Log in
Register|
SKR ☆ 2009-02-03 06:53 (5946 d 06:05 ago) Posting: # 3175 Views: 4,261 |
|
|
Dear All, we are planing a two stage group sequential design to conduct a 2-way crossover bioequivalence study, and this study is going to be conducted in multiple sites. Here are some doubts regarding this issue. 1. how the statistical model should be framed? 2. which CI to use 90 % CI or 95 % CI or still more stringent? 3. As the study will be conducted at defferent sites, should we have show Treatment*site interaction effect also? 4. If study will complete in two groups, should our model consider site and group effects both? 5. If study passes in interim analysis (without second group), which effects should statistical model include? And if its proceed with second group, which effects should statistical model include? If any body conducted such a study, Please let me know some useful info..atleast some reference. Thanks in advance. Waiting for your valuable reply. Thanks, SKR |
|
MGR ★ India, 2009-02-05 11:32 (5944 d 01:25 ago) @ SKR Posting: # 3181 Views: 3,810 |
|
|
Dear SKR, As it is a two stage group sequential design, you may have to go for an alpha spending fuction (likely O'Brien-Fleming method). May be the following links helps you to get an idea on the study design. http://www.cytel.com/papers/usefun.pdf http://www.cytel.com/Products/East/example_02.asp#2 http://www3.interscience.wiley.com/journal/113398744/abstract http://www.biostat.wisc.edu/Clinical_Trials/pdf/ron_talk.pdf http://www.cytel.com/papers/SymposiumNJ2002.pdf http://biostat.hitchcock.org/BSR/Analytics/Sequential/SpendingBounds.asp Note: I am not the expert in this topic, but we have to wait for the comments from the experts. Thank you, — Regards, MGR |
|
SKR ☆ 2009-02-05 12:06 (5944 d 00:51 ago) @ MGR Posting: # 3182 Views: 3,327 |
|
Ing. Helmut Schütz