Bioequivalence and Bioavailability Forum

If we take the example of a cross over study the standard balanced design has 2 periods, 2 sequences (A-B and B-A) with two treatments(test v reference) and an even number of subjects with a wash out between treatments. Now if there is only one sequence used for the crossover what is the best way to treat the data from the point of view of confidence intervals approach. I realize this is a non standard approach.

i have EquivTest and WinNonlin.


Angus