I got a question about sample size. I really do not know how to calculate it. Here is the question:

Doctors are trying to enroll patients with a disease(i do not know). the study design is like this:

Treatment A then --->Treatment B
                    OR --->Treatment C

Treatment B then --->Treatment A
                    OR --->Treatment C

Treatment C then --->Treatment B
                    OR --->Treatment A

For treatment A, the good response is 70%; for B is 55% and 60% for C. 10% difference between treatments are clinical significant.

The problem is this is not 3*3 cross-over design, that is if patient got treatment A, he(she) probably get treatment B or treatment C, and same for other two groups.

Anyone can help me calculate the sample zise?

Thanks in advance.

[此贴子已经被作者于2008-7-30 3:48:40编辑过]