Continued verification of pharmaceutical & biologic manufacturing processes is often performed using univariate statistical process control (SPC) charts with conventional mean +/- 3 SD limits. In cases where the number of attributes is large (e.g. 10 or greater) the false alarm rate (Type I error) for the overall system of charts is significant, resulting in overreaction to individual events and fulfillment of the dangers cited in the 2011 FDA process validation guidance. This has significant consequences for the utilization of resources in investigating such events as well as the potential for undermining the credibility and effectiveness of the monitoring program.
This webinar presents an alternative approach whereby more appropriate SPC limits may be applied to each attribute so as to minimize the system-level Type I error rate yet simultaneously detect unintended process variability.
Learning Objectives:
- Understand the significant potential for false alarms when using SPC charts for process and product monitoring.
- Learn how to minimize false alarms by adjusting the way statistical limits are calculated to account for anticipated process changes.
Contributors: Jeff Gardner is President & Principal Consultant of DataPharm Statistical & Data Management Services based in Cleveland, Ohio. DataPharm provides statistical consulting, data management, and analytics talent recruiting services to pharmaceutical & biotech manufacturers throughout the US and Europe. The company focuses on helping its clients optimize existing manufacturing & laboratory processes as well as drive the efficient evolution of next-generation processes to reduce costs, increase output, and ensure consistency of output.
Prior to launching DataPharm in 2012 Jeff worked as a non-clinical statistician for multiple companies in the pharma/biotech sector including Abbott, Amgen, and Boehringer-Ingelheim. In his professional roles he has served as the lead statistician responsible for ensuring manufacturing process quality and high performance of laboratory test methods. An avid SAS programmer, Mr. Gardner has also during his career developed and validated proprietary applications for automating data collection, analysis, and reporting. He received his BS in Mathematics from Marquette University in Milwaukee, Wisconsin in 1999.