Oscar Question

[paulag1955]

Okay, I'm not good with numbers, so if someone could explain this as they would to a five year old, I'd be grateful.

In the OSCAR sidebar, what to the columns 95% and 99.5% actually mean? 

Also, with regards to Flow Limits, what is the Median column actually telling me? Mine is always zero or a very small number, like .01 or .02, which doesn't make sense to me based on my understanding of what the median represents.

Thank you in advance for your extremely dumbed down explanations.

[Crimson Nape]

The values are percentiles.  The term "Median" is the same as a 50%-tile, or the mid-point of all data samples after sorting.   These values are calculated by sorting the data from the lowest value to the highest value, then reports the calculated value at the specified location.  The 95%-tile value means 95% of all sample data is at or below this number.

Example:
A group of numbers:
1  1  99  4  2  5  8  4  3  45

Sorted:
1  1  2  3  4  4  5  8  45  99

Average:  17.2
Median:  4  (mid-point of all data samples)
95%-tile:  74.7 (95% of all data is at or below this value)
99.5%-tile:  96.57 (99.5% of all data is at or below this value)
The use of the 99.5%-tile, instead of displaying the maximum value, is to remove spikes that may be considered artifacts (a bad data points)

- Red

[paulag1955]

Thank you, that was just what I needed. Now just to make sure that I understand what you're saying about the median, in my example, the median of .02 means that half of the readings are higher than .02 and half are lower. Is that right? Looking at my chart, it's hard to believe, which is why I was so confused.

   

[Crimson Nape]

You are correct.  Half of all the data points in that mess are at or below the .02.

Also, since these questions are concerning OSCAR, I am moving this thread to the Software Forum.

- Red

[paulag1955]

Thanks again. I think I've got it now.

[paulag1955]

I need some clarification on how AHI is calculated. I'm trying to calculate my RDI by adding all my apnea, hypopnea, and RERA events, then dividing by the number of hours used. What I'm finding is that for some days, this results in an RDI that's lower than my AHI. I don't understand how this can happen and think I must be misunderstanding what all is going into the AHI calculation. Can anyone help me out? 

Again, poor number sense, so please dumb it down as much as possible.

TIA.

[Gideon]

The Easy calc

RDI= AHI + RERAS
Using the index of the RERAS.

I suspect you have a math error of some type as RDI cannot be lower than AHI
Make sure that you are counting all the Central events as well.

[paulag1955]

That's what I thought, which is why I'm confused now. I'm using data that I exported from OSCAR into an Excel spreadsheet. I deleted all of the "Count" columns that had entries of 0 for every night, which leaves me with CA, OA, H, UA, and RE. That should be everything, right?

[paulag1955]

Here's a screenshot of what I'm seeing. I verified that the numbers match raw data. You can see that there are several where the RDI is lower than the AHI, but are clearly rounding errors. However I've got seven days where the calculated RDI is more than a thousandth lower than the AHI.

   

And this is how the RDI event count is calculated.

   

[Crimson Nape]

Based off your highlighted example, the RDI is correct.  It seems the AHI is off by 1 event.  The AHI's value seems to be based on 14 events.  Is there is something in the "P" column that is hidden?

Number of events listed: 13 (8 + 3 + 2)
Hours : 7.87
AHI: 7.87 * 1.79 = 14.0873 (14) events
RDI: 7.87 * 1.65 = 12.9855 (13) events

- Red