Input Needed for New Encore Pro Analyzer

Guest · Post by **Guest** » Wed Nov 15, 2006 9:21 am

I’ve been helping James Skinner review and debug his new Encore Pro Analyzer, a successor to Derek’s My Encore which works with Respironics new M Series CPAPs. In the course of this review, we’ve come across two questions which we’d like to throw out to the cpaptalk community.

1. Does anyone know whether the reported time in apnea in Respironics' Encore Pro software includes only time in non responsive and obstructive apneas, or does it also include time in hypopneas?

2. How should we (actually, James) calculate multi-day averages. For example, when calculating an average AHI for a week, should he average the daily AHIs or, alternatively, sum all of the the apnea/hypopnea events for the week and divide by the total number of hours slept. We think the latter method more accurately describes what happened during the week (nights in which there were only a few hours slept don’t distort the overall average), but would like to know what others think.

So, please let us know your thoughts by either responding to this thread or PMing me directly.

greyhound · Post by **greyhound** » Wed Nov 15, 2006 9:23 am

That was me, if you want to PM a response.

DreamStalker · Post by **DreamStalker** » Wed Nov 15, 2006 10:08 am

Do not have an answer to question number 1 but as for number 2:

The arithmetic mean (average) should be about the same calculated either way.

I guess it all depends on semantics ... calculate according to 1) if you want to calculate average AHI for last 50 hours or 60 hours or 80 hours, etc. ... or 2) whether you wish to calculate average AHI for last 5 days or last week or last 10 days, etc.

Distortion of the overall "average(s)" for short sleep sessions or naps does not, as you said, occur. However, the interpretation of those averages may be biased towards the lower end in that apneas are most common during REM stage which may often not be achieved during short sleep/nap periods.

ozij · Post by **ozij** » Wed Nov 15, 2006 12:48 pm

For question no. 2: The only way to keep short sessions from skewing the results, is to go back to summing apneas (not AHI) and dividing the overall total of apneas by overall total of hours spend.

If you sum AHIs , and give an AHI=5 from a one hour session the same weight you give an AHI=1 from an 8 hour sessions - you'll be getting skewed results. A session average will be ((5+1)/2 will give an AHI=3, however a total apnea divided by total hours (5*1 + 1*8 )/(1+8 ) =13/9 = 1.44

So simply averaging AHI's on a session basis is not a good idea, especially or people who - for instance - take a daily nap.

I use Silverlining, and export the data to Excel. when I want to know my trends, and how they are affected by pressure, I instruct my Excel chart to exclude any session shorter than 1 hour.

O.

_________________

CPAPopedia Keywords Contained In This Post (Click For Definition): AHI

greyhound · Post by **greyhound** » Wed Nov 15, 2006 1:18 pm

Thanks, Ozji. That's what I was trying to describe as my "alternative" (and preferred) calculation methodology. Even ignoring naps, it seems to better describe overall AHI if some evenings you slept 6 hours and others 9 hours.

ozij wrote:For question no. 2: The only way to keep short sessions from skewing the results, is to go back to summing apneas (not AHI) and dividing the overall total of apneas by overall total of hours spend.

CPAPopedia Keywords Contained In This Post (Click For Definition): AHI

jskinner · Post by **jskinner** » Wed Nov 15, 2006 4:12 pm

ozij wrote:For question no. 2: The only way to keep short sessions from skewing the results, is to go back to summing apneas (not AHI) and dividing the overall total of apneas by overall total of hours spend.

I'm pretty much convinced now that this is the proper way to go. Unless there are convincing arguments to the contrary I will convert the code to use this method for the next release.

greyhound · Post by **greyhound** » Wed Nov 15, 2006 4:17 pm

Now, if there was only some way of confirming that Encore Pro records only time in non responsive apneas and obstructive apneas as time spent in apnea. In other words, the time spent in apnea does not include hypopnea time.

Anyone know for sure?

jskinner · Post by **jskinner** » Wed Nov 15, 2006 4:20 pm

I have another question to add to the discussion list:

3. How long does a night need to be before I include it in an totals & averages. For example I have a couple nights stored that are only a few minutes (maybe I just got a new machine and tried it for a few minutes in the morning while setting it up) I really don't want this to add a day when averaging. For example if I was calculating events/day adding an extra day could make a big difference in the result particularly if I hadn't really slept that day and had no events to add.

