Statistical Quality and Process Control in Biopharmaceutical Manufacturing-Practical Issues and Remedies

Area of Science:

• Biopharmaceutical Manufacturing
• Process Control
• Statistical Quality Control

Background:

• Biopharmaceutical manufacturing relies on statistical quality and process controls (SQC and SPC) for process improvement.
• Typical SQC and SPC methods may not adequately address the unique characteristics of bioprocess data.

Purpose of the Study:

• To highlight characteristic features of bioprocess data and their impact on SQC and SPC applications.
• To evaluate the performance of control charts (I-chart) and process performance index (Ppk) with bioprocess data.
• To provide remedies for robust biopharmaceutical process monitoring.

Main Methods:

• Simulated bioprocess data with inhomogeneity, nonstationarity, autocorrelation, and outliers were used.
• Analysis focused on standard deviation estimates for 3σ limits and Nelson's rules.
• Comparison of five approaches for treating censored data (≤LOQ) and their impact on 3σ limits and Ppk.

Main Results:

• Inhomogeneity, nonstationarity, autocorrelation, and outliers significantly impact SQC and SPC applications.
• Estimates of within and overall standard deviation affect 3σ limits and sensitizing rules.
• Censored data treatment influences 3σ limits and Ppk estimates.

Conclusions:

• Applying standard SQC and SPC without accounting for bioprocess data's atypical nature increases false alarm rates.
• This can negatively affect batch release decisions, leading to the rejection of good batches.
• Robust control charts require addressing specific bioprocess data challenges for efficient monitoring.