Leak rate Respironics - Fisher combo

I am using Fisher&Paykel HC431 with Respironics M-series CPAP set for 15 cm.

The Fisher user instructions suggest that the "leak rate" (they refer to it as Bias Flow) at 15 cm should be 48 liters/min.

But the Encore Pro (1.7.128) reports an "Average leak" of 35 liters/min. for the whole night. How can that be?

Am I misinterpreting something?

You are not missing anything. Just how odd Respironics math can be at time. My mask at my average pressure of 11 and leak rate of I forget exactly, around 35, I often get total leak showing 28 L/min...

Just look at the leak line which is probably pretty straight and not worry about it. The suggested vent rate for any mask always has a plus or minus to it and I suspect that if we dug deep enough Respironics leak reporting also would have a plus or minus to it.
Enough plus or minus and we get something that doesn't make sense.

dataq1 wrote:Am I misinterpreting something?

The leak rate number can be arbutary. The exaust holes in the mask are not always mthat exact but the way the machine reads it is not exact either. I have three machines I alternate each month or so. Using the same mask and hose one machine gives me a reading of 25 l/s, the second machine runs 30 l/s and the third gives me readings of 32 l/s. The important thing is how much it varies from the norm. Watch the graph in EP for leaks which will show as peaks above the standard line and try to keep this to a minimum and you should be good no mater what the exact number is. I am using EP 1.8i.49. Yours is off a bit more than I normally see but since it is lower it should be okay. Apparently leak calibration is not an exact science.

Jerry

Thanks everyone, the take away from this is that the leak rate calibration may be a bit off- I shouldn't be concerned about the "baseline"

What I'm trying to reconcile is why sometimes leaks don't show up on the detail leak graph. That is, how significant must a leak be to be reported/detected. The reason I'm interested in that is because I believe that leaks can cause inaccuracy in event (OA/H) detection.

OK, so last night was a good night, however I woke up early because my full face mask (Fisher&Paykel HC431) was leaking between my cheek and the silicone seal. I refer to this leak as a 'flutter' as it vibrates as it "passes gas".

It let this go on for about 10 minutes (because I wanted to see if it would show in the report for the day). Download the data and the leak rate for that time (when the leak woke me) was still about 40 l/m and more importantly no greater than it had been for the prior 4 hours.

Ok, that would suggest to me that this 'flutter' is is fairly insignificant as it applies to the leak rate (in spite of what it "feels" like... fairly large volume of air escaping at the leak site.

Now, four hours earlier the leak rate graph displays a period of leak lasting about 15 minutes, with a peak at 75 l/m.

So, if the leak that woke me up is insignificant (that is was not detected/reported) then earlier leak that peaked at 75 must have been significantly greater than the 'flutter' that wakes me.
If that is the case, how significant a leak is required before you would start to view OA events skeptically?
[img][IMG]http://i172.photobucket.com/albums/w3/d ... -21-11.jpg[/img][/img]

The leak on that report you have posted (I got to it using the url addy since you have extra IMG brackets, it doesn't show active in your post, remove one of each of the opening and closing IMG brackets and it should work)

Is very short termed. If you see events during that time it could be that the leak was big enough to let them sneak by. We have never been given a certain number where it is definite that the leak has impaired therapy. We are just told "large leak" and Respironics actually have a very generous large leak definition of above 80 it seems. I have seen 80 on my reports and it not be flagged as large leak but 100 does. Go figure.

Your overall leak line is quite good and I wouldn't worry about those little blips that show up or even don't show up on the reports.
We can't see close enough into the minutes to see what happened at that time.

Pugsy wrote:The leak on that report you have posted (I got to it using the url addy since you have extra IMG brackets, it doesn't show active in your post, remove one of each of the opening and closing IMG brackets and it should work)

Thanks Pugsy, I'll get the hang of it soon!

Am I correct in my understanding that leaks might affect the reliability of the Respironics algorithms to properly detect OA and H events? And if so, does that only apply to large leaks (whatever Respironics considers a "large" leak)?

