Four Years and Three Myths Later
In the four years since the ICH outlined the concept of design space in its Q8 guideline, pharmaceutical companies – despite depending on innovation for their livelihood – have been slow to adopt Quality by Design (QbD). One year ago in this publication, two colleagues and I identified what we saw as the most common reasons that life sciences organizations resist QbD: the belief that it won’t work in pharma, the worry that it is unaffordable, and the fear of change (“Quality by Design – Shortening the Path to Acceptance”, Snee, et al 2008). We also offered some suggestions for overcoming those widespread sources of resistance.
But at the bottom of that resistance there may lie something deeper: fundamental misunderstandings about the nature of QbD. In fact, in a recent webinar survey of pharmaceutical professionals, 40% of respondents identified “lack of understanding of QbD” as their biggest problem with it. After four years, some of these misunderstandings have achieved the status of myths in the pejorative sense – widely held beliefs that are simply untrue. hree such misconceptions about QbD stand out:
* QbD is different from Process Analytical Technology (PAT).
* QbD is for drug development only, not for the manufacturing of in-market products or technology transfer.
* Pharmaceutical companies already do Quality by Design, though not in name.
As in the ancient story of the five blind men touching different parts of an elephant and then producing conflicting and incompatible accounts, these misconceptions add up to a fragmented, disjointed, and incomplete understanding of Quality by Design. A part is taken for the whole, and the unity that permeates all of the parts – the elephant in its entirety – remains unseen.
How do we find our way out of the darkness? First, by correcting these misconceptions. Second, by understanding the critical elements of QbD in terms of its simple, underlying principle: improving process understanding and control in order to reduce risk. Additional discussion of concepts, methods and tools of QbD and the improvement of R&D processes can be found in Heilman and Kamm (2008), ICH (2005, 2007), Juran (1992), Kamm (2007), and Snee, et al 1985, 2005).
PAT and QbD
Confusion about the relationship between PAT and QbD is based, in part, on a reductive understanding of PAT. Since the FDA released “PAT – A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance” in 2004, discussions of its significance have focused largely on its application of in-line instrumentation and sometimes simply equated PAT with the hardware used to monitor production processes. But a technology-only understanding misses the real point of PAT, which, said the FDA, is based on the principle that “quality cannot be tested into products; it should be built-in or should be by design.”
ICH’s Q8 document made it clear that the way to design-in quality from the first is to understand “design space,” which it defined 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.” In other words, QbD is not only an integral part of PAT but in fact subsumes PAT under the larger principle of designed-in quality. Further, The FDA made clear from the beginning that the goal of PAT is not the technological automation of process control, but comprehensive process understanding. As the PAT guidance says, “a process is generally considered well understood when (1) all critical sources of variability are identified and explained; (2) variability is managed by the process; and, (3) product quality attributes can be accurately and reliably predicted over the design space established for materials used, process parameters, manufacturing, environmental, and other conditions.”
Development – and Beyond
Because the PAT initiative and the concept of designed-in quality appeared at first to be focused on drug development, the misconception arose that QbD was also primarily for use in developing drugs and the manufacturing processes that would be used to produce them. But the FDA was also encouraging improved process understanding and continuous improvement of manufacturing processes for products already in the market. Manufacturers may have been further inhibited from applying QbD to in-market product manufacturing processes because they feared that changes to already approved processes would require new regulatory filings – a fear that the risk-based approach to regulation, which was the aim of new GMP initiatives, was designed to allay. In fact, QbD and design space can be used to gain further process understanding with in-market products and improve the process within the parameters of the existing regulatory filing.
QbD applies to all phases of the product lifecycle: drug development, scale-up, manufacturing both for new and in-market products, and technology transfer of processes and products to other sites. In drug development, for example, the potential for an understanding of design space to lead to robust manufacturing processes is, as noted, perhaps the most widely understood aspect of QbD. But QbD offers another advantage here. Because development usually takes place over a long period of time and often involves many people, it can sometimes result in misunderstandings, errors, and redundant activity. QbD keeps an understanding of design space in the forefront of development efforts, providing the coherence and continuity that trial-and-error approaches lack. QbD can also help maintain “lifetime development” – the continued gathering of data after a product is already in production. Because only a limited amount of data accumulates during initial development, additional data, framed by the product’s design space, can be helpful in deepening process understanding and continuously improving the process.
In manufacturing of new products as well as in-market products, QbD helps determine where the greatest risks and opportunities lie. Manufacturers can then strategically address those risks and opportunities – not simply throw technology at them or deal with them after the fact of failure. Similarly, in technology transfer, it is difficult to efficiently and effectively transfer a process that is not well understood (Snee 2006, Alaedini, et al 2007). Whether transfer takes place between two sites, two companies, a company and a third-party manufacturer, or even from R&D to a pilot plant or commercial facility, an understanding of design space confers the ability make transfers faster, more compliantly, and less expensively.
We Already Do QbD
Pharmaceutical manufacturers expend an enormous amount of resources in their unflagging efforts to assure quality, achieve regulatory compliance, and produce drugs as cost-effectively as possible. Further, they employ techniques and technology that entail a great deal of scientific sophistication and operational complexity. In addition, many apply some of the elements of QbD and PAT or engage in practices, such as a lifecycle approach to validation, that closely resemble aspects of QbD. They may also use risk assessment in quality assurance and sometimes selectively use multi-variable analysis when they encounter process problems. Given these practices and a high level of scientific sophistication, it’s not surprising that such organizations may feel that they are already doing QbD.
