Why doesn't APAP respond to apneas?

Thanks, Rested Gal!
OK, and what about you, -SWS who never needed IFL1 turned off: is that a baquet hall or a dancing hall on your chart?
O.

you guys keep making fun of them little chairs, like I said they have been around a long time and mean NOTHING on Bev's report, only thing the mean is the machine is trying to find the lowest pressure possible; pasted right out of the Remstar Auto Brochure (page 3 of 6 of the Remstar Auto "Classic" brochure, yes I save those kind of things).

The Science of Finding the Right Pressure.
At the core of the REMstar Auto’s ability to adjust to a patient’s pressure needs is a “multi-level” algorithm.
Working primarily by measuring instances of flow limitation, the REMstar Auto utilizes systematic, subtle
changes in pressure and assesses the impact of these changes on the patient’s breathing patterns. These proactive
changes in pressure and assessment of flow limitation are continually searching for the best possible
pressure level for the patient. Supporting this primary analysis are secondary analysis parameters devoted to
detecting major respiratory events and responding appropriately. These secondary analysis parameters
determine necessary levels of responseto eliminate events and position the patient with a more appropriate
pressure level. Additional levels of analysis are devoted to elevated – or “big leak” – levels and variable breathing.

then it says below the fancy dancy graph in the center:

Throughout the night the Proactive Algorithm of the REMstar Auto subtly searches for the best pressure level for a
patient’s needs by cycling through the Pcrit, Popt and Ptherapy modes. Pcrit determines the critical pressure where
the airway begins to collapse. Popt determines the optimal pressure level for a stable airway. Ptherapy provides an
extended period of constant pressure at the ideal level.

Note: It measures these values as pressure drops after stepping over the chair, ole split_city would like this, they monitor for these right where he showed they occurred

You guys crack me up, you have never played musical chairs before LOL

Bev: you mentioned you have James Skinner Analyzer program in your first post in this thread, can you post the Variable breathing report?

Snoredog, quoting the Remstar Auto Brochure (page 3 of 6), wrote: <snip>

Ptherapy provides an extended period of constant pressure at the ideal level.

I see no extended period of constant pressure at the ideal level in Bev's APAP chart.

O

ozij wrote:Do you assume a search for Pcrit to start at minimum pressure? I wouldn't.

According to Respironics that routine search for Pcrit always starts at the previous Pther value. And Pther is nothing more than whatever therapeutic pressure happened to be sustained pending the results of the previous pressure probe or responsive pressure routine. Respironics describes the probe for Pcrit as always being a downward pressure search. It's not even algorithmically possible to search downward for Pcrit when the machine already happens to be at the minimum pressure prescribed by a pressure range.

That means the scheduled probe has no incentive to look for an even lower pressure. The machine simply won't go below the APAP min pressure under any circumstances. Therefore a routine to search for an even lower feasible pressure is expressly prohibited by the logic of the algorithm never going below min pressure. For this scheduled routine to even be allowed, the algorithm must first satisfy the prerequisite logic that a lower pressure search is possible within the constraints of the APAP's min pressure setting.

ozij wrote:

-SWS wrote: So the algorithm doesn't score FL, for lack of all the other probability components of frank FL scoring criteria. But it decides that it needs to pressure probe upward until that flow wobble goes away. And by, golly, pressure goes up and up quite a ways before that flow-amplitude wobble spontaneously subsides. In general it would be possible for the algorithm to commence that same upward search routine for any subset of probability-based FL precursors---not just the amplitude parameter. That means slight obstruction or transluminal airway collapse (as opposed to respiratory controller oscillation) is a more likely case across the obstructive patient population.

Once again, agreed, and this is what I tried to say about the algorithm's response in my posts from Wednesday.

I pulled your Wednesday point out because I think it is correct and central to what is happening on Bev's charts and the charts of several other posters. Rested Gal's up-and-down Encore Pro "pressure chairs" as well as her 420e FL over scoring hint at two manufacturers' interpretations of the same less-than-frank precursor signs of FL----the same probability-based interpretation/misinterpretation of the same subtle FL probability component(s) embedded in the flow signal. What would the FL detection specificity be for any of those constituent probability-based wave components when an insufficient subset occurs? Mediocre at best I would suspect, since the algorithm doesn't even have enough probability-based flow signal components to even score an individual frank FL event!

ozij wrote:

Snoredog, quoting the Remstar Auto Brochure (page 3 of 6), wrote: <snip>

Ptherapy provides an extended period of constant pressure at the ideal level.

I see no extended period of constant pressure at the ideal level in Bev's APAP chart.

There's no basis to expect that Pther must be maintained for any extended period of time. Pther is nothing more than "the new therapeutic pressure that has been determined after a probe". That Pther is simply the new therapeutic pressure value that would result after either: 1) the scheduled Pcrit/Popt/Pther probe or 2) a responsive Popt/Pther probe---meaning a non-scheduled probe resulting from just a spontaneous and preliminary hint of FL, or 3) a therapeutic pressure value previously determined by a higher-priority control layer in the algorithm.

