Multivariate Statistical Analysis Principles and Applications in Dynamic Method Development

Abstract

Multivariate statistics is an extension of univariate statistics, combines both the mathematical modelling and spectroscopic, chromatographic analytical techniques to yield better results and helps in understanding of the hidden facts in a complex system of analytical techniques. It is a branch of applied statistics and is a form of multivariate statistics for designing, planning, and evaluating controlled experiments to predict the influence of multiple variables that have a combined impact of the desired outcome or response. Four different applications of pharmaceutical analysis are taken as case studies to evaluate the application of DoE in method development/ optimisation process with specific type of designs viz., 2- level factorial design, combined mixture design and dual design with historical data design are experimented in different case studies. Experimental design application is proven with detailed process of implementation in dissolution method development, assay method extraction solvent optimisation, combined assay and RS method development and impurities method development. Application of DOE in dissolution method development for Ezetimibe tablets are studied to evaluate the application of DOE in discriminatory dissolution method development. It is observed that pH of dissolution media, SLS concentration are having negligible positive effect on all the responses whereas RPM and dissolution media volume are having significant positive effect on all responses. The model is applied to predict possible dissolution methods that will yield necessary results of discriminatory dissolution method. Studied, DOE as an effective tool for optimising the extraction solvent to yield maximum filtration volume through 0.45 µ membrane filter.