Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> "Computerized health monitors built into watches, jewelry, and clothing which diagnose both acute and chronic health conditions are widely used. In addition to diagnosis, these monitors provide a range of remedial recommendations and interventions."

> MOSTLY RIGHT

I'll quibble with this one and call it mostly wrong. Yes, a very narrow and specific set of possible cardiac abnormalities can be detected (with significant error rates) by the Apple Watch's optical (PPG) and electric (ECG) heart rate sensors. Other vendors have functionally equivalent capabilities.

But medically speaking, smartwatches are probably decades away from independently diagnosing anything, let along recommending an intervention (i.e. a drug or surgery). Kurzweil's terms have specific meanings in medicine and current technology is nowhere close to that.

What exists today (or, in 2019) is a very narrow use case and far from the broad and general scenario Kurzweil envisioned. And there is no realistic roadmap to achieve any more than incremental development in this area - PPG and ECG were "low-hanging fruit" in some sense, and medically actionable wearable sensors for pretty much any other medical signal present orders of magnitude larger engineering challenges that few people are even seriously working on. Hence, "mostly wrong."



I agree with your assessment of MOSTLY WRONG but I think that you are overly pessimistic about the timelines:

> But medically speaking, smartwatches are probably decades away from independently diagnosing anything, let along recommending an intervention

Continuous PPG in its current form can diagnose acute and chronic changes in our health condition but the vendors are reluctant to do so. As examples, the newer PPG devices with SpO2 can detect acute and chronic changes in oxygen saturation (critical with COVID-19), Heart Rate Variability (HRV) can indicate infection associated stress, HRV should be used as a proxy for breathing rate, and the Machine Learning derived sleep cycles can indicate sleep disorders.

The data is useful now but the off-device analysis necessary seems to missing in action.


What about the blood sugar and insulin devices now available for diabetics?


Those aren't integrated in watches, jewelry or clothing, as Kurzweil suggested. In fact the whole trend around wearable devices never really seemed to reach a sophistication necessary for the mass market, just like 3D printed clothing. So far we've got Google Glass which disappeared as quickly as it arrived, smartwatches that were mostly a question of time, and some fitness products like those running shoes with Bluetooth. But the actually challenging, futuristic part is still more a topic for blog articles in the maker community than a serious product category.

Even with those insulin devices the biggest innovations were made by some of its users themselves as far as I'm aware(see "looping").


That’s a marvel which is ubiquitous in developed counties. However, there’s been too much hype around the wearables. Manufactures have foreseen that progress in smartphones will inevitably stall and promised wearables as the next big thing. And, as it turns out, “smart watch” is still dumb as hell, at least when it comes to health.


It's definitely wrong.

> But medically speaking, smartwatches are probably decades away from independently diagnosing anything

But it's because we won't do it. There's nothing technical or hard to make this right.

I don't believe it's hard to pick up a heart attack using a wearable. But I don't think many heart attacks are picked up with wearables.

Record everything, if something is different see a doctor should both meet this requirement and be highly beneficial.

A smartwatch will just communicate. That could be to an online database or a person.

If the intent is around privacy and no doctor or external thing is ever consulted and not recommended to be consulted. Then, yes it's decades away.


> I don't believe it's hard to pick up a heart attack using a wearable. But I don't think many heart attacks are picked up with wearables.

> Record everything, if something is different see a doctor should both meet this requirement and be highly beneficial.

Cue massive amounts of false alarms.


There's a lot of technical problems to doing this. We're still not great at getting heart rate or blood oxygen. Sure, you get a decent reading but that's with lots of averaging and the temporal resolution is decently low. Just because it's hard to maintain a good monitoring position when active let alone adjusting for things like skin tone/optical transparency.


Totally agree, the limitations here aren’t technical, they are regulatory. The barrier to entry for medical devices (FDA approval) is so incredibly high it greatly slows down development on health related devices. Apple has an entire team dedicated to regulatory stuff for the one tiny health feature they offer on the Apple Watch.


It may slow it down, but for good reason: it's outright dangerous to sell someone a device you claim to help monitor their health that does no such thing. Even worse, selling people's private health details (as sketchy as they may be) to whoever wants, or just making them public for hackers, is a much bigger detriment than any tech on the horizon can bring as a positive.


ITT: non-medical people saying something medical is easy.


There are large interpersonal physiological differences due to so many variables that it makes getting high accuracy with machine learning models hard. Only way I see this accuracy improving is by calibrating those models with good reference data for individuals.




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

Search: