Overview of Definitions and Regulatory Guidance Supporting DoE and MVA

Since the inception of the Pharmaceutical cGMPs for the 21st Century - A Risk-Based Approach and the FDA’s PAT — A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance , regulated industries have been provided with a golden opportunity to apply state of the art process analysis equipment and modern quality control systems to their processes, with the fundamental goal of better process understanding and knowledge management.

Overview Design Monitor & Control GenX products Ask our experts Contact us

The basic premise is, better understand your processes, improve quality and main result will be a reduction manufacturing cost. Fundamentally, cost cutting as a first step does not lead to improved quality, quite the opposite! Studies have actually shown that this leads to higher manufacturing costs due to more scrap and increased product recalls!

Processes are multifaceted, i.e. they cannot be monitored and controlled by looking at single process variables, or many process variables independently. The Six Sigma initiative has raised industrial awareness of the need for better monitoring systems, but fundamentally lacks the appropriate tools for complex process analysis. This is where the FDA PAT framework guidance lists the four important aspects of PAT as being (in order).

• Multivariate tools for design, data acquisition and analysis
• Process analysers
• Process control tools
• Continuous improvement and knowledge management tools

Multivariate tools encompass both Design of Experiments (DoE) and Multivariate Data Analysis (MVA) routines, which take into account all process/formulation variables simultaneously, but most importantly, their interaction. It is the interaction of variables that leads to most process failures and this is one of the major pitfalls of univariate Statistical Process Control (SPC).

The diagram below briefly describes the advantage of MVA over univariate analysis for the simplest, two process variable case.

MVA over Univariate Analysis

When pH and Temperature are plotted together, a linear relationship exists between them, i.e. they are dependent on each other. The grey region in this plot shows the allowed regions for univariate control charts. The abnormal (red) point would therefore be allowed in a univariate situation, but not allowed in a multivariate control strategy (defined by the ellipse around the plotted points). Only MVA tools could pick up this process fault.

Tools such as DoE can be used to define and establish what is known as the Design Space. The concept of design space was first introduced in cGMP’s for the 21 st century and later elaborated on in the International Conference on Harmonisation (ICH) guidance on Pharmaceutical Development (Q8) . Q8 defines the Design Space as

"The multidimensional combination and interaction of input variables (e.g., material attributes) and process parameters that have been demonstrated to provide assurance of quality. Working within the design space is not considered as a change. Movement out of the design space is considered to be a change and would normally initiate a regulatory post approval change process. Design space is proposed by the applicant and is subject to regulatory assessment and approval."

Design Space Representation

This definition fully encompasses the use of DoE and MVA tools for defining the design space. In particular, two important notations are described. The first are Critical to Quality Attributes (CQA’s). These define the attributes of a product that are deemed most critical to the final quality and effectiveness of the drug substance. CQA’s are monitored and controlled by isolating those process parameters that most influence them. These are defined as Critical Process Parameters (CPP’s). Implementing strategies to monitor and control CPP’s within the Design Space ensures that the CQA’s for the product are met i.e. this is known as the Desired State . Therefore " PAT and QbD is process centric, not a product centric approach to quality "

Next     Multivariate tools for defining the Design Space