An APAP Shootout (sort of) on Academic Journal

[Slinky]

Where, oh where, are the studies on the benefits of educating and equipping patients to take responsibility for their own treatment, under the minimally required supervision of medical professionals?

Yer, kiddin', right? WHY on earth would the professionals want their patients to be THAT educated or knowledgeable?? They don't even like those of us stumbling around TRYING to educate ourselves! Thank goodness my doctors aren't in that league - BUT - time is so limited w/my sleep doc when he's only available to me one day out of every 3 weeks. *sigh*

Thank you, Chuck. I keep trying. *sigh* I just get frustrated w/all the mistakes and bumbling I do along the way. Some days I feel dumb as a rock! Its em-bare-a$$ing.

[GoofyUT]

[quote="sleepergal"]I won't buy a cpap or apap machine that doesn't give me detailed treatment and results information, period. As long as Respironics allows the software to be sold to the patient and Resmed doesn't, I'll purchase Respironics equipment and not Resmed. Resmed may be the better machine for some, but as long as they maintain their current "no software to patients" policy, I'll pass. I suspect that the improved compliance, attention to leaks, mask results, etc. that I achieve with full access to the Respironics software greatly outweighs any potential benefits I might achieve by a different algorithm.

Where, oh where, are the studies on the benefits of educating and equipping patients to take responsibility for their own treatment, under the minimally required supervision of medical professionals?

[Slinky]

In addition to which, if one is innovative one can get the software AND the necessary reader for the Resmed products very reasonably despite Resmed's stupid new policy. I was able to thanks to advice and help from fellow PAPpers.

AND I like the easily accessible data on the LED screen, a clear, good-size, easy to read LED screen. The Respironics Auto w/C-Flex (pre-M series) window is next to impossible to read and the info not as detailed as the Resmed S8 Elite's.

Which takes us back to the algorhythm, whatever, w/different strokes for different folks. Unfortunately, given all that transpired during the loan of that Respironics RemStar Auto w/C-Flex, we have no way of knowing how well I would or wouldn't respond to the Respironics algorhythm vs the Resmed's S7 Elite's algorhythm and they are the only two I've had any experience with.

I'd LIKE to know more about the Puritan Bennet and others out there but ... given the problems I've had just getting these two, heaven only knows if it is worth it to try to get the opportunity to try any other manufacturer's machine. Of course, I want the machine that will give "me" the best results but goshes ....

[Snoredog]

hate to burst your excitement bubble, but this is an OLD report, it keeps rearing its ugly head here every year or so when someone over uses google.

The data from it is a joke. No human being responds like that iron lung so the data in this report is totally misleading. The machines ability to respond to events as suggested in this report is even a bigger joke.

Then this wasn't a study at all, the selection criteria is unknown, there was no peer reviewed or double-blind study done here, just some geeks that got together who built an iron lung in a box to bench test the various machines.

Fact: NONE of the machines listed above can detect the difference between a central apnea lasting 40-seconds and a central apnea lasting the same duration. In both cases, there is no airflow.

It claims the Resmed Spirit is the most sensitive of the bunch, that is true you might as well just use CPAP. Even Resmed acknowledges their machine is not very good at detecting events and limits any response above 10cm using the A10 algorithm. If pressure is at or above 10cm it won't respond to any apnea event without a flow limitation for fear it may be central. But then they forgot how the snore circuit worked.

Puritan Bennett even admits their 418P (now discontinued machine tells you how old this report is) and the 420E both equipped with a more sophisticated pneumotach sensor is only about 60% accurate at best at detecting a central apnea and it is supposed to be the best autopap on the market for detection. That sensor is sensitive enough to detect cardiac oscillations to determine if an event is a central.

That means even on a good day the P&B machine can only detect 6 out of 10 events accurately. With centrals, if it makes a mistake with only 1 of those it can cause a cluster of others.

