Hacker News new | comments | ask | show | jobs | submit login
Tracking Blood Sugar (jain.name)
171 points by troydavis 9 days ago | hide | past | web | favorite | 79 comments





I've been using one of these for about 6 weeks, and I find it is not very accurate, but the convenience that allows checking the readings very often nearly makes up for the huge error bars. The libre freestyle 14-day is usually within 15 mg/dl when the reading is between 80 and 100, but above that range, it is usually off by 20 to 30 percent of the correct value (not 20-30 mg/dl) and mine has never read much higher than the finger-stick test, only lower. The lower numbers make me inclined to believe it, but they are simply not true, the meter gives only about one significant digit of accuracy throughout its range, and it is pretty far off or refuses to produce a reading when my blood sugar goes below 60 mg/dl. Still, being able to see at a glance whether one is 100 or 200 is quite useful. Medicare pays about $500/month for the attachable sensors, which are extremely convenient and non-problematic so far.

There is a far better wearable system ready to go but not in production, a combined sensor and automated insulin dispenser, that has passed human trials, but no one wants to go into production with it because of liability issues. Given that people are riding in self-driving cars and landing in aircraft on autopilot, each of which are also potentially fatal and involve a larger set of variables, I wonder if and how long it will take to progress in this regard.


The libre is meant for monitoring the trend of the glucose, used together with a fingerstick reader. You can calibrate the values if you use xDrip to get better accuracy.

If you're interested in using a closed loop, take a look into AndroidAPS or Loop, how to build one by yourself. It's all open source and works really well. Been a user for some years now. You just need to compile it by yourself and own the right hardware.


> The libre is meant for monitoring the trend of the glucose, used together with a fingerstick reader.

Then it gets to be an issue with cost and convenience, as hauling around both systems with lancet device and strips and insulin and syringes and glucose tablets is a nuisance for anyone who does not carry a purse. Test strips cost around a dollar each (if one eschews the likes of Walmart), and it is very difficult to get Medicare to pay for more than 3 per day. I do not know if they will pay for any now that they are paying the higher monthly cost of the libre sensors.


Are you sure that the reading were 20-30 percent off and not just delayed? Libre and other sensors measure sugar level in the extracellular fluid and it takes some time before glucose gets there from blood.

The delays make the comparison difficult. Also if it gets too high or low then people react and quickly change the slope of the graph, then the sugar level in the extracellular fluid would not reach the same extremum as the sugar in the blood.


If the monitor problem were only a delay problem, my libre monitor would read higher than the finger stick when my glucose is heading downward, but I have never seen that.

> and I find it is not very accurate

i have the same experience, the Libre was very inaccurate for me as well.

> There is a far better wearable system ready to go but not in production, a combined sensor and automated insulin dispenser, that has passed human trials, but no one wants to go into production with it because of liability issues.

Well there's this, even if it's not a "100% pancreas replacement" yet: https://www.medtronicdiabetes.com/products/minimed-670g-insu...

"Automatically adjusts your basal (background) insulin every five minutes based on your CGM readings."


Here's a second non-diabetic who tracks their glucose with the same non-invasive device and started analyzing it.

Blog post: https://richardsprague.com/notes/cgm/cgm_analysis/

Slides from presentation last week: https://www.slideshare.net/richardsprague/quantified-self-se...

He also tweets about it: https://twitter.com/sprague


It seems to still be invasive: https://www.freestylelibre.us/support/faq.html#faqTabstab-3

> The sensor is applied to the back of the upper arm with a simple, disposable device called an applicator. When the sensor is applied, a small (5mm) filament is inserted just under the skin, and held in place with a small adhesive pad.

'Inserted just under the skin' means it's still invasive, but that sounds nicer than regular needle-sticks so perhaps 'less-invasive' would be a more accurate description.


Per the people I know who have tried it, including the author of this post, it's totally painless. One can debate whether or not a filament you can't feel is invasive, but for all practical purposes it's the same as no filament.

For clarity: the measurement you pasted is the length of the filament. The width/thickness is ridiculously tiny. Basically, this product has the same UX as a nicotine patch.


