Application of linear mixed effects models to the evaluation of dissolution profiles.

The performance of linear mixed effects models for the comparison of dissolution profiles is examined. This type of model is frequently used by statisticians, but is rather unknown to people that work in dissolution laboratories. Hence, an extensive theoretical part was introduced to make the methodology more accessible. Firstly, repeated measures ANOVA is discussed, followed by the "real" linear mixed effects models. The theory is applied to two types of dissolution data: one corresponding to an immediate and another to a slow release formulation. We tried to use as much as possible the standard settings of the statistical software (S-plus). Suggestions are given to solve problems encountered during model fitting. It was found that the statistical limits are much more discriminative than the similarity factor.