Product and process development to Good Manufacturing (GMP) principles has been the subject of regulations and guidances for around 40 years, whereas the concept of Quality by Design (QbD)¹ is a mere seven years old. Furthermore, QbD is a concept that is entirely voluntary, not mandated by any regulatory agency and not adopted by non-ICH (International Conference of Harmonisation – www.ich.org) country agencies. On the other hand, the product development process and the applicable validation requirements are now a standard element of GMP regulations globally. We shall look at the reasons why the FDA considered it necessary to update its process validation guidance in 2011 and how this was influenced by and aligns with the QbD paradigms.

Despite much hype, guidance documents from ICH² and regulatory agencies, numerous conferences and a good number of mock examples for QbD implementation, there is little published about where industry really stands with regards to adopting and implementing QbD. It takes time to interpret regulatory guidances, implement the appropriate structures and processes, so it is obvious that currently there cannot yet be any new drug applications in which the QbD concept was applied throughout the entire drug lifecycle.

A QbD approach does not replace the need to comply with GMPs; it is essentially a complimentary methodology, aimed at delivering enhanced product and process knowledge, and providing additional benefits.

FDA’s Process Validation Guidance and QbD

Regulatory inspections reveal consistent issues with compliance to healthcare regulations, particularly the lack of a lifecycle approach with regards to process validation. The 2011 FDA Guidance for Industry on Process Validation³ specifically addresses the lifecycle approach and aims at integrating the QbD concepts. The necessity for this guidance was not industry’s reluctance or inability to adhere to the regulations, but rather a reflection of the changed circumstances in industry, particularly the effects of globalization and virtualization. To fully understand the evolutionary development of process understanding, one needs to look back in time several decades. Figure 1 depicts the relationship between compliance and process understanding. Especially in the “Age of the 3 Batches”, the tendency by industry was to refrain from including much of the process knowledge and understanding in the CMC sections of the drug applications, only providing the very minimum of data. The regulators have to share some of the blame for this as they were encouraging such an approach. That FDA’s view has changed is clear from the new guidance where the concept of “establishing documented evidence” is replaced with “establishing scientific evidence.”

With the tendency to outsource much of the manufacturing, and even development, activities, knowledge is no longer readily available. Thus, the FDA’s drive to force industry to re-establish an adequate knowledge pool is fully justified. Some industry veterans may say that this is what happened with the introduction of the GMPs. That attitude misses the point that the tools at our disposal today are vastly different to those 20 or 30 years ago. In fact, some of the expectations from the FDA’s process validation guidance could hardly be achieved without the help of sophisticated automated systems for process control, measurement and analysis. Taking this one step further is the concept of Design Space in the QbD world. Multivariate analysis, application of statistical controls and Design of Experiments would be impractical, if not impossible, without the computing power of today. So from this perspective, regulators are not merely addressing concerns for the wellbeing of the patients; they are also addressing the evolving technological environment, encouraging modern methodologies and continuous improvement.

The European Perspective

As is the case for much of the healthcare rules and regulations, there is only some degree of harmonization between the two leading agencies, the U.S. FDA and the EMA. One of the key differences lies in the required number of validation batches, with the European regulators still satisfied to accept three consecutive validation runs without a specific rationale. FDA representatives have made it consistently clear that they will not accept any number of validation batches without a scientific rationale for that number. This must not be confused with a statistically significant number, which may be 30 or more batches, a number that is simply prohibitively expensive for any company.

Both agencies embrace the lifecycle approach, with the EU regulation following the long-established qualification / validation approach, using the DQ, IQ, OQ and PQ (Design, Installation, Operational and Process Qualification) terminology. The FDA guidance does largely away with that terminology, relying instead on terminology favored by U.S. engineers, taken from the ASTM 2500 standard. This U.S. approach is interpreted by some companies as requiring less quality unit oversight over engineering activities, something that carries a level of risk that should not be underestimated.

