ResScan 7.0 now available! [old version]

[Gideon]

Most likely rescan is using the entire dataset. OSCAR tries to eliminate the extremes that occur when starting or shutting down to provide closer to true stats

[ScottZZZ]

Most likely rescan is using the entire dataset.  OSCAR tries to eliminate the extremes that occur when starting or shutting down to provide closer to true stats

I'm not sure that explains it....  First, the difference is substantial.  ResScan is consistently lower (i.e., -0.83cm on Median, -0.81 on 95%, using the data I provided earlier).  I'm bad at statistics, but it doesn't seem like OSCAR tossing a few seconds of extremes (like the false-zero values from my AS11 in the first few seconds) would account for that magnitude in the median/95th percentile values.  (The "Max" and "Min" are different stories, since they are not percentiles but rather absoute values.)

Second...  Those -0.8 differences were from days where I had EPR at 1cm.  I just searched back and looked at five days in early November when I had EPR at 2cm.  And viola, ResScan's statistics for those days are consistently around -1.6cm for Median and 95th (i.e., almost exactly double the difference as for days where EPR was 1cm).

Based on that, I'm still theorizing that ResScan's statistics are including pressure observations during both Inhale and Exhale, since the variance "followed" the EPR change.  

Can you confirm whether OSCAR's pressure statistics are based on Inhale pressure only?  If yes, then I'm sticking to my theory that ResScan is reporting lower pressures due to including Exhale/EPR data.   On the other hand, if OSCAR includes Exhale pressures in its calculation, then my theory doesn't hold up. 

[GuyScharf]

I will pass this issue on to pholynyk as he is our ResMed expert. When I look at your comparison table, not everything makes sense to me. I will have to compare with your data to see if I can explain the differences based on known session handling differences between ResScan and OSCAR.

Could you make a zip of your SD card and upload it to my Dropbox folder? Please include your forum name in the zip file name so we know who the file is from.

[GuyScharf]

Can you confirm whether OSCAR's pressure statistics are based on Inhale pressure only?  

OSCAR is reporting pressure data supplied by the CPAP machine, and the same data is used for graphing and statistics. 

For an AS11, the statistics panel should show two lines: "Pressure" and "EPAP".  The relationship between the two lines should be approximately "Pressure = EPAP + EPR". There will be some variation because we're talking medians etc. here.

For your "pressure" section in your report, are you using Pressure or EPAP? Are you taking your data from the OSCAR Daily page, or somewhere else? And is ResScan calling the statistics "Pressure/IPAP" or "EPAP"?

[Crimson Nape]

ScottZZZ - A post in another thread has triggered a question here. Are you using the ramp feature?

[ScottZZZ]

OSCAR is reporting pressure data supplied by the CPAP machine, and the same data is used for graphing and statistics. 

For an AS11, the statistics panel should show two lines: "Pressure" and "EPAP".  The relationship between the two lines should be approximately "Pressure = EPAP + EPR". There will be some variation because we're talking medians etc. here.

For your "pressure" section in your report, are you using Pressure or EPAP? Are you taking your data from the OSCAR Daily page, or somewhere else? And is ResScan calling the statistics "Pressure/IPAP" or "EPAP"?

Yes, I understand OSCAR is getting pressure observations from the machine.  My question was about how it calculates the Pressure statistics.  I *assume* since OSCAR has a separate line on the Daily page for "Pressure" and "EPAP" that the "Pressure" line statistic is based on IPAP pressure only.  The difference I see between the OSCAR and ResScan "Pressure" statistics is making me think that ResScan is combining all the IPAP and EPAP observations for its resulting "Pressure" statistics.

The ResScan "Statistics" page makes no mention anywhere of EPAP or EPR.  It has a single line labeled "Pressure cmH2O", with values for Median, 95th percentile, and Maximum. I am comparing the Median and 95th percentile values on that line (viewing range set to 1 day) with values on the OSCAR "Daily" page (the Median and 95% values on the "Pressure" line). Both OSCAR and ResScan label this "Pressure", so I expected them to match -- but the ResScan "Pressure" statistic is consistently lower, and the variance (while not the same magnitude as EPR) does seem to correlate with the EPR setting.

The ResScan values are consistently about 0.8cm lower than the OSCAR values with EPR at 1, and consistently about 1.6cm lower with EPR at 2.  That's why I am theorizing that ResScan is combining the pressure observations during both IPAP and EPAP for a single "Pressure" statistic, whereas OSCAR is using only IPAP. 

ResScan *does* properly show the EPR setting on its "Settings" page, along with other machine settings. The ResScan detail pressure graph has a single trace, and the values depicted by that trace match closely with OSCAR's IPAP line.  There is no EPAP trace shown on the ResScan chart (could be an option somewhere that I missed, but I don't see it).

I will attach a snapshot of a ResScan Statistics page for reference.

My purpose with this discussion is to have an explanation -- or at least a logical theory, supported by some data -- why the "Pressure" statistic from ResScan is consistently and significantly different from the "Pressure" statistic being reported by OSCAR.  My current theory is they are both correct -- but that OSCAR's pressure statistic is based on IPAP observations only, and ResScan's pressure statistic is (likely) based on the combined IPAP and EPAP observation set, making it an "average" of IPAP/EPAP and therefore lower than the "Pressure" reported by OSCAR.

I suppose I could turn off EPR completely for a night, and then compare.  But I really like EPR, and the data I have with EPR-1 and EPR-2 seem to support my theory.

