Hello Guest, Welcome to Apnea Board !
As a guest, you are limited to certain areas of the board and there are some features you can't use.
To post a message, you must create a free account using a valid email address.

or Create an Account


New Posts   Today's Posts

EERS Experiment Data (sherwoga)
#81
EERS Experiment Data - Seq #19 Design #20
12/18/2019 EPR(Min Pressure),EERS pair = 1(6),6 

           
Post Reply Post Reply






Donate to Apnea Board  
#82
EERS Experiment Data - Seq #20 Design #18
12/19/2019 EPR(Min Pressure),EERS pair = 2(7),6

           

This completes one night of data at each for the four corners of the center box in the Experimental Design.  I will process data in the next few days.  

In the meantime, this setting still has flow limits, but they appear to be smaller.  I'll start with my data evaluation on Flow Limits.
Post Reply Post Reply
#83
RE: EERS Experiment Data - Seq #20 Design #18
(12-20-2019, 07:31 AM)sherwoga Wrote: 12/19/2019 EPR(Min Pressure),EERS pair = 2(7),6

This completes one night of data at each for the four corners of the center box in the Experimental Design.  I will process data in the next few days.  

In the meantime, this setting still has flow limits, but they appear to be smaller.  I'll start with my data evaluation on Flow Limits.
I did evaluate the data for the Center Square Corners yesterday as promised.  The data collected on the 19th looked like it might be another outlier, so I repeated that combination again on the 20th, last night.  Here are the charts.  The AHI, CA and OA did come back down.  Even so, the CA's are perhaps overstated.  One OA just as I was waking up for a bathroom visit and the 5 CA's in the 30 minutes or so afterwards.  I don't go back to sleep immediately, but I really don't believe all of these 5 events were Non Sleep CA events.  Furthermore, they don't seem to be so clustered.  I remain puzzled about these and how much eliminating them skews the data.  

           

I have been in contact with a statistician that I used while working.  She is going to help me evaluate the data, but not until after Christmas.  So I am holding off publishing any results, but I really don't expect them to be earth shattering.  My situation appears to be poorly suited for this experiment, but we just need to wait and see.  Also, I have not yet evaluated the CA data after first eliminating the "Non Sleep" CAs.  I am under obligation to Fred to do so, especially after he spent so much time helping to identify which events to discard.  I will need to evaluate the events as "count" data, not continuous.  Among other things, I need my statistician for that.  

In the meantime, I'm not sure where to go next.  Some options:

1) Continue checking the four corners of the inner square to get balanced statistics (same number of runs at each corner of the inner square as I have for the corners of the outer square).  Not really necessary in a designed experiment, but maybe useful.  
2) Check some of the edge combinations that have not been checked yet.  See discussion below.
3) Go back to the settings that I was using up until the time I went to my new pulmonologist for a second opinion.  That happened in early October of this year.  It marked the beginning of this journey of discovery:

  1. using the flash card to download my data to Oscar software, 
  2. the Oscar software, 
  3. the amazing resources at this forum, 
  4. finding out that many of my high AHI values reported by MyAir (during almost 2 years of faltering use of my CPAP) were masking the existence of many CA events, 
  5. the possibility of using the EERS to eliminate CA events, 
  6. even the need to eliminate them, 
  7. the abundance of flow limitations observable in my data, and 
  8. the really unusual flow pattern that my data exhibits).  
I was able shortly after my initial consult to transmit to the pulmonologist the report of my PSG, which had been done nearly two years earlier on Nov 22, 2017.  When he had reviewed it, he was very directive (by telephone message) that I not change anything.  Instead I started down this path of using, and experimenting with, the EERS and I've been changing everything.  My next appointment with him is in the middle of April, unless I change it.  In the meantime, I need to develop the case for what my treatment should look like.  Do I need greater "pressure support"?  Will the EERS play any role in my future treatment? Do I need a 2nd, more involved PSG to justify a more expensive Bi-Level Pump? So at this point I do envision going back to the conditions he dictated and collecting a significant number of nights of data using Oscar to really elucidate the inconsistent nature of my CA events and to document the flow limits/flow pattern behavior more completely.  