This whole report is totally meaningless especially if you are trying to use it to select a machine, probably every machine included in the report has upgraded their firmware and no longer functions the same as it did in that report even if it was an iron lung.

[Slinky]

Well, if the information is so outdated, WHY is it inspiring comments in CHEST in 2006??

Chest. 2006 Aug;130(2):343-9.
Comment in:
Chest. 2006 Aug;130(2):312-4.
Bench evaluation of flow limitation detection by automated continuous positive airway pressure devices. PMID: 16899831

A quick PubMed search only brought up two 1996 abstracts in favor of the autoPAP. I didn't take the time to hone my search criteria.

Am J Respir Crit Care Med. 1996 Sep;154(3 Pt 1):734-40.
Automated continuous positive airway pressure titration for obstructive sleep apnea syndrome. PMID: 8810613

Sleep. 1996 Jul;19(6):491-6.
Computerized adjustable versus fixed NCPAP treatment of obstructive sleep apnea. PMID: 8865507

I do VERY MUCH agree that an "artificial" lung is not particularly good for truly evaluating "real life" response to the various PAP machines' responses to events and appears to be poor "science". They need to have multiple persons w/sleep apnea try each machine over an equal period of time for each person to come to any adequate conclusions.

[Snoredog]

[quote="SamCurt"]Actually there was another APAP algorithm shootout in the same issue of the same journal, and the editorial was also on these two articles.

I would try to summarize the editorial, but not the second shootout since the latter has funded by a company related to Weinmann (manufacturer of SOMNOplus masks and SOMNOsmart APAPs).

[dsm]

Just wanted to add some thoughts re this.

Last year my respiratory surgeon passed a copy of this report to me when I visited him. I mentioned it here but didn't summarize it because I was pretty sure that if I did, there would be several folk who would immediately criticise its conclusions because the report's comments were somehat scathing of a well known (& popular) brand.

Some folk did criticise the approach to research these guys did - the criticism centering on 'how can a mechanical breathing machine in a lab setup in anyway relate to humans with all their variations'.

The *only* way scientists can conduct meaningful research is to lay down a common yardstick (their lab machine) and to use internationally accepted airflow waveforms that are acknowledged by the medical profession and that is exactly what they did. These airflow waveforms were originally devised by a medical researcher based on actual patterns of real people's breathing in his specialist practice and these then became the accepted international standards for testing breathing equipment with. They may not be 100% perfect but are 100% better than nothing. If they didn't adopt known and accepted airflow waveforms then their research would be pointless and easy to criticise.

So yes, the mechanical breathing simulator isn't a real human with all the normal human variations but it was fed known patterns of airflow that are recognised in the medical industry and thus have some common meaning in terms of interpreting results.

In a way I guess I am pre-empting an attack on this research paper as I think it will come. What I do hope is that people reading about these international research projects will appreciate that the many differences between how Auto algorithms are implemented is the very reason the industry needs some common and widely accepted yardstick measurement as the basis for doing any assesments.

DSM

[rested gal]

I do VERY MUCH agree that an "artificial" lung is not particularly good for truly evaluating "real life" response to the various PAP machines' responses to events and appears to be poor "science".

I think you're absolutely right, Slinky. Poor science, indeed.

Tests of how autopaps perform when presented with artificial patterns produced by a "breathing machine" are useless, imho, in providing useful information about what those same autopaps would do if they were receiving constant FEEDBACK from a living person.

Or as -SWS put it so well:

viewtopic.php?t=1715&postdays=0&postorder=asc&start=15

On page 2 of that topic discussing an older test using a breathing machine to assess autopaps' responses,
-SWS explains the inherent flaw in such tests using artificial breathing machines. Points that still apply to even the most recent "breathing machine" tests:

"most AutoPAPs will: 1) sense a sleep event, 2) administer an initial pressure-response based on that event, then 3) calculate patient airflow to determine required subsequent pressure adjustments. That study fails to show any patient response whatsoever to any of those initial pressure increments. That study also fails to accurately simulate pressure incremental reiterations that are based on patient feedback. The study only plays a non-responsive sleep-event "loop" to further demonstrate how each AutoPAP responds to a test dummy that doesn't respond at all to pressure. Or in other words: the study breaks the patient-to-machine feedback loop that would be absolutely crucial to any of those AutoPAP algorithms."

