Use of experimental design in the pharmaceutical industry.

Statistical modelling and experimental design (SMED) are essential tools for the development and understanding of complicated products and processes. SMED allows efficient experimentation in which all or a large subset of factors are together varied over a set of experiments, in contrast to the traditional approach of varying only one at a time. An overview of the SMED methodology and the generalization of statistical design to multivariate design is presented. The following examples illustrating the use of these methods are discussed: (1) use of factorial designs to improve drug solubility; (2) testing the robustness of an analytical method; and (3) use of multivariate design to select the solvent in analytical method development.