Apple Watch Series 8 Suppliers Developing Blood Glucose Monitoring Components

[mikemj23]

Here is a question for you all. Would you buy this if the measurement were only qualitative instead of quantitative?

Apple may bypass the FDA by making a clinically irrelevant metric around Blood Glucose, and just report "Low", "Normal", "High", and not report the actual value to the user. They'd likely add text that this is for "informational use only", since you don't want diabetics to rely on this to administer insulin to themselves based on what the watch said. They may collect that data for themselves though, sending it back to Apple headquarters for algorithm and Machine Learning (NN) development until they can pass quantitative determination through the FDA.

This is my fear and it would not be a buy for me if this is how it's implemented, but again, I have much better tools (Dexcom g6 and soon g7) at my disposal that I use to dose my insulin and my pump uses, to a certain degree, to regulate my blood sugar. I would likely buy the watch because I am an annual upgrader of the Apple Watch and would not even look at the blood sugar application.

In this case, you are really looking at two groups. Non-diabetics who would think it would be useful for a general picture of glucose trends in somewhat of a gimmicky way and diabetics (T1 and T2) who would love to have it to make every day life changing decisions with. I could totally see Apple, at least initially, going the qualitative route to bypass the FDA's stringent controls on approving a non-invasive device to make dosing decisions.

 

[mikemj23]

This comes up a lot but as a type 1 diabetic and someone who works in the field of clinical diagnostics, this is not going to provide equivalent function to an FDA regulated medical device. Apple has been keen to avoid that burden. It will be targeted to type 2 diabetics and will be a retrospective target based evaluation without precise numbers.

You can expect the initial implementation to be similar to how they currently do other stuff. e.g. today you were 80% in normal blood glucose range and 20% out of range.

Doing otherwise would require Apple to go through FDA regulatory processes as a medical device and demonstrate many requirements as well as opening themselves up to massive liability as well as causing problems for them when they do software updates as they would need to do extensive validation of the blood glucose calling functions every update.

Apple implicitly markets the apple watch as a quasi-medical device (health device) but be assured that it is not a medical device and Apple does not want it to be one.

This is spot on.

 

[AndyMers]

"Non-invasive" would mean no blood sample.

They need to make the SE version "invasive only"?

 

[jlbraun]

If Apple cracks noninvasive blood glucose, it goes beyond just "selling a lot of devices".

Diabetes is a $237B/year market. Managing diabetes with their own devices will suck so much money out of healthcare into Apple's pockets while making it cheaper to manage.

Apple will then start opening their own hospitals, offering "Healthcare+" at $599/yr.

 

[Theblacksaint24]

Forgot to that will be expeny, very expensive.

 

[svish]

Should be a worthy upgrade with the new sensors. Maybe the rumoured redesign that was expected this year may come true with series 8.

 

[aperantos]

They need to make the SE version "invasive only"?

As those of us who test already use invasive methods we would not only be okay with that, we would probably trust it more that it should be the Plus version!

 

[Luke MacWalker]

But it'll be welcome if it works. At the moment Apple fails to get embedded TouchID working ...

What does make you think they are even trying? ?

 

[axcess99]

This is my fear and it would not be a buy for me if this is how it's implemented, but again, I have much better tools (Dexcom g6 and soon g7) at my disposal that I use to dose my insulin and my pump uses, to a certain degree, to regulate my blood sugar. I would likely buy the watch because I am an annual upgrader of the Apple Watch and would not even look at the blood sugar application.

In this case, you are really looking at two groups. Non-diabetics who would think it would be useful for a general picture of glucose trends in somewhat of a gimmicky way and diabetics (T1 and T2) who would love to have it to make every day life changing decisions with. I could totally see Apple, at least initially, going the qualitative route to bypass the FDA's stringent controls on approving a non-invasive device to make dosing decisions.

Would still be valuable for T1's that don't have CGM if it just provided an advisory "you *may* be very low" type feedback; say <70.

 

[mikemj23]

Would still be valuable for T1's that don't have CGM if it just provided an advisory "you *may* be very low" type feedback; say <70.

Maybe. If it's wildly inaccurate then you may not need to treat a perceived low or it may be too late. You could be right though.

 

[JBinPDX]

Here is a question for you all. Would you buy this if the measurement were only qualitative instead of quantitative?

