> it sounds like the real news is just how difficult this problem is - even for Apple
I've worked in this space for 5 years now, and this is not news to any of us. Today, a tiny, noninvasive device, that is accurate enough to make insulin dosing decisions, is pure science fiction. It's probably not impossible, but we'll literally be able to use stem cells to regrow human pancreases and straight-up cure people of diabetes before this tech becomes a reality. (We're much further along that path.)
What is interesting about this though is the possibility of using this device for non-diabetics who don't need extremely accurate glucose readings. A rough approximation of someone's blood glucose levels is enough to learn how your body reacts to different types of foods and exercise, and would be helpful for amateur athletes who can't afford to purchase real continuous glucose monitors, as well as pre-diabetics for whom CGMs aren't covered by insurance.
I'm fascinated by the fact that this seems incredibly hard, yet not so impossible that Apple is trying.
Is there an ELI5 of the basic principle of how it would work (or at least the direction being pursued), and what the main obstacle is, and why we think we'll be able to overcome it to some degree?
Besides athletes, I can't help but think it could be a huge step forwards for weight loss as well. I'm aware of the concept (for non-diabetics) of only eating once your blood sugar is below a certain level, and never eating so much that it goes over a certain higher level.
Replying to my own comment: I found a 2021 paper that goes over all the different technologies being considered, and the wide variety is frankly staggering. Table 10 lists ten different principles under consideration (fluorescence, spectroscopy, etc.) with their various pros/cons. Not ELI5, but as close as you might get:
Seems like spectroscopy seems particularly promising, and according to the article, that's the approach Apple is taking. (Makes sense for a watch.)
And from the paper, the challenges with spectroscopy are signal interference -- scattering, water absorption, and distinguishing between glucose and other sugar components. So it ultimately sounds like a fiendishly complex signal processing problem. Which is exactly the kind of problem that is very very hard yet maybe not impossible.
The IMHO gold standard book to get an understanding of the history of non-invasive glucose monitoring and the different technological approaches is "The Pursuit of Noninvasive Glucose: Hunting the Deceitful Turkey" by John Smith: https://www.researchgate.net/profile/John-Smith-241/publicat...
We - at DiaMonTech - working on a solution with mid-infrared spectroscopy and showed good accuracy in a published paper in the "Journal of Diabetes Science and Technology". We plan another clinical trial this year, but yes, this took some time. We started in 2015 and still have some work to do.
Important nuance here is the difference b/w the parent poster's "accurate enough to make insulin dosing decisions" criteria (extremely difficult+risky)
vs
"useful enough signal for general population / those at risk of type 2 diabetes / those with diabetes who don't require insulin treatment / etc" (much larger market for Apple w/ greatly reduced requirements)
^^ This - the ECG & Blood Oxygen levels are "good enough" for the public to consume and track, but perhaps not quite at the level of monitors used in medical settings to make care decisions.
Yeah I don't know about ECG but I checked my (non apple) watch with a real finger-clip SpO2 from the doctor and while it's pretty accurate under highly controlled situations (sitting still and positioned perfectly) it's really lousy under real-world conditions, moving around etc.
And I didn't just check the bog standard 98% measurement but also tried to hold my breath etc.
I can't speak for the Apple Watch, but the O2/BPM readings on some other watches are often dropped/delayed with the operating system taking priority. I was trying to use the heart rate data for a pet project and found its practically useless, on Samsung's watches its just enough data to show a few fancy numbers and thats about it.
On the contrary, the Apple Watch ECG is quite accurate, not quite enough to replace the system the doctor makes me wear, but enough that apple can get a general idea of any defects and tell you to go to the doctor. The "regular" hear rate sensor isn't anywhere near good enough to be taken seriously, but its leaps ahead my previous samsung.
I dream of the day where my Apple Watch can replace the myriad of annoying wires and sensors I have to wear to monitor myself. Even then, the Apple Watch has greatly helped me out when I've fallen and need assistance, and I've gifted several to my family just in case.