From a design perspective definitely invasive. Design strategies and regulations change greatly for products which penetrate the skin. But you are correct in that it is a patient friendly device without the associated burden of lost invasive devices.

From an FDA perspective invasive also. Otherwise you wouldn't need a prescription.

I have a Guardian 3 sensor from Medtronic, which I've been using for about a year. The initial insertion isn't "painful", but can definitely be uncomfortable, and there might be a bit of blood that leaks into the sensor cavity. Once it is in, there is no discomfort.

The adhesive tape which is needed to attach it to your skin is actually more uncomfortable than the sensor.

Combined with a 670G pump it has been a huge benefit for me. My A1C of 9.0 is now down in the 6's. My pump broke while I was on vacation for a week, and my control was clearly worse with pens and finger sticks than the pump and sensor.


> perhaps 'less-invasive' would be a more accurate description

I prefer "minimally invasive"... You can't just remove the device and stick it back on, but it's good enough that a truly non-invasive device would be a marginal improvement for me.


If you travel a lot it’s invasive in that you can’t really wear it through full body scanners.

I wear a FreeStyle Libre every day, AMA. It's very good for T1s, of course.

It's reliable enough - the biggest downfall is that it's around 15 minutes behind real-time, ie, what you can measure with blood. Sometimes it can feel 'stuck' too, and change rapidly between a short amount of time.

The second model, which I am waiting to try, includes Bluetooth alerts if a trend high or low is detected. This may make it even more useful.


What do you mean by 'second model'? Like a new version of the FreeStyle Libre?

I just got my first CGM, FreeStyle Libre 14 day, 2 days ago :) The relief I've already experienced from not having anxiety about where my blood sugar is at has been incredible. That said, being able to set alerts for lows/highs and not having to manually scan the sensor would be huge improvements.

The way I was checking blood sugar before (finger pricks) it seemed like I'd gain relatively little by using a pump, but now I'm thinking a pump may be indispensable if used in combination with a CGM: Since I can now see how my blood sugar is changing in near real-time, if I have errors in dose size, I'm going to want to be able to make small correction...


Yes, v2 is coming soon or is now out :)

Drink water if it feels to be stuck. Usually a symptom of dehydration.

Alrighty, thanks!

I’ve been using my Libre with Miao Miao, Spike and Nightscout for getting a real time feed for a while now. It took some time to get a feel for how the sensors behave, but I’m happy with the setup - it’s certainly good enough for my goals.

I really like the real time data aspect which means I can access it easily on other devices.

I also still regularly use the official phone app to get a reading via NFC which suits certain circumstances for me and I find pretty accurate most of the time.


Are there any devices that track insulin levels directly?

That’s what I’d be most interested in. Folks conjecture that it’s your insulin response that causes weight gain rather than the blood sugar increase. They’re not as correlated as you’d expect!

Source https://www.google.com/amp/s/optimisingnutrition.com/2015/05...


Summary: I have no citation but my understanding is that measuring insulin levels needs to be done from a blood draw, so a Freestyle Libre-type device wouldn't work.

The Freestyle Libre and regular continuous glucose monitors measure glucose in the interstitial fluid under the skin - not blood. This is why the numbers can be wildly different to a blood glucose test done via finger-prick.

From what I've seen while researching very low carb diets, insulin levels are incredibly important and influence a lot of body chemistry. For example, someone without type 1 diabetes can have a normal blood glucose but only because they're producing an enormous amount of insulin. A large amount of insulin in the blood will cause fat to be stored, so the patient puts on weight.


Sounds like what you’d be interested in is insulin sensitivity, the ability to control blood sugar with the minimal amount of insulin possible. Sadly this is also a huge pain in the butt to measure, and the only way I’ve heard is to spike your blood sugar and time how long it takes to return to normal.

I personally use a keto mojo to measure blood glucose and blood ketone bodies. I’m not diabetic so I don’t need to use it, but I’m a fan of being able to tell how far into ketosis I am.


Your example is someone without diabetes but referred to as a patient?

My comment wasn't clear; apologies. I was referring to someone who had blood drawn to test for and diagnose high insulin levels. I'd classify anyone having that test in a clinical setting to be a patient.