I should also be able to collect data at a Sweet Spot in my experimental design.  See the bottom of Post #69 for a schematic/picture of the design.  That brings me back to the paragraph above labelled as "2)".  Which edge combination should I begin working with?

Sleeprider wrote in #51
Quote:RE: EERS Experiment Data (sherwoga)
(12-05-2019, 10:28 AM)slowriter Wrote: Wrote:My assumption is that EERS would have no impact on FL, but the pressure settings would.

EERS increases tidal volume, which means the user is breathing more deeply with each breath. That in turn affords a greater opportunity for flow limitation as airway resistance increases with the increased flow rate.  The solution, that is not being used in this case, is pressure support or EPR.  By using pressure support, the larger inspiratory flow rate is supported or made easier, and flow limitation can be reduced.  The whole idea with EERS is to reduce CA to enable the use of pressure support, so I'm a bit surprised the therapy has not moved in that direction.

Sleeprider
Sleeprider is here quoting slowriter from Post #44.  There was a lot of discussion during that period about flow limits and flow patterns, as well as, the minimizing of CA's.  While I don't pretend to fully understand what either is saying, it does appear that they are promoting using the highest EPR setting available to me, which is 3 cm H2O.  So my thinking is that the Sweet Spot, if one exists using my current pump (the ResMed AirSense 10 Autoset) and the EERS, would be at EPR = 3 and EERS set at 6 inches, maybe 12 inches.  The 6 inch EERS seems more promising since Sleeprider advocates against increasing Tidal Volume too much.  I might have 1 or more runs at the 3/6 combination that occurred before I started the experiment, but I can't guarantee that I was using the EERS correctly in the period leading up to the experiment.  I was stumbling around trying to understand the EERS.  So I think one option for tonight and a couple of nights in the near term would be this combination of EPR at 3 and the 6 inch EERS.  It is a combination that Fred Bonjour advocated for right at the start.  If it proves to be a good combination, then my justification to my pulmonologist for a different pump would include a set of multiple night data at this combination to be the comparison for that data generated at the settings that he wanted me to NOT change.  

Sorry for the length of this post.  Any comments are appreciated.
Post Reply Post Reply
#84
RE: EERS Experiment Data - Seq #20 Design #18
(12-21-2019, 05:04 PM)sherwoga Wrote: ... I think one option for tonight and a couple of nights in the near term would be this combination of EPR at 3 and the 6 inch EERS.  It is a combination that Fred Bonjour advocated for right at the start.  If it proves to be a good combination, then my justification to my pulmonologist for a different pump would include a set of multiple night data at this combination to be the comparison for that data generated at the settings that he wanted me to NOT change.

I was assuming earlier the above; that you would need EPR 3 to get the FL as low as possible, and that six inches of EERS would likely allow you to do that without increasing CAs.

I was also assuming you would want to collect some data at that setting to see how effective it was at reducing your FL. If the answer was "not enough," then the bilevel would be the next step (assuming the EERS works at reducing your CAs, that is).
Caveats: I'm just a patient, with no medical training.
Post Reply Post Reply






Donate to Apnea Board  
#85
RE: EERS Experiment Data - Seq #20 Design #18
(12-21-2019, 05:14 PM)slowriter Wrote:
(12-21-2019, 05:04 PM)sherwoga Wrote: ... I think one option for tonight and a couple of nights in the near term would be this combination of EPR at 3 and the 6 inch EERS.  It is a combination that Fred Bonjour advocated for right at the start.  If it proves to be a good combination, then my justification to my pulmonologist for a different pump would include a set of multiple night data at this combination to be the comparison for that data generated at the settings that he wanted me to NOT change.

I was assuming earlier the above; that you would need EPR 3 to get the FL as low as possible, and that six inches of EERS would likely allow you to do that without increasing CAs.

I was also assuming you would want to collect some data at that setting to see how effective it was at reducing your FL. If the answer was "not enough," then the bilevel would be the next step (assuming the EERS works at reducing your CAs, that is).
I used the setting of 3 for EPR, Pressure Range from 8 to 16 and the 6-inch EERS.  Had a good night and felt pretty good when I got up.  

