At Pugsy's request, I've added this post about PR Snore Data to this thread so it doesn't get "lost"
The question about whether the VS2 data from PR xPAPs is important came up (again) on another forum today. It's been a while since I've seen questions about VS versus VS2 data over here, but since I took the time to write a detailed post over there, I thought I'd post the same thing over here.
VS and VS2 events are the two ways that the PR machines score snoring. We don't know exactly what the difference in criteria used for scoring VS and VS2 snores are on the PR machines, but it turns out we know quite a bit about how the PR machines and Encore (the official software) use the VS and VS2 events.
PR machines and recording snore data
If a PR machine is running in fixed pressure mode, only VS2 data is recorded. In other words, if you are using fixed pressure on a PR machine, the only available snoring data is the VS2 data.
If a PR machine is running in auto adjust mode, both VS and VS2 data is recorded. But only the VS events cause pressure increases.
Anecdotal evidence is that the criteria that PR machines use to score VS2 data is pretty sensitive, and as a result VS2 data can record a lot of "non snoring" stuff as snores. There are reports on other CPAP forums that people have been able to track excessively high numbers of VS2 snores to sleeping with snoring dogs and the hose rubbing against the sharp edge of a night table or headboard. Nonetheless, if snoring is a potential problem, then the VS2 data can be important.
SleepyHead and how it uses the PR snore data
SleepyHead does not compute the VS2 index in the same fashion as Encore does. Since Encore is the official
software for the PR machines, this is a potential problem with SH data. Note that JediMark is aware of this, but at the time he was writing SH, we didn't have any ideas at all about how PR machines recorded their snore data, let alone how Encore computed it's "VSI" index. Heck, we didn't even know there were two forms of snoring data before JediMark started writing SleepyHead.
In looking at lots and lots of his own data, Sleepy Pete made the observation that when there is a whole string of VS2's scored, they are scored in 2 minute intervals
. In other words, the closest together two VS2 events can be is 2 minutes. (See viewtopic.php?f=1&t=93689&p=982257#p982152
). That observation underlies how SleepyHead draws the Snore graph for a PR machine.
SleepyHead does two things with the VS2 data: It creates the VS2 Index in the usual way:
- SleepyHead VS2 Index = (Number of VS2 events)/(run time)
SleepyHead also draws a "Snore Graph". Each VS2 event comes with both a time stamp (when it happened) and a mysterious "number". You can see these numbers if you click on the Events tab in the left side bar and look at the individual events. What those numbers numbers actually mean and how they are determined is not known to us because PR is pretty hush, hush about the algorithms its machines use. However, the common assumption is that the higher the number, the worse the snoring is in some (undetermined) way. But remember that the VS2's are scored in two minute intervals, and hence JediMark's Snore Graph is a "step" graph formed by drawing a two minute horizontal line segment for each VS2 snore; the height of the line segment is the number attached to the VS2 snore.
So visually speaking, the more activity there is in the SH Snore graph and the larger the y
-axis numbers on the Snore graph, the worse the snoring is.
The VS events have no number attached to them. Hence JediMark had no way of turning the VS data into a visual graph. He reports a SleepyHead VS index computed in the usual way:
- SleepyHead VS Index = (Number of VS events)/(run time)
If the VS event tick marks are overlaid on the pressure graph
, it's easy to see that the VS events are the snoring events that lead to increases in pressure when a machine is running in Auto mode.
If your SleepyHead VS = 0.0 night after night, it's tempting to think that you are not snoring at all. But it is critically important to remember VS snores are NOT recorded if the machine is running in a fixed pressure mode. Hence if you use a PR DreamStation PRO
, a System One PRO
, a DreamStation BiPAP PRO
, or a System One BiPAP PRO
, your SleepyHead VS = 0.0 regardless of how much snoring you are doing at night. If you run a PR Auto CPAP or Auto BiPAP in fixed pressure mode, then your SleepyHead VS = 0.0 regardless of how much snoring you are doing.
In other words, if you are using fixed pressure on a PR machine, the only available snoring data is the VS2 data.
Encore and how it uses the PR snore data (Part I)
The SleepyHead VS2 snore data shows up as tick marks in the Encore Event table and is used to compute the Encore VSI, although not in the usual way. The SleepyHead VS snore data does NOT show up in the Encore Event table and it is not important in computing the Encore VSI.
When you download data into Encore, you also get wave flow data (flow rate data in SH) for just the last night before the download. The wave flow data is NOT a standard part of the Encore Daily Detailed Data, but by clicking on the appropriate icon, you can look at it and print it off. When looking at an Encore wave flow, the OAs, CAs, Hs, and SleepyHead VS evens show up as tick marks. But the SleepyHead VS2 events do NOT show up as tick marks on the Encore wave flow. The Encore wave flow data also includes a pressure curve, and it is clear that the SleepyHead VS events cause PR machines to increase pressure.
