Bioequivalence of brand-name and generic medicines

To test if the two formulations are equivalent we should use the same procedure employed in Section 2.7.4, where we made paired comparisons of the anal h ical capabihties of Arrhenius and Berzelius. Applying Eq. (2.28) to the values in the last column of Table 2A.3, we have 

Since this interval includes the zero value, we can consider the two medicines as bioequivalent, at least regarding the area below the curve. One point deserves comment, though. When we treated the data of Arrhenius and Berzelius, we saw that the paired comparison interval was much narrower, because the variation owing to sample origin had been eliminated. In the present example, the data behave in another way. The standard deviation of the differences is similar to the standard deviations of the averages of the two treatments. In Fig. 2A.5 we can see both the change in the average values and the similarity of the dispersions. We can also notice that the points seem to deviate a little from a normal distribution, but this should not be a cause for concern. As we have already said, the -test is quite robust to such deviations. 

---