Query on higher pressure

[mdh235]

Folks,

It's time for a basic statistics lesson folks.

Thank you for that, robysue, hopefully, some folks who've been making incorrect statements about the 95% will read your post, and stop spreading misinformation.

btw, you almost make math fun

I don't know if you're referring to me, but if so, don't worry about it.

I'm done.

[zonk]

I thought it was the pressure that I spend 95% of the time at or below.

Yes what I thought too, at or below for the 95% of the night

From DeVilbiss IntelliPAP AutoAdjust clinical overview

95% Percentile Pressure
90% Percentile Pressure
Percentile pressure is provided as a number and as a graph. Percentile pressure indicates that therapy pressure was at or below this pressure for 90- or 95-percent of its usage time. Percentile pressure can be used to help choose fixed CPAP pressures. If the AutoAdjust device operates at the upper pressure setting for more than 10% of the therapy time, this may be an indication that the upper setting should be increased. *Refer to Pressure Plateau Time

Pressure Plateau Time
Pressure plateau time is provided as a number and as a graph. Pressure plateau time is the percentage of time therapy was at the upper prescription limit. If the time is 10% or higher, the pressure has been railing at the upper pressure and this may be an indication the prescription setting should be increased to allow the therapy to move to a higher pressure than is currently allowed.

[diamaunt]

Thank you for that, robysue, hopefully, some folks who've been making incorrect statements about the 95% will read your post, and stop spreading misinformation.

I don't know if you're referring to me, but if so, don't worry about it.

I'm done.

fwiw, I wasn't.

[diamaunt]

Thank you for that, robysue, hopefully, some folks who've been making incorrect statements about the 95% will read your post, and stop spreading misinformation.

Others said the same thing about 95th percentile pressure in other threads but what RobySue say does carry more weight than the rest of us uneducated lots

Now about "S9 Sleep Quality", that was misinformation on my part but unintentionally, actually its not misinformation because misinformation when you say something knowingly to be untrue. That was what people were saying here and other forums by topnotch posters until you came along and said something different and debunked the myth

I think of misinformation as simply 'incorrect information'. perhaps I'm wrong about that.

I didn't know what the 95th percentile pressure meant until I read robysue's blog. (I don't think I even too statistics in school, and that was a LONG time ago) but, I do think it's important to know what it means.... and how meaningless it is by itself.

so your 95% and max was, say 12cm. that could mean you spent a few minutes at 12cm, or it could mean you spent most, or all the night at 12cm. there's just not enough information in that one (or even those two) numbers to know.

as to the s9 sleep quality you mention, and as you mentioned in another post, "say something enough times and it becomes 'fact'", even if it's wrong. I appreciate your looking into it for yourself, and helping debunk that myth

[Sleepster]

You are AT or BELOW the 95% number for 95% of the time. And you are ABOVE the 95% for NO MORE THAN 5% of the time.

Leaving out the phrase "NO MORE THAN" makes no significant difference in the meaning of the statement for the actual data sets involve thousands of numbers that are not exact.

In fact the OP's claim that the 95th percentile pressure equals the maximum pressure is an example of this very thing. They are equal only in the sense that when rounded off, and when the experimental errors associated with the sensor readings are ignored, the reported numbers are identical.

(This is an old argument that goes on all the time in academia between those who deal with pure mathematics and those who deal with applied mathematics. Those in fields that make use of applied mathematics make statements that cause pure mathematicians to cringe. It's always been this way and it will always be this way. In this case I can't possibly fathom a reality in which the OP has been mislead by my claim. Nevertheless I will try to more careful in the future.)

[Sleepster]

so your 95% and max was, say 12cm. that could mean you spent a few minutes at 12cm, or it could mean you spent most, or all the night at 12cm. there's just not enough information in that one (or even those two) numbers to know.

Not in the case of pressure readings on CPAP machines. And lots of other things, too. The pressure readings come in at regular intervals of time, like if you have a data collection rate of 10 Hz you are collecting one number every 0.1 seconds.

In this case you spend approximately 5% of the night at a pressure of 12 cm. And you spend approximately 95% of the night at pressures below 12 cm.

