|
kaushalraval ☆ India, 2008-12-22 09:07 (6390 d 21:24 ago) Posting: # 2959 Views: 3,790 |
|
|
We want to do five-way crossover study using two reference and three test formulations? Can anyone please explain how anova will be constructed? Thanks |
|
boonchai_l ☆ Thailand, 2008-12-24 09:56 (6388 d 20:36 ago) @ kaushalraval Posting: # 2969 Views: 3,010 |
|
|
First, it's the most important, I think you must concern about ethic issue or the total of blood that you draw from subject. Second, are you sure that you review study objective completely? Because if you use this design the maximum advantages will occur when you compare all combination but if not it's useless to select this design. For example if you only want to compare between all tests and all reference (don't want to compare among tests or among reference) I suggest you should select incomplete block design, because it spend less than 5 periods. However the incomplete block design also can compare all combination but you may have many difference sequences (it depends on the number of period that you want). If you select complete crossover design (someone call latin square design) it means you have to spend 5 period and 10 difference sequence. Next, for longtime period or many periods trial, you have to concern many things such as the characteristics of drug and its disposition, availability of subject and dropout rates. About, how anova will be constructed? I think it's difficult to explain in hear. |
|
kaushalraval ☆ India, 2008-12-24 10:20 (6388 d 20:12 ago) @ boonchai_l Posting: # 2970 Views: 2,997 |
|