I did go to bed a bit late and awoke at 4:30 for a trip to the bathroom after only about 5.5 hours.  I did not go back to sleep.  CPAP shows it was in use for a total of 6.5 hours, but I was awake for the last hour.  Only 1 CA event occurred during that last hour.  All event totals are quite low. AHI 1.07, OA 0.46, CA 0.46, and HI 0.15.  Flow limits are not absent but they are low.  Respiration rate never went bonkers, but the two times that it increased do correspond to greater flow limit activity and to poorer flow rate patterns.  Median tidal volume was 740, consistent with my experiment results (more on that later).  

I plan to use these conditions for "several" nights unless I hear convincing input that I should do something different.  After I have accumulated 5 to 10 nights of data I will try to summarize it. 

In the meantime, I will get to interact with my statistician and perhaps be able to first publish data from the experiment.
Post Reply Post Reply
#86
RE: EERS Experiment Data (sherwoga)
It's time for an update on progress.  I just want to keep those that have been interested in this project informed that I am still very much committed to keeping my word about this effort.  

Original 17 nights of data for the designed experiment were acquired between Nov 29 and Dec 15.  This included one repeated night to improve statistics.  These nights correspond to 4 acquisitions at each of the four corners of the design surface represented schematically here.The four corners are marked with "pluses".  

   

I've done some data analysis, but waited until after the holidays to get together with my statistician, Martha H, who will help me with that task.  We met Thursday, Jan 2, for the first time.

An important part of any designed experiment is "checking" the center point of the design.  For that I would need to run a EPR, EERS settings combination of 2.5 cm H2O and 9 inches of EERS.  That is not technically possible for this design because my CPAP has EPR settings limited to the whole numbers 0 (EPR off), 1, 2, and 3.  And, the  Corr-A-Flex II tubing used to make the EERS comes in 6 inch segments, limiting the settings of EERS to 0 (no EERS), 6, 12, and 18.  

I had intuitively determined that the next best thing to checking a center point would be to check the combinations at the corners of the inner square of the designed experiment (see schematic above).  These corners, marked with open O's, correspond to EPR(Min Pressure), EERS combinations of 

1(6), 6; 
2(7), 6; 
1(6),12; and 
2(7),12.

I completed those four nights of data with one repeat (total of 5 nights) between December 16 and 21.  At that point I really felt any further work should wait on the statistics input from Martha. 

On Jan 2, I explained the experiment executed so far to Martha, she agreed that a center point does not exist.  But she was also very quick to point out that the average of all my data (for any given output) for the setting combinations at the four corners of the inner square is in fact equal to checking the center point.  While that was an expansion of my thinking (because it will be averages that I compare) it also suggests that it would be better to base the average for the inner square on the same number of nights of data as I am using for the original designed experiment.  So I am committed to acquiring another 12 nights of data for those inner-square corners (three more nights at each of the four combinations).

Therefore, starting tonight, I will be collecting an additional 12 nights of data at the inner square corners, postponing any further data analysis. 

Also, some of my preliminary data analysis put in question the data acquired for one of the outer corners of the design schematic.  So I also plan to collect more data (up to four nights) at that point. I collected the first of those last night and will randomize the other 3 with the 12 referred to above.  So I suspect the next milestone in my progress to occur on Jan 19.  

Starting on Dec 22, I have been collecting data I plan to use the next time I see my pulmonologist that is related to, but not strictly a part of this designed experiment. I will not comment further on that data in this update.  