Recall this particular control layer in the algorithm has virtually bottom priority. That means all the higher priority control layers in the algorithm can and will very quickly leave Pther behind for whatever pressure value those control layers deem more suitable. Pther doesn't have to be maintained for an extended period if the patient's spontaneous events or precursors just so happen to keep the responsive and especially higher-priority control layers of the algorithm busy.

ozij wrote:OK, and what about you, -SWS who never needed IFL1 turned off: is that a baquet hall or a dancing hall on your chart?

Bookmark placed... I need to find my old data amidst an assortment of hard drives and computers. Recall that I was off CPAP for more than a year because of trigeminal neuralgia.

----------------------------------------------------------------------

A comment about Pcrit or "critical pressure" relative to transluminal collapse. These are generic concepts in the sleep industry that are taken from the Starling Resistor model that sleep science routinely employs to understand flow-limitation related collapse of the human airway. Respironics and PB/Tyco both account for Starling-Resistor modeled (transluminal) critical pressures in their respective algorithms.



Snoredog, don't dance so much on the illusory surface of logic when it comes to marketing text---that superficial surface of text where manufacturers necessarily dance in communication with a public comprised of people who are generally unfamiliar with the science and technology behind their products... I have yet to see anything definitive in your argument to preclude those pressure chairs from occurring exactly as ozij and I describe. Nada! Zip! Zilch!

And the logic that an algorithmic probe for a feasible lower pressure from APAP min pressure baffles me. Respironics very clearly states the purpose of that scheduled pressure-searching routine is for the purpose of determining a lower feasible pressure. Lower from APAP's min pressure? It ain't gonna happen!

-SWS wrote: Snoredog, don't dance so much on the illusory surface of logic when it comes to marketing text---that superficial surface of text where manufacturers necessarily dance in communication with a public comprised of people who are generally unfamiliar with the science and technology behind their products... I have yet to see anything definitive in your argument to preclude those pressure chairs from occurring exactly as ozij and I describe. Nada! Zip! Zilch!

Well, that so called marketing text happens to display the same graph as you put up only in color. It matches perfectly as to what is happening during low activity periods on Bev's charts. There is NO apnea or hypopnea taking place to trigger the A/H controller to take control over that circuit, there may be variable breathing happening causing the Variable Breathing controller to look at it, but we cannot see that happening on a Encore report. There is no event, not even snore for the snore controller to to take control.

So during those periods of no activity the machine probes for a more optimal lower pressure, it does that by increasing pressure holding for 30 seconds, monitoring for change, increasing pressure again then decreasing the pressure where it monitors for change as the pressure decreases. If it detects an event such as a Hypopnea it exits the routine and hands control over to the A/H or other controller.

But you and ozi keep searching for it, you'll find it one day.

Snoredog wrote:But you and ozi keep searching for it, you'll find it one day.

Ah, don't take it personal... ozij and I are just heading in the wrong and opposite direction with the rest of science...

Snoredog wrote:So during those periods of no activity the machine probes for a more optimal lower pressure, it does that by increasing pressure...

Search for a lower optimal pressure from APAP's min pressure? I think we'll have to politely agree to disagree on this point my friend. Don't take it personal.

-SWS wrote:

ozij wrote:But you and ozi keep searching for it, you'll find it one day.

Ah, don't take it personal... ozij and I are just heading in the wrong and opposite direction with the rest of science...

Snoredog wrote:So during those periods of no activity the machine probes for a more optimal lower pressure, it does that by increasing pressure...

Search for a lower optimal pressure from APAP's min pressure? I think we'll have to politely agree to disagree on this point my friend. Don't take it personal.

Well then increase pressure to 15 and completely mask it, because that is what it will do. It is bouncing along the Minimum pressure and for this circuit to work it has to increase pressure and monitor what happens from that increase on the down side. It is only moving up by a total of 2 cm. Total working range is only 14 to 17, it needs at least 2 cm for that circuit to work, mask it out and you might as well put it in straight CPAP mode at 15 so you see a straight line across the report and cannot see what is happening. What residual apnea events seen on that report are probably central, that is if the machine was smart enough to determine that, and it is not.

She is on average having fewer events under current settings than was seen in the lab report bilevel settings, but hey if it ain't broke let's fix it anyway.

but hey if it ain't broke let's fix it anyway

Symptomatically something is very broke---it may or may not be related to xPAP therapy.

And what might the measurement specificity and sensitivity be for FL or even hypopnea on single-channel flow-sensor based APAP machines? Far less than perfect, which is tantamount to being therapeutically broke for many patients.

And what might the sensitivity and specificity be for machine-related excessive cortical arousals? Exactly a big fat broken zero in all cases, meaning we can never be certain when xPAP therapy is or isn't broken..

but hey if it ain't broke let's fix it anyway.

Bev is tired and having long apnea that concern her. We were trying to find clues in her data for that.

