Thomas Gsponer

Dr. Thomas Gsponer, CSL Behring
Thomas Gsponer studied mathematics at the Swiss Federal Institute of Technology in Lausanne (EPFL) and got his PhD in statistics in 2004 from EPFL. Main area of interest is Bayesian statistics which he applied successfully in pharmaceutical industry and epidemiology between 2004 and 2012. Since 2012 Thomas Gsponer is working at CSL Behring in the quality department of the manufacturing site in Bern.

Abstract
Economic pressure and globalization affect the pharmaceutical industry. A possible strategy in this environment are site-to-site technical transfers where manufacturing processes are transferred from one manufacturing site to another. At the same time, pharmaceutical industry undergoes a shift in paradigm: the understanding of quality evolves from 'Quality by Testing' to 'Quality by Design'. This shift is also reflected in the increasing number of regulatory guidance documents (e.g. ICH Q8, Q9, Q10, Q10). The message is clear: understanding process variation is key to success (quality). Tools like exploratory data analysis, design of experiments and statistical process control are essential to achieve Quality by Design.

With site-to-site technical transfers, several challenges arise, in particular concerning data management and statistical process control. For two products transferred to other sites, we developed a pilot data warehouse. Together with Statistica Enterprise Server and R CSL Behring has established a powerful system for cross-site process monitoring. We illustrate the system by means of examples and show how Bayesian model selection methods support statistical process control and facilitate knowledge exchange and process understanding.