Any input on this plan is welcome, but I primarily want to, as stated above, assure you that I'm still working on this.  In the meantime, I hope you all had a wonderful (well slept) Christmas and New Year and I look forward to more dialog with you.
Post Reply Post Reply
#87
RE: EERS Experiment Data (sherwoga)
Interesting stuff and you are certainly diligent in finding significance in your data. A bilevel would be nice to have access to for this. What is your impression so far? Has EERS been a worthwhile pursuit to reduce CA for you?
Sleeprider
Apnea Board Moderator
www.ApneaBoard.com

____________________________________________
Download OSCAR Software
Soft Cervical Collar
Optimizing Therapy
Organize your OSCAR Charts
Attaching Files
Mask Primer
How To Deal With Equipment Supplier


INFORMATION ON APNEA BOARD FORUMS OR ON APNEABOARD.COM SHOULD NOT BE CONSIDERED AS MEDICAL ADVICE. ALWAYS SEEK THE ADVICE OF A PHYSICIAN BEFORE SEEKING TREATMENT FOR MEDICAL CONDITIONS, INCLUDING SLEEP APNEA. INFORMATION POSTED ON THE APNEA BOARD WEB SITE AND FORUMS ARE PERSONAL OPINION ONLY AND NOT NECESSARILY A STATEMENT OF FACT.
Post Reply Post Reply






Donate to Apnea Board  
#88
RE: EERS Experiment Data (sherwoga)
(01-04-2020, 05:52 PM)Sleeprider Wrote: Interesting stuff and you are certainly diligent in finding significance in your data.  A bilevel would be nice to have access to for this.  What is your impression so far?  Has EERS been a worthwhile pursuit to reduce CA for you?

Yes, I believe EERS has reduced my CA significantly, but it doesn't seem to help with flow limitations.  The data I collected between the 12/22/2019 and 1/2/2020 was all at 6 inches of EERS, EPR at 3, and minimum pressure at 8.  This point is on one of the edges of the design, but is where I think the sweet spot will be.  That characterization is based more on intuition than on data analysis, so I may have to adjust later on.  During that period, my AHI never went over 5.  That data serves at least two purposes.  Fred Bonjour had requested that data as any other user of the EERS would likely start with the shorter length.  And, I plan on sharing that data along side data at the settings most recently prescribed by my pulmonologist, where the CA index was inconsistent, but often quite high (as high as 19 in one case).  And the AHI was rarely below 5.  

As to whether the EERS is the best solution for my situation is uncertain.  The Bilevel pump, if it worked to eliminate CA's might also improve my flow patterns and/or reduce flow limits.  Not sure if the latter is necessary, but I continue to be too sleepy through much of my day.  

Thank you for your continuing interest.
Post Reply Post Reply
#89
RE: EERS Experiment Data (sherwoga)
I've been assuming you'd want to move to a bilevel, and that EERS would allow you to do that.
Caveats: I'm just a patient, with no medical training.
Post Reply Post Reply
#90
RE: EERS Experiment Data (sherwoga)
PRELIMINARY REPORT OF Designed EERS/EPR Experiment Results

I finished acquiring the data for the EERS/EPR Designed Experiment on the night of Jan 18.  Since then I have been analyzing the data.  While the process is probably not done, I do have some results to report.  My purpose in sharing this preliminary report is to generate dialog with anyone interested, that dialog probably helping to direct my efforts forward.  

By way of what I hope is a brief review: I varied the EPR setting on my CPAP and the length of EERS tubing.  The EPR settings were limited to the whole numbers 0 (EPR set to off), 1, 2, and 3 cm H2O.  The EERS was either not used at all corresponding to an EERS volume of zero or was constructed using 6, 12, and 18 inch lengths of 22 mm diameter tubing.  These lengths of tubing correspond to about 0, 58, 116, and 174 cc (or mL) of volume inserted right at my mask in the CPAP delivery hose that it not swept by flowing air. Those numbers are not exact but will suffice for now.  They represent the approximate amount of exhaled air (used air) from one breath that I re-breathed in the next breath taken when the corresponding length of EERS tubing is in my delivery hose.  

EPR and EERS tubing length were the only two factors in my experiment and since each factor could take on 4 values, there are 16 possible combinations of the factors.  I've shared the attached figure before, but share it again here for convenience and with minor modifications for clarity. Each point on the diagram represents one possible combination of EPR and EERS settings, referred to hereafter as a "Treatment Combinations" or TCs.  

