st: xtgee, margins, and postestimation of odds ratios

[Chuck Cleland <ccleland@optonline.net>]

xtgee outA cov1 cov2 cov3 c.TIMEU##i.tx, link(logit) corr(unst) 
family(binomial) robust

  I then used margins to describe the outcome for each treatment 
condition and time point, adjusting for covariates:

margins tx, at(TIMEU=(0 .3333333 .6666667 1))

  I would like to get point and interval estimates for odds ratios at 
each of the four time points (comparing treatment and control). This is 
the closest thing I know of using margins:

margins r.tx, at(TIMEU=0) at(TIMEU=.3333333) at(TIMEU=.6666667) 
at(TIMEU=1) contrast(nowald effects)

  But the contrast is the risk difference rather than the odds ratio.  
Does anyone know how to obtain point and interval estimates of odds 
ratios with margins or some other approach?

Thanks,

Chuck Cleland

--
Chuck Cleland, Ph.D.
NYU College of Nursing
726 Broadway, 10th floor
New York, NY 10003
tel: (212) 992-9417 (Tu, W, Th)
tel: (732) 512-0171 (M, F)

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/