Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple Monitoring Blood Glucose

mondaynote.com

Post reply on HN
Print view View original

111–120 of 142 posts

Re: Apple Monitoring Blood Glucose

#111

Post by GordonS »

swsieber wrote: post #102
JustSomeNobody wrote: post #77

Earlier quoted context omitted.

Generally speaking, if you're leading an unhealthy lifestyle, buying a new gadget will not change that. I'm sure there's plenty of anecdata to counter this, though.

If I had a device that let me know the health impact (from an insulin sensitivity standpoint) of a food I ate without invasive stuff, I'd use it in a heartbeat, and I'd listen (except for cheese - if it said cheese would bad, I'd ignore that bit).

The good news is that almost all cheeses have not effect on blood glucose levels!

Re: Apple Monitoring Blood Glucose

#112

Post by mahyarm »

dominotw wrote: post #73
cs2818 wrote: post #64

Earlier quoted context omitted.

I don't have any data on hand to back this up, but I don't think younger people are necessarily averse to monitoring their own health. I think personal health monitoring devices fit well with the devices the younger generation already uses to help maintain their identity and concept of self. I agree if it's poorly implemented or marketed it won't be adopted, but Apple probably has a shot at getting it right.

I agree. I've only seen middle aged office workers wear stuff like Fitbit, posture correction, sleep tracking type of things. Young people don't have to worry about 'sitting is new smoking' .

There is a nerdy segment that loves tracking all of this health data forever and ever.

There is also the segment that just wants to lose weight, and that can apply at all ages. A continuous glucose monitor is also a great way to monitor keto diets or similar. Keep yourself under a certain glucose level continuously and your body will be forced to use fat sources!

Re: Apple Monitoring Blood Glucose

#113

Post by quickben »

wapz wrote: post #40
dingo_bat wrote: post #33

Earlier quoted context omitted.

Isn't it a bit scary to run the auto-insulin code on an android phone?

I do not have an insulin pump and I'm presuming you don't with my response. The pumps I have seen make noise when they are pumping and also have an alarm (I'm not sure if that's from the pump or the reader) if your levels are on either extreme. I'm guessing malfunctioning code would only execute once by pumping too much or not pumping at all and the user would figure it out very quickly.

I'd argue the users can't identify anything if they are rendered unconscious:

https://en.wikipedia.org/wiki/Diabetic_coma

You definitely don't want anything buggy with medical devices.

Re: Apple Monitoring Blood Glucose

#114

Post by Balgair »

charlesism wrote: post #108
korethr wrote: post #107

This article talks about type 2 diabetes, but I'm wondering why a non-invasive continuous blood glucose monitor would not be just as useful and desirable for type 1s as well. A friend of mine has type 1, and more than once, she's expressed hopeful anticipation for a future combined non-invasive blood glucose monitor and insulin pump -- effectively an external artificial pancreas to take over the functionality of her…

It would be very desirable, but all of the many companies who have tried to get accurate results non-invasively have failed. With Apple's resources, perhaps they'll lick the problem, but I wouldn't bet on it. So to answer your question: Probably because if a Type 1 Diabetic uses a non-invasive glucose monitor to determine their insulin dosage, much of the time they will overdose, and possibly fall into a coma or die.

A co-worker of mine is doing a similar kinda thing but with circulating tumor cells, so I have some 3rd hand knowledge of these kinda tests.

Fact of the matter is that the human body is, chemically and pathologically speaking, fucking bulletproof. Our bodies are just super freaking good at maintaining homeostasis. This means that a lot of tests like blood-glucose, cancer diagnosis, etc that are trying for non-invasive blood testing just have a HUGE parameter space to contend with. Apple's idea may work with calm light-skinned office workers, but will it work with sweating dark-skinned day laborers? African Americans are the population most at risk for DM-II. How about if your blood pressure goes all wonky? What about someone going through a growth spurt? Does it work when you put your arm in motor oil? What about people without arms? Lives aren't 'edge cases'; your device MUST cover all those real people with families and children. I'll believe their device when it passes FDA approval and the clinical data is out there for us all to see and argue over.