The link Mikesus posted on page 1 to the older study no longer brings up the PDF. As best I can remember, this was the first "Bliss/Eiken" test of autopaps with a breathing machine. Later, Todd Eiken repeated those tests for yet another useless (imho) look at more modern autopaps. And most recently a French team did the same sort of tests...probably the test SamCurt found in "Chest" 2006.

I don't think anything useful about autopaps or comparison of autopap performances comes out of any of these autopap assessment "tests with a breathing machine." The autopaps are not receiving the crucial patient feedback which would enable the autopaps to determine what they will (or won't) do next.

The worst thing about such artificial tests is that they seem to impress a lot of doctors.... on the one hand, turning them anti-autopap; and on the other hand causing some to believe one brand gives better responses to various sleep disordered breathing patterns than another. I'm sure ResMed reps are well armed with colorful slick charts to plop down on doctors and DMEs' desks, showing quick response to artificial breathing machine "events."

[Rabid1]

You're right RestedGal.

Any experiment on a non-responsive subject (machine) is inherently flawed.

It's somewhat like testing drugs in a test tube instead of living creatures.

[rested gal]

Some folk did criticise the approach to research these guys did - the criticism centering on 'how can a mechanical breathing machine in a lab setup in anyway relate to humans with all their variations'.

Actually, if you'll read carefully -SWS's criticism of the mechanical breathing machine/autopap tests the problem is not in what waveforms are being generated by the breathing machine. The problem is in lack of feedback from a living person, with all the dynamic changes that can occur in a person's breathing. The autopaps depend on living breathing feedback to "decide" what to do next.

The *only* way scientists can conduct meaningful research is to lay down a common yardstick (their lab machine) and to use internationally accepted airflow waveforms that are acknowledged by the medical profession and that is exactly what they did. These airflow waveforms were originally devised by a medical researcher based on actual patterns of real people's breathing in his specialist practice and these then became the accepted international standards for testing breathing equipment with. They may not be 100% perfect but are 100% better than nothing. If they didn't adopt known and accepted airflow waveforms then their research would be pointless and easy to criticise.

Well, to my way of thinking, that particular research is pointless. Precisely because of the lack of "real" feedback from a "real" person as the test proceeds.

Think about it this way. In the link I posted to the old discussion, -SWS also said this, regarding the usefulness of assessing autopaps presented waveforms by an artificial breathing machine:

"There must be a patient response (simulated or real) to truly test the algorithm. To rely on lobbing one obstrutive sleep event (even repeatedly) is like assessing world-class tennis players using only a serving machine. It just doesn't make for any sort of useful comparison in my opinion."

What I do hope is that people reading about these international research projects will appreciate that the many differences between how Auto algorithms are implemented is the very reason the industry needs some common and widely accepted yardstick measurement as the basis for doing any assesments.

What I hope is that doctors will recognize the inherent FLAW in such tests. Autopap algorithms' performances depend upon second by second, minute by minute FEEDBACK from live people. Not a "lobbed" repetitive waveform no matter how complicated or simple the waveform itself. It really doesn't matter how "accepted" as an international standard any waveforms are for the purposes of describing various types of sleep disordered breathing waveforms.

What matters is an unbroken feedback LOOP between a living, breathing patient and any autopap. That's not there in any artificial breathing machine tests no matter whose sleep lab produced the waveforms from "real people."

