This chapter is published for information. It is an introduction to the use of design of experiments (DoE) for the screening and optimisation of analytical data.
The objective is to provide guidance on good practice as well as to set out the regulatory framework and critical aspects that need to be addressed.
This chapter complements general chapter 5.21. Chemometric methods applied to analytical data and general chapter 5.28. Multivariate statistical process control.
Chemometrics is defined in general chapter 5.21, where the first part indicates the possibility of designing or selecting ‘optimal performance experimental procedures, using techniques generally known as ‘design of experiments’.
In contrast to the traditional approach of exploring one variable at a time, the application of multivariate chemometric approaches makes it possible to develop suitable and high-performance analytical procedures by providing a rational way to explore problems defined by multiple variables and introducing a systematic approach for data evaluation.
A valuable feature of chemometric analysis is the production of graphical representations and numerical outcomes that allow the evaluation, exploration and interpretation of analytical data.
Depending on the intended purpose, numerous DoEs covering screening, optimisation and prediction may be used.
The aim of this general chapter is to introduce basic principles of design of experiments (DoE) as one of the areas of use of chemometrics, along with multivariate data analysis.
It focuses on the development, optimisation and assessment of the robustness of analytical procedures (rather than manufacturing processes).
By providing a list of possible applications in the field of pharmaceutical analysis, as well as comprehensive consideration of good practices related to DoE, this chapter aims to align the technical expertise for the use of DoE with regulatory expectations as set by the pharmacopoeial principles and ICH guidelines.
){kind=link}