It is usually the upper limit if the 95% Confidence Interval for
whatever the parameter is. However sometimes the upper limit of normal
is a cut-off, above which the "target disorder" has a high enough
probability of being present for it to be worth acting on the test
result - eg PSA of 4
Toby
In message <[log in to unmask]>, Mary Hawking
<[log in to unmask]> writes
>Interesting reading.
>*Does* anyone have any data on the upper limit of normal anything - and
>how it is scientifically derived? ;->
>
>http://seattletimes.nwsource.com/news/health/suddenlysick/
>
>MaryH
--
Toby Lipman
Chair, North of England Faculty Board, RCGP
|