D. Pre-Covid he was in school (in the USA.) Only one potentially serious incident when the staff accidentally gave him 2X the insulin he was supposed to receive for lunch (this was when he was on a lousy Medtronic pump/CGM combo.) Fortunately the staff realized their mistake and called us immediately. We live a few minutes from the school and so a crisis was averted with a stack of glucose tabs before his blood sugar dropped dangerously low. We've homeschooled him since Covid (he has no desire to go back and socializes with other kids outside of school.) He's now on a Dexcom / Tandem pump + iPhone mini (for transmitting Dexcom blood sugar levels). We have three SugarPixels (https://customtypeone.com/products/sugarpixel https://customtypeone.com/products/sugarpixel) - one for our bedroom, one in the kitchen and one in my office. A game-changer for us was the free app called SugarMate. It integrates with Dexcom and will call you with an automated message when blood sugar level goes below a specified threshold. This helped us sleep through the night knowing if his blood glucose level went dangerously low we would be woken up by the automated phone call.
bleair 4y ago
May I suggest you look at this project that combines the Dexcom with your pump along with software. It is not a simple turn it on and automatically everything is perfect, however it is by far the best option right now for type 1 diabetics if you are willing to put in the effort to set it all up - which from all the testing and finger sticks indicates you are. https://loopkit.github.io/loopdocs/ https://loopkit.github.io/loopdocs/
- veryfancy 4y ago
  +1. Been running this for my 7yo son for several years now. Substantial quality-of-life gains.
- pimeys 4y ago
  There's another project for Android, mainly developed in EU called AndroidAPS: https://androidaps.readthedocs.io/en/latest/ https://androidaps.readthedocs.io/en/latest/ It has a quite nice support for different devices already, and the installation and usage is a tiny bit easier due to Android having much nicer side-loading experience compared to iOS. The project Discord has lots of tech-minded diabetics discussing... I've been using this software for multiple years now, and it has completely changed my life to the better.
- grahar64 4y ago
  Getting algorithms like loop that are available in New Zealand AND have a bunch of evidence for being safe with toddlers is difficult. At the moment we are using CamAPS which has been used a lot in the UK with toddlers but as far as I know Sam is the only one on it in New Zealand
Jarmsy 4y ago
First of all - my sympathies - it's hard enough managing this condition in yourself, and controlling it for someone too young to communicate must add a whole extra level of difficulty. What's the current thinking on a low carb diet for T1 infants? I'm T1 and during the periods when I've eaten very low carb I had amazing control - once I got my basal adjusted right I was using only tiny amounts of Humalog and my curve was the flattest it has ever been. It felt more like very occasional nudges to something that wanted to move straight, rather than constantly batting back and forth between 2 lines. That said, I do eat moderate carb these days, because I find that I can still manage good enough control - after decades of practise and with close monitoring and a fair bit of work - and it's more convenient for eating with others, not to mention cheaper. I found my energy levels and concentration on low carb were good after a while, but the process of switching is tough, and it's not something you can easily dip in and out of. If my highest priority was easier and more reliable control with minimal treatment though I would absolutely choose low carb. I realise that for a growing child, the nutritional concerns would probably be rather different though.
- devilbunny 4y ago
  You would definitely need to find the right pediatric endocrinologist with dietician support - preferably a dietician who has constructed and managed true ketogenic diets (for epilepsy, as the primary goal thereis to produce maximum ketone bodies as antiepileptics, not just cut carbs) for a good while. Those can be structured to provide all the necessary nutrients for growing children; infants would be hard even for someone with a lot of experience. For teenage and adult T1’s, simply being low-carb is often enough to greatly improve glucose control.
- vallassy 4y ago
  This is the comment I was going to write. I am not a parent and I have no experience with diabetes, but in my simple mind, if insulin's main role in the body is to control blood glucose, and the main trigger of blood glucose is carbohydrates... why not restrict carbohydrates? As with parent comment, it would just eliminate the blood glucose level rollercoaster that diabetics are otherwise expected to endure. You probably would not be able to find a nutritionist that would recommend such a diet for a toddler though, given the current mainstream recommendations on meat, fat and low carbohydrate diets.
toddm 4y ago
You are obviously a wonderful father and Sam is lucky to have you. That's all I wanted to say.
jablongo 4y ago
A great write up. Fate would have it that my name is Sam and I am T1D as well! You should sign up for the mailing list of the application I am currently working on https://replica.health/ https://replica.health/. We are building a context aware search and recommendation system for T1D data. If CamAPS writes its data to apple health like dexcom does you guys might find our tools useful for exploring his data and for diabetes decision support.