   

In a designed experiment you want to be as aggressive as possible in order to learn as much as you can for the effort you put into the experiment.  Doing so may mean you can avoid taking data at many TCs or perhaps learn enough to know which of those "other" TCs are best to test.  

For this situation the "aggressive" approach corresponded to making measurements at the four corners of the diagram marked with "black" circles.  I acquired data four times at each of those four TCs for a total of 16 nights of acquired data in the original design.  (In the interest of scientific integrity, I admit that I acquired an extra four nights of data for one TC that had very large variability.  I ended up with a minor reduction in variability that did not change results using the last four nights and rejecting all of the first four nights.  There was one other run rejected for cause and repeated.)

You also want to make measurements at the center point of the factor diagram to determine if the response of an output parameter to the factors is linear or curved.  Unfortunately, that center point TC does not exist for my equipment.  It would have been with an EPR of 1.5 and a EERS tube length of 9 inches.  I could use neither.  In consulting with a statistician, I learned I could make measurements at the four interior points on the diagram marked with "white" circles and take the average as the next best possibility to measuring the center point.  So I spent another 16 nights acquiring four sets of data at each of these interior TCs. (Again there was one night of data rejected for cause and replaced with a repeated run.)

Outputs With No Significant Trends (effects) 
 
1.      Obstructive Apnea Index
2.      Median Pressure
3.      Corrected AHI (All central apnea events when I was clearly or quite probably awake or waking up were discounted manually.  Fred Bounjour helped with this process.)

Outputs With Trends

4.      Median Tidal Volume

TV increases with both EERS tube length and EPR, but more so for the EERS Tube length.  The trends also appear to be curved, not linear.  The result for EERS is probably the most significant and will probably surprise no one as multiple moderators have indicated that TV will go up with EERS.  Is anyone surprised that TV also goes up with EPR?  

5.      Median Minute Ventilation

MV increases with EERS tube length only.  Effect for EPR did not pass the statistics tests.

6.      95% Flow Limits

Median Flow Limits were 0.00 for all TCs, but the 95% numbers were usable.  I found that 95% flow limits decrease as EPR increases.  No effect from EERS tube length.  Conclusion: I need greater EPR to minimize flow limits, which is what I think moderators have been saying all along. 

I have not analyzed Maximum Flow Limits yet, as I have assumed they would be much more subject to random variation.  Perhaps I should test that assumption by doing the analysis.  I think Slowriter might agree.

7.      Median and 95% EPAP

Both Median and 95% EPAP decreased with increasing EPR.  EERS tube length had no effect on either.  The surprise is how perfectly these outputs behaved.  When plotted, the EPAP value vs the corresponding EPR setting produced almost perfect straight lines with negative slopes and with linear regression coeffieicnts (R^2) equal to 0.996 and 0.976 for the median EPAP and 95% EPAP, respectively. Furthermore, the slope was more negative for the 95% EPAP measurements than for the median EPAP measurements.  See graphs from Excel.  

           

For those uninitiated in statistics, this is remarkable linearity implying excellent system performance and measurements (which I suspect is entirely a function of the CPAP and OSCAR, i.e., my erratic breathing had no impact on median and 95% EPAP).  The change in the slope is also noteworthy.  One possible explanation might be related to changing vent rate with line pressure.  FoxFire shared the manufacturer's graph of Vent Flow Rate vs Mask Pressure in Post #38 of the parent thread to this one, Fabrication Of An Enhanced Expiratory Breathing Space (EERS).  I recommend you take a minute to look at that graph.  

If there is some other abundantly obvious explanation, please enlighten me.  Or if you can articulate the impact of line vent rate on EPAP dependency on EPR setting better than I have, please do. 

Further, my mask pressure is nowhere near consistent.  Are Median and 95% EPAP independent of mask pressure?  The linearity observed for EPAP measures of population data seems to say that they are.  If so, why Huhsign ?


