Bioequivalence and Bioavailability Forum

Dear xy (what a nick!)

❝ […] it is impossible to separate the period effect from the effect. But the following FDA guidance says:

❝ "A study can use a randomized crossover (e.g., S followed by S+I, S+I followed by S), a one-sequence crossover (e.g., S always followed by S+I or the reverse), or a parallel design (S in one group of subjects and S+I in another)."

»

❝ Is this true?

You are quoting correctly. I don't have the slightest idea what was going on in the minds of FDA's gurus when they suggested a »one-sequence crossover design«. Of course you can evaluate such a study, but imagine a situation when there's no 'true' PK interaction, but only a period effect (like: sunshine in period 1 and a thunderstorm in period 2). What will you get?
If you really want to follow this track, see this post - just the other way 'round (i.e., you assume than any effect is due to treatment and there's no period effect).
Since FDA also suggests a cross-over as one of the alternatives why don't you do it the 'easy way'?