If you're right, snoredog, raising both minimum and maximum in the same amount will keep the algorithm bouncing merrily along the minimum because it will then have enough of a range to raise the pressure which - according to you - it will do in order to find out if it can drop the pressure to lower than the minimum....

O.

Deleted

-SWS wrote:

but hey if it ain't broke let's fix it anyway

Symptomatically something is very broke---it may or may not be related to xPAP therapy.

And what might the measurement specificity and sensitivity be for FL or even hypopnea on single-channel flow-sensor based APAP machines? Far less than perfect, which is tantamount to being therapeutically broke for many patients.

And what might the sensitivity and specificity be for machine-related excessive cortical arousals? Exactly a big fat broken zero in all cases, meaning we can never be certain when xPAP therapy is or isn't broken..

Well what about the lab report? Appears they couldn't keep those cortical arousals at bay in the lab setting for very long either, they only tested her from 12/8 to 25/21 and still came to no clear conclusion. So what pressure combination produced the lowest RDI and fewest cortical arousals? Let's look at the lab report;

I/E - RDI - SaO2
16/12: 6.3 SaO2=84
20/16: 9.2 SaO2=82
23/19: 7.0 SaO2=81

I think you have to ignore the RDI and look at the oxygen level, she has to be above 89%, so based upon that titration table, that happens at 19/15. Step above that with EPAP=16 and oxygen drops down to 82 again, unacceptable in my opinion. While RDI climbs back to 33, oxygen remains higher.

While that lab report suggests she fits the criteria for CSDB fact is she is doing very well on the current machine. AHI doesn't really lie, the machine might confuse a central event for an obstructive, it will still count and include those in the final score. Her final scores have been .9, 1.4 and 2.0, if those scores include both central and obstructive events that is great and I doubt you will improve upon that.

If they couldn't find her optimal pressure in the lab, what makes you think the Adapt SV will find it?

She is still going to need enough EPAP pressure to keep her obstructive events in check. Even her lab report shows it becomes a train wreck at 12/8. In fact it remains a train wreck up until she gets to 19/15. Get a cm or two above that and her oxygen levels again go south with no real improvement to RDI. She doesn't appear to be having clusters or long periods of centrals, in fact they are so few in frequency they don't even trip the NRAH circuit on the current machine. She doesn't have a lot of leak skewing things in fact her leak control efforts are excellent.

Her doctor has already tried her on the other obvious pressures with not much success. She could try increasing the Minimum to 15 and see what happens but if the frequency of the total residual events increase then she would have to go back down.

Ehreesejr, since your question and machine constitute an entirely different topic than this thread's discussion, I'm afraid your question may get lost here in the shuffle. Would you mind posting that same set of questions in a new thread? This thread is already swimming with more issues than most of us can keep track of toward the thread's intended purpose.

Snoredog, great points. I'll place another bookmark here because I'd like to delve more into your excellent rhetorical questions about cortical arousals, what constitutes "doing well" by definition, even the necessary EPAP you suggest (which may or may not be necessary with adaptive servo for some possible scenarios in etiology). Bookmarked because client work now calls...

In the meantime, I also had a thought about those scheduled upward pressure probes from min APAP pressure: if they are scheduled and routine, we should see them on all "activity free" Encore charts. Any chart that has little or no VS/FL/H/A activity going on should theoretically have those "pressure chairs" if those were routine and scheduled. But over the years we have seen quite a few "activity free" Encore charts where those "pressure chairs" just never happen. Those boring charts just sit at min pressure forever. Routine upward "pressure chairs" should be scheduled in those cases but are not. I'm still thinking that when we see the upward pressure chairs, we are seeing an upward pressure probe that is launched from probability-based FL components in the flow signal.

Anyway, as always, what cool discussions this message board manages to put together on a routine basis! THANK YOU!!!

-SWS wrote:...Any chart that has little or no VS/FL/H/A activity going on should theoretically have those "pressure chairs" if those were routine and scheduled...

Don't forget Variable Breathing has a control circuit, while it doesn't report VB on Encore reports, it does if you use James Skinners' program against the database.

Are those chair pressures representing Variable Breathing? We don't know, reason I asked Bev if she had Skinners software to run that against the data.

One thing about that algorithm, if it is not using one of the higher controllers addressing a particular event it goes to that probing to see if it needs to reduce pressure, if it cannot do that due to a self imposed bottom it can only run it bouncing off the bottom. If an event matching one of the other controllers gets its attention it stops with that probe and gives control to back to that higher control circuit.

I'm trying very hard to get some client work done and my good friend Snoredog comes up with an excellent hypothesis IMO.

That's one great possibility to explore IMHO. If you get a chance before I do, you might want to dig through the patent description for the multiple VB controller pressure-response patterns to see if you can find a pattern match with what's happening on Bev's charts. I'm vaguely thinking the VB controller initially does only one of two things: 1) holds pressure or 2) reduces pressure-----as opposed to raising pressure.

If we do happen to find a match, we might be witnessing the algorithm's response to periodic breathing. A few other possibilities as well....