Bioequivalence and Bioavailability Forum

❝ For statitical anaylsis Semi-Replicated (TRR,RTR,RRT). Some pesrons do the followings:

❝ model= Sequence Period Treatment Sequence*Treatment as fixed effects.

❝ Random : Subject(Sequence) Treatment*Subject(Sequence).

❝ CI for T/R ratio derived using Treatment*Subject(sequence) as the error term.

❝

❝ Is this true under FDA or EMA guidelines.

• The FDA’s model is over-specified for the partial replicate design and might not converge → study done, no result. For details see this article.
  
• The EMA prefers ‘Method A’ (all effects fixed) over ‘Method B’ (subjects random).