Unreliable data during “compliance” period?

[dataq1]

You think you can extrapolate like that?

No extrapolation involved.
Wakemenot was quite explicit in his hypothetical.

"all 20 {events} occurred in the one hour I was actually sleeping"

Whatever was happening while he was still awake (reading or whatever) can't (or should not) be considered valid events for the purposes of AHI calculation.

The problem is that the device ASSUMES he is sleeping, so divides by 4 instead of 1 to calculate the AHI in his hypothetical.

He is correct, and I agree. Please explain the basis of your disagreement.

[ozij]

For example: If I am wearing mask for four hours but only asleep for one; does this mean my 5 AHI is actually 20 (e.g assuming all 20 occurred in the one hour I was actually sleeping but software divides it by 4 total hours on?

Thanks.

The machine divides the number of events by hours of use.
The machine does not know in which of those periods you were asleep and in which you were awake.
OSCAR is used by people who care about their therapy to try to tease out that information from the graph that shows breathing patterns.

You do not have OSCAR info, so your question is a mathematical question.
This board is about therapy, so let me reformulate your question:
If the machine aggregates hours of use, regardless of whether I was asleep or not, can I use the AHI number to improve my sleep therapy?
The answer is no.
Can I know what my breathing disturbances are while I'm asleep without software that shows the breath flow?
The answer is no.
If I use the machine just for proof of compliance e.g. while I'm awake can anyone use its AHI data to improve my sleep therapy?
The answer is no.

[dataq1]

I agree with each of your answers to Wakemenot’s reformulated question.

However I answered the question he actually asked. His question is appropriate for this forum b/c while it may be “mathematical”, it requires some knowledge of how devices calculate AHI.

I stand by my answer to Wakemenot, but don’t understand why that answer generated such derision.

[ChicagoGranny]

However I answered the question he actually asked. His question is appropriate for this forum b/c while it may be “mathematical”, it requires some knowledge of how devices calculate AHI.

I stand by my answer to Wakemenot

"THE INFORMATION PROVIDED BY DATAQ1 IS NOT INTENDED NOR RECOMMENDED AS A SUBSTITUTE FOR PhD PROFESSIONAL MATHEMATICAL ADVICE."

[lynninnj]

You think you can extrapolate like that?

No extrapolation involved.
Wakemenot was quite explicit in his hypothetical.

"all 20 {events} occurred in the one hour I was actually sleeping"

Whatever was happening while he was still awake (reading or whatever) can't (or should not) be considered valid events for the purposes of AHI calculation.

The problem is that the device ASSUMES he is sleeping, so divides by 4 instead of 1 to calculate the AHI in his hypothetical.

He is correct, and I agree. Please explain the basis of your disagreement.

Have you ever looked at a chart?

Like EVER?

Or do you refuse to do so because you don't think any of us should be looking at them?

And IF you have looked at one, have you noticed clusters of events? Or are they all evenly spaced?

Heres a zero AHI hourly snippet: https://sleephq.com/public/7345d0bc-222 ... 4bfc4aaf7b

Do you think this means I have zero AHI?

How about this? Does this mean I have 8 AHI? https://sleephq.com/public/845f8087-d15 ... b7395ad6ff

I had better hopes for you but.....

[lynninnj]

I agree with each of your answers to Wakemenot’s reformulated question.

However I answered the question he actually asked. His question is appropriate for this forum b/c while it may be “mathematical”, it requires some knowledge of how devices calculate AHI.

I stand by my answer to Wakemenot, but don’t understand why that answer generated such derision.

Because by now, you should KNOW better.

[dataq1]

Because by now, you should KNOW better.

Let's start over/try again:
Wakemenot's hypothetical is saying :
I had my device on for 4 hours. (three hours of which I was trying to become acclimated to mask wearing by reading)..
'I was asleep for only one hour, during that hour my device detected 20 events.
but my device assumed that I was asleep for 4 hours, so it reported my AHI as 5. But shouldn't the AHI be closer to 20 for that period.

Would that "5" be incorrect, he asks.
To which I say, yes, you are correct.

The device should be dividing events by the number of hours in actual sleep (or as near enough as possible). Hours spent reading, watching TV, trying to become acclimated to mask wearing should not be used to calculate index values. (but the algorithm doesn't know that you are not trying to sleep.

[lynninnj]

Because by now, you should KNOW better.

Let's start over/try again:
Wakemenot's hypothetical is saying :
I had my device on for 4 hours. (three hours of which I was trying to become acclimated to mask wearing by reading)..
'I was asleep for only one hour, during that hour my device detected 20 events.
but my device assumed that I was asleep for 4 hours, so it reported my AHI as 5. But shouldn't the AHI be closer to 20 for that period.

