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

How on-the-wrist sleep apnea detection works

empirical.health

Post reply on HN
Print view View original

21–30 of 170 posts

Re: How on-the-wrist sleep apnea detection works

#21

Post by carlmr »

tootie wrote: post #6

I wouldn't put a ton of stock into this, but it could still be useful. The WatchPAT device is purpose-built for this test and includes a full fingertip oximeter and while it's FDA-approved, it's still not very accurate. It's also a notoriously tricky condition to diagnose and can't really be done just by looking at numbers. I think there's value in the test being accessible to so many more people, it will likely have…

>can't really be done just by looking at numbers

Depending on the numbers, why not?

If it's sensor accuracy a single night might not be enough, but if it detects the condition averaging over a few weeks even unreliable measurements can be made more reliable (given they're just noisy).

Re: How on-the-wrist sleep apnea detection works

#22

Post by runjake »

I'm not an qualified expert, but I spend a lot of time reading about and using this stuff and I'm extremely skeptical sleep apnea detection is very accurate without the pulse oximeter functionality.

But gosh, I hope this is accurate across the Watch's wide user base.

I wouldn't say it's likely, but I do wonder if the pulse oximeter is being used surreptitiously. In any case, if you could use the pulse oximeter, this functionality would immediately become much more powerful.

Re: How on-the-wrist sleep apnea detection works

#23

Post by jrockway »

lysace wrote: post #8

The gatekeeping in the CPAP field is so insane. 1. Yes, we should regulate manufacturers intensely. 2. No, we shouldn't regulate to whom companies should be allowed to sell them. The defaults are good enough for most. Provide an easy to use control for max pressure and people will figure it out. Perhaps a knob? I claim: During the past decade, the risk of being maltreated by a personally set-up CPAP unit has been far…

I'm just happy that the "service mode" screens are easy to get to. Changing the ramp-up from 10 minutes to 45 minutes made a big difference for me.

(I'm honestly surprised someone doesn't make the vendors require a 1000 character passphrase that locks you out of the device after 2 failed attempts.)

Re: How on-the-wrist sleep apnea detection works

#24

Post by smelendez »

GaggiX wrote: post #18

The new Apple Watch 10 has an 18-hour battery life, so if you use it while you sleep, you won't have much left. I have no idea how some people can charge their smartwatch 1 or more times a day.

People probably use it principally as a fitness (and sleep?) tracker, and charge it while they work

Re: How on-the-wrist sleep apnea detection works

#25

Post by a_vanderbilt »

GaggiX wrote: post #18

The new Apple Watch 10 has an 18-hour battery life, so if you use it while you sleep, you won't have much left. I have no idea how some people can charge their smartwatch 1 or more times a day.

I find that integrating it into my routine made my current watch (SE1) battery a non-issue. I wear it starting in the morning, charge it on the way to work, and then to bed. Repeat ad infinitum. Battery usage at night is way lower in my experience, and it usually survives til morning or close to it.

Re: How on-the-wrist sleep apnea detection works

#26

Post by dwighttk »

cromka wrote: post #13
brandonb wrote: post #3

Earlier quoted context omitted.

I suspect they also need oxygen saturation (introduced in the Apple Watch Series 6 in 2020) in order to make their overall sleep apnea detection algorithm work. Given that it only works with the Series 9 or above, it's possible they're also using the the temperature sensor as another input into their algorithm.

They explicitly only mentioned accelerometer. If they used more sensors, in particular temperature sensor, it's hard to imagine they wouldn't mention in, especially that it would justify them limiting the feature to Series 9 and 10. However, reading on PAT (Peripheral arterial tone) and existing FDA-approved devices using that, they indeed use oxygen but not temperature. See for example "WatchPAT ONE" device.

Perhaps they aren’t worried about justification

Re: How on-the-wrist sleep apnea detection works

#27

Post by swatcoder »

abtinf wrote: post #10

I’ve always wished I could put Apple Watch into a “constant oxygen monitor mode” for sleep. That seems to be the obvious way to assess sleep apnea. Anyone know why that isn’t the method?

There's a ton of cleverness involved in keeping an extremely compact device with radios and sensors running for tens of hours on a tiny battery while making it seem like it's always awake and ready to go when you need it to be.

But the reality is that everything that needs any kind of power is being aggressively idled down and time sliced and being made sample-based as can be gotten away with. You can easily get a sense of this with heart rate monitoring, if you dig into Apple Health and observe the pattern of samples it takes.

Almost certainly, running the sensor at such a high frequency that you could call it "constant" would drain the battery long before you woke up.

Re: How on-the-wrist sleep apnea detection works

#29

Post by brandonb »

jerlam wrote: post #9
brandonb wrote: post #3

Earlier quoted context omitted.

I suspect they also need oxygen saturation (introduced in the Apple Watch Series 6 in 2020) in order to make their overall sleep apnea detection algorithm work. Given that it only works with the Series 9 or above, it's possible they're also using the the temperature sensor as another input into their algorithm.

Did Apple resolve the patent issue that disabled oxygen saturation in the Series 9? Although it's possible that Sleep Apnea detection uses the oxygen saturation sensor but doesn't show its results directly.

It wouldn't surprise me if they did the latter. The oxygen saturation sensor is fundamentally just another optical sensor with a different wavelength of light -- so if you feed all three wavelengths, plus the accelerometer, into a neural network, you may be able to pick up on oxygen-related physiological signals without explicitly processing an oxygen timeseries.

Re: How on-the-wrist sleep apnea detection works

#30

Post by runjake »

abtinf wrote: post #10

I’ve always wished I could put Apple Watch into a “constant oxygen monitor mode” for sleep. That seems to be the obvious way to assess sleep apnea. Anyone know why that isn’t the method?

1. Because of the lawsuit around the pulse oximeter intellectual property, as other commenters mentioned.

2. That would require more battery life than the Apple Watch offers.

There are third-party options that do constant monitoring that either directly or indirectly support Apple Health. You'll need some additional intelligence, in the form of an app or something, to analyze that data. WatchOS won't do it at present.

Post reply on HN