121 points by grahar64 4y ago

This is a follow up post from “The Unreasonable Math of Type 1 Diabetes” (https://maori.geek.nz/the-unreasonable-math-of-type-1-diabetes-8c96bdf5b7fb https://maori.geek.nz/the-unreasonable-math-of-type-1-diabet...). It is more about the long term complications of T1D and the way we have adapted our strategies to minimize them.

vertis 4y ago
It's an absolutely amazing read. Not really a useful comment, because I lack anything useful to say. ¯\_(ツ)_/¯
- grahar64 4y ago
  It’s better than the pro eugenics discussion above this comment. Cheers :)
agileAlligator 4y ago
- iLoveOncall 4y ago
  That's kinda the feeling I got from reading the post. Eugenics is a part of science deemed immoral and forever shunned because associated with the nazis, but maybe it would be good if we had made enough progress to avoid babies that come near death 4 times a day every day and require their two parents to be dedicated to their care 24/7.
  - consp 4y ago
    I've had this argument before: If you argue eugenics then you should really look into your family history and be 100% sure no-one has any "defects" anywhere. Most people who argue it have either needed some form of intervention somewhere in their lives to fix a "defect" and in other cases one of their predecessors and would they be unwanted in the eugenics-fanatics eyes. (Reflection usually works quite well with when someone is reasoning for eugenics)
    - iLoveOncall 4y ago
      Well that's a pretty week argument. Yeah sure if we killed every baby to be that had a defect I probably wouldn't exist. But I wouldn't care, because I wouldn't exist. Note I'm not saying to kill those unborn children, I'm saying that if this medical field had not been completely put to a stop following WW2, we wouldn't have to kill unborn children but we could do in utero interventions to prevent those conditions from developing. "Kill the babies" is the current state of eugenics because it has been at a full stop for more than half of a century. That's my point.
    - agileAlligator 4y ago
      Eugenics doesn't necessarily mean kill the baby. It meant that in the past only because science wasn't advanced enough. Techniques to modify the human genome exist; we could potentially "fix" the bad genes without killing the child.
  - borski 4y ago
    How many diabetic scientists and engineers have there been that have changed the world for the better?
    - iLoveOncall 4y ago
      Probably not as many as diabetic terrible people? I'm not saying to kill those unborn children, I'm saying that if this medical field had not been completely put to a stop following WW2, we wouldn't have to kill unborn children but we could do in utero interventions to prevent those conditions from developing.
      - valarauko 4y ago
        How, though? It's not like we've had the ability to diagnose the majority of these diseases in utero until fairly recently, nor have we the means to intervene, even now. CRISPR as a tool has existed for just over a decade, and we're still very far from the ability to reliably modify fetuses, even in lab animals, and it's been only possible for, what, 5 years now? Lab research did not stop after WW2, and even then we've barely gotten to this point. The ability to sequence the whole genome and find mutations for the major disorders is still a far off goal. The majority of disorders are now known to be polygenic, affected by hundreds of different gene variants, with each population group seemly affected by different groups of genes. Nor have we settled on a definitive group of gene variants that can reliably be identified to be deterministic of any widely distributed disorder, outside a small handful like CF and DMD.
    - agileAlligator 4y ago
      I am not saying we should let diabetic babies die at birth. I am simply reflecting upon the fact that their genes would have been removed from circulation in the past; this is no longer the case. If we disallow nature from taking its course, it will have consequences on "genetic quality". Following that train of thought and assuming we don't want to kill genetically malformed babies, an alternate route must be chosen. Better in-utero screening, CRISPR to fix the deformed genes, etc.
      - 4y ago
- datapolitical 4y ago
  Except type one diabetes is only partially affected by genetics. A woman with type one diabetes has less than a 4% chance of passing it to her child. The biggest risk right now to our genetic success is overwhelmingly the fact that the people most likely to reproduce are the ones with the least amount of education.
  - mschuster91 4y ago
    > The biggest risk right now to our genetic success is overwhelmingly the fact that the people most likely to reproduce are the ones with the least amount of education. Idiocracy wasn't a dystopia, it is a documentary... Sarcasm aside: A low amount of education is not really related to genetics (some forms of inheritable mental health issues aside). There are, however, extremely strong correlations with poverty: - poor families tend to live in areas with underfunded and thus lower-quality public schools, whereas rich people have better schools and can afford after-school private education - poor families tend to have issues with food security, which was brutally shown by covid as schools had to stay open simply so that the children could have at least one warm, somewhat healthy meal a day - poor families often tend to send their children off to work as soon as they can and are legally allowed to, simply to help make ends meet. That leaves them less time for homework and learning (leading to lower grades), and not being able to pursue higher education at all because having a job to survive is more important, which in turn locks them out of higher-paying careers and sets them up for being poor for life
    - mlyle 4y ago
      > A low amount of education is not really related to genetics Intelligence in adulthood is overwhelmingly heritable. Environmental and upbringing effects are small. Twin studies, unrelated sibling studies, etc, demonstrate very large effects. (Oddly enough, intelligence in childhood does seem to be very influenced by environment, but by adulthood your genetic destiny takes hold). Intelligence is also the biggest single predictor of your child's educational attainment-- even bigger than SES. Of course, it's not really clear what one can do with this information that is both ethical and useful. A confound: severe neglect and abuse is known to permanently reduce intelligence, and this is much less represented in adoption studies because of parental screening and social worker follow-up.
- whostolemyhat 4y ago
  What a stupid thing to say
- grahar64 4y ago
  Pro Eugenics comment to a post about keeping my kid alive. Wow!
  - sunmag 4y ago
    These comments and thoughts will come at you. You will quickly find reasons to dismiss them as noise. - father of T1D son.
  - agileAlligator 4y ago
- piokoch 4y ago
  I think we no longer aim at purely physical strength. In the old days person who was not able to work on the field, in the garden, have 10 kids, hunt, fight was useless. This is no longer the case, maybe those kids when they grow up will be scientists or artists, maybe they will make our lives much better.
robocat 4y ago
Perhaps add “article author” or something similar, as it is not obvious on some of your comments. Thank you for your write up about such a personal topic.
shagie 4y ago
(hmm, visited link)... ahh. For other people passing through the comments, that post was on HN several months ago https://news.ycombinator.com/item?id=30376777 https://news.ycombinator.com/item?id=30376777 with 673 points and 304 comments. There is a lot of interesting material in there to read too.

