How can it be?

[BleepingBeauty]

Last night was my first decent night's sleep since my surgery on the 11th. Here's my data for the night:



I love seeing a nice, flat leak line. But my question is this: How can my machine record a lower leak rate than the intentional flow rate of the mask (i.e., less than zero)? I'm using the Hybrid at 12cm, and the leak rate should be 42 lpm.

Considering Respironics reports leak in increments of 7 lpm, I can see that most of my night was spent right at 39 lpm, so I understand why my average leak is reported as 38.93. But with regard to all those dips down to 32 lpm, I just don't understand how the mask can leak less than what it's designed for and am wondering if this is just another example of Respironics' screwball math.

I'm certainly not complaining. Just scratching my head over this... Any input is appreciated.

[FatiguedMe]

Hey BB!
Good to see you back! You prolly been back already and I missed it! Love ya, miss ya and glad to see you are doing well.
I can't help here haha as this technology is over and under my head LOL I'll just join ya in scratchin and see what others have to say!
Love ya
Susan
P.S. If you feel up to it, can you look at my post Friday night report? How are you getting your report to show up big? Help me please?!?

[Muse-Inc]

I'm interested too! My Hybrid vent rate is higher than any ResMed, so I use their highest and expect to see a reported leak matching the vent rate overage, but my reported leak is often close to 0 much less than I expect! Makes no sense to me scratching head along with the others. Is the vent rate for masks that off, er, variable? Doubt it's the machines.

[plr66]

I'm not sure what the answer is here, but when it's happened to me, I suspected it meant I had blocked the exhale vent to some degree or another (not good) making the intentional leak rate less than it should have been.

[cinco777]

I just don't understand how the mask can leak less than what it's designed for and am wondering if this is just another example of Respironics' screwball math.

The EncorePro (EP) Average Leak # on the Sleep Therapy Daily Details report is sourced from the 30-second time-stamped (T-S) Leak Rates written by the Respironics machines to the SmartCard (and which EP stores in an SQL table for that type of machine: CPAP, APAP, ...). This EP Average Leak # is a weighted average of all the T-S records for a sleep session. The weighted Average Leak calculation is correct, therefore, the EncorePro Math is Correct.

Please note that I make a differentiation between "Respironics Math" and "EP Math". For me, "Respironics Math" is only used for the machine's LCD reported numbers. In prior posts, I reported that the Respironics LCD math is accurate within reporting granularity. For me, Respironics "Machine Design" determines which Leak Rate is recorded to the 30-second Time-Stamped records that it writes to the SmartCard. This "Design" is where the lower than intentional mask leak rates are originating.

I have observed Leak Rates that are less than my intentional leak rate of 33 LPM at 8 cmH2O for my Flexifit FF #431 mask. These lower than 33 LPM leak rates (32 LPM or 25 LPM) are reflected in some of the T-S 30-second records. WHY the machine records these 32 LPM and 25 LPM leak rates for my mask is unclear. My earlier speculation/theory was that our machines did this to represent leak rates between the 7-apart granular leak rate values of 18, 25, 32, 39, 46, ... I have not developed any new theories for this phenomenon in recent months. I suspect that Velbor will also have some comments on this subject of Leak Rates being recorded/reported that are less than the mask's intentional leak rate. BTW, the intentional mask leak rate has its own variability based on manufacturing tolerances (vent holes in the mask), and whether any of these vent holes are covered as a result of the mask being pressed against a pillow, bed covering, etc.

[carbonman]

I just don't understand how the mask can leak less than what it's designed for and am wondering if this is just another example of Respironics' screwball math.

I think those are sampling or search markers.
Nothing to worry about.

Simply the doped silicon intelligence expecting to expect something.....
when there is nothing there to expect.

Don't worry, be happy!

[Wulfman]

Hi BB.

Welcome back.

My take on the "lower-then-intentional" flow rate is that individual machines can vary somewhat in their recording characteristics. One of my machines (newest firmware version) gives me leak rates in the lower 30's and the rest of the machines are in the high 30's to low 40's (which is about what it SHOULD be).

The little "zipper" deviations in your leak line have been discussed before, but nobody (to my recollection) has had a concrete explanation for them. Mine almost always look like yours from last night. To me, it's just "normal" for the Respironics machines when viewed in the Encore programs.

