Prasad, I have to assume you have specified a different model; The above table came from a BE performed as per FDA guidance;
Fixed terms : int+Sequence+Trt+Period
Random/repeated terms : Sequence*Subject
The EMA requests a different model where "Fixed effects, rather than random effects, should be used for all terms."; this leads to intra subject variances not being calculated; nothing to do with the software used - a result of the model specified. there have been more indepth conversations regarding this on other forums e.g.;
http://forum.bebac.a...=DESC#top4660/>To expand on my answer above, most probably you have a replicated design; for those cases CV is not output.
Since we have 2 variances, there are 2 random effects specified.
The first one is the residual variance (could be used for CVintrasubject calculation)
The second one is the subject based variance (could be used for CVintersubject calculation). That is, the subject is specified as a pure random effect (see EMA Q&A for details).
Basing on that:
CVintra = 100 * sqrt[ exp( Var (residual)) - 1]
CVinter = 100 * sqrt[ exp( Var(seq*subject)) – 1]
Please also review the WNL User Guide
Simon.
Edited by Simon Davis, 03 July 2018 - 04:46 AM.