Calculating geometric mean ratios by subgroups

[Kisty Le Daore]

I am trying to calculate geomeric mean ratios of antibody concentrations by HIV status for a vaccine vs placebo trial where the vaccine has 4 components and the placebo 2.
My syntax is fitint glm outcome [in my case HIV status] variable [log10 antibody concentration for each vaccine component], f(bin) l(logit) eform
Is there an easier way to do this than putting in each vaccine component separately?

I then want to compare the geometric mean concentration of each of the vaccine components (continuous variable) in vaccine vs placebo by HIV status (binary)
I am not sure how to do this

thanks all

[Nick Cox]

I had to search for information on fitint, which is from SSC. I don't think it is often used and I don't recognise the authors' names as active here.

If you don't get a better answer you may need to back-up and give both example data and the fitint results as there isn't much to go on in #1. I am completely unclear on what data structure you have, although people in medical statistics may have a much better idea.

Geometric means aren't difficult as even if they are not directly available the recipe exp(ave(log()) is quite easy to code.