Challenges in Inferring Statistical Results from Scale-Down to Manufacturing Scale

• Activity Number: 114
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Daniel Coleman*+
• Companies: Genentech
• Keywords: DoE ; lab-scale ; manufacturing

Abstract

The 2011 PV guidance defines three stages corresponding to levels of knowledge. The first stage, Process Design, concerns the development of the commercial manufacturing process, through small lab-scale studies and scale up. The second stage, Process Qualification, assesses the process design in relation to reproducible commercial manufacturing. Due to the many constraints on manufacturing systems, e.g., cost, almost all multifactorial designed experiments (DoE) used for process design and characterization during Stage 1 are performed in scale-down systems. This talk describes the challenges in inferring statistical results from these experiments to manufacturing scale and implications on Stage 2. Inference relies on both a mechanistic understanding of the systems as well as an empirical comparison of the systems at different scales with limited data from the manufacturing system. Examples from both biologics and small molecule manufacturing are discussed.