   

[ScottZZZ]

ScottZZZ - A post in another thread has triggered a question here.  Are you using the ramp feature?

Negative -- Ramp always OFF for me.

[GuyScharf]

Yes, I understand OSCAR is getting pressure observations from the machine.  My question was about how it calculates the Pressure statistics.  I *assume* since OSCAR has a separate line on the Daily page for "Pressure" and "EPAP" that the "Pressure" line statistic is based on IPAP pressure only.  The difference I see between the OSCAR and ResScan "Pressure" statistics is making me think that ResScan is combining all the IPAP and EPAP observations for its resulting "Pressure" statistics.

The machine reports both EPAP and IPAP (aka. pressure), from which OSCAR plots data and computes statistics for each one. The OSCAR pressure median calculation is the median of all IPAP measurements reported, while the EPAP median calculation is the median of all EPAP measurements reported by the machine.

The current version of OSCAR does not discard unhelpful 0 values at the beginning of a session. Previous versions discarded 40 seconds (I think) and the next version will discard 5 or 10 seconds to get rid of those useless zero values. I doubt that is your problem.

Is it possible that ResScan is reporting Mask Pressure statistics, which is high resolution data that in fact contains both inhalation and exhalation values? OSCAR is using "set" pressure values for statistics. It graphs Mask Pressure does not make statistical computations on it. The difference is that the "set" pressure is what the machine wants to deliver to the mask, while the mask pressure is what is actually delivered. Normally those are about the same unless there are very high leak values and the machine cannot deliver the pressure it is trying to deliver.  If you look at a chart where the hose has been disconnected, you may see a normal "set" pressure but the mask pressure may fall to zero because the machine is unable to push enough air to deliver any pressure.

In the EDF files, both Mask Pressure and Set Pressure are labelled simply "Pressure." They are in separate files, so the names do not conflict, but it does demonstrate that in at least some places ResMed just uses the word "pressure" to refer to mask pressure.

[ScottZZZ]

Is it possible that ResScan is reporting Mask Pressure statistics, which is high resolution data that in fact contains both inhalation and exhalation values?

Perhaps…. But if using mask pressure vs set pressure was the explanation for a difference between the two softwares, why would that difference increase proportionally to an EPR change?  I would expect an EPR change to affect mask pressure on both equally, yet increasing EPR (reducing EPAP) reduced the “Pressure” reported by ResScan but not the “Pressure” reported by OSCAR.  

Remember, the 0.8cm difference between ResScan and OSCAR at EPR=1 became 1.6cm difference at EPR=2.  That makes me think ResScan is combining IPAP and EPAP into a single averaged value, because doubling EPR also doubled the shortfall on ResScan (the Median/95% statistics).

But I’ll take a look at the mask pressure charts, and see how it compares to the daily “Pressure” statistic reported by ResScan to see if there is a possible correlation there.  I should be able to post some findings from looking at that tomorrow.

[ScottZZZ]

I spent some more time this morning investigating the difference between ResScan's and OSCAR's "Pressure" statistics.

I can only conclude they are "different", with ResScan's "Pressure" statistics being lower when EPR is used (with magnitude of difference related to magnitude of EPR).   I cannot conclude how the statistics reported by ResScan are actually calculated, or what specific data is used in their calculation.

@GuySharf -- You speculated above that perhaps ResScan's "Pressure" statistics were based on "Mask Pressure".  That's entirely possible.  The ResScan "Pressure" statistics look like an "averaging" of IPAP and EPAP pressure (thus moving lower with EPR increase).  That could, indeed, be explained by ResScan using "Mask Pressure" for their statistic, since "Mask Pressure" generally ends up being between IPAP and EPAP in my data. 

I had speculated that perhaps ResScan was combining pressure observations during IPAP and EPAP to create their "Pressure" statistic.  That theory would also explain the lower "Pressure" statistic (compared to OSCAR's "Pressure" statistic) with EPR.

But it also appears that ResScan isn't calculating ANY of the values on its "Statistics" page.  Instead, all values on that page are being read directly from the PAP machine's "Summary Data" (i.e., as calculated/stored by the PAP machine itself).  ResScan is just displaying those numbers, not calculating them from the detail session data.  To verify that:  I created a copy of my SD card, deleted all the sessions (i.e., empty "DATALOG" folder), opened ResScan, created a new patient profile, and imported "All available data" from the card.  All the daily "Statistics" are viewable, without ANY detailed session data.  ResScan cannot be calculating its "Statistics" from the detailed data here, because there is no detailed data available.

My conclusions, at least regarding AS11 data with ResScan7 (it could be different with other machines) are:

• ResScan7 reports "Pressure" statistics that don't match the "Pressure" statistics reported by OSCAR.
  
• The ResScan7 "Pressure" statistic is decreased when EPR is used. (i.e., reduced EPAP pressure reduces the ResScan "Pressure" statistic)
  
• The ResScan7 "Pressure" statistic is not calculated by ResScan7 using detail session data. Rather, it is obtained from the PAP machine's summary data file, having been calculated and written to the card by the PAP itself.
  
• The ResScan7 "Pressure" statistic may be based on "Mask Pressure", or some averaging of IPAP/EPAP, or something else.  Only ResMed knows for sure. 
  

That's probably as close as I can get to having an answer when a sleep doc asks me "Why are the pressure statistics from our ResScan and your OSCAR different?"   Rescan's "Pressure" statistic comes from the PAP machine and is some unknown combination of IPAP/EPAP, and therefore affected by EPR, whereas OSCAR reports IPAP and EPAP separately.