The Freestyle Libre requires a prescription (why?) — is there another similar system that does not?

No. In the US, all similar products require a prescription. This is by far the best available option (no fingerstick calibration, longest-lasting sensor, easy mobile app, etc.). Basically, get this one. It’s $80 including 1 month of sensors.

I believe the thinking is that people need a professional to discuss results with. I understand that perspective, but I don’t agree with it, and I definitely don’t agree that having someone to discuss results with should outweigh the increased access/adoption of over-the-counter sales. This is the first product that makes it practical - even easy - and completely painless for non-diabetics like me to track blood glucose informationally, so perhaps this will change. That said, the FDA has made other overly-conservative decisions, like disallowing fingerstick A1c tests (which don’t require fasting or a full venous blood draw) for first-pass screening of diabetes: http://www.onsitehealthdiagnostics.com/blog/a1c/, https://twitter.com/troyd/status/1018150306125758465.


Thanks for clarifying.

As a side note, I think this requirement is completely bogus. You should be able to get any "read-only" medical device you want, anytime you want, especially in a country where there is no universal healthcare. I see a contradiction where on one hand the authorities say: we have to protect you so that you don't harm yourself and on the other the same authorities say: your health is your responsibility and we want nothing to do with it.


You need a prescription because it's a continuously attached invasive medical device. You don't need prescriptions for finger stick devices so that's the only difference between them.

"Read-only", but "mutable" in C++ const sense with slightly bloody side effects :)

Dexcom has a similar sensor on the market that’s approved for insulin dosing decisions. I don’t know if any of the other continuous monitors have that particular feature which might be important to those who take fast insulins.

Dexcom G5 and G6 are the best available CGM systems for type 1 diabetics. What they add over the libre is a Bluetooth transmitter over the sensor, giving values to your phone every five minutes. Used together with xDrip and AndroidAPS it can easily safe the life of a person using insulin and not waking up to a hypoglycemia, me included. Or work as a part of a closed loop system.

The problem with Dexcom is the price. A sensor should be replaced every week and costs 80€, the transmitter turns off after three months and costs between 300-400€. If you understand what you're doing, you can easily extend the sensor lifetime to 3-5 weeks and transmitter to 6-8 months.

My take to this is, every T1 deserves a Dexcom system from the insurance. It easily saves your life and makes the control much easier.


My kids transmitters get flaky after about 3 weeks, and I didn’t think you could reset an expired sensor on the G6 the way you could with the G5?

Should be possible with xDrip. Using the G5 still, so no real life experience.

Some sensors get noisy sooner than the others. I noticed they work longer during the winter when its cooler. Last one worked for 4 weeks quite good, but in the summer they stopped working in less than two weeks.


I realized I wrote weeks - the transmitter is good for 3 months and gets flaky in the last couple weeks.

I found some pictures of someone replacing the battery on the transmitter of a G6. This is.... not really practical. [1]

What would be interesting is if you could pull out the internals and 3D print a new enclosure that had an actual user serviceable battery.

[1] - https://seemycgm.com/2018/07/12/g6-transmitter-battery-repla...


The transmitter has a counter inside and will stop working after a certain day count. xDrip has a feature where one can reset the counter back to zero and I've been using them at least six months with no trouble.

Without changing the battery? Are you using G5 or G6?

I can understand why the sensor will expire, but why does the transmitter expire? Shouldn’t it just require changing a battery?

Because of money. I know dentists who can change the battery. It is possible, but requires the right tools.

Didn't realize that. I occasionally use a finger stick A1C to compare to my lab test and it's very accurate (within 0.2 units).

It does seem like a helpful tool to screen.


That seems a bit strange.

Here in the UK, I don't have a prescription (despite being type 1 diabetic) but occasionally self-fund the devices directly from Abbott. I get a VAT excemption so it costs ~£45 per unit, which lasts 2 weeks.


Maybe to get insurance to cover the cost?

If I'm not a diabetic but just want to track my blood glucose, will a doctor still prescribe me one? Will insurance balk at it? Any estimates as to final cost?