ECG is a relatively easy thing to measure, so that single axis they measure is quite accurate. But the conclusions one can make from that is limited (you basically take a video from a single point). Doctors use 12 axes instead to get a real-time, 3D vector of the heart.
To be honest, “real finger-clip” sensors are similarly fuzzy in their results - after all they work the exact same way. Maybe using percentiles is not ideal as it gives a sense of more accuracy than what makes sense.
They don't work quite the same way. The finger clips shine through the entire finger. The illuminator is on one side and the sensor on the other. The wrist-based sensors rely on reflections through the skin which is a lot more iffy.
To be fair, the Verily folks did design the algorithm that predicts where your blood sugar will be in X minutes in Dexcom CGMs, which lives on. Dexcom now owns the patent.
> nothing beats an actual blood sample
Technically, interstitial fluid can get you pretty dang close, which is considered minimally invasive.
to say google discontinued a project after a short period of time does not exactly speak to the level of difficulty or if that difficulty was even the reasoning for the discontinuing. someone at G could have just gotten bored or moved to a different department, or a new darling employee stole focus, or whatever other reasonings we've heard when G projects are shuttered.
Nothing beats a blood sample, but interstitial fluid (which is invasive kind of) is good enough for insulin dosing. Most CGMs (Dexcom, Freestyle, etc) work this way. They don't sample blood, they sample the fluid between cells.
The latest generation products have an MARD of around 8%, which is good enough to make insulin dosing decisions.
They don't currently send data directly to Apple watches (it goes through the iPhone first), but that's largely a regulatory approval issue, not a technical one.
> I've worked in this space for 5 years now, and this is not news to any of us.
The reason it's your comment I'm replying to is because you indicate you actually worked in this field.
Whenever these discussions surrounding noninvasive monitoring of blood solute concentrations come up I keep wondering why anyone is surprised taking a hard route is hard, and why there seems little attention to seeking easier routes of noninvasive optical sensing.
Please do shoot my comment if it is nonsensical, so that I can stop caring next time I read such a discussion, but I would like a mechanistic explanation why the lips wouldn't be easier.
From an a priori perspective, it would certainly be desirable to have any type of noninvasive monitoring in the form factor of a watch. There is nothing wrong with dreams or hope. Yet to me it sounds like a dream was interpreted too literally: starting from some hypothetical dreamer exclaiming "imagine testing concentration X or Y, but noninvasively! sure wearing a device would may come with some inconveniences, but if the advantages outweigh the inconveniences, it would find a niche" and "after all people have been wearing watches too".
An idea that started as an analogy: just like a personal wearable watch had advantages compared to an immobile pendulum clock, perhaps we can devise the analogous personal wearable device that performs the duties of some large inconvenient measuring setup, workflow, discipline,....
I can't shake the feeling that the analogy has been taken too literally, as if the original dream insists the blood glucose meter has to be integrated into the watch and located on the same or similar place on the body...
Blood is red.
Lips are red.
The color of the lips is the color of the blood.
Acknowledging the inconvenience of mounting a device or perhaps an optical fiber to your lips, it may constitute a more tractable milestone.
What is the reason all the 'failed' startups avoid the lips and make life hard on themselves by choosing to do spectroscopy (in the widest sense: plain, Raman, ...) through the highly variable noisy medium of skin (different pigmentations by genetics, different pigmentations by varying exposure to sun within an identical individual, different states of skin like sweating, the intinsic noise incurred by imaging through a robustly thick layer of skin, scattering like mist, ...).
It almost looks like attempting an artificially troubling tour de fource, akin to painting a masterpiece... with your toes. I think its possible eventually, fundamentally speaking, but it doesn't sound like a wise choice to make, at least not initially.
(I wrote failed in quotes, because in reality after bankruptcy IP is transferred, experience was gained by employees, some of which end up at different companies in the same space, others end up in new disciplines where they gain a new applicable insight that nobody previously connected to blood glucose monitoring, etc...)
