Dear All, Few months before I attended webinar concerning replicated designs using Phoenix WNL. Must say amazing piece of work it is. In Webinar, Box plot method was suggested based on Studentized intra-subject residuals for EMEA to prove higher CV is not caused by outlier. Is that method suggested by EMEA than any other method like Cooks distance and hotelling T[sup]2[/sup] applicable for this? I wanted to go with replicated study aproach for EMEA submission. but confused regarding Outlier method since I dont have SOP on this method? Is it Ok if I mentioned this box plot method on protocol only? Thanks in Advance. Regards, Tushar
Dear Tushar, The EMEA Guidelines do not mention any specific method but just says that outliers should be reviewed: " The applicant should justify that the calculated intra-subject variability is a reliable estimate and that it is not the result of outliers ". The proposed intra-subject studentized residuals box plots is one way of determining outliers but not the only one. This method is described on: Page 224 in the second edition of Chow/Liu book – Design and Analysis of BA and BE studies, equation 8.2.9. Hope this helps, Ana
Hi Tushar, I asked EMA’s panelists at the joint EGA/EMA-workshop (London, June 2010) which one of the many statistical outlier tests they would prefer. The answer was clear: “We don’t want statistical tests.” So – as a joke – I replied “What do you expect instead. Box plots?” I aimed it as a joke because box plots are essentially nonparametric – a topic which in the past gave EMA the creeps. Remember that EMA removed the comparison of t[sub]max[/sub] from the GL just to avoid nonparametrics? To my great surprise they said “Yes, box plots are a great idea.” Stupid, but see the quote from the Q&A-document (page 21) published after the workshop: On a case by case basis, a study could be acceptable if the bioequivalence requirements are met both including the outlier subject (using the scaled average bioequivalence approach and the within-subject CV with this subject) and after exclusion of the outlier (using the within-subject CV without this subject). An outlier test is not an expectation of the medicines agencies but outliers could be shown by a box plot. This would allow the medicines agencies to compare the data between them.