I enjoyed the Authors blog post history of:

Tracking Blood Pressure

Tracking Blood Sugar

Chocolate Popcorn

Panna Cotta


Plus Smoothies and Protein Shakes. There's clearly a pattern here :)

> The sensors worked as advertised

I mean... yes. They're a prescription medical device used by Type 1 diabetics to stay alive.


Unfortunately this means less than you'd think. See other comments on this thread on the subject of accuracy. Also, from my understanding, the current (still wanting) state of accuracy is the result of large improvements during just the past few years. Prior to that, CGMs were still prescription medical devices used by Type 1 diabetics to say alive.

And more unfortunately, it has been my experience that information about managing blood sugar given to diabetics by doctors is almost uniformly quite bad (I've received the same sort of overly simplistic, misguided instructions from 4 or 5 different doctors myself; now I just nod and let them finish speaking). As a result, nearly every other (non-engineer/scientist) diabetic I've been met has very bad numbers (A1C: three month average blood glucose)—as I would have too were I to follow my doctors' advice. Instead I have very good numbers which are a result of research and reasoning I had to do myself.


That's troubling. Which doctors have you been speaking to? I've always gone to see endocrinologists specializing in diabetes and have never had a doctor give me overtly bad advice.

Always general practitioners. I tend to just visit them when I need a prescription now. I haven’t visited an endo since I never understood the purpose: I suppose they are more expensive and I can just use the internet. My A1c is generally high threes, so I’m not sure what else they could offer. I’ve asked general practitioners and they have never been more concrete than: endocrinologists can help you figure out which pump is right for you. But, internet.

Edit! High fives, not threes. See my comment below.


I had the same issues until I found a doctor who also has a type 1 diabetes. After I started going to her practice, my A1c dropped from 7.5 to 5.5-6.1 and I learned so much about keeping a good control.

The one I was seeing before her once started yelling at me for having such a bad control, so I seriously started looking for a better practice.


Wow - if you're Type 1 and have an A1C in the "high threes"...you're definitely one of those exceptional patients. Your A1C is probably better than some non-diabetics!

I could see why you have a pretty high threshold for the quality of advice you want. Not sure it's a typical experience, though.


It seems dangerous to me. One unit of insulin too much and you're in diabetic coma territory.

> My A1c is generally high threes

Wait, really? Your average bg is in the low 60s? I've never heard of that before.


My bad, I actually haven’t checked it in a couple years because every time I went in my doctors assured me I had very good scores - one said better than they’d ever seen. So I haven’t changed my patterns and visit doctors as infrequently as possible.

I just checked my last score which was 5.8. IIRC I’ve always been between 5 and 6 since I started directing my own treatment.


This is Hacker News. There are lots of people obsessing about things that aren't a problem.

I'm curious about this. Theoretically healthy people shouldn't care about their blood sugar since your body normally regulates your blood glucose using insulin and other hormones.

If you're wondering if you may be pre-diabetic then wouldn't an A1C test (home tests available and are very accurate) be a better diagnostic test?


That way you can detect a problem, yes.

But HBA1C can sometimes still be good, even for people who have high & low glucose peaks.

Maybe because the average value will still be good if you have equal peaks "up" as "down"? i don't know..


It is absurd that two (not one) doctor's scripts are needed to buy a sugar testing device. I've been trying to get this for my mother but it has been an awful experience in terms of their willingness to sell. I wish a company that actually wants to make money would soon offer an alternative.

They are opaque on pricing, do not answer questions about where to buy, do not clearly state what kind of prescription they need.


My mother had the Freestyle Libre and liked it. Now using a Dexcom which works with her Apple Watch. I just bought her a Dario finger-stick glucose meter for her birthday which plugs into iPhone lightning port ($40!).

https://imgur.com/a/24FhAON

http://intro.mydario.com/DarioGlucoseMeter/about.php


There is a German company offering personalized food recommendation based on measuring blood sugar: https://www.millionfriends.de/ (Google Translate: https://translate.google.com/translate?sl=de&tl=en&u=https%3...).

I have the Medtronic MiniMed pump, which includes its sensor that the pump uses for automated basal dosing. Accuracy is similar to what others have described.

How well does the automated bits work? I've considered a pump in the past for that reason, but was always afraid what mistakes could happen.

