Smart watches and sleep apnea ?

VVV wrote: ↑

Thu Jan 04, 2024 7:46 pm

That may not have answered your questions, but I am so excited about the results my Apple Watch has helped me achieve, I wanted to express this.

Thank you. This was incredibly helpful. I will pass this info to my coworker. I have an apple watch given to me by my kids but its a series 3 and has none of the extras she was asking about.

VVV wrote: ↑

Thu Jan 04, 2024 7:46 pm

babydinosnoreless wrote: ↑

Mon Jan 01, 2024 5:03 pm

Does anyone use these smart watches for the blood oxygen and ecg and are they worth the cost ?

7. ECG. My heart is healthy, so I am not a good judge of this function. I did print out a couple of ECGs to show to my PCP at an annual visit. I told him there was a concern that my ECG graph was not perfect. He said everyone's heartbeat has some irregularities and mine were of no concern. (He was impressed with the quality of the report.)

I've not had much experience with the ECG function on the Apple watch, but I wanted to say that my cardiologist is very impressed with the quality and stability of my Kardia device. The basic 3 lead was just 60.00, and the 6 lead was 90 something. I have a couple of minor issues with arrythmias and a horrible family history. It was a good investment.

Janknitz wrote: ↑

Thu Jan 04, 2024 5:44 pm

But, in my experience the random readings are not always accurate by a long shot. For example, my AW 6 (the same model used in the studies cited) recorded O2 sats down to 84% during sleep

IMO, there is something that we're not accounting for and that's the possibility of artifact readings.
I could be mistaken, but most of the clinical studies I perused seem to rely on an average number of readings before saying "the current O2 level is..."
Some researchers have used the average of 10 or 15 individual readings before concluding that the "current" O2 level is X.
(what we don't know is if those individual readings are at 1/sec or 4/sec), but what we can surmise is that some span of time passes over the course.

I am in possession of several finger-tip oximeters, one that is a "consumer" (not a watch-type) type and one that is a "medical-grade". Both display and record at one-second intervals.
If I examine the one second readings ( right thumb and right middle-finger) the instantaneous readings can vary by 2 percentage points over a ten second period reading for each meter, but if I average the 1second readings on each device and then compare the two meters, they are within 0.3 percentage points.*


* not an outragous difference when using two different "finger sites", simultaneously.

IMO, there is something that we're not accounting for and that's the possibility of artifact readings.

I don't think that the AW can give an artifact reading. If for any reason during the 15 seconds something is not right, the entire report of blood oxygen level fails, and it's not recorded.

I believe the way it works is that it measures for 15 seconds and must have an algorithm that produces an average of the readings within the 15 seconds. If you aren't positioned correctly or move too much, it will stop the reading altogether and indicate an error. Other times it will measure for the entire 15 seconds, but then the following message appears: "Unsuccessful Measurement: Blood oxygen could not be measured. This can be caused by your band being too lose or too tight, your watch being too low on your wrist, or other factors." It does not record unless the measurement was successful. So it's unlikely to record artifacts, while a finger pulse oximeter, measuring and recording every second, can record an artifact.

Janknitz wrote: ↑

Fri Jan 05, 2024 8:01 pm

IMO, there is something that we're not accounting for and that's the possibility of artifact readings.

I don't think that the AW can give an artifact reading. If for any reason during the 15 seconds something is not right, the entire report of blood oxygen level fails, and it's not recorded.

I believe the way it works is that it measures for 15 seconds and must have an algorithm that produces an average of the readings within the 15 seconds.

IIMO the averaging (which we are assuming) is the most plausible. Just so we are clear, when I mentioned artifact I was referring to a sequence that might encounter one or two values that were 4 or 5 units lower or higher. Averaging would accommodate those artifacts. On the other hand a series of 15 values, that includes one or two zero or near zero readings, should be considered a error.. Just guessing, but that would make sense to handle it this way.

That said, it's not surprising in the least that that a wrist measure might be different than finger value. The research reports indicate that there is a difference between fingers and a difference between dominate hand and non-dominate hand. One interesting observation is that the reason that there is a difference between thumb and finger (of the same hand) is because there is much more capillary activity available in the thumb than the index finger (because the thumb is "fleshier"). Following that reasoning, the wrist (fleshier compared with the finger) might be a better location to observe color changes in the blood.

See:
Comparison of SpO2 values from different fingers....

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4627972/