Some related observations:

I carefully checked the Minimum EPAP, as well, as if it were a reasonable output to test and found that it was 5 for all TCs. Hence the variation, both random and nonrandom, was zero.  In hindsight this makes sense as I was told to keep the EPR and minimum pressure settings on my CPAP in pairs for which (Minimum Pressure – EPR setting) = 5.   Minimum EPAP had to be 5 by definition for all TCs.  The exercise admittedly served as a revelation for me, a revelation that I should perhaps have anticipated. The exercise also served to test my systems.  First and foremost, this non result vindicated my procedures and execution of acquiring my data.

I did a similar exercise with what OSCAR reported for the minimum pressure for each TC and again found vindication of my systems but no surprising result.  The minimum pressure went up very linearly with EPR setting and again this is function of the TC settings for minimum pressure in combination with the paired EPR settings.  EERS tube length had no effect. 


8.      Corrected CAI

I eliminated clear airway events that definitely or very probably occurred while I was awake. Doing so left CAI counts that were very small for all TCs suggesting that there would be nothing left but random variation.  (I feel also that doing so impugns the integrity of the measurement system.  The decision to discount an event was sometimes iffy!) There was, however, still a statistically significant trend of decreasing CAI with increasing EERS tube length.   Conclusion: I want some EERS, but the benefit on CAI will be very small since I’m already seeing such low CAI’s anyway.  The need for the EERS would be dictated by it's impact on Flow Limitations and/or on the shape of the flow patterns.  The latter is at this point a very subjective output for which I have no quantitative data.  

This is an area for further discussion.  In post #37, JoeyWallaby gave an excellent discussion of EPR and PS in the context of using the EERS.  Others, especially Slowriter, added to the discussion in subsequent posts.  If I understood JoeyWallaby correctly, one experiment I could do is to incrementally increase my minimum pressure setting (without EERS) to see if and at what minimum pressure I experience the "significant" onset of CA events, say consistently greater than 5.  Then once I’ve created the problem with increased minimum pressure, see if I could add the 6 inch EERS and again eliminate the CA events. I would of course want to monitor both OA events and Flow Limitations throughout such an experiment and be prepared to abandon the experiment if necessary.  One reason I see for doing this experiment is that it might help develop the case that I really do need a bi-level or ASV pump.  Thoughts are welcome!
  
9.      Uncorrected AHI

For this output I took the AHI readings as reported by OSCAR, i.e., I did not eliminate any CA events.  I surprisingly do see a trend.  Uncorrected AHI decreased with increasing EERS tube length.  This again suggests the need from some EERS. But the result doesn’t seem consistent with the trends seen for Corrected CAI and the lack of trend seen for Corrected AHI.  Very puzzled here  Huh Thinking-about .

sherwoga
Post Reply Post Reply






Donate to Apnea Board  


Possibly Related Threads...
Thread Author Replies Views Last Post
  Fragmented Sleep, FLs, Suspected UARS, REM Related. 5 Months of Optimization + EERS. pineh 9 411 08-19-2020, 08:23 AM
Last Post: Sleeprider
  Not Diagnosed, But Let The Experiment Begin rsandhu 17 533 08-04-2020, 11:10 AM
Last Post: ar14
  Viewing Old Data Without Importing New data from an SD Card mjrPdqkLguUgGLE 5 367 05-23-2020, 05:18 PM
Last Post: mjrPdqkLguUgGLE
  Dreamstation CPAP vs Aircurve BiPap data? Data over time for Insp/Exp? Desiree 12 675 05-20-2020, 10:00 PM
Last Post: Sleeprider
Question New here, loaded data to OSCAR - only summary data swampbastard99 5 535 03-10-2020, 01:43 PM
Last Post: Crimson Nape
Exclaimation [Health] Interpreting sleep study results, EERS enhanced expiratory rebreathing space JoeyWallaby 221 8,378 12-29-2019, 08:12 AM
Last Post: JoeyWallaby
  hey TORONTO! I need Sleep Dr for UARS, FL's, EERS etc. cary1 5 385 12-23-2019, 04:42 PM
Last Post: Sleeprider


New Posts   Today's Posts






About Apnea Board

Apnea Board is an educational web site designed to empower Sleep Apnea patients.