Getting measure of blood pressure
From Dr Bernie Bedford, Southampton
I would be astonished if some blood pressure checks are not clustered around 150/90 (News, 2 March).
A single blood pressure reading is not a static measurement. It is subject to significant variability during the course of 24 hours and within a single consultation.
It is generally higher when patients visit hospital, and lower when you give the patient five minutes to relax, listen to the waves lapping on an imaginary desert island and support their arm while taking their BP.
If the pressure is within target, why take another reading? If it is above target, why not give them a minute or two longer to relax and overcome the white-coat bit as we are instructed to do by insurance companies.
Is it illegal to take one or two readings to allow the patient to relax and gain control of the consultation?
Blood pressure readings are subject to regression to the mean so clustering towards the target is almost bound to occur, since we aren't going to bother too much rechecking when readings are within target. This is simple basic statistics.
The problem, of course, is choosing to use a single random variable as a proxy for accurately predicting long-term outcomes. An average would surely be a far better predictor of potential long-term harm.