Machine interpretations of events flawed?
I've been looking at some of the events over the last several months and wondering how/why the machine flags some things and not other things.
I realize that some of these have "time" criteria (e.g. 10 seconds), or amplitude criteria, etc. It seems to me that the machine can flag and "fix" obvious events, but some of the other things either should be events that are "fixed" by pressure pulse or increased pressure, or are not events at all and should not be counted as such in the stats.
Looking at last night's data I see some confusing stuff, maybe one of the experts on board here can help me understand. Am I missing something?
-- Image 2 & 4 - The CA#001 and OA#003 flow rates look basically the same, the machine initiated a pressure pulse for the CA like it was supposed to, but since it interpreted the other waveform as an OA it didn't initiate one.
-- Image 3 - Flow rate doesn't really look like anything other than slightly irregular breathing at the OA point (should it be FL?) and the FL looks like normal breathing.
-- Image 5 - What is this? Just irregular breathing? The first jiggly part of the flow rate wave is flagged as VS2, the second jiggly part - nothing.
06-20-2018, 07:24 PM
(This post was last modified: 06-20-2018, 07:24 PM by EddieR.)
RE: Machine interpretations of events flawed?
Looks like the images got corrupted somehow...and I can't edit it or delete it for correction.
Post edited...Moderator
I've been looking at some of the events over the last several months and wondering how/why the machine flags some things and not other things.
I realize that some of these have "time" criteria (e.g. 10 seconds), or amplitude criteria, etc. It seems to me that the machine can flag and "fix" obvious events, but some of the other things either should be events that are "fixed" by pressure pulse or increased pressure, or are not events at all and should not be counted as such in the stats.
Looking at last night's data I see some confusing stuff, maybe one of the experts on board here can help me understand. Am I missing something?
-- Image 2 & 4 - The CA#001 and OA#003 flow rates look basically the same, the machine initiated a pressure pulse for the CA like it was supposed to, but since it interpreted the other waveform as an OA it didn't initiate one.
-- Image 3 - Flow rate doesn't really look like anything other than slightly irregular breathing at the OA point (should it be FL?) and the FL looks like normal breathing.
-- Image 5 - What is this? Just irregular breathing? The first jiggly part of the flow rate wave is flagged as VS2, the second jiggly part - nothing.
RE: Machine interpretations of events flawed?
Image 2: CA#001
[attachment=6897]
Image 4: OA#003
[attachment=6898]
Image 3: FL#006 & OA#002
[attachment=6899]
Image 5: Not Flagged
[attachment=6900]
RE: Machine interpretations of events flawed?
Most of your examples show moderate flow limitation. I agree that the OA flagged in image #3 is incorrect.
You need more minimum pressure, and if you had a Resmed, i would increase EPR. Flex is less reliable, but I would turn that on and set it to 1. This much flow limitation would respond nicely to pressure support, but since we have a Philips, well, it is what it is. I would change pressure to 9.0 to 14 Flex at 1, or possibly 9.5-14.
RE: Machine interpretations of events flawed?
I think it is best to view event flags as just definitions that correlate with breathing. No definition is perfect. The trend is what's important. Every event is calculated based on the prescribed criteria but that does not necessarily mean that all events are equivalent.
The nice thing about sleepyhead is that you can see every breath and assess whether it looks like "a series struggle for air and sleep" or "just a cough and roll-over".
RE: Machine interpretations of events flawed?
(06-20-2018, 09:41 PM)Sleeprider Wrote: Most of your examples show moderate flow limitation. I agree that the OA flagged in image #3 is incorrect.
You need more minimum pressure, and if you had a Resmed, i would increase EPR. Flex is less reliable, but I would turn that on and set it to 1. This much flow limitation would respond nicely to pressure support, but since we have a Philips, well, it is what it is. I would change pressure to 9.0 to 14 Flex at 1, or possibly 9.5-14.
Thanks Sleeprider,
I've been working my way up with the min pressure and down with the max. Started at standard prescription setting of 4 - 20. It's been at 8.5 - 14 for 21 days now. Flex has been at 1 for quite awhile, 0, 2, & 3 just didn't work for me at all. 90% pressure has been around 10 - 10.2 for the last month or more. Max pressure hasn't been above 14 ever, so I'll keep it dialed in there, min is moving up to 9.0 tonight.
I appreciate the help!