bigmattystyles 4y ago

Amazing read, really well written - prior to reading, it seemed odd to me that a simple PID insulin or sugar pump armed with a glucose sensor could not mimic a pancreas. I didn't realize there were so many variables. It still seems from my engineering mindset that a PID device should be able to do this, but I'm starting to understand why it's not so simple now.

bsder 4y ago
This could be engineered around--mostly the limitations are liability. If you, the patient, take a glucose sensor, an insulin pump and tie them together with an app, you don't get to sue somebody if something fails and you die. If a company takes all three of those, packages them together, and sells them, then that company is now liable when one of the pieces fails and you wind up dead. The sad part is that AI/ML would be incredibly useful for all this. People could take pictures of their meals, upload their sleep and exercise data, and weigh themselves, feed all that into a cloud algorithm that could determine how you are likely to react and dial your insulin plan into your pump while feeding the glucose readings back to the cloud. Sadly, I suspect none of this is going to happen in the US due to the liability. Maybe China will pull this off.
- bayesian_horse 4y ago
  I don't want to call your comment ignorant, but you are lacking a lot of the information necessary to see why it's not that easy. For starters: Insulin is quite long acting. Maybe 4 hours for some injected "short-acting" types. Regular insulin (like the one from the pancreas) acts for maybe 6 hours. The ratio of how much insulin you need for a given weight of carbs doesn't stay constant. If you inject too much insulin, if you raise your carb intake, that leads to a higher ratio. Physical exercise modifies the ratio in a short-term fashion. Inflammation, like from a cold or other infection, can raise the ratio significantly. You could model the situation with a bunch of differential equations, and that's what a lot of scientific Most of these variables can't even be estimated precisely enough because there is too little data for too much variation. ML/AI is good at using information in the dataset that is hard to describe or see for Humans. But it can't use information that just isn't in the input data, it's not magic. For Type-2 diabetes there are relatively simple rules that work remarkably well. Type-1 is a lot harder, partly because a lot of the "buffering" of the system dynamics is missing.