Re: Apple Monitoring Blood Glucose

#115

Post by npunt »

aedron wrote: post #59

Apple's recent obsession with health monitoring is weird from a branding standpoint. Losing appeal with young people is the nightmare of any technology company, becoming known as that gadget for geriatrics is another level of dejection. I read that this focus is particularly driven by Tim Cook's fascination with the space. It sort of confirms my idea of him as not a very image conscious CEO, especially in contrast wi…

Athletes are the branding antidote to these concerns - they're extreme users in their own right and love to track all the things. One step down from the halo-effect of athletes is the 'Athleisure' segment of the fashion market (think Lululemon and the like), which shows broad interest in healthy living in desirable/cool demographics that have high correlation with Apple purchasers.

On the contrary, going into health will reinforce the brand, not dilute it. Apple has always been an aspirational lifestyle brand. Given the growing scope of computing in our lives - work, social, leisure, health - the lifestyle that Apple sells is only going to get more complete and compelling.

There's also a growing trend of QS and tracking health data points, and increasingly not just among geeks. You can see it in the nutrition space as well, where people are optimizing food, getting bloodwork, etc. These fields are ascendant and glucose tracking matters there too.

Re: Apple Monitoring Blood Glucose

#116

Post by sctb »

nabla9 wrote: post #42

> I'm excited to see where it goes. BREAKING NEWS FROM THE FUTURE (Reuters): Hundreds of thousands Apple users die within hours after IoT glucose measurement + insulin pump device overdoses insulin during WannaDie cyberattack. Apple urges all surviving users to update the system immediately. Surviving users and family members of deceased will get 10 year full Apple subscription and special $20,000 memorial gift card…

We detached this subthread from https://news.ycombinator.com/item?id=14391166 and marked it off-topic.

Re: Apple Monitoring Blood Glucose

#117

Post by harigov »

weego wrote: post #17
harigov wrote: post #4

Earlier quoted context omitted.

I think machine learning coupled with increase in sensitivity of the signals could make a difference.

Yeah let's get some AI in there to help out, maybe building the front end in react native will solve some problems no one though of. We could get the data with graphql to solve some issues.

Except, it actually does. ML is just advanced statistical modeling which performs complex computation over large amounts of data to give out better results. Because Apple can afford to do things at scale and use their marketing prowess to buy sensors at an economical price, they can afford to use multiple sensors to give out a reasonably accurate metric. All the technologies that you mentioned do solve problems they are intended to solve. Why so much negativity?

Re: Apple Monitoring Blood Glucose

#118

Post by harigov »

sillysaurus3 wrote: post #10
harigov wrote: post #4

Earlier quoted context omitted.

I think machine learning coupled with increase in sensitivity of the signals could make a difference.

Why would machine learning be able to overcome a lack of statistical power?

Did you miss the part about sensitive sensors? A single sensor may not be enough to classify accurately but multiple sensors measuring different things can sometimes work. Such sensor fusion has worked incredibly well for IMU's and also in self-driving cars.

Re: Apple Monitoring Blood Glucose

#119

Post by harigov »

rubidium wrote: post #74
harigov wrote: post #4

Earlier quoted context omitted.

I think machine learning coupled with increase in sensitivity of the signals could make a difference.

hahaha. Siri can't even parse "What time does the Indianapolis 500 start?" It gives me movie start times.

They are totally different things. Sure, Siri can't parse that and yet we have working* self-driving cars and autopilots for planes.

Re: Apple Monitoring Blood Glucose

#120

Post by shas3 »

aedron wrote: post #59

Apple's recent obsession with health monitoring is weird from a branding standpoint. Losing appeal with young people is the nightmare of any technology company, becoming known as that gadget for geriatrics is another level of dejection. I read that this focus is particularly driven by Tim Cook's fascination with the space. It sort of confirms my idea of him as not a very image conscious CEO, especially in contrast wi…

I am deeply skeptical that anything other than R&D and patents will come out of this move by Apple. I am bearish on the prospects of using technology in healthcare. Primarily for these reasons:

1. Regulation: it is very hard to innovate at the pace of the consumer-electronics industry while satisfying FDA regulations.

2. For general purpose applications, signals and measurements are highly noisy and sensors are still deeply in the academic realm, working only in very special use-cases, as opposed to 'in the wild'. It is a different matter for special applications such as Type-I diabetes or other diseases because devices are critically important and patients are willing to tolerate discomfort (of wearing probes, pumps, etc.)

3. In general, predictive power of algorithms tends to be lower for non-genomic health care applications because of variability across individuals, errors in installing devices, rarity of certain types of data for training and validation, etc.

4. Some of the key benefits are hard to grasp for users because of the slow time scales over which things like weight loss occur, and also due to the multi-factorial/multi-component nature of human well-being.

Post reply on HN