Would that "5" be incorrect, he asks.
To which I say, yes, you are correct.

The device should be dividing events by the number of hours in actual sleep (or as near enough as possible). Hours spent reading, watching TV, trying to become acclimated to mask wearing should not be used to calculate index values. (but the algorithm doesn't know that you are not trying to sleep.

Dude-again I ask have you ever looked at these graphs?

Did you bother to look at the two graphs I linked so you could understand why what you are saying is bullshit?

Or are you saying,”no that data does you no good at all”?

Sample size of one hour? Really?

[dataq1]

Did you bother to look at the two graphs I linked so you could understand why what you are saying is bullshit?

Actually I did, but I don't know why you think it applies to the discussion.

Isn't the rule of thumb that events that are reported while awake (particularly in this case - reading, etc ) are not valid.
If they are not valid then the "while awake" time period should not be considered when calculating the indices.

Should the sample size be limited to one hour? Should a one hour sample be considered representative of 6 to 10 hours?
Probably not, but that was the condition that Wakemenot question dictated.

The bottom line here is: the Cpap THOUGHT the subject was asleep for 4 hours so calculated the AHI as 5. (20 divided by 4)
BUT the subject was really only sleeping for 1 hour, and all 20 events happended during that 1 hour, so the AHI was 20 (20 divided by 1)

No bullsh*t involved, just discount awake events and awake hours.

[lazarus]

It is possible for one simple question to have multiple levels of meanings and implications, depending on which aspects and deeper meanings one chooses to focus on.

Dataq1's answer, as I read it, is only that the OP is basically on the right track of understanding, in general, based on a broad hypothetical, how home-treatment machines do math and report it.

The other answers from other posters are validly focused on deeper implications, practical considerations for improving outcome, addressing the meaningful troubles the OP revealed, stressing the importance of seeing actual data, a brief spelling lesson, and the long-term education of the OP.

Cool.

I enjoyed this thread and would like to assure the OP that I, for one, in my opinion, believe everyone reading this thread looks forward to the day that actual data gets posted.

I'm hopeful that PAP will be as effective as possible once optimized--even if we all do occasionally seem to talk a bit past one another now and then here at the zoo.

[zonker]

...here at the zoo.

[lazarus]

...here at the zoo.

I'm not ashamed that I have studied at the feet of Zonker.

Didn't realize it was copyrighted. Sorry, sir.

[chunkyfrog]

Humor is the gentlest form of rebuke--sometimes.

[Tec5]

to dataQ: what you are saying is bullshit?

Lynn, Data is actually correct - so, NO BULLSHIT there. Straightforward answer to the question asked, nothing added would be "yes, sir you are correct".

We can speculate about why Wakemenot asked the question - my take on it is that he was asking for confirmation that hours that are not actually asleep should not be counted in AHI computation.

BTW, personal attacks and vendettas ("you should know better") amongst forum members may not serve to encourage Wakemenot to participate. He already warned " Don't yell at me about posting".

Take a DEEP breath!

[lynninnj]

Did you bother to look at the two graphs I linked so you could understand why what you are saying is bullshit?

Actually I did, but I don't know why you think it applies to the discussion.

Isn't the rule of thumb that events that are reported while awake (particularly in this case - reading, etc ) are not valid.
If they are not valid then the "while awake" time period should not be considered when calculating the indices.

Should the sample size be limited to one hour? Should a one hour sample be considered representative of 6 to 10 hours?
Probably not, but that was the condition that Wakemenot question dictated.

The bottom line here is: the Cpap THOUGHT the subject was asleep for 4 hours so calculated the AHI as 5. (20 divided by 4)
BUT the subject was really only sleeping for 1 hour, and all 20 events happended during that 1 hour, so the AHI was 20 (20 divided by 1)

No bullsh*t involved, just discount awake events and awake hours.

first off, you haven’t even seen the data. You have no idea if there were 20 events in the one hour that the person was supposed to be awake. you can’t just arbitrarily divide a number by four and say oops it’s this.

To boot, you also can’t say that it has ANY usefulness when trying to determine AHI.

You’ve been here long enough to know that when someone posts a chart, they are usually encouraged to keep settings a certain way and come back and post another chart each day for several days.

I can’t tell if you were just ignorant to this or being intentionally disruptive.

A shitty sample is a shitty sample no matter how you want to spice it up and serve it. At the end of the day you are still serving up shit.

Like I said: you should know better by now.