OA clusters align perfectly with leak spikes

[ozij]

I can't help questioning the validity of the data. Again I had a night where I laid awake for a long time and recorded 5 distinct OAs. Yes, I can throw that data out, but it really calls into question OAs that are recorded when I am asleep,

Not at all. Its a case or garbage in garbage out. The waking breathing patterns are garbage for the algorithm, and therefore, whatever it reports is garbage as well. A tool is good only when it fits the circumstances in which it is used. The fact a Ferrari will sink when attempting to race a sailboat on the sea ocean does not call into question its quality on the race course on land.

especially when they show a nearly 1.0 correlation with leak spikes.

A correlation, even of 1, does not indicate causality. There is no way for you to disprove the following scenario: While you sleep, whenever you have an apnea, you move your head, thereby dislodging the mask, and causing a leak spike.

I might even argue the only reliable indicator of a night's therapy is how one feels.

You've argued that in the past...

Obviously good numbers are good, but are bad numbers always bad?

Obviously bad numbers are bad. But are good numbers always good?


O.

[-SWS]

Obviously bad numbers are bad. But are good numbers always good?

Wakefulness can reveal great AHI and snoring numbers! Similarly SDB-disturbed sleep presenting little or no time in REM and deep stages can in some cases reflect a misleadingly low AHI---as if treatment and sleep were optimized.

Great rhetorical question IMHO.

[roster]

Interesting discussion. But one basic point I am convinced of is about the accuracy of the machines. The flow detectors can reliably detect air flow and absence of air flow; the timer can accurately count elapsed time; and the software can accurately calculate absences of air flow of ten seconds or more duration and record and report them. Anyone of these elements by itself is a very simple matter with today's state of technology.

You do have to remember that the machine has been instructed to record an absence of air flow of ten or more seconds as an apnea without concerning itself about the patient's state of sleep/awake.

[ozij]

Interesting discussion. But one basic point I am convinced of is about the accuracy of the machines. The flow detectors can reliably detect air flow and absence of air flow; the timer can accurately count elapsed time; and the software can accurately calculate absences of air flow of ten seconds or more duration and record and report them. Anyone of these elements by itself is a very simple matter with today's state of technology.

You do have to remember that the machine has been instructed to record an absence of air flow of ten or more seconds as an apnea without concerning itself about the patient's state of sleep/awake.

The machine interprets the interruption based on the pattern of the breaths that happened before the interruption. It does not automatically tag any 10 second interruption as either an apnea or a hypopnea. The interpretation given to the data gathered, depends on the context in which it was gathered. And that assumes the context is the orderly breathing of sleep. It does not identify sleep vs. wake -- that's why it make mistakes in interpretation.

O.

[roster]

The point in another thread and repeated here by Jason is that "false positive OAs" are in the report. I continue to maintain that a tick mark on the apnea chart means you did not breathe for ten or more seconds and it is an accurate reporting. Whether you were awake or asleep, the tick mark means your were "breathless".

[ozij]

But it does not mean you had an obstruction.

[Jason S.]

Interesting and worthwhile discussion. However, my point is if I look at my nightly AHI and see a 3.2, that by itself is not necessarily meaningful, because I might have recorded 5 OA's while awake, thus falsely escalating my nightly AHI. Secondly, I can't completely trust what is recorded later in the night, as I might have been laying awake after a bathroom break and recorded another OA, but I can't verify the timing of the events the next morning. Finally, while I can't "disprove" that OA's cause my head to move which results in a leak spike, I can argue that random head movements not associated with OA's should also result in a leak spike, but that is not what is observed in my data.

[jnk]

viewtopic.php?f=1&t=42793&p=391253#p391267

. . . Not all obstructive apneas are sudden and/or complete airway closures. But the apneic closures that are both sudden and completely occlusive are specifically the events that will present a very sudden step or spike in flow-impedance. And those sudden flow-impedance spikes in turn will create an instantaneous pressure spike that can more easily breech interface seals.

I distinctly recall Berthon-Jones commenting (in one of his patent descriptions) that instantaneous pressure spikes induced by sudden apneic closures are prone to breeches at the interface seal. . . .