Bioequivalence and Bioavailability Forum

Dear All,

Plz give your valuable suggestion on below mentioned situation.

For reference replicated design, in case of missing observation, if the reference treatment data is available in two periods for more number of subjects, but both test and reference treatment data is available for lesser number of subjects (i.e. subjects data is available for all three periods) then on what no. of subjects the Statistical evaluation should be done ? i.e. if we calculate reference variability (or standard deviation) on available R-R data, then should this variability be considered in Scaling & calculation of 95% upper bound or it should be done on lesser number of subjects completing both test and reference treatment and having all three period data?

---

Edit: Category changed. [Helmut]