If you want to get even better system using AndroidAPS, get the DanaRS pump, which is the only manufacturer allowing unofficial use.

Before you must do some reading and understand the concepts. OpenAPS and AndroidAPS are state of the art systems and you can do even better than the basal setup using the oref1 SMB algorithm. I know countless of diabetics using it and we all have much better and easier control in our therapy.

https://openaps.org/ https://github.com/MilosKozak/AndroidAPS


Pretty well. You actually have to recalibrate it with traditional blood readings at least twice a day (and when it sees extremes) or it goes out of auto mode, and when it hits certain boundaries, it also turns off auto mode. It's worked great, and in a few months, I went from an A1C around 12 to around 7-8.

When I lived in Canada I used to occasionally see infomercials about blood sugar monitors that didn't require pricking fingers - are those considered to be any good or is it some kind of infomercial scam? If they are considered to be mostly accurate, why can't that technology be used instead of having a 5mm needle in your body, which probably feels uncomfortable, particularly for the elderly.

Monitors that don't need a finger stick can take blood from other body areas, usually the forearm. The theory is there are less nerves in places other than the finger.

The result is not considered as accurate compared to finger sticks but it's an option for people that can't really stick their fingers I guess.

They're mostly the more modern ones that don't need much blood. Also the ones I've purchased have a special attachment for the lancet device to accommodate the different location.

The ads do seem to imply that you don't need to use a lancet but that's just marketing and advertising. You still need to use a lancet to get a drop of blood.


My son use it since a year and measures are accurate for normal glucose concentration but not for higher ones. The device also includes a classic finger reader for double checks. The French healthcare reimburse 2 devices per month, which is perfect.

I tested my blood sugar with strips for a few days. One wierd thing I noticed for some meals is I’d be at 98 blood sugar or so before the meal but within 30 minutes after a moderate meal I’d be at 78 or so. Seemed really backwards.

Digestion uses a surprising amount of energy, this is why people tend to feel sleepy after carb-heavy meals. And diabetics need to plan carefully.

>why people tend to feel sleepy after carb-heavy meals.

No. This is well understood science. The fatigue you feel soon after eating is caused by postprandial hyperglycemia, or a “spike” in blood sugar soon after eating.[1] You can, and I have tested this after overeating carb heavy meals and that's exactly what's going on.

I'm prediabetic, so I watch this carefully as I have a family history of diabetes.

[1] https://www.diabetesselfmanagement.com/nutrition-exercise/nu...



I know someone who could benefit (either psychologically or medically) from a sort of holter monitor for her blood glucose levels. Are you the person who wrote this? If so how did you get your prescription?

I’m not the author, but the author simply asked their primary care doctor for one and I know others who have done the same. Next time you’re at the doctor, just explain that you want to be proactive about health and ask for one.

Make sure to get the 14-day version: https://www.freestylelibre.us/system-overview/freestyle-14-d.... It’s about $80 (cash to a pharmacy, not through insurance) and includes 2 14-day patches.

Over the next year or two, performing at least a few weeks of real-time monitoring per year will become standard practice among those who consciously monitor their health. With the Libre, doing a few weeks per year is no harder, and not much more expensive than a home blood pressure cuff. It’s not even all that expensive to do year-round.

(I don’t think this should require a prescription, but in the US it currently does.)


It's a waste of money and time for the average person that eats right and exercises. It borders on obsessional behavior.

You may be able to get this soon, too:

https://www.pkvitality.com/ktrack-glucose/

It works the same way. It's 2x more expensive but it's a watch.


> Fasting glucose and A1C tests done over the past few years show me at low risk for diabetes. But these tests don’t necessarily capture how much blood sugar “spikes” after meals–something that may be a better predictor of disease risk.

It's hard to take the author's concern seriously when he is at low risk. A better take is that he's just curious.


I'd be curious to know how he measures stress levels.

I don't even know how I'd go about measuring stress levels, but I did notice that there were two instances where my blood sugar staid elevated for much longer than it should have after lunch, and both times there was a somewhat stressful/non-routine situation (an interview, and riding a scooter across town).

What about availability of this outside the US?



Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: