Question on Average 90% Leak for REMStar M Series

Can someone please help my understanding of the summary figures on an Encore report. Looking at the summary page I pretty much understand what the green average leak line is. I think I also understand the red maximum leak, which is the largest leak rate experienced on a particular night (averaged across the report period)? Feel free to confirm! I don't think I quite understand 90% leak or average 90% leak. Is that by any chance the average leak for 90% of the time under treatment? I couldn't locate this in my help screens. Thanks in advance!

why waste your time on this? - the leak graphs are the important thing on the daily report

Well, mostly because my medical provider seems to care about that page. I'd just like to know what it is. I assume you don't know either.

Naturally, I use the daily reports.

OK. I'd love it if you would explain, but I'll eventually find out if you'd rather not. Please accept my apology. I'm not spoiling for a fight.

Scarlet834 wrote:Can someone please help my understanding of the summary figures on an Encore report. Looking at the summary page I pretty much understand what the green average leak line is. I think I also understand the red maximum leak, which is the largest leak rate experienced on a particular night averaged across the report period)? Feel free to confirm! I don't think I quite understand 90% leak or average 90% leak. Is that by any chance the average leak for 90% of the time under treatment? I couldn't locate this in my help screens. Thanks in advance!

What is an average to you? I ask because of what you have said above.

Once you can tell me what you understand the average to mean and what you think 90% leak means, we can go from there.

Do you see average 90 % leak on any one day reports?

I just took off on a trip and my reports are not on this laptop, so forgive me if my memory is inaccurate.

No, there is no 90% leak on any 1 day report. Only Large Leak (in minutes, I think) and average leak. When I print a full report from Encore Pro, one of the first pages summarizes various parameters over multiple days in a single graph. There is a graph for CPAP pressure, one for AHI, etc. One of the graphs is for leaks.

So, there is a green line on the graph, a yellow line on the graph, and a red line on the graph. The green line is labeled "average leak" which I assume is across those multiple days. The red line is labeled "maximum leak". Again, I suppose that shows the maximum leak that took place each day across the multiple days. There is a single figure at the right side of the page for each of those figures, which I think is all the leaks in whatever category added up and divided by the number of days. That's what I meant by "average" - the mathematical average of all the days. I admit that I haven't tried to duplicate the math to check.

And the yellow line (at least I remember it as yellow) is labeled "90 % leak". This is the terminology that confuses me, because I don't see it used on the daily graphs. I haven't a clue what it is. I gave you a wild guess in my first post. I really don't know.

I realize this won't make a whit of difference in how I handle my treatment. It just bugs me that I can't figure it out.

Average is sum divided by total instances. Now with respironics version of arithemetic it can be just the average of the x days of data but in all honesty it really should be some kind of weighted average where the nights that you used the machine more get counted more. Duplicating Respironics math (I have posted many times that it sucks) is always interesting - sometimes they get things right - very scary.

A 90 % leak / pressure refers to the largest number you are at or below for 90 % of the session. It is really a 90 %ile. This means that 10 % of the session you were at and above this mark.

Again with respironics math, the average 90% leak is probably just the unweighted average of the 90 % leak values off that graph.

Let me know what else. I was confused in reading you first post that you weren't sure what an average was and I needed to know what level audience to write to.

Got off a plane not long ago and won't be on the forum much this week, but just wanted to take a second to say thank you. What you explained about the 90% is what I suspected but sure couldn't find anywhere!

The Respironics LCD displayed System Leak (L/min) for 7-Day avg and for 30-day avg are simple averages.

The EncorePro (EP) AverageLeak on the Sleep Therapy Daily Details page is a weighted average.

The EncoreProAnalyzer (EPA) 90% Leak is, I believe, a weighted average. I am away from home tonight and will look at this EPA 90% number in more detail tomorrow and report my findings. I don't use this number in monitoring my therapy so I never looked at it in detail.

The EPA Average Leak is a weighted average. EPA uses a somewhat different calculation than EP if there are multiple EP reported sleep sessions.

The EncoreProAnalyzer (EPA) 90% Leak is, I believe, a weighted average. I am away from home tonight and will look at this EPA 90% number in more detail tomorrow and report my findings. I don't use this number in monitoring my therapy so I never looked at it in detail.

I have to retract this statement.

Before I explain why, I want to make an admission. I do not use the 90% Average Leak # from EP or the EPA 90% Leak # in monitoring my sleep therapy. The leak time-line graph has all the information that I need to assess the impact of leaks on my getting a restful night's sleep. The Leak graph provides me a contiguous view of how my leak rate is progressing from the time that I turn my blower on until I turn it off. I can tell, from viewing the leak graph, if I am experiencing leaks that compromise my therapy, if my mask is secure (leak line is flat), if a leak may have preceded or followed a sleep event, how long I was fiddling with my mask, if my mask seal is wearing out and needs replaced, if I need to resume my nightly shave before bedtime regimen, and other leak-related questions that I may want to answer. Others who do not use a FF mask like me can use the leak graph to determine if they are mouth breathing or lip fluttering. Just to be clear, all the associated leak #s are a distant second to the value that I receive from the leak graph.

My belief that the EPA 90% Leak # is a weighted average is correct BUT the EPA 90% Leak # does not fit the definition of what a 90% profile means. The 90% Average Leak # in the EncorePro report, at least during my 7-months reporting history, is reported correctly for single day reporting. It is always reported as one of the set of possible leak values recorded by my Respironics machine: 4, 11, 18, 25, 32, 39, 46, 53, 60, ... (all numbers 7 apart starting with a leak value of 4). Please note that my EP report show this number with one decimal which is incorrect formatting as all possible leak values are integers (no decimal points). A 90% profile is, by definition, the amount of time - 90% - spent at or less than a determined value, a value that is possible, does occur, and is measurable and recordable. For leak purposes, the 90% number must be one of the set of possible leak values: 4, 11, ... EPA for a single session will report the same 90% number as EP. For multiple sessions, EPA will average the 90% Leak #s from EP in an undermined way and end up with #s that have decimals and vary from the possible leak values (4, 11, 18, ..), often by one or more digits. Proclaiming a 90% Leak profile # that is not possible does not fit the definition of a 90% profile. Simply put, how do you spend 90% of your time less than or equal to a Leak value that does not exist?

Addendum: EP reports for multiple days will also average the 90% Leak #s in an undermined way, leading to Average 90% leak #s with decimal points. Please note that I have multiple day reports only for my vacation days as I use multiple smartcards each day while at home, one for each of my sleep sessions.

I hope this explanation addresses some of the confusion surrounding this 90% leak #. I don't use it. If someone wants to use it, please be aware of its limitations. I welcome your questions and comments on this explanation, especially if you have differing views on its value.