Major discrepancy between ResMed and Sleepyhead AHI

Can anyone explain what may be going on here?

Up until today there have been minor differences between the results shown on the dial of my APAP (or on myAir) and the Sleepyhead data. Usually it was likely just differences in rounding, or for a while I chose on my Sleepyhead results to eliminate any sessions shorter than five or fifteen minutes. Then, after I wanted the total on Sleepyhead to show as close as possible to what my doctor would see on the compliance reports, I changed that back to show all sessions.

But last night after the first sleep session, I noticed that the AHI on the dial showed as twenty point something. I made a mental note, as that was considerably worse than most of my results with this new mask in the previous five days. After the second session, the AHI was even worse, twenty six point something. And this morning the total night shows an AHI of 10.2.

But my Sleepyhead report shows something totally different. After the first session, AHI was 12.74, after the second session (so results of both sessions 1 and 2) it was 7.58 and the total night shows as 2.73. There are no events recorded on Sleepyhead at all during the second and third sessions, which doesn't seem at all likely from my past. The total time shown is identical between ResMed and Sleepyhead.

Is Sleepyhead having some problems - or is it just targeting me?

I've included this as a separate topic rather than putting it with my usual thread because it didn't seem that related to the rest, but just in case someone thinks there is something you want to check back about here's the link to the rest of my story. viewtopic.php?t=171740

And here's last night's Sleepyhead report:

Sleepyhead is easy. look at the number of events flagged (and it's just showing flags that are recorded in the data by your machine) and divide by the number of hours.

It's hard to get a good count just by the tick marks, since some are close together and might be more than 2, but the count looks reasonable. for that AHI over the number of hows shown. If you doubt the flags, then zoom in and look at the breaths, and look for any flatlines that are ten seconds or longer.

As to what the machine/myair says? who knows what it's counting on... since it doesn't say when it's reporting period starts and stops.

You can get ResScan if you wish and see what it shows.
Send me a PM if you want it. Windows only software though. I am betting with all the sessions turned on in SH that it will match ResScan with only maybe a tenth difference since SH goes to 2 decimal points past 0 and ResScan only goes to 1 decimal point.

The onscreen (the LCD of the machine)...you need to verify the reporting period first. Make sure you are in a 24 hour reporting period and not something else.

My Air..who knows about it.

palerider wrote: ↑

Thu Aug 02, 2018 9:32 pm

It's hard to get a good count just by the tick marks, since some are close together and might be more than 2, but the count looks reasonable. for that AHI over the number of hows shown. If you doubt the flags, then zoom in and look at the breaths, and look for any flatlines that are ten seconds or longer.

It isn't even necessary to try to count the tick marks. The events tab lists them with time for each. It clearly shows 25 events, not counting large leak events. (See chart below.) So the 2.73 AHI makes sense if I really only did have the 25 events. (Time on all three shows 8 hours 48 minutes. The Sleepyhead adds an additional 4 seconds.)

But if my math is correct, the number of events in the total night should have been approximately 90 according to the LCD on the machine and myAir. Where are they? It looks like Sleepyhead just stopped being able to "see" events after about 10:15. With as many events as I normally have, it just seems highly unlikely that it could be true that I had none after that time. I do sometimes have fewer in the early morning hours than early in the night, but this is just too extreme to believe.

I'm curious enough about this that I put in a call to the CRT at my DME. She wasn't in this afternoon, but will return my call. I want to find out if she can see any more details about events that night that might put some light on what happened. However, unless a strange difference like this occurs again, I'll probably just call it a mystery and move on if she can't provide any more information to explain such a major discrepancy.

palerider wrote: ↑

Thu Aug 02, 2018 9:32 pm

As to what the machine/myair says? who knows what it's counting on... since it doesn't say when it's reporting period starts and stops.

The machine counts from noon to noon as one day in both the LCD display and in myAir. At least it does since about two months ago. My DME supplier sent the machine out with the clock two hours off, so at first if I took a nap between 10 AM and noon, it got recorded as part of the next night's record instead of the previous night. I haven't had that problem since I corrected the clock. The times between the LCD display and Sleepyhead have coincided.

I meant to see if things coincided more closely last night, but wouldn't you know it, that was the night I forgot to put the SD card back in my APAP until the middle of the night. Maybe tonight.

Pugsy wrote: ↑

Thu Aug 02, 2018 10:08 pm

You can get ResScan if you wish and see what it shows.

Thanks for alerting me that there even is such a program. I'll pass for now. Don't know what the cost is and I see that it needs a doctor's prescription. Plus I'd probably need my BF's assistance in getting it installed and started. Sounds like a lot of hassle if it's only something I use a few times. Generally, the Sleepyhead has been working well for me and mostly it has given similar results to what I see on the display.

However, if this turns out to be a problem that repeats several times, I'll look into it more.