DSM, I truly do understand your desire for measurements, yardsticks, etc. But, given the ever changing dynamics of any one person's sleep disordered breathing, their response to what autopap does, the autopap's response to the change in the person's airflow....back and forth...each determining what happens NEXT in a live situation.... well, I just don't think artificial breathing machine tests are useful or meaningful at all.

If anything, they may even be detrimental when mistaken ideas arise about what machines A, B, and C will do with a live person in the presence of this or that kind of flow limitation, hypopnea, apnea...even internationally accepted "examples" of such events.

It's still going to be like a tennis serving machine, lobbing the ball (the waveform - no matter how realistic, accepted, etc.) again and again at each autopap. The test is missing the crucial breathing feedback from a living person. At best it is a pointless exercise in measurements, imho.

[dsm]

Some folk did criticise the approach to research these guys did - the criticism centering on 'how can a mechanical breathing machine in a lab setup in anyway relate to humans with all their variations'.

Actually, if you'll read carefully -SWS's criticism of the mechanical breathing machine/autopap tests the problem is not in what waveforms are being generated by the breathing machine. The problem is in lack of feedback from a living person, with all the dynamic changes that can occur in a person's breathing. The autopaps depend on living breathing feedback to "decide" what to do next.

The *only* way scientists can conduct meaningful research is to lay down a common yardstick (their lab machine) and to use internationally accepted airflow waveforms that are acknowledged by the medical profession and that is exactly what they did. These airflow waveforms were originally devised by a medical researcher based on actual patterns of real people's breathing in his specialist practice and these then became the accepted international standards for testing breathing equipment with. They may not be 100% perfect but are 100% better than nothing. If they didn't adopt known and accepted airflow waveforms then their research would be pointless and easy to criticise.

Well, to my way of thinking, that particular research is pointless. Precisely because of the lack of "real" feedback from a "real" person as the test proceeds.

Think about it this way. In the link I posted to the old discussion, -SWS also said this, regarding the usefulness of assessing autopaps presented waveforms by an artificial breathing machine:

"There must be a patient response (simulated or real) to truly test the algorithm. To rely on lobbing one obstrutive sleep event (even repeatedly) is like assessing world-class tennis players using only a serving machine. It just doesn't make for any sort of useful comparison in my opinion."

What I do hope is that people reading about these international research projects will appreciate that the many differences between how Auto algorithms are implemented is the very reason the industry needs some common and widely accepted yardstick measurement as the basis for doing any assesments.

What I hope is that doctors will recognize the inherent FLAW in such tests. Autopap algorithms' performances depend upon second by second, minute by minute FEEDBACK from live people. Not a "lobbed" repetitive waveform no matter how complicated or simple the waveform itself. It really doesn't matter how "accepted" as an international standard any waveforms are for the purposes of describing various types of sleep disordered breathing waveforms.

What matters is an unbroken feedback LOOP between a living, breathing patient and any autopap. That's not there in any artificial breathing machine tests no matter whose sleep lab produced the waveforms from "real people."

DSM, I truly do understand your desire for measurements, yardsticks, etc. But, given the ever changing dynamics of any one person's sleep disordered breathing, their response to what autopap does, the autopap's response to the change in the person's airflow....back and forth...each determining what happens NEXT in a live situation.... well, I just don't think artificial breathing machine tests are useful or meaningful at all.

If anything, they may even be detrimental when mistaken ideas arise about what machines A, B, and C will do with a live person in the presence of this or that kind of flow limitation, hypopnea, apnea...even internationally accepted "examples" of such events.

It's still going to be like a tennis serving machine, lobbing the ball (the waveform - no matter how realistic, accepted, etc.) again and again at each autopap. The test is missing the crucial breathing feedback from a living person. At best it is a pointless exercise in measurements, imho.

Rested Gal,

I think we are debating oranges & apples here.

If we look at the issues facing researchers in an environment where subjects such as people, have such varied responses and actions it is an outright impossibility to test every single person then produce a scientific report.

