CombiStats: Equivalence testing
CombiStats v7 offers the option to display statistics required for equivalence testing. In equivalence testing, you typically consider the difference or the ratio between slopes. The confidence limits of that value have to fall within predefined limits. These limits have to be defined by protocol on the basis of historical information.
Here is example 5.1.4 from the European Pharmacopoeia to illustrate what it will look like in CombiStats. In the options for the Analysis of Variance, there is an extra checkbox to enable equivalence testing.
The output will now include an extra table with equivalence statistics. This table appears after the analysis of variance and before the potency estimates. It shows the estimated slopes of individual samples with their 90% confidence limits. It also shows differences between slopes and ratios between slopes, again with 90% confidence limits. If no standard is included in the assay, no difference or ratio is shown. In assays where the doses are not log-transformed, intercepts are shown instead of slopes. If a standard is included, differences between intercepts are shown, but not ratios because that is meaningless.