The onscreen AHI at the first time you turned off the machine will show the AHI just for that session and it probably was high because there was some very dense clustering happening but while the average for that time frame is quite high...once that session ended and a new session started the AHI started downward....and if you had looked at the AHI again then it would be lowered because not much was happening.

MyAir reports sometimes gets whacky when there are multiple sessions.

On the off chance something went weird with the reporting software...unplug the machine and then plug it back in and let it reset itself. It's very similar to a reboot of the computer...cleans out the cobwebs.

SleepyHead can only report what the machine reports. SleepyHead doesn't do any flagging...the machine does that and SH only reports it. Nothing is showing up later in the night because either nothing happened or the machine didn't flag it for some reason.
So it has nothing to do with SH and everything to do with what your machine reported..

One remote possibility...a file got corrupted when being written to the SD card and made it difficult for SH to do an accurate reporting of what the machine did flag. That's really rare.

I would do the unplug thing and see what happens with another night...if you start seeing this happen often where the machine isn't flagging where you normally would see flags then there may be a problem with the software in the machine that does the writing to the SD card...or there might be a problem with the machine sensors.
Or a bad sector in the SD card causing a problem with the file being read by SH.

A one off fluke night if the unplug reset trick fixes things...shrug your shoulders and move on but just watch to see if it happens again.

ResScan is free...we just need to get you the link to download it but I wouldn't bother unless we see this happen again.
It's Windows only software...so if you don't use Windows...not much sense in getting it.
If that is the case and this does happen again...there's a way to get me a copy of that SD card and I can run it through my ResScan.

Pugsy wrote: ↑

Fri Aug 03, 2018 8:33 pm

The onscreen AHI at the first time you turned off the machine will show the AHI just for that session and it probably was high because there was some very dense clustering happening but while the average for that time frame is quite high...once that session ended and a new session started the AHI started downward....and if you had looked at the AHI again then it would be lowered because not much was happening.

But that was one reason I was very suspicious of the Sleepyhead report. In looking at the machine after the first session, the screen said twenty point something. The second time it said twenty six point something. That would imply that there were not only additional events in the second session, but that there were more per time unit. And the Sleepyhead report showed zero events for that session, as well as the third. I guess it is possible that I could be remembering wrong. I didn't write the numbers down. But I definitely remember being upset after the second session that the events per hour had gone up, not down.

I'll definitely unplug the machine and plug it back in tonight if that might re-set something.

There was one time before that I got a message that a file had been corrupted when I transferred the data from the SD card, but that didn't happen this time. Doesn't mean there wasn't a problem, though.

So ResScan is free? Surprise! There is a link "How to buy" on the web page advertising it. But free or not, unless it looks like I'd have reason to use it fairly often, it just sounds like more stress. LOL. I'm not that much of a computer person and I absolutely hate learning new programs. Take me back to my quilting instead.

ResScan by Resmed for Resmed machines vs open source Sleepyhead. Sleepyhead is prettier!

Uff Da wrote: ↑

Fri Aug 03, 2018 8:15 pm

palerider wrote: ↑

Thu Aug 02, 2018 9:32 pm

It's hard to get a good count just by the tick marks, since some are close together and might be more than 2, but the count looks reasonable. for that AHI over the number of hows shown. If you doubt the flags, then zoom in and look at the breaths, and look for any flatlines that are ten seconds or longer.

It isn't even necessary to try to count the tick marks. The events tab lists them with time for each. It clearly shows 25 events, not counting large leak events. (See chart below.) So the 2.73 AHI makes sense if I really only did have the 25 events. (Time on all three shows 8 hours 48 minutes. The Sleepyhead adds an additional 4 seconds.)

You didn't show the events tab the first time

Uff Da wrote: ↑

Fri Aug 03, 2018 8:15 pm

But if my math is correct, the number of events in the total night should have been approximately 90 according to the LCD on the machine and myAir. Where are they? It looks like Sleepyhead just stopped being able to "see" events after about 10:15.

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Spotting hypopneas manually is much harder.

Uff Da wrote: ↑

Fri Aug 03, 2018 8:15 pm

With as many events as I normally have, it just seems highly unlikely that it could be true that I had none after that time. I do sometimes have fewer in the early morning hours than early in the night, but this is just too extreme to believe.

Easy enough to check.

palerider wrote: ↑

Fri Aug 03, 2018 10:44 pm

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Oooooooooooooooooooh, you think I have patience, looking for ten second intervals over an almost seven hour period! Wrong! LOL. If I continue to have problems and can't figure it out otherwise, I'll do it, though. At least you've clued me in to how to look for it.

Pugsy wrote: ↑

Fri Aug 03, 2018 8:33 pm

One remote possibility...a file got corrupted when being written to the SD card and made it difficult for SH to do an accurate reporting of what the machine did flag. That's really rare.

