Head colds and oximeter statistics/ Managing SpO2 performanc

cnaumann wrote:The difference between accuracy and precision is very important

You are missing the proximate point. If the oximeter were precisely accurate, the difference between medians of 94 and 93 is still not significant.

The OP needs to look at the O2 graph and ask questions. What are the low points? What is typical high period and how long do such periods last? What appears to be artifacts?

Then fully examine the CPAP data.

ChicagoGranny wrote:

cnaumann wrote:The difference between accuracy and precision is very important

You are missing the proximate point. If the oximeter were precisely accurate, the difference between medians of 94 and 93 is still not significant.

The OP needs to look at the O2 graph and ask questions. What are the low points? What is typical high period and how long do such periods last? What appears to be artifacts?

Then fully examine the CPAP data.

*ding!*

sleepy1235 wrote:Posted a technical inquiry directed to those who have oximeters and instead got snide remarks from those who I doubt have oximeters or don't know what they are taking about.

Some people in this forum are looking for technical information.

sleepy1235,
This is an open venue. Most of the people here genuinely want to help others. Ignore those posts that you find offensive, or better -- read their past posts to get a truer measure of their more immediate response. If more often than not their posts aren't cruel, your question may actually unintentionally be funny without your realizing.

ChicagoGranny soundly sums up the problem. This is more cooking than chemistry, and even then, not all of us are on the same page with regards to maths and sciences. Why not post a couple nights of your oximetry? Someone might be able to offer you a clearer perspective from that.

ChicagoGranny wrote:

cnaumann wrote:The difference between accuracy and precision is very important

You are missing the proximate point. If the oximeter were precisely accurate, the difference between medians of 94 and 93 is still not significant.

The OP needs to look at the O2 graph and ask questions. What are the low points? What is typical high period and how long do such periods last? What appears to be artifacts?

Then fully examine the CPAP data.

I understand what you are saying, and I know that I am being really picky here, but I disagree when the discussion was about accuracy and precision. It comes down to the use of the term significant. Are we talking significant in a statistical way or significant in a clinical way? Even if something is not clinically significant, it still may be an interesting event.

Let’s say I have a precision instrument and make a number of measurements and come up with a well established average and standard deviation for those measurements and a histogram of those measurements produces a nice looking bell curve. Then I make another measurement that deviates more than 3 standard deviations from that established mean. I would call that amount of deviation ‘statistically significant’. It may be that the Gods of random noise simply came in sync and the measurement means absolutely nothing. However, it is more likely that one of my process parameters have changed.

Now let’s say that the instrument is this mythical precision pulse oximeter and I have established an overnight average spO2 level of 94.0% with a sigma of 0.2%. Then one night I measure an overnight average of 93.0%. That change is statistically significant; however, it is not clinically significant. In other words, there is not real need to worry about it, it does not in of itself indicate any kind of problem. But it does indicate that something has (likely) changed. It is reasonable to investigate what might have caused this change. Maybe it is due to congestion, maybe it is due to a new CPAP setting, maybe you have a tumor in your lung. It could just be noise and you may never find out.

TMI Naumann. Over and out.