At what value should I ignore a day all together. I had currently set it to 4 hours but maybe it should be smaller like under 1 hour? What do you think?

greyhound · Post by **greyhound** » Wed Nov 15, 2006 4:44 pm

Personally, I don't find the information on events per day very useful. Expressed as units per hour, it's very useful (AHI, etc.). Large leak minutes, for example, is not nearly as useful as a large leak minutes, expressed as percentage of the time slept.

To the extent that the multi-day averages are calculated as discussed earlier, the impact of the "small useage" times on these averages (and totals) is likely to be miniscule.

I guess what I'm saying is that there's no harm leaving all the data in, because the totals and averages won't be materially affected.

Ideally, it would be nice to override the date range to exclude specific dates. That would make it easier, for example, to compare days when different masks were used. But that's a different (and less important) issue.

cwsanfor · Post by **cwsanfor** » Wed Nov 15, 2006 7:12 pm

Ideally, it would be nice to override the date range to exclude specific dates. That would make it easier, for example, to compare days when different masks were used. But that's a different (and less important) issue.

This messge constitutes topic and scope creep, but is it reasonable to have a user-entered field that tracked mask type used, and maybe one or more charts such as AHI by mask? I realize this would mean writing to the database, which may not be a good thing, or adding another instance of SQL Server, which sounds like a lot of complexity. Just an idea. I've often generated an Excel sheet to do this, but charts are better.

jskinner · Post by **jskinner** » Wed Nov 15, 2006 7:30 pm

cwsanfor wrote: This messge constitutes topic and scope creep, but is it reasonable to have a user-entered field that tracked mask type used, and maybe one or more charts such as AHI by mask? I realize this would mean writing to the database, which may not be a good thing, or adding another instance of SQL Server, which sounds like a lot of complexity. Just an idea. I've often generated an Excel sheet to do this, but charts are better.

You guys like to keep me working don't you

jskinner · Post by **jskinner** » Wed Nov 15, 2006 7:36 pm

greyhound wrote:Personally, I don't find the information on events per day very useful. Expressed as units per hour, it's very useful (AHI, etc.).

The reason I like events per day chart is that I find for me they give me a better idea of how my night really was. For example if I slept for 6 hours at an AHI of 5 then I would have had 30 events. On the other hand if I slept 8 hours at an AHI of 4.5 then I had 36 events. Comparing on AHI alone the first night looks better, however if those events where bad and impacting the way I felt then its easy to see why by looking at the event chart why.

Just a different way to slice the data.

greyhound · Post by **greyhound** » Wed Nov 15, 2006 8:20 pm

jskinner wrote:
At what value should I ignore a day all together. I had currently set it to 4 hours but maybe it should be smaller like under 1 hour? What do you think?

You could always make it a user option.

dllfo · Post by **dllfo** » Wed Nov 15, 2006 9:06 pm

Encore Pro allows us to "dump" a day. If we had a real bad night, due to pain, or a baby crying, it might be nice to toss it out of the mix. It would tend to skew the data with unusual events.

And I use any day in which I slept over 4 hours. UNLESS I had something unusual happening. Construction next door, new puppy yelped all night and so on.

Thanks for all your hard work.....Dave

ozij · Post by **ozij** » Thu Nov 16, 2006 10:46 am

needingu wrote:I have this reading on my Encore data..Is there anything I have missed.
Pressure set at 11-15

OA 0.1 , H 0.8, VS 0.8 AHI 0.9, Large leak 0.0 min, 0.096 of night
Av. leak 30.82

MyEncore..
Apnea duration 5- 20 seconds
Sec. in apnea 137.0
Variable breathing 3.7%
Daily FI (flow linitations) 0.0 lim/hr
Daily Si( snoreindex) 0.9/hr

Compliance (hr. per night) 10.2 hr/da

I think you have an answer in the posted data:
1 apnea, 8 hypopneas total seconds in apnea 137 - there is no way that total could refer to the single apnea.

viewtopic.php?t=14648

O.