bayesian_horse 4y ago

I only had a short period of injecting insulin and can empathize. I also had to do some of that Type-1 circus to cats and dogs in a clinical rotation. If you ever get Type 2 diabetes, go on a fast for maybe two or three weeks once your insulin requirements have stabilized somewhat... I'd like to say ask a doctor, but they mostly think fasting is bad for you, just without any evidence to that effect. Of course, BGL testing is mandatory, and insulin dosages have to be adapted (reduced) quickly. But I got over insulin dependence twice using that method and I blame the second crisis on my failure to follow through.

bayesian_horse 4y ago
I'd like to emphasize that I know Type-2 isn't very comparable to Type-1. I don't want to lessen or cheapen the struggle of Type-1 patients. I can only guess how hard that must be. I just wanted to pass on some knowledge because Type-2 is vastly more common.

Benjamin_Dobell 4y ago

Well written, great to see this getting attention. My daughter is 4, and was diagnosed at 2. Her blood sugar today has already been both 16 mmol/L (288 mg/dL) and 1.6 mmol/L (28.8 mg/dL). It's a constant struggle. 1.6 is not a normal occurrence. She was 6.7 and bolused 15 minutes before lunch (as per usual) and for whatever reason today she dropped incredibly fast. Treated before she reached that low, but metabolism can only work so fast. Here in lies the challenge of being someone else's pancreas. Just a note to anyone reading this who doesn't know much about Type 1 diabetes - and in all honestly, why would you? However, if someone says they or their child has Type 1, please don't start talking about your experience with Type 2 or gestational diabetes. Don't get me wrong, they're serious medical illnesses in their own right. However, it's sort of like someone saying they have debilitating rheumatoid arthritis and responding with how you broke your arm this one time. A broken arm can for sure cause life-long problems, but they're not the same thing. And just as someone with rheumatoid arthritis can also break their arm, someone with Type 1 can in fact still get Type 2, or gestational diabetes (if they're capable of childbearing).

michaelgrafl 4y ago
I think it was a mistake to call both those conditions diabetes.
- grahar64 4y ago
  There are even more flavours of diabetes each with their own complications
consp 4y ago
Do note that all blood glucose meters get pretty inaccurate when it goes below about 4 mmol/l, depends a bit on the exact test used and varies wildly with environmental conditions. Consider it "low" from that point onward. The meter manufacturers will tell you it's always within the ISO bounds but that is under controlled conditions with new test strips accuracy goes out the window the lower you get and the less favorable the environment is. From my personal experience with bolus is that it can vary wildly with injection site especially for slower working insulin, I used to get massive hypoglycemia from time to time with actrapid simply to injection site changes.

godmode2019 4y ago

I'm not telling you how to parent, in saying that waking a growing boy in the night and limiting food during the day, may be more harmful in the long term. Childhood sleep and nutrition is probably the most important thing to a boys health. Forgive me if any offence is taken. Sometimes more data is not better and only leads to increased anxiety and over reactions.