Apple may bypass the FDA by making a clinically irrelevant metric around Blood Glucose, and just report "Low", "Normal", "High", and not report the actual value to the user. They'd likely add text that this is for "informational use only", since you don't want diabetics to rely on this to administer insulin to themselves based on what the watch said. They may collect that data for themselves though, sending it back to Apple headquarters for algorithm and Machine Learning (NN) development until they can pass quantitative determination through the FDA.

As a Type 2 diabetic with very good control (no meds, normal A1C), I'm allowed only 1 test strip per day by my insurance. Any more than that I have to pay for myself, and they are not cheap. I have to spend the extra money to buy them because one test/day is not enough to maintain that good control. Even qualitative measurements would be really helpful in knowing that it might be a good time to use one of those precious extra strips to see where I'm at. But that's with the caveat that "low," "normal," and "high" need to be clearly defined with some actual quantitive ranges. I need to know if what I consider "high" is what Apple considers "high."

That being said, I'd probably sit out the first watch with blood glucose monitoring and wait to see if it's even close to accurate first. I think it's probably going to be hard to buy without waiting months for delivery, because it will be such a game changer for so many people. I'll let others take it for a test drive first. I think it's years away in any case.

 

[AndiG]

What does make you think they are even trying? ?

There are just too many rumors and leaks of insiders to believe Apple doesn‘t try.

Long story short - Apple failed with the oxygen monitoring, from a medical point of view. It is likely that Apple will fail with glucose monitoring. Maybe the system could help to detect a type 2, but for controlling the the blood sugar of a type 1 it would have to deliver super accurate results.

 

[RVijay007]

Problem there, what is the range for Low/Normal/High? Because those targets would differ between a normal person and someone with diabetes.

You are absolutely right. This was the only thing I could think at the time, but now I fear it may be even worse than what I wrote. Namely, they will do something similar to the EKG and calling sinus rhythm or "irregular heart rhythm detected". So it may be just a binary state they call either "normal" or "high glucose detected". Still suffers from the "what is high?" question, but likely using something comparable to 100mg/dl. Or "normal" / "not normal" detected.

Reminds me of the "pizza" / "not pizza" classifier that Silicon Valley (show) poked fun at.

 

[aperantos]

The other problem with "what is high?" is that it not only varies by person but also throughout the day. For a non-diabetic being as high as 7.8 mmol/mol or 140 mg/dL could be normal after having eaten, but would be rather high first thing in the morning.

If Apple Watch lacks reasonable accuracy to be able show some kind of usable numerical value, then the only way I can see it working is by looking at trends and only reporting those. It would need to be regularly collecting data, but with that it should be able to determine someone's normal range, the height of peaks (and troughs), and how long they take to return to normal. With that you could say when someone's levels are low, normal, or high for them.

But I am not sure how much value that would be to anyone, beyond telling people when to seek medial help. You should never ask "can we do it?" without also asking "why are we doing it?"

 

[rui no onna]

The other problem with "what is high?" is that it not only varies by person but also throughout the day. For a non-diabetic being as high as 7.8 mmol/mol or 140 mg/dL could be normal after having eaten, but would be rather high first thing in the morning.

If Apple Watch lacks reasonable accuracy to be able show some kind of usable numerical value, then the only way I can see it working is by looking at trends and only reporting those. It would need to be regularly collecting data, but with that it should be able to determine someone's normal range, the height of peaks (and troughs), and how long they take to return to normal. With that you could say when someone's levels are low, normal, or high for them.

But I am not sure how much value that would be to anyone, beyond telling people when to seek medial help. You should never ask "can we do it?" without also asking "why are we doing it?"

Yep. I believe a numerical value is still more useful even if it's not super accurate. For example, there are plenty of diabetics who unaware they're having hypoglycemia. Having an alert on the watch when it drops below a certain target is a useful warning to check blood sugar level with a glucose monitor or if circumstances don't allow for it, to have a light snack.

 

[jz0309]

This comes up a lot but as a type 1 diabetic and someone who works in the field of clinical diagnostics, this is not going to provide equivalent function to an FDA regulated medical device. Apple has been keen to avoid that burden. It will be targeted to type 2 diabetics and will be a retrospective target based evaluation without precise numbers.

You can expect the initial implementation to be similar to how they currently do other stuff. e.g. today you were 80% in normal blood glucose range and 20% out of range.

