Sweeping changes instituted by FDA in recent years are beginning to have a tangible impact on the way pharmaceutical manufacturers conduct business. This time, though, it isn’t a compliance fire drill focused on electronic signature capture or electronic application delivery formats. Instead, it goes much deeper, transforming the way life sciences companies look at their research, development and manufacturing process engineering organizations.
FDA’s “Pharmaceutical Quality for the 21st Century—A Risk- Based Approach” squarely takes aim at the issue of product quality, this time from a lifecycle perspective. The first incarnation of this initiative, originally floated in 2002 by FDA as the “Pharmaceutical CGMP Initiative for the 21st Century—A Risk-Based Approach” introduced the industry to the idea of process analytical technology (PAT).
This latest incarnation reflects a more holistic and mature perspective, recognizing the fact that implementing risk-based approaches to product quality management in manufacturing is only one piece of the puzzle: one that shares multiple interdependencies with upstream processes in research, development and early-stage commercial production.
The evolution of FDA’s Pharmaceutical Quality for the 21st Century initiative has been a journey, not only for the industry it impacts, but for FDA itself. The agency has, to its credit, openly acknowledged that previous compliance prescriptions have had unintended side effects—namely, discouraging manufacturers from embracing new technologies and process improvements even though such improvements might positively influence product quality outcomes, reduce waste, improve yields and ultimately drive costs down for consumers. The costs and risks associated with change, as a consequence of stringent validation and regulatory re-filing requirements, have simply been perceived by manufacturers as too high.
On the flip side, it’s only fair to point out that exorbitant product margins have allowed big-brand life sciences manufacturers to quietly absorb the expense of excessive inventories, waste, production inefficiency and variability that would put even the savviest consumer electronics manufacturer out of business in a heartbeat.
That said, the winds of competitive change are finally taking their toll, and FDA, in collaboration with partners from industry and academia, is encouraging the application of science and engineering to drug product manufacturing, as well as to its own internal validation and quality management processes.
This internal- and external-facing effort, labeled by the agency as “Quality by Design,” represents a sea change for the conservative life sciences industry, whose relationship with FDA has been cordial at best, and also for FDA. The latter has been actively engaged for the past three to four years in what amounts to a significant internal transformation and public relations campaign to gently persuade industry that the agency truly can—and should— become a change leader and a steward.
What began with PAT some five years ago has matured into a broader vision of change for the entire industry. PAT hasn’t gone away, mind you. If anything, the Quality by Design (QbD) vision provides a conceptual framework that helps companies position PAT as part of a greater strategic initiative, rather than relegating it to an isolated process control exercise (which was sadly the fate of many early forays into the application of PAT in manufacturing).
Positioning PAT in a QbD World
The concepts behind FDA’s Quality by Design initiative are quite straightforward. The idea is that quality should be built into a product (versus tested only after the fact) with a thorough understanding of the product and process by which it is developed and manufactured, along with a knowledge of the risks involved in manufacturing the product and how best to mitigate those risks.
Sounds simple enough, but the fact is that getting to the “thorough understanding of product and process” is a tall order. This is where PAT comes in.
If QbD defines “what to do,” then PAT is a framework for “how to do,” offering a system for designing, analyzing, and controlling manufacturing through timely measurements (during processing) of critical quality and performance attributes of raw and in-process materials and processes, with the goal of ensuring that the final product performs as specified on various measures. PAT isn’t a single product or technology.
Rather, it’s an architecture that incorporates elements of instrumentation, real-time data acquisition and storage, analytics, and—in its most sophisticated incarnation—advanced process simulation and multivariate process control. Furthermore, PAT has a dual role to play.
In a cGMP context, it promotes process control, and ultimately, the holy grail of release by exception (or perhaps more accurately, release unless there’s an exception!).
In support of QbD, it can feed empirical data about product and process interactions back to upstream development processes.
Today’s focus continues to be primarily on improving product quality and compliance while driving down the cost of high variability. However, the allure of feeding knowledge garnered from the manufacturing environment back to product R&D—potentially reducing the risks inherent in tech transfer processes—has forwardlooking manufacturers exploring new ways to model product and process data, specifically for cross-organizational knowledge sharing and reuse.
