+- Apnea Board Forum - CPAP | Sleep Apnea (https://www.apneaboard.com/forums)
+-- Forum: Public Area (https://www.apneaboard.com/forums/Forum-Public-Area)
+--- Forum: Software Support Forum (https://www.apneaboard.com/forums/Forum-Software-Support-Forum)
+--- Thread: measurement uncertainty (/Thread-measurement-uncertainty)
measurement uncertainty - 0rangebear - 03-07-2016
Can anyone explain the variation in these reports?
RE: measurement uncertainty - DeepBreathing - 03-08-2016
Not sure which particular variation you mean - there are quite a few.
Compliance
ResScan reports the percentage of days with usage > 4 hours. I'm not sure what Sleepyhead reports for compliance.
Indices
ResScan truncates an index to one decimal place before using it in a calculation. Hence there are substantial rounding errors. SleepyHead reports to two decimal places and I suspect it doesn't truncate at all in its calculations. The outcome of this is that the SleepyHead numbers usually add up pretty closely while the ResScan will show cumulative errors over a large number of nights.
I think the fact that Resmed aren't too fussed about these errors is an indication that we shouldn't get too hung up on the exact numbers. They are an important indication of efficacy, but they are no the be all and end all.
RE: measurement uncertainty - 0rangebear - 03-08-2016
(03-08-2016, 07:45 AM)DeepBreathing Wrote: Not sure which particular variation you mean - there are quite a few.
Compliance
ResScan reports the percentage of days with usage > 4 hours. I'm not sure what Sleepyhead reports for compliance.
Indices
ResScan truncates an index to one decimal place before using it in a calculation. Hence there are substantial rounding errors. SleepyHead reports to two decimal places and I suspect it doesn't truncate at all in its calculations. The outcome of this is that the SleepyHead numbers usually add up pretty closely while the ResScan will show cumulative errors over a large number of nights.
I think the fact that Resmed aren't too fussed about these errors is an indication that we shouldn't get too hung up on the exact numbers. They are an important indication of efficacy, but they are no the be all and end all.
I have attached some more objective evidence of the measurement accuracy issue maybe there is something I am doing wrong with the software
I timed three night on my cell phone and there is a contestant 2 second deviation in rescan and more in sleepy head that will effect the compliance report.
in addition; on two of the nights the AHI and total time don't agree insight the Sleephead tool. On two different sleepy head tabs Daily and the Statistics screen give different numbers
see attached data from screen shots.
Rounding does explain some of it, but there is larger issue here that rounding
RE: measurement uncertainty - DeepBreathing - 03-09-2016
If you look at the ResScan summary graphs you will see there is a "usage" and "total usage" graphs which (on my machine) are typically two minutes different. Don't ask me why.
I'm not seeing any significant discrepancy between SH and RS in terms of the overall usage - they are the same to within a minute. SH does tell you the sleep / wake times and they are consistent with the durations in your charts. ResScan doesn't tell you this information, but you can extract it from the detailed chart. Have a look at what time you went to sleep / woke on the night of 6 March. According to SH you only slept 35 minutes compared with 4:10 hrs on ResScan. What does the flow rate chart look like on SH? Does it extend for 4 hours or 35 minutes?
Another thought - SleepyHead has the ability to ignore short mask sessions or combine several mask sessions into one. Got to Files | Preferences | Import and check that "Combine close sessions" and "Ignore short sessions" are turned off. Also check that the day split time is 12:00pm (ie noon). And check if the mask sessions are the same on both programs. On ResScan it's on the usage chart in the summary chart page.