(12-16-2013 05:40 PM)zonk Wrote: From encore basic glossary
Snore: A loud upper airway breathing sound during sleep, without episodes of apnea.
Vibratory Snore (VS) Index
The Vibratory Snore Index is the total number of vibratory snoring events per night divided by the hours of use
Yes, the Encore basic glossary says that. But---if you look at the snoring data in Encore and you look at the same snoring data in SH you can get very different impressions. And if SH is counting the number of VS and VS2's correctly, then the Encore VSI is just plain NOT equal to the number of vibratory snore events divided by the hours of use.
As an example: Here are screen shots of the same
night's data from way back when (Feb. 16, 2011), when I was still using my BiPAP in fixed pressure mode at 7/4 and had a particularly bad night for snoring:
First Encore's view of the night:
The Encore VSI = 118.3, which is one of my highest Encore VSI numbers in three years of data. Although what this means is not clear. The usage for this night was a bit over 5.5 hours. If we use the standard idea that Event Index = (number of events)/(run time)
to try to compute the number of "Encore VS" snores for the night, we wind up with an astonishing 650 "Encore VS" snores, which seems somewhat unlikely given the fact that the snore ticks are not so densely packed as to become solid red rectangles and there's a huge part of the night with no snoring recorded.
When we look at the same data in SH, we pick up different things based on whether we have the Details
or the Events
tab selected in the left side bar. When Details
is selected, we see this information:
Several things stand out in this picture of the data. First and foremost, all of the indices except the VSI
equal the corresponding Encore index once you round the SH indices to one decimal place. Next: the SH VSI = 0.0 even though some serious snoring is obviously going on. (Remember, the Encore VSI = 118.3 and there are a lot of tick marks on that VS2 line in the event table.) The Snoring statistics show nonzero numbers for the 90% and Maximum values, although the Max Snore = 25, which is still way, way smaller than the Encore VSI = 118.3. And the Snore graph clearly has some (serious) bumps in it.
Part of trying to reconcile the SH version of the data with the Encore version involves understanding that when JediMark stared working with the System One data, he discovered that it records two kinds of snoring data, which he called VS and VS2 snores.
What the heck distinguishes a VS snore from a VS2 snore? No one knows except the engineers at PR and they're not telling us.
And VS snores are recorded by System One machines ONLY when they are run in Auto mode. So if you are using a PR System One in fixed pressure mode, your SH VSI = 0.0 night after night after night, no matter how badly you are snoring. Your Encore VSI, on the other hand will NOT be 0.0 when you are using fixed pressure mode. And that's worth knowing if you are a PR System One user trying to make sense of the snoring data presented in SH.
Also note that in this night's data, the VS2 tick marks are not so densely packed in the SH event table simply because the table itself is larger. It's actually possible to try to count them. If you do, you should get around 60 of them. That's less than 1/10 of the expected number of "Encore Snores" if we assume that the Encore VSI is computed in the same way the Encore OAI, HI, CAI, RERA I, and FLI are computed. (Although the FLI = 0.0 is also misleading because a System One records FL data only when running in Auto mode.)
When we select the Events
tab in SH, we get this picture:
SleepyHead finds a record of 59 VS2s recorded in the data. (The list is too long to show them all.) Note that we can't simply divide 59 by the run time of 5.5 and get the Encore VSI that way: 59/5.5 = 10.72. If Encore were computing the VSI in the same way it computes the other indices (the OAI, CAI, HI, RE I, and FLI), then the Encore VSI = 10.7.
In all the data that I've carefully looked at (including this night), I've found the same calculation works for computing the Encore VSI. If you add up all those numbers in parenthesis by the individual VS2's in the SH version of the data and divide that sum by the run time, low and behold, you get the correct Encore VSI. In other words, it seems that
Encore VSI = (Sum of the VS2 snore numbers in SH)/(total run time)
Why Encore computes the official Encore VSI in this fashion is a mystery to me. I've found nothing in the Encore documentation that even hints at the fact that this is how the Encore VSI is computed or why it's computed in the way it is. What is clear is that once a set of persistent snoring starts in my data, there's a VS2 snore flag detected by Encore every two minutes with a number associated with that two minute interval.
Perhaps each SH VS2 is actually some kind of a conglomerate of multiple "snoring" events. But then the question goes back to "What is a snoring event
?" Is a "snoring event" a fixed amount of time (or a fixed number of breaths) in which snoring was detected where the number following a VS2 indicate how long the snoring lasted, and hence counts the number of "snoring events" recorded (at or near) that time stamp? Or does a single breath with a snore count as a "snoring event"?
Clearly SH does not represent the snoring data coming from a PR System One machine in the same way that Encore does. That's not a criticism of SH or JediMark, by the way. JediMark has made some reasonable attempts at presenting the confusing snoring data in a way that makes some innate mathematical sense, whence comes his decision to make SH VSI = (number of VS snores)/(total run time)
and to present the VS2 data graphically while not trying to compute a SH VS2 Index