Here's an interesting one that "jules" posted a year ago.

viewtopic/t39186/viewtopic.php?p=313405#p313405


Den

[BleepingBeauty]

Thanks for the responses, everyone.

Fatigued, I'll PM you when I get back from the surgeon's office, in a few hours. (Second post-op appointment is this afternoon, and I'll be leaving soon. Hope to be rid of the sling today!)

I'm not sure what the answer is here, but when it's happened to me, I suspected it meant I had blocked the exhale vent to some degree or another (not good) making the intentional leak rate less than it should have been.

I can appreciate that possibility, but not in my case, as I've only been able to sleep on my back since the surgery. (Hating it, too...)

cinco, thanks for sharing what you've learned. I don't have a head for statistics like you do, so I'm glad you're here and can offer your input on questions like mine.

Don't worry, be happy!

No worries, C-man. If my therapy data looked like what I posted above every night, I'd be a grinnin' fool all the time.

Hi BB.

Welcome back.

My take on the "lower-then-intentional" flow rate is that individual machines can vary somewhat in their recording characteristics. One of my machines (newest firmware version) gives me leak rates in the lower 30's and the rest of the machines are in the high 30's to low 40's (which is about what it SHOULD be).

The little "zipper" deviations in your leak line have been discussed before, but nobody (to my recollection) has had a concrete explanation for them. Mine almost always look like yours from last night. To me, it's just "normal" for the Respironics machines when viewed in the Encore programs.

Here's an interesting one that "jules" posted a year ago.

viewtopic/t39186/viewtopic.php?p=313405#p313405


Den

Hi, Den, and thanks.

I'm happy to be in your good company, with our similar leak lines. My leak is almost always pretty good, but my AHI sure does vary (and rarely is this low). I'll be thrilled if my therapy eventually becomes as consistently effective as yours obviously is.

[DoriC]

BB, nice to see you and your lovely leak line back again. Our own MSeries Auto/Aflex machine always reports leak rates in the mid to high 40s and our gently used backup, MSeries Auto/Cflex reports 40-42 which is what it should be. Naturally that's the one we're using now. So that confirms Den's "take" on the question.

[Velbor]

My earlier speculation/theory was that our machines did this to represent leak rates between the 7-apart granular leak rate values of 18, 25, 32, 39, 46, ... I have not developed any new theories for this phenomenon in recent months. I suspect that Velbor will also have some comments on this subject of Leak Rates being recorded/reported that are less than the mask's intentional leak rate. BTW, the intentional mask leak rate has its own variability based on manufacturing tolerances (vent holes in the mask), and whether any of these vent holes are covered as a result of the mask being pressed against a pillow, bed covering, etc.

Reporting granularity, sensor variability, mask variability and sleeping variability, as have been suggested, seem like more than enough!

Another possible contributor is breathing variability, in terms of depth and rate, and synchronization with the 30-second data reporting periods. Respironics "leak" reflects total system airflow, but to the best that we know, it is reported as some sort of "average". But what is the actual "data sampling" paradigm - what is the the machine taking an average of, and how often, and with what granularity (which need not necessarily be the same as the reporting granularity)? There has to be some finite "data interval." Every second? Every few seconds? Every tenth of a second? Not a clue. BUT, consider that we are inhaling and exhaling about 500cc, about every 6 seconds, though not linearly. The inhalation phase, when system pressure will drop slightly and airflow will be INCREASED to compensate, is shorter but steeper than the exhalation phase, when pressure will go up slightly and airflow will be slightly DECREASED. Over time it all "averages" out, but what the machine records for any particular "data interval" and what it reports as an "average" for the 30-second window, just might occasionally fluctuate. Yawning could produce significantly greater "instantaneous" flow changes. How all of this might affect leak reporting is unclear. Interesting, but almost certainly irrelevant. Just guessing.

Nice catch, though, BB, and Cinco777, in observing and noting this phenomenon.

[brain_cloud]

I love seeing a nice, flat leak line. But my question is this: How can my machine record a lower leak rate than the intentional flow rate of the mask (i.e., less than zero)? I'm using the Hybrid at 12cm, and the leak rate should be 42 lpm.

If you have a different mask with a known leak rate laying (lying?) around, you could do a short awake test session to see what the data is like for that mask. This will allow you to prune the branches of the theory tree a bit.