Log in
Register|
jag009 ★★★ NJ, 2014-08-05 21:34 (4342 d 09:12 ago) Posting: # 13337 Views: 7,414 |
|
|
Hi everyone, If you run a multiple (2)-group crossover BE study and there is a group effect, you run the stats separately for each study group and the results showed that one group pass BE and one fails. What will happen to your submission? Lets say both groups have acceptable sample size. Just curious if anyone has experience on this. Thanks John |
|
ElMaestro ★★★ Denmark, 2014-08-06 01:13 (4342 d 05:33 ago) @ jag009 Posting: # 13338 Views: 6,272 |
|
|
Hi John, ❝ If you run a multiple (2)-group crossover BE study and there is a group effect, you run the stats separately for each study group and the results showed that one group pass BE and one fails. What will happen to your submission? Lets say both groups have acceptable sample size. Just curious if anyone has experience on this. Fails how? Partial overlap with acceptance range = no problem (but if PE's are subjectively distant from each other you'll need plenty violins). No overlap with acceptance range = big, ugly, evil trouble. Anyways, a group effect itself isn't necessarily a showstopper. It's a test of another hypothesis anyway. What about the originally planned BE test? By the way, I sent you an email 11 days ago. Went into the spam folder, perhaps? — Pass or fail! ElMaestro |
|
Dr_Dan ★★ Germany, 2014-08-06 09:55 (4341 d 20:51 ago) @ ElMaestro Posting: # 13342 Views: 6,240 |
|
|
Dear John, I totally agree with ElMaestro. If you assess the groups separately you will have inflated 90%CI due to the lower number of subjects. Therefore a partial overlap with acceptance range should be no problem and to my experience this is the usual outcome. If you have no overlap with acceptance range you have shown for your product that (at least for this group) it is bioinequivalent. This group effect is a showstopper as long as you have no reasonable explanation for the different behavior of this group. Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan |
|
jag009 ★★★ NJ, 2014-08-08 18:49 (4339 d 11:57 ago) @ ElMaestro Posting: # 13354 Views: 6,164 |
|
|
Thanks ElMaestro and Dr. Dan, What if the drug is HVD? ElMaestro: Sorry your email was stuck in my junk/spam box. Did you get my reply? Thanks John |
|
Dr_Dan ★★ Germany, 2014-08-11 01:59 (4337 d 04:47 ago) @ jag009 Posting: # 13359 Views: 6,088 |
|
|
Dear John ❝ What if the drug is HVD? The same applies. Your 90%CI will be even wider due to the high variability of the drug and the low number of subjects and you will certainly get only a partial overlap with the acceptance range. Kind regards Dr_Dan — Kind regards and have a nice day Dr_Dan |
|
pash413 ★ India, 2014-08-18 10:26 (4329 d 20:20 ago) @ Dr_Dan Posting: # 13401 Views: 5,962 |
|
|
Dear All I would like to know the significance of group effect in statistical model. |
Ing. Helmut Schütz