REPEATED and random intercept models are not supported？
Closed this issue · 4 comments
Dear all.
As I know, repeated models and random intercept models are also commonly used in trials.
Additionally, covariance structure TYPE=FA0(2) is also needed.
So far, I cannot apply them with mmrm. If these features are already included in mmrm, Please let me know.
Thank you.
Hi @kunghiu ,
thanks for the feedback!
- Regarding random intercept, this is discussed in #300 please feel free to add there further thoughts
- For FA0(2) I understand from https://documentation.sas.com/doc/en/statcdc/14.2/statug/statug_mixed_syntax14.htm#statug.mixed.mixedcovstruct and the corresponding details that this is only useful in combination with random effects, because it is not positive definite when having more than 2 time points. Can you maybe say more here why you need FA0(2), maybe with an example?
Thank you for your reply, @danielinteractive .
In this FDA guidance, https://www.fda.gov/regulatory-information/search-fda-guidance-documents/statistical-approaches-establishing-bioequivalence-0, and many draft guidances before this one, they all use FA0(2) in proc mixed model. And this has become a standard procedure in the whole ANDA statistics. Softwares like winnonlin, they all mimic SAS and use FA0(2) structure.
However, if mmrm donot solve SAS REPEATED-like clause, only FA0(2) being added is valueless.
Thank you for your reply, @danielinteractive .
In this FDA guidance, https://www.fda.gov/regulatory-information/search-fda-guidance-documents/statistical-approaches-establishing-bioequivalence-0, and many draft guidances before this one, they all use FA0(2) in proc mixed model. And this has become a standard procedure in the whole ANDA statistics. Softwares like winnonlin, they all mimic SAS and use FA0(2) structure.
However, if mmrm donot solve SAS REPEATED-like clause, only FA0(2) being added is valueless.