bayesian_horse 4y ago
That you feel the need to give advice despite being monumentally unqualified to do so should give you pause. Your chosen name "godmode2019" speaks for itself. None of that comment is true except in the most general sense. You think you are giving advice, but the OP and other parents of Type-1 children have vastly more knowledge about diabetes, "childhood sleep and nutrition" and what it all means for their lives. Twice nightly blood sugar tests are unavoidable in many patients. No, you don't have to wake up the child necessarily. It's a simple finger prick at worst, so they might not even wake up or go right back to sleep. With modern sensors (which are less reliable) you only need to swipe the reader over the sensor.
- godmode2019 4y ago
  39min of rem sleep is not enough for any human. My Grandad has type 1 diabetes he treats it with a strict routine of eating at the same time every day and injecting at the same time, and testing only if he feels low. The article I read was a parent in panic, and my only advice could possibly be to try and chill out, which is almost impossible as its your son and you would want to do as much as you can. Which is why I felt empathy to the poster, I could feel his anxiety.
  - grahar64 4y ago
    Your grandad was following advice given before the 1980s when typically using only long lasting insulin (which was all that was available) and strict routines were the only way to survive and be healthy with T1D. You should read that DCCT citation I link in my post, it is a good read, about the impact of following the treatments of nearly 1,500 diabetics and their long term complications. DCCT lead to a lower HbA1c being the goal as it is the largest predictor of long term issues. There are many books like “think like a pancreas” and “sugar surfing” which explain and use the same methods I explain in the post. If you are interested. Also, I would love to meet a person who can write a multi page blog post in a panic. Like “help my car is on fire: a history of car fires”

braingenious 4y ago

Out of curiosity, does the G6 require no calibration over 10 days? The way it’s worded is kinds of confusing because it kind of sounds that way. From my T1D partner of many, many years: “Great job! You’ve learned so much! All I would say is you’ve learned a lot, but as he ages, you’ll have to toss that out and learn that again, and again, and again.” Solid A1C! Solid care! Kudos!

Benjamin_Dobell 4y ago
No calibration required for the entire 10 days. The one significant inconsistency with BGL between my daughter's regular meter vs. G6 CGM was when my daughter had paracetamol in her system. Dexcom do claim (in Australia at least) that the G6 is not affected by paracetamol - which is the opposite claim they made about the G5. It was actually incredibly dangerous since it was 3am and the sensor was indicating she was very high and she was in fact slightly below target. Glad I thought it seemed sus and chose to do a finger prick, rather than telling her pump to deliver a bolus. That one (scary) incident aside, we've had no other issues with the G6. In general I've found it to be much more reliable than even a regularly calibrated G5. Perhaps the paracetamol incident was a coincidence and the paracetamol didn't play a part, but at least personally, I'm careful with medicines that have been known to interfere with CGMs - otherwise I tend to trust the G6.

wjnc 4y ago

This is an amazingly good read. The author pretty much describes in full all the requirements for a system / model that could help his son. (And which they have developed on an n=1 basis.) And it’s a complex system. Food, type of food, time, insuline (two types), body response times and curves / ratios, a measurement model. And then after admiring his technical writing I come away with a deep respect and empathy for what they are going through.

PetahNZ 4y ago

Its almost uncanny the similarity I have to Graham. NZ, programmer, daughter T1 diagnosis about 2 years ago. A CGM and insulin pump were absolute game changers for us. We went with the Dexcom CGM (not funded) and t:slim pump (funded). The control IQ software in the pump is great and we have got her time in range to > 95%. One scary story to share though, as we started trusting her school/teacher aids to administer/supervise bolus and corrections (before she was on a pump) it only takes one small mistake to potentially kill her. She was had high BGL and needed a correction to take her back down to 7 mmol. There is a bit of math to do, but they made a mistake and corrected her to 0 mmol. Luckily because we had the Dexcom we were alerted. We tried contacting the school when we saw her dropping, but got no answer. We rushed to the school to find he delirious lying on the field alone. It could have been worse but we are grateful for the technology that we have that basically saved her. Needless to say, we don't trust anyone else with those calculations any more.

Benjamin_Dobell 4y ago
Holy hell! This is literally my nightmare. I'm so sorry to hear you went through that and so glad she is okay!
grahar64 4y ago
Hey, We do have a lot in common. We have not trusted anyone with Sam’s management yet. That would be my nightmare situation though, and more reason to put technology with safeguards into the loop.
- GiorgioG 4y ago
  It's been 8 years since our now 10 year old son was diagnosed with T1D. Pre-Covid he was in school (in the USA.) Only one potentially serious incident when the staff accidentally gave him 2X the insulin he was supposed to receive for lunch (this was when he was on a lousy Medtronic pump/CGM combo.) Fortunately the staff realized their mistake and called us immediately. We live a few minutes from the school and so a crisis was averted with a stack of glucose tabs before his blood sugar dropped dangerously low. We've homeschooled him since Covid (he has no desire to go back and socializes with other kids outside of school.) He's now on a Dexcom / Tandem pump + iPhone mini (for transmitting Dexcom blood sugar levels). We have three SugarPixels (https://customtypeone.com/products/sugarpixel https://customtypeone.com/products/sugarpixel) - one for our bedroom, one in the kitchen and one in my office. A game-changer for us was the free app called SugarMate. It integrates with Dexcom and will call you with an automated message when blood sugar level goes below a specified threshold. This helped us sleep through the night knowing if his blood glucose level went dangerously low we would be woken up by the automated phone call.

