Predictive Ppk calculations for biologics and vaccines using a Bayesian approach - a tutorial

Area of Science:

• Pharmaceutical Manufacturing
• Biopharmaceutical Process Development
• Statistical Process Control

Background:

• Speedy process development in biologics and vaccines can compromise robustness, leading to issues post-launch.
• Current methods assess process capability (Ppk) late in commercialization, hindering timely adjustments.
• Biopharmaceutical development often focuses on Critical Quality Attributes without fully assessing process robustness.

Purpose of the Study:

• To propose a Bayesian-based methodology for predicting manufacturing process performance at scale during development.
• To enable early identification of process vulnerabilities and areas for improvement.
• To enhance process robustness and knowledge prior to product launch.

Main Methods:

• Utilizing Bayesian statistics to incorporate limited development data, similar product data, and Subject Matter Expert (SME) knowledge.
• Formulating informative priors within the Bayesian framework to predict process capability.
• Applying the predictive Ppk approach at key development stage-gates.

Main Results:

• Demonstrated a method to predict long-term process capability using limited data and prior knowledge.
• Provided early insights into potential process vulnerabilities before commercialization.
• Enabled data-driven prioritization of development efforts for enhanced robustness.

Conclusions:

• A Bayesian approach can predict manufacturing process performance before commercial data is available.
• Integrating predictive Ppk assessments at stage-gates facilitates continuous process improvement.
• This methodology leads to more robust manufacturing processes and increased process knowledge at launch.