I know this is slightly off the mainstream of EBHC but nevertheless the same
principles of being aware of the meaning of things expressed as numbers and
asking the right questions apply.
Producing statistics of physician (or group or unit or whatever) performance
seems to be one of the great growth industries in medicine. Graphs of
performance in just about anything seem to be produced - usually with
something that looks at first glance like a normal distribution (and almost
never with any statistical addenda). But I would like to know whether we can
use them sensibly as anything other than pictures. In particular when I am
one of the subjects of the analysis how do I interpret my own performance?
I don't recall seeing this discussed in mainstream (UK based) journals so
probably either its a stupid question or there's no easy answer. I have some
ideas but I'm not a statistician.
So let's try a few scenarios:
1. Local radiology dept feedback usage stats for a procedure (say plain
lumbar spine) to 30 primary care practices. They circulate a bar chart of
number of requests /1000 registered patients and put a few values on the y
axis. I know which bar is my practice (and I know how many patients I have)
but I don't know who any other bar is (although I know the approximate total
number of patients of all the practices). There is no mean supplied although
the median is easy to see and because the distribution is fairly symmetrical
the mean is probably about the same. My practice is the second lowest user,
I want to know if I am behaving differently from most of my colleagues or
whether it is just a random thing. My practice population is somewhere near
the median of the distribution so it's not a very small sampling effect.
2. The above radiologists didn't limit it to plain L spine, but added in
abdo ultrasound, and barium enema. I'm the lowest user of abdo ultrasounds
but in the middle of the pack for ba enema. Do I pool them or analyse
separately?
There are a few more but perhaps I'll see if anyone bites on this first.
Chris
Dr Chris Burton
GP, Sanquhar, Dumfriesshire
Member of WestNet, the West of Scotland Primary Care Research Network
http://medicine21.com/heartGP
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|