Hi there,
I'm going to be comparing the effects of p different drug treatments on
whether or not hospitilisation occurs. It's a prospective observational
study and thus logistic regression with binary varaibles for the drugs plus
relevant covariates such as baseline severity is a logical choice. My
problem is that I suspect that the main drug I'm interested in (my clients)
is much more likely to be prescribed to severe patients than the other drugs
and the resultant multicolinerity problems might make it impossible to
accurately gauge the drugs effect (worried that there might not be enough
cases of this drug in the less severe categories for the model to be able to
assess its effect). I've proposed to my client that I therefore stratify the
sample based on severity and run separate logistics in each. I might well
find that my odds ratios for the client's drug compared to a competitor are
similar across the severity groups) eg 95% confidence intervals overlap.
What I'd like to ask is does anyone have any suggestions on how to perform a
statistical test for homogeneity across these statra on this odds ratio? and
assuming homogeneity is not rejected how to compute a pooled estimate?
Any help much appreciated - I''ll summarise and post the replies.
Thanks,
Steve
DISCLAIMER: The information in this message is confidential and may be
legally privileged. It is intended solely for the addressee. Access to this
message by anyone else is unauthorised. If you are not the intended
recipient, any disclosure, copying, or distribution of the message, or any
action or omission taken by you in reliance on it, is prohibited and may be
unlawful. Please immediately contact the sender if you have received this
message in error. Thank you.
|