Bioequivalence and Bioavailability Forum

Dear Ratnakar,

❝ […] whether we can consider the subjects who have not completed all four period (i.e. subjects completed 3 period, 2 or only 1 period of the study) or we have to consider subject who have completed all 4 periods of the study […]

Since replicate studies are usually evaluated with SAS Proc MIXED (see FDA Guidance “Statistical Approaches to Establishing Bioequivalence” Appendix E for code) or equivalent software you can include subjects with missing data. But you have to assume that these missings are missings at random.
For “missings at random” or “completely missing at random” Goooogle is your friend.

To my restricted knowledge no regulatory body will have concerns if you include subjects with missing data in your statistical analysis. Except of course that missing data are not the rule, which will shed doubt on the quality of the performance of your study.
Quite the contrary you will have to justify why you have not included certain subject.

The Draft of the EMEA guideline explicitly states:
“All treated subjects should be included in the statistical analysis, with the exception of subjects in a crossover trial who do not complete at least one period …”