As it turns out, many of these early adopters are taking their lead not from the esoteric design world, but from the ISA S88 models already employed by the process modeling, control and execution systems in their manufacturing plants.
You May Already Have the Tools at Hand
Most pharmaceutical manufacturers are heavily invested in batch execution and distributed control systems. These are the systems of record for the master and control recipes used by process equipment to manufacture the product. They provide a wealth of information about the process and product if appropriately instrumented.
We’re not suggesting that you run out and implement a massive PAT architecture overnight, but certainly start with a detailed and thorough analysis of the information from these systems to get a picture of historical performance, outcomes and capabilities.
Since other industries have tackled this problem before, identify vendors and partners with an established presence in other verticals that bring implementation and cross-industry expertise.
The pharmaceutical market can benefit greatly by leveraging the successes achieved, and lessons learned, by manufacturers in other more cost-competitive industries.
- Data Historian: If you don’t have a resident data historian to acquire what might turn out to be key variables, we advise you to get one. Environmental variables, process variables, high-fidelity analyzer data and even operatorrelated information, can all provide clues about process performance and lead you to a more robust system design. Look to providers of stand-alone historians— such as AspenTech, Automsoft and OSIsoft—in addition to your automation provider. Honeywell, for instance, supplies process automation and controls, but also provides building and environmental monitoring and control systems (that can impact product outcome) that feed into its historian product.
- Electronic Recording of Batch Records: Once you have a data historian, you’ll need a detailed audit trail of the production steps. This record should include material genealogy data, material movements, process steps, operator interactions, sampling events and their results. Many manufacturers continue to construct and maintain such audit trails on paper, but moving to an environment that can share information digitally requires an electronic batch record (eBR) system (sometimes mistakenly referred to as an MES in the pharma industry). Process control and automation vendors like ABB, Honeywell and Rockwell have wellknown product offerings in this area, as do software providers like Elan and Werum.
- Multivariate Analysis Tools: We also recommend that our clients invest in process analytics capabilities. Discovering the relationships between manufacturing process variables and product properties is a multivariate analysis problem, and vendors like Aegis Analytical, Intercim/Pertinence and Umetrics offer strong, and usable, capabilities in this area. For analytics that are linked in real-time to process automation, look to advanced process control and simulation providers like Pavilion Technologies (now part of Rockwell) and AspenTech.
- Best of Both Worlds: Keep an eye out for interesting partnerships. Automation and controls provider GE Fanuc partners with multivariate analysis software provider CAMO Software for the life sciences industries; Siemens likewise partners with Umetrics; and ThermoFisher and Pavilion Technologies are working together on integrated laboratory instrumentation and model predictive analysis. As the lines between the lab and the production environment blur, we anticipate many more such marriages between control, analytics and lab informatics providers.
Closing Thoughts
Leading manufacturers are already well on their way to building integrated quality and compliance into business process performance using QbD riskbased approaches and operations excellence initiatives. However, companies are still challenged with tying patient outcomes to QbD processes.
As a senior executive from one large pharma company put it, ‘We’ve gotten off-track—we are looking at the design space from a perspective of building a robust process in manufacturing, but we should look at it from the perspective of the patient. We should design a product to meet patient needs and produce it by using the right process to minimize patient safety concerns. I’m not sure we have our priorities in that order.”
About the Authors
Alison Smith brings more than 25 years of manufacturing software development, support, marketing and sales experience to her role as a research director in AMR Research’s Market Services group. Smith studies the productive application of information technology to the challenges faced by global producers as they seek to implement and evolve demanddriven manufacturing strategies.
Hussain Mooraj, also a research director at AMR Research, has more than a decade of experience in pharmaceutical and biotech consulting, strategic consulting and marketing. As part of the Value Chain Strategies team, Mooraj currently focuses his research on the pharmaceutical, biotech, medical device, and wholesale distribution business segments.