Doing otherwise would require Apple to go through FDA regulatory processes as a medical device and demonstrate many requirements as well as opening themselves up to massive liability as well as causing problems for them when they do software updates as they would need to do extensive validation of the blood glucose calling functions every update.

Apple implicitly markets the apple watch as a quasi-medical device (health device) but be assured that it is not a medical device and Apple does not want it to be one.

I agree, “gimmicky” trends (low - high) are not useful, so, in order to be useful for diabetics, it has to be accurate within the guidelines of the FDA. That means, it would have to be a medical device.
The “standard” AW will never become a medical device, it would be too much burden for a consumer yearly updated device, but, Apple could release a “special edition” with CGM - at a cost. My estimate would be $2000+, which for a diabetic is something people would shell out, your “average” AW customer not.
I do believe the technology is getting there, but in order to do CGM we are probably looking 2-3 years out

 

[AndiG]

Apple could jump in, help industry to develop smaller CGM sensors and connect it to their watches. At the moment a CGM sensor reports to the iPhone and the iPhone can send a message to the watch. This breaks if you don‘t carry the iPhone with you.

But i doubt this, cause at the moment you cannot even connect the heart rate monitor of your watch to a bicycle computer. Just because Apple doesn‘t want to …

 

[KindJamz]

As a Type 2 diabetic with very good control (no meds, normal A1C), I'm allowed only 1 test strip per day by my insurance. Any more than that I have to pay for myself, and they are not cheap. I have to spend the extra money to buy them because one test/day is not enough to maintain that good control. Even qualitative measurements would be really helpful in knowing that it might be a good time to use one of those precious extra strips to see where I'm at. But that's with the caveat that "low," "normal," and "high" need to be clearly defined with some actual quantitive ranges. I need to know if what I consider "high" is what Apple considers "high."

That being said, I'd probably sit out the first watch with blood glucose monitoring and wait to see if it's even close to accurate first. I think it's probably going to be hard to buy without waiting months for delivery, because it will be such a game changer for so many people. I'll let others take it for a test drive first. I think it's years away in any case.

Ask you doc for some strips. We hand out glucometers and strips to every patient that needs one or can’t afford it

 

[KindJamz]

Ask you doc for some strips. We hand out glucometers and strips to every patient that needs one or can’t afford it

I agree, “gimmicky” trends (low - high) are not useful, so, in order to be useful for diabetics, it has to be accurate within the guidelines of the FDA. That means, it would have to be a medical device.
The “standard” AW will never become a medical device, it would be too much burden for a consumer yearly updated device, but, Apple could release a “special edition” with CGM - at a cost. My estimate would be $2000+, which for a diabetic is something people would shell out, your “average” AW customer not.
I do believe the technology is getting there, but in order to do CGM we are probably looking 2-3 years out

So your saying that apple will sell watches and not make profit ???

Try closer to 10 years

 

[aperantos]

So your saying that apple will sell watches and not make profit ???

Try closer to 10 years

In the U.K. a Dexcom G6 annual subscription is around £1,900 (US$2,600) whilst sensors for Freestyle Libre is around £1,300 (US$1,800).

I do not know how those products are priced in the U.S., but here an Apple Watch with medical grade continuous glucose monitoring for US$2,000 would definitely be an amazing bargain.

 

[KindJamz]

In the U.K. a Dexcom G6 annual subscription is around £1,900 (US$2,600) whilst sensors for Freestyle Libre is around £1,300 (US$1,800).

I do not know how those products are priced in the U.S., but here an Apple Watch with medical grade continuous glucose monitoring for US$2,000 would definitely be an amazing bargain.

Watch would be far more expensive if it was accurate. Medical grade equipment is a whole more set of rules and regulations. I work in the lab industry and this isn’t happening anytime soon.at this point it’s still a pipe dream

 

[axcess99]

Watch would be far more expensive if it was accurate. Medical grade equipment is a whole more set of rules and regulations. I work in the lab industry and this isn’t happening anytime soon.at this point it’s still a pipe dream

Don't they have some sort of FDA approval for their ECG and a-fib functionality though? That didn't turn them into a "medical device".

Why woudnt/couldnt there be a similar classification used/created for a degree of CGM. As long as they didn't advertise it as a medical CGM and a quack-Miranda warning that the functionality should never be trusted and to check with your doctor. I don't see it as a significant blocker. Apple knows any such function as a non-consumable expense would drive huge adoption from the diabetic and pre-diabetic population as well as self-health gurus, no matter how crude or caveated.

