Is AHI Misleading?

[SA888]

Yesterday my AHI was 3.14 which I thought was pretty good but at around 3:00 the AHI hit 12.
Am I right that the AHI is really only the AHI for the last hour.
It really uses only the last hour's data.
Perhaps Oscar should show the Min ; Med ; 95% and Max for AHI to give added perspective.
Again last night the AHI showed 1.93 which is the first time I was below 2
However again at 3:00 the AHI was 6 and there were 34 RERA events and 17 Hypopneas
While I am making progress AHI 1.93 is misleading in that I am still having disturbed sleep.
Any thoughts on how to reduce the RERA and Hypopneas?

[MediocreOpie]

I don’t use that graph, but I would assume it’s an hourly value for a rolling average. The chart and number on the left is for your session or sessions (if you have more than one). So, if you sleep 8 hours and have 16 events for the 8-hour session, your AHI would be 2.00. I’ve had a few nights where I had multiple events over the course of an hour, but my overall was close to 1.00.

Hope that helps...at least, as far as I understand it.

[Melman]

The AHI and other event numbers listed in the left panel are averages for the night. OSCAR uses the data from the entire session to calculate he AHI. It iss calculated by dividing the number of events by the number of hours of sleep. Hypopneas are included in the AHI but RERAs are not.

[GuyScharf]

Am I right that the AHI is really only the AHI for the last hour.
It really uses only the last hour's data.

My understanding (from reading old posts and looking at my graphs) is that the AHI graph shows the number of events in the last hour.  This becomes especially easy to see if you have only a few events so there are not many contributing to the AHI graph.

I never look at this graph myself.  I can see my problems with the Flow Rate, Pressure, and Event Flags.

[Melman]

I almost always have and AHi of between 0.5 and 1.0 but seldom have events in the last hour. From the Apneaboard WIKI: "The AHI, as with the separate Apnea Index and Hypopnea Index, is calculated by dividing the number of events by the number of hours of sleep."

[Sleepster]

Am I right that the AHI is really only the AHI for the last hour.

No. AHI is number of events divided by number of hours for any given length of time. It could be all night, or it could be all month.

[sheepless]

daily tab statistics: ahi is total included events averaged over total sleep time. however, the graph in the daily tab is "running ahi for the past hour".

[mesenteria]

...
Am I right that the AHI is really only the AHI for the last hour.
It really uses only the last hour's data.
...

I don't know why you would conclude that the AHI, if it's intended to be a useful and veridical statistic, would ignore the 200+ events in the previous four hours of your sleep and then only take the two events from your last hour...?  The resultant AHI of 2.0 would be off by a factor of 100!!  

For such a 'night's' report to be meaningful, the data must include all events during which you were being treated, and if you were wearing the mask for eight hours, the result should be averaged over those same eight hours....no?  So, an AHI statistic is (Total number of events)/Hours of machine usage.  Necessarily, that makes it an averaging statistic, or it provides the 'mean' for the distribution of events during the total time of monitoring...the night.

It's a useful statistic, but it doesn't tell the entire story.  For example, suppose you slept eight hours, but you only had events in one of those eight hours.  It happens that you had 24 of them within a 10 minute period, but then no events prior and none until you awakened.  Your AHI for that night would be a balmy 3.0, well under the 'treated' 4.9, and some would feel they were well taken care of.  But a person looking at that one hour would want to understand why that strong cluster of events happened, and how to preclude their potentially debilitating effect in future nights.  After all, 24 events per hour means something bad was going on, something best controlled for or avoided if the machine were being pushed to its limits to control.  So, the AHI is just the night's average, but one should occasionally review the OSCAR data and events to learn what they are, and if they can be controlled and reduced in frequency, particularly if they 'gang up on the hapless sleeper' during a short period.

[MediocreOpie]

...but one should occasionally review the OSCAR data and events to learn what they are, and if they can be controlled and reduced in frequency, particularly if they 'gang up on the hapless sleeper' during a short period.

My wife gives me grief about looking at my data each morning, but this is exactly why I look at it. What happened? When did it happen? Can I figure out why it happened?

Having just been officially diagnosed with central sleep apnea a little less than a week ago, it scares the bejesus out of me. I am on a 90-day compliance period now and will have to go sit down with my doctor in August to discuss my future. During that discussion, I hope to see if we can get some answers. I read pretty nasty stuff can be a cause for central apnea and, as far as I know, I do not have brain damage and my ticker is still ticking good. I may fall into that group of idiopathic CSA sufferers; at least, I certainly hope I do.

[sheepless]

is there a method to the madness of putting you on apap which doesn't treat central apnea? do you also have obstructive apnea and the thought is that the ca is pressure induced / treatment emergent?

the literature is replete with references to all the horrible conditions associated with central apnea. however, I was diagnosed with central apnea as a young man in 1987. in my (very limited) experience, idiopathic or what I call 'native' ca is common.

start your own thread if you haven't already; post some charts.