Yes, I agree that the picture I posted would be considered a pretty "good" night as far as minimum of leaks, looks like a pretty good seal. But what I'm befuddled with is why the "fluttering" of the seal doesn't seem to show up as a leak of some significance. (These flutter events are sufficiently loud so as to wake my wife, who elbows me until I get the mask resealed -LOL).

The encore pro software is all that I have, so I can't really look as closely "minute to minute". However I do seem to recall that there is a way of extracting data from the SQL database so that it can be examined much more closely. I just can't remember how to do that right now.

The fluttering is simply too small and too short of duration to make any impact that we can see on the reports. Sometimes I wish Respironics had given us a "line in the sand" like ResMed does, but all we can go on is more what we see. Large leaks are really the only thing to worry about, those minor flutters that don't show up and even an occasional blip to say 50 L/min won't impact the ability to respond to events or score them. Large leaks never being fully defined of course but I have seen 75 and 80 and it not be flagged as large leak. We have to remember that we also have to subtract the vent rate to. Then we have to add in the overall inconsistencies with Respironics math and how they report the leak rate. I have seen an overall leak average (quite straight line) be stated at 28 L/min and the vent rate for my mask alone is 34 L/min so go figure that one. Best thing is to look at the line in general and if a lot of events occur at the same time, then it might be big enough leak to impact data, of course it could also be that the events caused the leak too. We just don't know for sure which came first.

Mr Skinner's tool might be what you are thinking about getting some extra data. I used to use it but I haven't in a long time.
It will work with Pro 1.8 but not the later versions of Pro.

Encore Pro Analyzer. I don't have an active link. Might be one somewhere in one of the posts here. I don't remember what data we could pull from it.

Pugsy wrote:Mr Skinner's tool might be what you are thinking about getting some extra data. I used to use it but I haven't in a long time.

Actually Pugsy, I have used EncorePro Analyzer, but I was referring to actually extracting data from the SQL server. James Skinners program is very nice for displaying the statistics, but does not reveal the raw data. What I mean by that is (for example), a graph that will detail exactly what the leak rate was at 2:10:30 am, then what the leak rate 30 seconds later, etc.

So I spent some time yesterday and rediscovered how to get that kind of data. It's a bit complicated, but anyone who knows how to work in excel should be able to obtain that data. I would explain further, if anyone is interested.

dataq1 wrote: Actually Pugsy, I have used EncorePro Analyzer, but I was referring to actually extracting data from the SQL server. James Skinners program is very nice for displaying the statistics, but does not reveal the raw data. What I mean by that is (for example), a graph that will detail exactly what the leak rate was at 2:10:30 am, then what the leak rate 30 seconds later, etc.

So I spent some time yesterday and rediscovered how to get that kind of data. It's a bit complicated, but anyone who knows how to work in excel should be able to obtain that data. I would explain further, if anyone is interested.

I have no clue how to use Excel. Never bothered to learn because never needed it. I can see where some people like to dig deep into the reports and that is fine especially if they are of the mind set to do it. I am a bit lazy when it comes to all that. I did use Skinners tool for a while to play around but that is pretty much where it ended. Breaking it down the way you have into minute seconds might answer the question I have always had. Which comes first the event causing the pressure spike which might cause a leak or the leak which might let an event sneak by?

I might want to play with it but if Excel is hard to learn.... I don't know if I can handle two complicated tasks at once. In fact I am not sure I even have Excel on this computer....

[quote="Pugsy Breaking it down the way you have into minute seconds might answer the question I have always had. Which comes first the event causing the pressure spike which might cause a leak or the leak which might let an event sneak by?
[/quote]

If I get a chance, maybe I can look at some of my traces and see if it's possible to resolve that question

Cheers
-bob

Pugsy wrote:Encore Pro Analyzer. I don't have an active link. Might be one somewhere in one of the posts here.

http://james.istop.com/EncoreProAnalyzer/