Large differences in AHI between Oscar and DeVilbiss [reason: see post #7]

[Thanassis]

Hello,

I am a newbie to this forum (just registered today) and also new user of Oscar software.
I downloaded my data from DeVilbiss DV64 and imported it to Oscar today.
I see a great difference between what Oscar indicates as daily AHI and what DeVilbiss recorded
as AHI for the same day (for example 15.25 in Oscar compared to 2.25 in DeVilbiss).
Does anyone know to explain this difference?

Thanks in advance
Thanassis

[DeepBreathing]

Hi Thanassis. Welcome to Apnea Board.

Just to clarify - when you say you got a reading of 2.25 in DeVilbiss, is this what is shown on the machine's screen or are you using DeVilbiss software?

There are known problems with the DV64 in Oscar and this may account for the discrepancy you're seeing. It would help greatly if you could post some daily charts for us to look at. Organise your charts as shown on this page: http://www.apneaboard.com/wiki/index.php...ganization

To attach images to your post, see here: http://www.apneaboard.com/wiki/index.php...pnea_Board

[SuperSleeper]

In addition, it's possible that the timing isn't the same - ie-- the CPAP machine's "day" isn't set the same as OSCAR's "day", (usually from 12 noon to 12 noon).  If that's the case, you're going to get different AHI readings, depending on the sample time period.

[Thanassis]

   
Thanks for your welcome,

When I refer to the DeVilbiss intications I mean what I read on the machine's screen as well as the report
produced by using the online DeVilbiss report Generator ,
[attachment=11631][attachment=11631]

I attach report of the same day to clearly see the differences

[DeepBreathing]

Looking at the Oscar output you can see that there are a LOT of flow restrictions and hypopneas happening, which aren't picked up in the Devilbiss report. DeVilbiss DV64 machines are only partially supported by OSCAR. There are issues with the graphing of data and calculation of AHI and other parameters may contain errors. However I suspect there may also be an issue with the way the hypopnea is determined - there is no one universally accepted definition and various manufacturers use their own set of parameters (eg whether it's a 40% or 50% reduction of flow).

I'm going to send a link to this thread to the Oscar development team, who may have some more detailed knowledge. In the meantime, can I ask you to do three things for me:

1. Go to this page: http://www.apneaboard.com/wiki/index.php...ganization and organise your daily chart according to the instructions there. That will give us a more valuable set of data to work with.

2. On the daily page, just above the big AHI number there are tabs for Details, Events, Notes and Bookmarks. Click on the Events tab, expand the hypopnea and apnea results and give us a screenshot of that.

3. Finally, zoom in on the hypopneas in the flow chart so we can see the actual wave-forms and try to determine if it's a genuine hypopnea or something misinterpreted by Oscar.

Sorry to give you so much "homework" but it will really help the development team and hopefully also allow you to get better treatment.

[Thanassis]

Thank you for your suggestions!

Taking into account your directions I submit

1. A screenshot of Devilbiss report generator for 27/04/19
2. The main (rearranged) page or Oscar report for the same day
3. A screenshot of the expanded hypopneas in the Events tab together with  
   zooming in on the hypopneas in the flow chart to observe the actual wave-forms

I hope that  I have helped you!

[Heyns]

The DV64 implementation in OSCAR is still incomplete. The data representation is not correct yet. This is a known issue at present. The "severity flags" associated with events (hypopneas, periodic breathing, etc) are not yet correctly implemented in OSCAR. There are also still synchronization issues with the data. This will hopefully be addressed in subsequent releases of OSCAR.

In the meantime, it is still helpful to observe the tendencies of the graphs and events displayed in OSCAR. However, the values displayed should not count for too much at present. Rather believe the values from your DV64 machine for now (until the DV64 compatibility has been addressed fully in OSCAR).

OSCAR will display many more event flag incidences than are actually used by the DV64 machine for the calculation of e.g. AHI. This is not incorrect, as the DV64 data structures do contain those flags with associated severities. However, they should still be be grouped and handled correctly within OSCAR (in a similar way as within the DV64 machine) to obtain similar AHI values. This is true for all values calculated using the event data.

[Thanassis]

Thanks a lot!

I will follow your very useful forum and wait for the full incorporation of DeVilbiss machines
in Oscar!

[Gideon]

Try something that may give you better info.

Under File/Preferences/CPAP  there is a "Custom CPAP Event Flagging"  
set up  #1 (you could use #2) as 80% and 10 Seconds
set up  #2 as 50% and 10 Seconds

These could match up better.

Here are standard definitions.
Event Definition Apnea, Hypopnea, Flow Limitation

• Apnea: 80% to 100% reduction in airflow for >= 10 seconds
  
  
• Hypopnea: 50% to 80% reduction in airflow for >= 10 seconds
  
  
• Flow Limitation: <50% reduction in airflow for >= 10 seconds
  

[Jarm]

The DV64 implementation in OSCAR is still incomplete. The data representation is not correct yet. This is a known issue at present. The "severity flags" associated with events (hypopneas, periodic breathing, etc) are not yet correctly implemented in OSCAR. There are also still synchronization issues with the data. This will hopefully be addressed in subsequent releases of OSCAR.

In the meantime, it is still helpful to observe the tendencies of the graphs and events displayed in OSCAR. However, the values displayed should not count for too much at present. Rather believe the values from your DV64 machine for now (until the DV64 compatibility has been addressed fully in OSCAR).

OSCAR will display many more event flag incidences than are actually used by the DV64 machine for the calculation of e.g. AHI. This is not incorrect, as the DV64 data structures do contain those flags with associated severities. However, they should still be be grouped and handled correctly within OSCAR (in a similar way as within the DV64 machine) to obtain similar AHI values. This is true for all values calculated using the event data.

---

The DV64 implementation in OSCAR is still incomplete. The data representation is not correct yet. This is a known issue at present. The "severity flags" associated with events (hypopneas, periodic breathing, etc) are not yet correctly implemented in OSCAR. There are also still synchronization issues with the data. This will hopefully be addressed in subsequent releases of OSCAR.

In the meantime, it is still helpful to observe the tendencies of the graphs and events displayed in OSCAR. However, the values displayed should not count for too much at present. Rather believe the values from your DV64 machine for now (until the DV64 compatibility has been addressed fully in OSCAR).

OSCAR will display many more event flag incidences than are actually used by the DV64 machine for the calculation of e.g. AHI. This is not incorrect, as the DV64 data structures do contain those flags with associated severities. However, they should still be be grouped and handled correctly within OSCAR (in a similar way as within the DV64 machine) to obtain similar AHI values. This is true for all values calculated using the event data.