The Past Decade of Quality by Design: A Perspective Overview

By Brad Swarbrick, owner, Quality by Design Consultancy

Brad Swarbrick, owner, Quality by Design Consultancy


In 2011 when the US Food and Drug Administration (US-FDA) published their guidance document on Process Validation (PV), this represented the first real attempt by a regulatory authority to mandate the implementation of the Quality by Design (QbD) approach to process validation. By mandate, although the PV document is only a guidance document, the US-FDA stated that moving forward, all new submissions must be based on a QbD approach.

What this translated to, compared to the traditional approach to process validation, is that the three batch validation system is no longer acceptable and a Continuous Verification approach must be adopted. This is where industry has to look deeper than just looking at increasing QC testing (this approach is known as “Quality by Testing” and has no part in the new guidance). The original cGMPs for the 21st Century document (the foundation guidance document for QbD) makes the following, well known, statement,

“Quality cannot be tested into products; it must be built in or by design”

The QbD initiative was introduced around the same time as the Process Analytical Technology (PAT) initiative and it was originally envisioned that QbD and PAT would be tightly interrelated in order to implement state of the art process monitoring technology and modern knowledge management systems in a scientific, risk based approach. This approach would ultimately achieve the vision of the PV guidance document. The key point of PAT is that it provides “timely measurements” that relate to product quality (as the product exists in the process) and this is the only way to achieve Continuous Verification. But PAT alone cannot do this, therefore the International Conference on Harmonization (ICH) prepared four key documents as a framework for implementing QbD into both primary and secondary manufacturing operations. These documents are the ICH Q8, Q9, Q10 and Q11 guidance on Pharmaceutical Development, Risk Assessment and the Pharmaceutical Quality System (PQS).

QbD is often referred to as a Paradigm Shift from the dark old days to a modern approach to pharmaceutical manufacturing. The definition of a Paradigm Shift is provided below for reference,

“A fundamental Change in Approach or Underlying Assumption”

The QbD Paradigm Shift

Based on the above definition, QbD is a major paradigm shift compared to the traditional manner in which pharmaceutical processes are validated and monitored. The recent approval of the Continuous Manufacturing System (CMS) at Vertex Pharmaceuticals by the US-FDA represents the first of its kind for a total QbD manufacturing system. A CMS is by definition a QbD process for the following reasons,

1. The process is designed to adapt to changes in raw materials so that it can produce consistent product within the Design Space established for the process.

2. PAT is used to monitor the passage of intermediate product from one operation to the next, with rejection systems in place to remove out of specification sub-batches.

3. A CMS allows for complete batch tracking and 100% continuous verification of each batch manufactured.

As with any batch or continuous manufacturing process, the only way to achieve timely measurements is through the implementation of PAT tools, such as Near Infrared (NIR) spectroscopy, Raman spectroscopy and Particle Size analyzers that operate in an in-line manner. Tsuchechnology must be capable of measuring one or a number Critical Quality Attributes (CQAs) in a manner where proactive, not reactive quality decisions can be made, and  within the design space.

PAT tools such as NIR Spectroscopy (NIRS) and Raman spectroscopy typically rely on the methods of Chemometrics to predict quality attributes from the spectra generated. These models can be qualitative, i.e. classification or trending based, or quantitative. In order to make these predicted values useful in continuous verification strategies, the outputs from the spectrometer and chemometrics models must be fed into an Advanced Process Control (APC) system. These systems agglomerate, store, and use the data to make real time quality decisions on the process.

A key concept in the QbD approach is Knowledge Management and its utilization. In the past, process information was stored in farms of servers in a manufacturing facility only to be accessed when a regulator came onsite to audit a facility. Under the QbD approach, this information is readily accessible through web-interfaces and other specific graphical user interfaces (GUIs). However, the “Big Data” approach of plotting raw data in pretty ways just doesn’t cut it with complex manufacturing data. This is where multivariate analysis is required to turn the data obtained into information that is then turned into business benefits.

There have been great advances made in the development of miniature, cost-effective, and accurate PAT tools, such as the ultra-compact MicroNIR™ Spectrometers from Viavi Solutions Inc. (California, USA), ProcessPulse software from CAMO Software and  PAT Software Management, synTQ™, by Optimal Automation Ltd., to name a few. The availability and effectiveness of these tools are aiding in enabling this transformation to QbD principles in the pharma manufacturing industry.

Using CMS as an Example of QbD in Practice

A CMS system joins all of the unit operations of a solid dose operation into a steady state manufacturing system. In order to understand how to join the process together, the unit operations are individually assessed using tools such as Design of Experiment (DoE) and the correct PAT instrumentation is placed into equipment where representative samples can be measured. The most common technology used is NIRS, but other technologies are suitable for specific cases.

NIRS is used to assess raw material processing characteristics and is typically the first point of assessment for any process development strategy. From there, materials are fed using volume feeders in the right proportions into pre-blenders where NIRS can be used to assess uniformity. Methods such as roller compaction or twin-screw granulation can be used to produce particles with precise characteristics for optimal blending properties and these too can be measured with NIRS and particle size analyzers to ensure that the right particle size distribution is achieved such that there are no issues at compression and moisture content is being assessed to reduce any possible issues with microbiological activity.

NIRS can also be used to assess blend uniformity in continuous blenders or mounted into the feed-frame of a tablet compression machine to understand the total uniformity during compression. Finally, NIRS or Raman spectroscopy can be used to assess coating uniformity and thickness so that the Quality Target Product Profile (QTPP) of the product is ensured.

In order for a CMS to be of most benefit, the whole process much be orchestrated in a way that allows for feed forward, feedback process control to be implemented. This is where the data agglomerator and orchestrator is used to ensure the entire process is run in a steady state of operation. This is what QbD is, not the use of a single analyser in isolation or bringing the laboratory to the process, but the usage of technology and control systems to ensure the product manufactured is of highest quality and where early event detection is used such that the process can be adjusted, within the design space, to prevent the onset of a process/product deviation.

MicroNIR PAT Spectrometer, world’s smallest NIR spectrometer with high performance to price ratio bringing versatility and new options relative to the traditional bulky, fiber-optic-coupled and expensive NIRS tools,  for process monitoring and control on batch or continuous manufacturing lines.