SH data leaks: 95% reading

[lorraineg57]

I thought I read that the 95% meant that you'd spent 95% of the night at that level? My leak data consistently has me at a 95% level higher than it looks like it should be...

According to this, my 95% leak rate was 31.20...unless I understand incorrectly (which is entirely possible), that means that 95% of the night my leaks were at 31.20? It doesn't look that way on the graph!

[Xney]

no it means if you chop your night by time, minutes for example, sort them by the leak number for that minute, 95% of the night was lower than 31. 5% of the night was 31 and higher.

[lorraineg57]

no it means if you chop your night by time, minutes for example, sort them by the leak number for that minute, 95% of the night was lower than 31. 5% of the night was 31 and higher.

Ah, okay...still too high though then. Doesn't the graph look like 95% was lower than like 27 or around there?

[archangle]

I think you're right. Looking at my results, the 95% leak rate on SH is flat out wrong. It doesn't match the graph, and it disagrees with ResScan.

[sol]

I thought I read that the 95% meant that you'd spent 95% of the night at that level?

At or BELOW 95% of the time

[Denial Dave]

I may be mistaken, but I think there is a setting in SH, that allows you to ignore short instances from showing up on the charts.

so, it will display the data in the table, but not on the charts.

As Sol indicated, the 95% mark means that the number was "at or below XYZ for 95% of the time"

[archangle]

I may be mistaken, but I think there is a setting in SH, that allows you to ignore short instances from showing up on the charts.

so, it will display the data in the table, but not on the charts.

As Sol indicated, the 95% mark means that the number was "at or below XYZ for 95% of the time"

You're probably thinking about the "ignore short sessions" option. I don't think there's an option to ignore short leaks.

[Pugsy]

Looking at my results, the 95% leak rate on SH is flat out wrong. It doesn't match the graph, and it disagrees with ResScan.

SleepyHead bug in 95% leak calculations. I have seen it and in fact posted about it quite a while back. Sometimes just a little off and sometimes off by quite a bit. Some people not off at all.
Mark was aware of it and since I did the Patch my ResMed leak data seems to be spot on. Maybe off a little but not more than 1.0.

[Denial Dave]

you're right!! it was the short sessions setting that I was thinking about.

Dave

[Xney]

I think you just need to see what and why and how often those leaks are happening. It may or may not be a problem. If you're just barely above 30 LPM for a bit each night, I don't really think it's a major problem.