Then there's the social issue of the first iterations of the tools potentially working better for less pigmented skin...
> It's probably not impossible, but we'll literally be able to use stem cells to regrow human pancreases ...
I have some serious question marks surrounding the near-term viability of stem cell regrown organs.
As an unethical thought experiment consider freshly born twins, and removing their pancreas and putting each in a chamber where you would otherwise grow the artificial pancreas. How fast would you notice harmful levels of sugar purely from observation of the kept-alive pancreas? (without cheating and using knowledge of healthy blood sugar levels). If it would take many many experiments or very long durations to detect these problems (for glucose or otherwise), then one doesn't really have the engineering confidence of how long it would last, what the side effects are, ... i.e. blanket statements that stem cell derived organs will arrive earlier than affordable democratized noninvasive sensing could be true, could be false, but isn't proven by a blanket statement. The burden of proof doesn't lay with the reader.
> ... and straight-up cure people of diabetes before this tech becomes a reality. (We're much further along that path.)
Which type of diabetes?
Type 1:
The root of the problem is an unsuitable autoimmune response, during V(D)J the immune system rolls the dice to construct a generator of randomly chosen but from then on specific antibody. Due to natural selection library of segments for V(D)J recombination has a strong bias for resulting in antibodies that match foreign pathogens and potentially endogenous cancers while simultaneously also having a strong bias for not matching normal endogenous or symbiotic features. So usually nothing goes wrong.
We all carry a different but similar library of segments for V(D)J, so our immune systems display slightly different statistics for expressing antibodies.
Imagine an individual has the flu and its immune system ramps up the production rate of new generators of antibodies, each generator corresponding with a randomly chosen but from then on fixed type of antibody, usually safe but not guaranteed. Perhaps this individual has bad luck and besides generating suitable antibodes for the flu, it has also generated a low probability antibody that happens to attack the pancreas. Over time the pancreas starts to degrade and fail. The person now has type 1 diabetes. The pancreas damage is a symptom, not the cause.
Even if you grow a new pancreas, and replace the damaged one with the freshly grown one, the immune system cells generating the regrettably bad antibodies are not filtered out of the body, so after a while the type 1 diabetes will recur! Your proposal does not straight-up cure diabetes type 1
(how hard or easy does it sound to locate and filter the specific immune system cells that result in the autoimmune disease without also removing those that represent your acquired immunity of diseases the individual was exposed to throughout life? if such a thing were easy, curing HIV would be easy by simple filtration of every last HIV genome from a human body, so it does not sound that easy at all)
The reason it needs to be a watch (I have also done several years of research in this field, by the way) is that devices only provide a benefit if you use them.
There are decades of research with all kinds of wild wearable devices in all kinds of form factors that can do amazing things. They have gone over like a lead balloon with the consumer.
Socially, it is acceptable to wear exactly one piece of technology on one wrist. People with very dangerous medical conditions might be convinced to wear a well hidden device under their clothes but it's a big ask and compliance won't be as high as you'd expect.
That's just how people are, so as engineers we have to make products to fit that.
Imagine selecting diabetes patients at random, and asking them one by one:
"You are currently already by necessity accepting the monitoring of blood glucose invasively with device X, here is a less invasive device Y that sits under your lip, or perhaps has a fiberoptic cable running up to your lip. Would you switch to this lip-oriented device today? Or would you rather profusely refuse to use this device Y until some more distant unspecified future date where the technology Z finally robustly works through thicker layers of skin while continuing to use the more invasive device X?"
I don't claim but tend to believe they would take improvement Y over X while waiting for Z.
You actually claim they wouldn't.
I say let the patients speak for themselves.
Perhaps your claim is correct and no suitably large niche of patients would switch to the lip-oriented device.
Perhaps you are wrong and you have confirmed my suspicion of the miscommunication between patients and engineers.
Your rebuttal does not prove compliance of a less invasive lip oriented device would be lower than current invasive devices.