But such piecemeal efforts provide neither the comprehensive understanding of design space – all of the possible permutations of critical variables that will produce an in-specification result – nor the predictive power such understanding confers. With an understanding of design space, manufacturers can know in advance how changes in inputs will affect each other and affect product and process performance characteristics. In other words, quality – an in-specification result – can be designed into the process, not empirically derived from test batches or the like.
The Critical Elements of QbD
The underlying theme that runs through the clearing up of all three misconceptions above can be simply stated: greater process understanding. Viewed through the lens of process understanding, it is clear that PAT and QbD are interdependent; that QbD’s power is equally applicable to development, manufacturing, and technology transfer; and that piecemeal application of statistical techniques and of analytic technology do not add up to the scientific understanding of processes that QbD seeks.
From this perspective also, an enumeration of the critical elements that go into this advanced process understanding should render QbD easily understandable [Figure 1]. These critical elements are, in essence, the building blocks of the kind of process understanding that QbD envisions:
* Critical Quality Attributes (CQAs) are the critical process output measurements linked to patient needs.
* Critical Process Parameters (CPPs) encompass the process input (API and excipient), control, and environmental factors that have major effects on the CQAs.
* Raw Materials Factors include the stability and capability of raw material manufacturing processes that affect process robustness, process capability, and process stability.
* A Process Model provides a quantitative picture of the process based on fundamental and statistical relationships that predict the CQA results.
* Design Space is, as previously noted, the combinations of input variables and process parameters that provide assurance of quality.
* Process and Measurement Capability tracks process performance relative to CQA specs and provides measurement repeatability and reproducibility regarding CQAs.
* Process and Measurement Robustness is the ability of the process and measurement system to perform when faced with uncontrolled variation in process, input, and environmental variables.
* Process and Measurement Control includes the use of control procedures, including statistical process control (SPC), to hold the process and the measurement system on target and within the desired variation.
* Failure Modes and Effects Analysis (FMEA) of the CPPs, including raw material variables, identifies how the process can fail and, after appropriate controls and fixes are in place, the areas of the process that remain at greatest risk of failing.
* Risk Level is a function of the design space, FMEA results, and process and measurement capability, control, and robustness.
It is in reduced risk that all of these critical elements converge. As Hall of Fame quarterback Johnny Unitas said, “There is no risk of an intercepted pass, if you know what you are doing.” QbD provides the process understanding that enables you to know what you are doing, resulting in greatly reduced risk.
With the ability to determine precisely where the greatest risks and opportunities lie and address, life sciences organizations can realize the many operational, business, and financial benefits that QbD generates, including:
* The ability to design new products and processes and bring them to fruition faster, with fewer setbacks at critical stages such as scale-up, validation, and transfer
* Processes performing on target and within specs at minimum cost with fewer defective batches and fewer deviations
* Greater flexibility in process operation
* Greater regulatory flexibility based on a science-based approach to risk management
* Ability to continue to optimize and improve the manufacturing operation without facing additional regulatory filings or scrutiny
* Faster time to market and reduced rework, resulting in reduced costs and increased revenues
Instead of incrementally improving unit operations that, in isolation, may have little effect on overall process performance or quality, manufacturers can adopt a holistic approach to QbD that is applicable at every stage of the product lifecycle. When combined with sound deployment methodologies, improvement management systems, and appropriated infrastructure, QbD dispels misconceptions with the most convincing corrective of all – a sustainable, significant increase in value.
About the author: Ronald D. Snee is Principal of Process and Organizational Excellence at Tunnell Consulting in King of Prussia, PA. He earned a doctorate in applied and mathematical statistics from Rutgers University, New Brunswick, NJ. Snee has published 4 books and more that 195 articles on process improvement, quality and management. His work has been recognized by 28 professional awards and honors including America Society for Quality’s Shewhart and Grant Medals and American Statistical Association’s Deming Lecture Award. He can be reached at Snee@TunnellConsulting.comCopyright Ronald D. Snee 2009
References
Alaedini, P. R. D. Snee and B. Hagen (2007) “Technology Transfer by Design – Using Lean Six Sigma to Improve the Process”. Contract Pharma, June 2007, 4-9. Heilman, C. (2008) “Coming to a Biotech Near You: Quality by Design Part 1: Its Applicability in Biopharmaceuticals”; “Part 2: Design Space in Development and Manufacturing (with J. Kamm)”. International BioPharm, May 2008, July 2008. ICH (2005) “ICH Harmonized Tripartite Guideline: Pharmaceutical Development, Q8,” Current Step 4 Version, November 10, 2005. ICH (2007) “Pharmaceutical Development Annex To Q8,” November 1, 2007. Juran, J. M. (1992) Juran on Quality by Design, Free Press, New York, NY. Kamm, J. (2007) “Can You Win the Space Race?” Pharmaceutical Manufacturing, May 2007. Snee, R. D., L. B. Hare and J. R. Trout (1985) Experiments in Industry – Design, Analysis and Interpretation of Results, Quality Press, Milwaukee, WI Snee, R. D. (2006) “Lean Six Sigma and Outsourcing – Don’t Outsource a Process You Don’t Understand”, Contract Pharma, October 2006, 4-10. Snee, R. D., P. Cini, J. J. Kamm and C. Meyers (2008) “Quality by Design – Shortening the Path to Acceptance”, Pharma Processing, February 2008. Snee, R. D. and R. W. Hoerl (2005) Six Sigma Beyond the Factory Floor – Deployment Strategies for Financial Services, Health Care, and the Rest of the Real Economy, Financial Times Prentice Hall, NY, NY.