If you look at the wave flow for a PR machine running in fixed pressure mode, there will be NO snore tick marks because the SleepyHead VS events are NOT recorded in fixed pressure mode and the SH VS2 events are not overlaid over the wave flow data in Encore.
Encore and how it uses the PR snore data (Part II)
I'm a long time PR BiPAP user. I also have a known snoring problem: On my bad apnea nights, I snore. According to my hubby, most of the time my snores are not really loud, but they are persistent on the nights when I snore. And there's some anecdotal evidence that on my worst nights, snores may contribute to the number of wakes I experience. Hence I've spent a fair amount of time looking at snore data in both SH and Encore trying to sort out just what the VS and VS2 data tells us.
I knew right away (long before SH came out) that the Encore VSI was NOT computed in the same way all the other Encore indices were computed. In other words, by looking at Encore data, I knew:
- Encore VSI not= (Number of Encore VS events)/(run time)
And this bugged me for a very, very long time: I really wanted to know what the heck that Encore VSI might mean since my Encore VSI varied tremendously from night to night in my early days of PAPing. And after SH came along, the PR snoring data as presented in SH never even came close to matching up with the PR snoring data as presented in Encore. And that also bugged me (and continues to bug me.)
One assumption that I made in doing my analysis is that PR's programmers are privy to information about what the VS2 data actually means and what the criteria for scoring a VS2 snore is. PR's programmers are the ones who wrote that abomination called Encore, but for its kludginess, Encore does present the Daily Data in a pretty nice format, except for lack of the Flow Rate data being immediately available. And since sleep docs rely on Encore to evaluate how effective PR machines are treating their patients, I'm inclined to give some real consideration to how Encore presents the snore data.
It took me a very long time to backwards engineer how
Encore computes the so-called Encore VSI. What was clear to me even before JediMark started writing SH was that Encore's VSI was NOT computed in the usual way. In other words:
- Encore VSI not= (Number of Encore VS and/or VS2 events)/(run time)
As soon as JediMark had a test version of SH (SH 0.7ish, if I recall), he had discovered what we now call VS snores, but not VS2 snores. And it quickly became apparent that
- Encore VSI not= (Number of SH VS events)/(run time)
It was also apparent by looking at SH Event Tables and Encore Event Tables, that the VS events JediMark was scoring were NOT the same as the VS events Encore was using in its Event Table and the Encore computation of a VSI.
JediMark kept looking at and analyzing more PR data, and soon found a second type of snore data, what we now call VS2. And the SleepyHead VS2 events in SH match up precisely with the VS events in an Encore Event table. But it was also clear that:
- Encore VSI not= (Number of SH VS2 events)/(run time)
Eventually I figured it out: The numbers
attached to each VS2 event are what is used to compute the Encore VSI. Encore's programmers have chosen to add all the VS2 numbers
together and then divide that total by the run time. In other words:
- Encore VSI = (SUM of the numbers attached to each SH VS2 event)/(run time)
All of which raises the question: What is the meaning of the Encore VSI?
And I've hit a bit of a brick wall in trying to answer that question in a definitive manner. I have a working hypothesis for the answer to that question, but I have no way to verify whether my hypothesis is accurate or wildly off the mark.
Robysue's hypothesis about the SH VS2 numbers and the Encore VSI computation
I strongly suspect the number that is attached to each SH VS2 event is the number of individual breaths in the last two minute window that are considered to be "snores"
by the PR xPAP machine. What criteria the PR xPAPs use to decide that an individual breath is a snore is unknown to me; it's got to be something in the shape of the Flow Rate data, but whatever the subtle changes are, they're not enough to pick out visually.
The main reasons that I've come to the conclusion that those numbers attached to the VS2 events are likely the number of individual snores in a two minute window are:
- The fact that in a long, persistent stretch of snoring, the VS2's are always scored in two minute intervals, and
- The maximum VS2's that I see in my data (and others data) seems to be roughly twice the RR rate, and 2*RR = maximum number of breaths in a two minute period.
If my assumption about the meaning of the SH VS2 numbers is correct, then meaning of the Encore VSI
also becomes both clear and intuitive. Recall:
- Encore VSI = (SUM of the numbers attached to each SH VS2 event)/(run time)
If the SH VS2 snore numbers are counting individual snores in a two minute window, then
- (SUM of the numbers attached to each SH VS2 event) = total number of snores during the night
and that means
- Encore VSI = the average number of snores per hour of run time
and, since "run time" is a (crude) approximation of "sleep time",
- Encore VSI = approximation of the average number of snores per hour of sleep
Note: Edited to give Sleepy Pete the credit he deserves for noticing that VS2 snores are recorded in two minute intervals. Thanks to palerider for pointing that out to me.