bleair 4y ago

May I suggest you look at this project that combines the Dexcom with your pump along with software. It is not a simple turn it on and automatically everything is perfect, however it is by far the best option right now for type 1 diabetics if you are willing to put in the effort to set it all up - which from all the testing and finger sticks indicates you are. https://loopkit.github.io/loopdocs/ https://loopkit.github.io/loopdocs/

veryfancy 4y ago
+1. Been running this for my 7yo son for several years now. Substantial quality-of-life gains.
pimeys 4y ago
There's another project for Android, mainly developed in EU called AndroidAPS: https://androidaps.readthedocs.io/en/latest/ https://androidaps.readthedocs.io/en/latest/ It has a quite nice support for different devices already, and the installation and usage is a tiny bit easier due to Android having much nicer side-loading experience compared to iOS. The project Discord has lots of tech-minded diabetics discussing... I've been using this software for multiple years now, and it has completely changed my life to the better.
grahar64 4y ago
Getting algorithms like loop that are available in New Zealand AND have a bunch of evidence for being safe with toddlers is difficult. At the moment we are using CamAPS which has been used a lot in the UK with toddlers but as far as I know Sam is the only one on it in New Zealand

Jarmsy 4y ago

First of all - my sympathies - it's hard enough managing this condition in yourself, and controlling it for someone too young to communicate must add a whole extra level of difficulty. What's the current thinking on a low carb diet for T1 infants? I'm T1 and during the periods when I've eaten very low carb I had amazing control - once I got my basal adjusted right I was using only tiny amounts of Humalog and my curve was the flattest it has ever been. It felt more like very occasional nudges to something that wanted to move straight, rather than constantly batting back and forth between 2 lines. That said, I do eat moderate carb these days, because I find that I can still manage good enough control - after decades of practise and with close monitoring and a fair bit of work - and it's more convenient for eating with others, not to mention cheaper. I found my energy levels and concentration on low carb were good after a while, but the process of switching is tough, and it's not something you can easily dip in and out of. If my highest priority was easier and more reliable control with minimal treatment though I would absolutely choose low carb. I realise that for a growing child, the nutritional concerns would probably be rather different though.

devilbunny 4y ago
You would definitely need to find the right pediatric endocrinologist with dietician support - preferably a dietician who has constructed and managed true ketogenic diets (for epilepsy, as the primary goal thereis to produce maximum ketone bodies as antiepileptics, not just cut carbs) for a good while. Those can be structured to provide all the necessary nutrients for growing children; infants would be hard even for someone with a lot of experience. For teenage and adult T1’s, simply being low-carb is often enough to greatly improve glucose control.
vallassy 4y ago
This is the comment I was going to write. I am not a parent and I have no experience with diabetes, but in my simple mind, if insulin's main role in the body is to control blood glucose, and the main trigger of blood glucose is carbohydrates... why not restrict carbohydrates? As with parent comment, it would just eliminate the blood glucose level rollercoaster that diabetics are otherwise expected to endure. You probably would not be able to find a nutritionist that would recommend such a diet for a toddler though, given the current mainstream recommendations on meat, fat and low carbohydrate diets.

toddm 4y ago

You are obviously a wonderful father and Sam is lucky to have you. That's all I wanted to say.

jablongo 4y ago

A great write up. Fate would have it that my name is Sam and I am T1D as well! You should sign up for the mailing list of the application I am currently working on https://replica.health/ https://replica.health/. We are building a context aware search and recommendation system for T1D data. If CamAPS writes its data to apple health like dexcom does you guys might find our tools useful for exploring his data and for diabetes decision support.