Bioequivalence and Bioavailability Forum

Dear all!

I have a question regarding randomization and I would be happy to read your suggestions. The study design is as follows:

• I have two test items (T1 and T2) and one control item (C)
  
• two treatments can be tested simultaneously (within the same subject) leading to two separate readouts
  
• per subject (n=4, pilot study), between 8 and 13 pairs of applications will be performed (= 16 to 26 applications) so it will start with two simultaneous applications, then the next two and so on (maximum 13)
  
• it cannot be extended to three simultaneous applications

so we have the following possible pairs of applications: T1+C, T2+C, T1+T2

How should randomization look like for a subject?

• a 2:1 randomization using only T1+C and T2+C
  
• or a 1:1:1 randomization using all three possible pairs?

Am I correct that there is a risk, that the study will not be balanced as the number of pairs are not fixed and will vary between 8 and 13? is this a problem?

Thank you very much for your help!

Kind regards,
Alex