(Note that I explicitly insert "approximately", but in the vast majority of interactions it's assumed that when we are dealing with real data that was collected using real measurements, everything that's stated is an approximation. Pedantic quibbling aside. )

Now if we wish to be more precise we express the results along with their uncertainly. Suppose we have an uncertainty of 0.1 cm in our pressure readings, and 0.2% in our percentile determinations. Then we can say as an example, the following:

(95.0 ± 0.2)% of the time we have a pressure below (12.0 ± 0.1) cm.

(95.0 ± 0.2)% of the time we have a pressure at or below (12.0 ± 0.1) cm.

There's usually no need to be this pedantic about it, though.

[Sleepster]

Perhaps I'm unclear about what the 95% number means. I thought it was the pressure that I spend 95% of the time at or below.

Your thinking is correct.

[Sleepster]

I think of misinformation as simply 'incorrect information'. perhaps I'm wrong about that.

To me misinformation is something that's created in attempt to lead or mislead people into thinking or doing something.

On the other hand incorrect information could have been created with no such attempt in mind, it could[/quote] be simply the result of an error.

I see no one on this forum intentionally spreading misinformation, but I do see some incorrect information being posted. For example I am not guilty of the former, but I am guilty of the latter. I know no one who is perfect and never makes mistakes. We're all human, myself included. No one makes mistakes on purpose.

[diamaunt]

Not in the case of pressure readings on CPAP machines. And lots of other things, too. The pressure readings come in at regular intervals of time, like if you have a data collection rate of 10 Hz you are collecting one number every 0.1 seconds.

In this case you spend approximately 5% of the night at a pressure of 12 cm. And you spend approximately 95% of the night at pressures below 12 cm.

(Note that I explicitly insert "approximately", but in the vast majority of interactions it's assumed that when we are dealing with real data that was collected using real measurements, everything that's stated is an approximation. Pedantic quibbling aside. )

Now if we wish to be more precise we express the results along with their uncertainly. Suppose we have an uncertainty of 0.1 cm in our pressure readings, and 0.2% in our percentile determinations. Then we can say as an example, the following:

(95.0 ± 0.2)% of the time we have a pressure below (12.0 ± 0.1) cm.

(95.0 ± 0.2)% of the time we have a pressure at or below (12.0 ± 0.1) cm.

There's usually no need to be this pedantic about it, though.

the *mistake* in statement that was perpetuated earlier in this thread was:

Perhaps I'm unclear about what the 95% number means. I thought it was the pressure that I spend 95% of the time at or below.

Right, so you spend 5% of the time above that pressure. But as you said the 95% pressure equals the maximum pressure, so you are spending 5% of the time at your maximum pressure.

it does *not* mean that they are spending 5% of the night at the max pressure.

if max pressure = 95% pressure, then *at least* 5% of the night was at max pressure, but it could well be 10%, 20%, 30% or even 90% of the night at max pressure.

you only have *two* data points, and they are both the same, in the case of mdh235, 12.6cm.

if his machine ramped up to 12.6 immediately upon his first event and stayed there all night, he'd still have 12.6 95% and 12.6 max.

it doesn't matter if there's 20 data points, or 57,600.

[diamaunt]

I think of misinformation as simply 'incorrect information'. perhaps I'm wrong about that.

To me misinformation is something that's created in attempt to lead or mislead people into thinking or doing something.

On the other hand incorrect information could have been created with no such attempt in mind, it could be simply the result of an error.

I see no one on this forum intentionally spreading misinformation, but I do see some incorrect information being posted. For example I am not guilty of the former, but I am guilty of the latter. I know no one who is perfect and never makes mistakes. We're all human, myself included. No one makes mistakes on purpose.

thank you for your information, however, I find that dictionaries do say that it is a a noun that's based on the transitive verb, 'misinform: to give incorrect information to'. eg: http://www.thefreedictionary.com/misinformation

though some do add "especially that which is deliberately intended to deceive." which indicates that it is not *only* deceptive information that is misinformation.

Merriam-Webster helpfully only defines 'mis-' as:
1 a : badly : wrongly <misjudge>
b : unfavorably <misesteem>
c : in a suspicious manner <misdoubt>
2: bad : wrong <misdeed>
3: opposite or lack of <mistrust>
4: not <misknow>

then lists a plethora of words that have that prefix.