[R-sig-ME] help with mixed modeling

[Daniel Cher]

I need help with a meta-analysis project. The fundamental design is this:

*          3 trials, 2 of which are randomized trials

*          Continuous outcome measure (pain score), assessed at multiple time intervals (see plot below).

*          I'd like to calculate the effect size across all time points in months 1 to 6 taking into account: 1) repeated measures within subjects, and 2) random effect across sites
 I'm an R programmer, and have used gls<https://stat.ethz.ch/R-manual/R-devel/library/nlme/html/gls.html> from the nlme package. However, I get quickly lost when it comes to LS means and appropriate covariance structures.
 I appreciate any advice or reference to a person who can help.

Daniel Cher, MD
Vice President of Clinical Affairs
Tel: 408-207-0700 x2233
Mob: 650-269-5763
FAX: 888-243-4502
dcher at si-bone.com<https://bl2prd0510.outlook.com/owa/redir.aspx?C=ZquXLWlJAUucgyZxRxKb4zmkXsdVH88IFjGENZQYTO-GuKZ_ASTriHzF7kSzRDs8HnsZfZz4H5Y.&URL=mailto%3adcher%40si-bone.com>

Notice: This message, including any attachments, may con...{{dropped:18}}