When newscasters broadcast opinion polls they take a 'sample' & use selected formulae to come up with a result that usually includes a margin of error. And agreed they don't always get things right but they get them right enough that the world uses polls as better than no input at all and in general the polls have a significant impact on the targets of the polls.

Science on mendical equipment is no different. They can only take the equivalent of polls. To argue that the polling or research is flawed merely because one doesn't like the outcome, is highly suspect.

Re Autos & breathing, the only scientificly responsible thing to do when testing medical equipment like this is to stick to known accepted models else the research has no common ground by which results can be judged.

Your case comes across to me as this (please point out if I have misinterpreted it)

1. No one should conduct scientific research on Autos unless they test every person who uses one (well how about every second person ?, perhaps every 3rd person ?), just how many people should they test & how real is it that they can do that ?. Should these people match internationally accepted breathing patterns documented for testing purposes, if yes they why not just use the patters through a simulator ???

2) The outcome of any research no matter how many degrees or qualifications the researcher's have is meaningless if they don't do their
tests of humans rather than simulators. Science has no right to try to set standard yardsticks for accurate testing.

3) The breathing algorithms the scientific community has accepted (R.. series etc:) don't mean anything and have no value & thus all scientific research using them is flawed.

In addition to the above points, can you explain how you if you were a researcher wanting to produce a report, that won't be laughed at or ignored, would go about testing Autos ?

I do understand your point that Autos *when for individual use* are best assesed by an individual themselves when it is for their own use. But scientists don'r have that luxury & have to test against accepted norms. This is where I think we are debating apples & oranges.

DSM

[DreamStalker]

I do VERY MUCH agree that an "artificial" lung is not particularly good for truly evaluating "real life" response to the various PAP machines' responses to events and appears to be poor "science".

First I must preface my following comments with the fact that I have not read all of the iron lung journal articles but I have briefly browsed a couple of them. Also, It probably is not an apples-to-apples comparison but as a scientist, we often fall back to simplicity as the most appropriate approach to modeling a complex system.

I work with a team of scientists who develop models to predict groundwater resources 50 years into the future. One of the unknown parameters that we must incorporate into the models is the occurrence of drought(s). Obviously, we cannot input the exact spatial and time characteristics of real future droughts. Using the paradigm of Occam’s razor, we gather the available drought data and statistically analyze it for representative droughts to input into our models.

I view the iron lung as a representative parameter for modeling the APAP algorithms and so I don’t exactly agree that their research attempts are “poor science”. As Einstein once said -- “Things should be as simple as possible but not any simpler”.

[rested gal]

When newscasters broadcast opinion polls they take a 'sample' & use selected formulae to come up with a result that usually includes a margin of error. And agreed they don't always get things right but they get them right enough that the world uses polls as better than no input at all and in general the polls have a significant impact on the targets of the polls.

Science on mendical equipment is no different.

There is no comparison between sampling from opinion polls and the setup of breathing machine tests. Talk about apples and oranges!

To argue that the polling or research is flawed merely because one doesn't like the outcome, is highly suspect.

Suspect in what way?

[rested gal]

As Einstein once said -- “Things should be as simple as possible but not any simpler”.

Good quote, Roberto. In the case of the artificial breathing machines, I think the people who set the tests up did go "too simple" to yield any truly meaningful results about what autopaps actually do when used by sleep-disordered breathing people.

Gotta have that continuous feedback loop from a live person (perhaps someday a simulated "live" person will be in the testing loop?) There's no point in lobbing repeated waveforms at the autopaps, even if the waveforms were generated from live people. Without a live person (or simulated live person) in the loop giving feedback to the autopap, it's impossible for autopaps to perform the way they are designed to perform, imho.

[Goofproof]

What they need to do, is to write the real truth. Model XYZ XPAP, is superior at treating ABC'S Iron lung simulator machine for Apnea. Then they would have a useful conclusion. Jim