Bioequivalence and Bioavailability Forum

We have done a BE study in normal volunteers using equal numbers of male and female subjects. I ran the BE data analysis in the usual way in Phoenix WinNonlin 6.3 and obtained the confidence intervals.

Since we suspect a gender difference for this drug I then included gender in the model as recommended by Helmut (see below)

"Have a column denoting gender (0/1, f/m, whateveryoulike...) and map it in the BE setup as "Classification".
Model Specification (Fixed Effects):
sequence+period+treatment+gender+treatment*gender

Example from one of my studies (6×3 Williams’ design):"

I repeated my data analysis for BE and I found that the confidence intervals were slightly different. The outcome for the BE assessment was not changed.

The ANOVA p value was 0.08 so the gender effect was not significant. Under what circumstances would one use the BE data analysis with gender in the model.

Angus

---

Edit: Category changed. Plese don’t use all capitals in the future. [Helmut]