←back to thread

330 points wglb | 1 comments | | HN request time: 0.217s | source
Show context
crazygringo ◴[] No.41841006[source]
There's definitely a huge variance in blood pressure readings depending on posture, relaxation, arm position, recent activity, etc. If you buy a blood pressure monitor, it's really interesting to see how "random" a single reading at the doctors' is, and how large your fluctuation throughout the day is.

That being said, it really makes me wonder about studies that correlate blood pressure with other things. Is the blood pressure really being measured "correctly" in all those studies? Or not?

In other words, if your "true correct" blood pressure is lower than what the doctor normally takes, but then a lot of the studies are based on real-life "incorrect" higher blood pressures, then don't you similarly want an "incorrect" higher reading for consistency? Or are the studies always really done with far more accurate blood pressure readings, where the patient sits still for 5 min beforehand, keeps their legs uncrossed, is totally free of stress and anxiety, didn't exercise beforehand, etc.?

replies(18): >>41841065 #>>41841066 #>>41841097 #>>41841378 #>>41841596 #>>41842242 #>>41842355 #>>41842512 #>>41842978 #>>41842987 #>>41843132 #>>41843291 #>>41843843 #>>41843882 #>>41844009 #>>41844704 #>>41845158 #>>41848507 #
hombre_fatal ◴[] No.41841097[source]
As long as blood pressure variance is randomized then you’re getting a signal, and that happens when everyone is measured in a similar way (or randomized to different ways). You don’t need perfect precision.

What you don’t want is to make everyone who, say, smokes wear the cuff while standing up while everyone else gets cuffed lying down.

replies(1): >>41841152 #
1. crazygringo ◴[] No.41841152[source]
> and that happens when everyone is measured in a similar way (or randomized to different ways)

But that's exactly the issue -- that the similarity or randomness is one way for one study, and another way for another study, because of culturally different sets of nurses and/or patients.

In other words, you're right it's not affecting results within a study, but it makes comparison between studies questionable.

And it makes it equally questionable whether a study's results apply to you, if your signal is 10 units off of a study's signal, and occurs across a cutoff that determines whether you should take a medication or not.