There has been plenty of political pressure in the US to turn plenty of medical devices into pseudo-regulated OTC devices. Just look how professionally fitted/calibrated hearing aids are now being undercut by glorified in-ear amplifiers.

 

[mikemj23]

Don't they have some sort of FDA approval for their ECG and a-fib functionality though? That didn't turn them into a "medical device".

Why woudnt/couldnt there be a similar classification used/created for a degree of CGM. As long as they didn't advertise it as a medical CGM and a quack-Miranda warning that the functionality should never be trusted and to check with your doctor. I don't see it as a significant blocker. Apple knows any such function as a non-consumable expense would drive huge adoption from the diabetic and pre-diabetic population as well as self-health gurus, no matter how crude or caveated.

There has been plenty of political pressure in the US to turn plenty of medical devices into pseudo-regulated OTC devices. Just look how professionally fitted/calibrated hearing aids are now being undercut by glorified in-ear amplifiers.

Not sure on "huge adoption from the diabetic" community if not FDA approved. I can only speak for myself as a T1 that it would be essentially useless as I have FDA approved, more accurate devices that I use to manage my diabetes. I do think it would be adopted to a certain degree by some who are curious about glucose trends in a postprandial sense. "Crude and caveated" (in other words, gimmicky) scares me to death, but I recognize I may be in the minority.

 

[aperantos]

Apple knows any such function as a non-consumable expense would drive huge adoption from the diabetic and pre-diabetic population as well as self-health gurus, no matter how crude or caveated.

As a diabetic a crude device would be of no use to me, it would do nothing to negate the need to test properly for insulin dosing nor the requirement for testing when driving. I am not really sure how it would help any diabetic, even with pre-diabetes if someone only wants a device to say if their diet puts them above a certain level, they would need a device to accurately recognize that level.

And how crude is crude?

The stricter F.D.A. standard (for home testing) is a 15% margin of error (95% of the time), and the international standard ISO 15197:2013 is 20% accuracy at 5.6 mmol/L and above, or ±0.83 mmol/L or 15 mg/dL below this (which is a margin 15% at 5.6 mmol/L).

For a non-diabetic person the normal fasting range, from hypo to high, is normally regarded as being 4–5.6 mmol/L or 70–100 mg/dL. So even the margins of those standards are pretty big, fortunately most glucometers far exceed those minimum requirements.

But it means if Apple managed a 20% margin of error, which is far from crude as until 2016 it was the F.D.A. standard, then a reading in the middle of that normal range could be reported as either hypo or high. That is of no use for managing even pre-diabetes, and could be damaging if people act on lower readings when they are at the higher end of the margin.

I have to agree with mikemj23 that such a thing could be scary.

 

[rui no onna]

The stricter F.D.A. standard (for home testing) is a 15% margin of error (95% of the time), and the international standard ISO 15197:2013 is 20% accuracy at 5.6 mmol/L and above, or ±0.83 mmol/L or 15 mg/dL below this (which is a margin 15% at 5.6 mmol/L).

For a non-diabetic person the normal fasting range, from hypo to high, is normally regarded as being 4–5.6 mmol/L or 70–100 mg/dL. So even the margins of those standards are pretty big, fortunately most glucometers far exceed those minimum requirements.

Most? I think you're giving glucose meters more credit than they deserve.

Are Blood Glucose Meters Accurate? New Data on 18 Meters

Results from the Diabetes Technology Society’s Blood Glucose Meter Surveillance Program identifies only six out of 18 meters that passed. Did yours make the cut?

diatribe.org

But it means if Apple managed a 20% margin of error, which is far from crude as until 2016 it was the F.D.A. standard, then a reading in the middle of that normal range could be reported as either hypo or high. That is of no use for managing even pre-diabetes, and could be damaging if people act on lower readings when they are at the higher end of the margin.

Difference there is Apple Watch numbers will theoretically be provided through continuous monitoring. How many non-insulin dependent diabetics actually test multiple times a day? I'm Type 2, non-insulin and I just test in the morning and before bed. That's it. In between, it's a major hassle to test.

I can see how results from an Apple Watch could be helpful as an early warning for severe lows and highs that would otherwise go undiagnosed. Kinda like a nudge to check/verify blood sugar levels with a blood glucose meter and then take steps to correct (or avoid in future).