How can the results of the multiple logistic regression analysis be expressed as
LRs? This is currently one of my problems and I am very interested in an
explanation on the methods to be used. '
Usually LRs are supposed to be independent (independent Bayes) or the LRs are
calculated for sequential tests where the post-test odds of the first test are
used as pre-test odds for the following test, but symptoms are present
concomitantly and they are not independent. Are there any ways to adjust the LR
of a number of symptoms as can be done with logistic regression for the ORs?
Greetings
Roland Andersson
[log in to unmask]
Atle Klovning wrote:
>Often these LRs are based on mutliple logistic regression analyses, which
>may be the explanation of why you couldn't do the simple maths....
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|