I would do the unplug thing and see what happens with another night...if you start seeing this happen often where the machine isn't flagging where you normally would see flags then there may be a problem with the software in the machine that does the writing to the SD card...or there might be a problem with the machine sensors.
Or a bad sector in the SD card causing a problem with the file being read by SH.

Now these possibilities make sense. MyAir gets the information from the machine by wireless, so it makes sense that it and the display on the machine are identical. But Sleepyhead gets the information from the SD card, so if there is anything wrong with the card or the writing to the card, that could account for the difference.

Last night's results were close to the same between the machine and Sleepyhead. AHI 6.59 on Sleepyhead and 6.5 on the machine. Hours used 5:28:01 for Sleepyhead and 5.5 on the machine. Not enough difference to be concerned about. Let's hope this continues.

Uff Da wrote: ↑

Sat Aug 04, 2018 12:08 pm

palerider wrote: ↑

Fri Aug 03, 2018 10:44 pm

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Oooooooooooooooooooh, you think I have patience, looking for ten second intervals over an almost seven hour period! Wrong! LOL.

It only takes a few minutes... zoom in so that you can see breath shapes, then just scroll sideways till there's breaks. It's not nearly as tedious as it sounds on the surface.

palerider wrote: ↑

Sat Aug 04, 2018 1:29 pm

Uff Da wrote: ↑

Sat Aug 04, 2018 12:08 pm

palerider wrote: ↑

Fri Aug 03, 2018 10:44 pm

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Oooooooooooooooooooh, you think I have patience, looking for ten second intervals over an almost seven hour period! Wrong! LOL.

It only takes a few minutes... zoom in so that you can see breath shapes, then just scroll sideways till there's breaks. It's not nearly as tedious as it sounds on the surface.

I got this advice in another thread and laughed it off as well. But just last night I watched a video at freecpapadvice YouTube channel that shows just how easy and quick it can be. The video creator is well known around here, is a sleep technician, and I got a big kick out of his reactions to what he was seeing.

djams wrote: ↑

Sat Aug 04, 2018 3:02 pm

palerider wrote: ↑

Sat Aug 04, 2018 1:29 pm

Uff Da wrote: ↑

Sat Aug 04, 2018 12:08 pm

palerider wrote: ↑

Fri Aug 03, 2018 10:44 pm

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Oooooooooooooooooooh, you think I have patience, looking for ten second intervals over an almost seven hour period! Wrong! LOL.

It only takes a few minutes... zoom in so that you can see breath shapes, then just scroll sideways till there's breaks. It's not nearly as tedious as it sounds on the surface.

I got this advice in another thread and laughed it off as well. But just last night I watched a video at freecpapadvice YouTube channel that shows just how easy and quick it can be. The video creator is well known around here, is a sleep technician, and I got a big kick out of his reactions to what he was seeing.

Well, dang it, you could have left a link. I looked for maybe half an hour and couldn't find the video you might have been talking about. Unless for some reason that kind of link isn't allowed around here. I really haven't checked that kind of thing out.

I learned something new today...

how to quickly scan through the flow rate zoomed in at my preferred scale.

Go zoom in so that you have the scale you want to see...then use the left or right arrow button on your keyboard to scroll and move the flow rate. Easy peasy. I kept looking for a scroll button in SH but there isn't one...just use the arrow...left or right.

I didn't look at Jason's videos either..figured it out all by myself once I got a little hint.
But here's the link to Jason's SH videos if anyone wants them.
Good stuff.

http://freecpapadvice.com/sleepyhead-free-software

Uff Da wrote: ↑

Sat Aug 04, 2018 5:38 pm

djams wrote: ↑

Sat Aug 04, 2018 3:02 pm

palerider wrote: ↑

Sat Aug 04, 2018 1:29 pm

Uff Da wrote: ↑

Sat Aug 04, 2018 12:08 pm

palerider wrote: ↑

Fri Aug 03, 2018 10:44 pm

SH just reports the events that are flagged by the machine in the data.

If you doubt it, it's easy enough to zoom in, and look at your breathing on the flow chart, if there are any flatlines that are more *at least* 10 seconds, then either the machine was not recording the data properly, or SH is missing the flags. (which seems strange).

Oooooooooooooooooooh, you think I have patience, looking for ten second intervals over an almost seven hour period! Wrong! LOL.

It only takes a few minutes... zoom in so that you can see breath shapes, then just scroll sideways till there's breaks. It's not nearly as tedious as it sounds on the surface.

I got this advice in another thread and laughed it off as well. But just last night I watched a video at freecpapadvice YouTube channel that shows just how easy and quick it can be. The video creator is well known around here, is a sleep technician, and I got a big kick out of his reactions to what he was seeing.

Well, dang it, you could have left a link. I looked for maybe half an hour and couldn't find the video you might have been talking about. Unless for some reason that kind of link isn't allowed around here. I really haven't checked that kind of thing out.

Sorry - was rushed, on my phone.
https://youtu.be/FMkBatUVu7s