RESEARCH ARTICLE Application of Multivariate Tools in Pharmaceutical Product Development to Bridge Risk Assessment to Continuous Verification in a Quality by Design Environment Simeone Zomer & Manish Gupta & Andy Scott Published online: 14 September 2010 # Springer Science+Business Media, LLC 2010 Abstract An important aspect of a quality by design approach to pharmaceutical product formulation and pro- cess development is continuous quality verification. This is an innovative way of validating the process where manufacturing performance is continuously monitored, evaluated and adjusted as necessary. For new drug products, the body of knowledge accumulated through the development cycle and formalised via risk assessment forms the natural basis of this activity. This paper shows how multivariate tools can be used as part of a continuous quality verification approach for a new drug product relying on the information that summarises the control strategy, i.e. the subset of critical variables selected via risk assessment and the related proven acceptable ranges determined during developmental studies. Keywords Quality by design . Multivariate analysis . Continuous quality verification . Risk assessment . Data trending Introduction Quality by design (QbD) is an approach to product development and manufacturing that is currently advocated by the Food and Drug Administration (FDA) and other regulatory agencies around the world [1, 2]. The essence of QbD is that quality should be built into the product with a thorough understanding of the product and process by which it is developed and manufactured along with a knowledge of the risks involved and how to best mitigate those risks. For the pharmaceutical industry, the adoption of QbD represents both an opportunity and a challenge. The manufacturer may be granted greater regulatory flexibility if proven that product quality is adequately controlled and the related manufacturing processes are well understood. However, this approach constitutes a shift from a traditional way of working for a conservative industry [3] and there are relatively few examples yet available to provide a framework [4–7]. Much of the debate focuses on the most appropriate determination of a design space, defined as “the multidimensional combination and interaction of input variables and process parameters that have been demon- strated to provide assurance of quality”. Moving within the design space is not regarded as a change and does not require regulatory pre-approval. Because regulatory agen- cies provide freedom to the applicant on how to best define the design space, depending on the application and on the volume and characteristics of the data that are generated at a given stage of the product life cycle, a variety of approaches are proposed, including mechanistic models [8], response surface methodologies based on design of experiments (DoE) [9], Bayesian approaches [10] or multivariate methods linking in a feed-forward control scheme the process [11]. A brief review of some of the S. Zomer (*) : A. Scott Chemometrics Group, Process Understanding and Control, Pharmaceutical Development, GlaxoSmithKline, New Frontiers Science Park (South), Third Avenue, Harlow, Essex CM19 5AW, UK e-mail: simeone.2.zomer@gsk.com M. Gupta Product Development, Pharmaceutical Development, GlaxoSmithKline, Upper Merion R&D Laboratory, 709 Swedeland Road, King of Prussia, PA 19406, USA J Pharm Innov (2010) 5:109–118 DOI 10.1007/s12247-010-9085-z