Optimal Design Development for a Chemical Kinetic Model with Random Batch Effects: A Multi-Way Collaboration

• Activity Number: 206
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Quality and Productivity Section
• Author(s): Ketil Tvermosegaard*+ and John Joseph Peterson and David Woods
• Companies: GlaxoSmithKline and GlaxoSmithKline and University of Southampton
• Keywords: nonlinear mixed effect model ; hierarchical model ; quality by design ; statisical consulting ; pharmacokinetic model

Abstract

This presentation reviews the collaborative and technical aspects of developing methodology and computer code for constructing an optimal design for a class of chemical kinetic models with random batch effects. Chemical kinetic models can play an important role in pharmaceutical manufacturing. In the pharmaceutical industry, there has been a push by financial incentives and industry regulators to make better efforts to design quality into manufacturing processes. While optimal design can aid in building better models for less cost, appropriate methods and computer software were not available due to the particulars of this class of problems. Recent success in developing optimal designs for pharmacokinetic models, led us to believe that our (similar but different) problem was solvable, but not without a deep understanding of optimal design theory and considerable effort in highly technical statistical computing. A practical solution to this problem required a triumvirate of different entities with appropriate knowledge, focus, and skill. This talk will outline how these worked together to produce a practical solution to an important industry problem.