Diabetes falls under dangerous medical conditions, and they have a variety of wearable devices now for continuous glucose monitoring that are popular. They're invasive though, can allow other people to learn about your medical history just by looking at you (depending on what you're wearing/which device you have), can complicate intimacy and contact sports, and in general the patients would be much happier if a smart watch could do it without specialized hardware.
There is a lot (a lot) of history here that I think would help you test your assumptions on this.
> There is a lot (a lot) of history here that I think would help you test your assumptions on this.
A 2017 survey found only 32% of the US population wears a watch daily, so for a large fraction wearing a glucose monitor in the form of a watch would also constitute a change, and correlate with diabetes patients. Under the lip or monitoring the inside of the cheeks, etc may be even less conspicuous than wearing a watch. Perhaps you might wish to test your assumptions too, and again, neither me nor you should be the arbiters in this matter: the patients ought to be the ultimate arbiters.
I don't understand why a person would contort themselves into less defensible or self-contradictory stances. It would seem engineers or others involved in the field would welcome the suggestion that perhaps there is a more tractable way, that might simultaneously be more acceptable by the patients themselves.
Observe I don't accuse you of causing the miscommunication. Everyone is prone to following the herd, including me.
While it isn't proven the patients would prefer a miniaturized device under the lips (or somehow reversibly attached to a teeth in a way that it doesn't interfere with eating, kissing, etc..) let us suppose for a moment that it actually turned out to be the case, after a survey or after actual trials.
In this case it's not the fault of the average employee or engineer in the field that they were designing thick-skin CGM devices before the survey or trials were done.
The real question in that case is: who was or wasn't aware before the survey that the red mucous membranes were easier? If not why not? It's obvious these tissues are redder due to better visibility of blood. Those that were aware before the eventual survey: did they yes or no tell their superiors? why not? if they did, they are absolved, but the line of questioning now transfers to the superior. Why didn't they order a survey asking for patients concerns about the idea, and if they would consider switching to such a product if properly executed? Why didn't they consult the patients?
> prove compliance of a less invasive lip oriented device would be lower than current invasive devices
I can't prove it, but I think it actually makes a lot of sense. The interstitial CGMs like Libre/Dexcom are invasive (brief pinch on application) but they last up to 2 weeks and you can set and forget. I'd think any other non-invasive device that requires carrying it around, running wires, and being even a little pro-active would lose on convenience alone.
You could get some traction I'm sure, but the majority of people are going to gravitate towards the easiest solution, and applying a patch once every 10-14 days is pretty darn easy. Even if Apple could check glucose I'd still prefer a CGM patch so I don't have worry about always wearing the watch.
That's the thing, people will wear an accessory, but different people will wear different accessories and will absolutely refuse to wear others.
It makes it very hard for a business when, for instance, very few men are willing to wear your bangle or people can't wear your glucose monitoring hat inside nice restaurants.
It's likely that different wearable forms will become more common and accepted, but that is one of those predictions you hear every year for the last thirty years and it's true to an extent but also there is a long way to go.
Doesn't seem like a big problem since enthusiasts and early adopters are precisely the type of folks who would most likely break established norms in using the product.
Maybe one day it will be common for men to wear bangles on each arm, or for hat wearing to be common indoors via their actions.
For example, no one carried around headphones out-and-about before the Walkman.
I've worked in this space for 5 years now, and this is not news to any of us. Today, a tiny, noninvasive device, that is accurate enough to make insulin dosing decisions, is pure science fiction. It's probably not impossible, but we'll literally be able to use stem cells to regrow human pancreases and straight-up cure people of diabetes before this tech becomes a reality. (We're much further along that path.)
What is interesting about this though is the possibility of using this device for non-diabetics who don't need extremely accurate glucose readings. A rough approximation of someone's blood glucose levels is enough to learn how your body reacts to different types of foods and exercise, and would be helpful for amateur athletes who can't afford to purchase real continuous glucose monitors, as well as pre-diabetics for whom CGMs aren't covered by insurance.