Both regulators insist on post-launch data that assures ongoing compliance and validated state. This information may have been provided as a summary in an annual report in the past, but the expectation now is that current data on trending is available, so immediate action can be taken to rectify excursions. Companies struggle to implement this requirement for all their processes and in a consistent manner. This is partly caused by a lack of systems and in part by ambiguity to the extent such trending data has to be available.

Fitting Process Validation into QbD

It is necessary to voice a word of warning to those keen to embark on QbD: Unless process validation in the traditional way is fully and properly implemented, there is absolutely no use in attempting QbD. Some of the serious issues still observed in industry today include:

• Operators “forgetting” to take samples
• Equipment or instrumentation not qualified or calibrated
• Specifications being changed while validation is ongoing
• No technology or process transfer process in place
• Unplanned validation, e.g. no scope defined, no deviation procedure in place
• Ineffective quality oversight by the quality unit

Bad practice does not get any better by giving it a different name.

However, where process validation is well and truly established, the step change towards Quality by Design involves the application of increased process understanding, e.g. from Design of Experiment (DoE) studies and effective use of risk management to define and steer the control strategy. Too many companies merely use risk assessments as an excuse to find ways for reducing (read: omitting) tasks, rather than to assure that higher risk process parameters are controlled in a tighter manner, thereby delivering product with the right quality attributes. Often DoE studies may lead to an extended development phase, but this is compensated for with a much higher assurance of success for the validation batches and the increased process and product understanding resulting in highly optimized processes.

This enhanced approach requires companies to consider the following prerequisites for success:

• In-house competence: As an example, few development departments employ statisticians, who are needed to design, explain and analyze DoE studies. The conclusions that will be drawn from these experiments seriously impact the outcome of the development phase and the success of bringing the product to market. This lack of expertise often proves challenging.
• Revision of the Quality Management System (QMS): Existing QMS are not designed to accommodate the QbD approach. For example, companies operating in ICH and non-ICH countries must stick to a traditional approach in the non-ICH countries as they do not yet accept QbD. Such a dual approach is rarely provided for in companies’ QMS.
• Willingness to invest: Current estimates are that it takes about $3 million to implement a QbD structure and the associated processes.
• A business plan: Investments of that magnitude for long-term projects with an estimated time to return of investment of five or more years requires top management buy-in. This in turn necessitates a well-rationalized business plan that identifies the benefits that can be reaped from QbD.
• Leveraging electronic tools: Integrated sophisticated software systems are a prerequisite to knowledge management; this is far beyond a mere data warehouse solution.
• Willingness to share process knowledge with the regulators: This is quite a paradigm change, as both good and bad experimental results must be shared with the authorities, through (new) drug applications / submissions or change / variation requests. These documents, typically in the Common Technical Document (CTD) format, will have to be significantly more detailed than those using the traditional approach to process and product validation.

The Business Case for QbD

Given the global economic perspective and the state of the pharmaceutical industry, any initiative that is geared towards process improvements and has the potential for regulatory flexibility is welcome. But as the adoption and implementation of QbD requires an upfront investment with RoI likely to take five to 10 years or more, industry enthusiasm is somewhat muted. Unless there is a real competitive advantage, uptake of QbD will remain slow.

At a recent FDA committee meeting, Yatindra Joshi, vice president of generics R&D at Teva noted that, when looking to adopt QbD principles, generic players must offset the efficiency gains against the risk the extra work will mean a competitor wins the race to be first-to-file. He added, “If you’re not first to file you may as well be last. Coming second can mean a $100 million opportunity withers to just a few million dollars. In this environment it can be hard to justify the extra initial time and work QbD requires”⁴. It is not surprising then that the key adopters of QbD can be found among the global top 50 pharmaceutical companies.

Business Opportunities for CMOs?

The vast majority of pharmaceutical companies are specializing in specific steps in the drug lifecycle, outsourcing many activities, including drug development, process validation and manufacturing. This is especially the case in the biotechnology industry, where many firms are the result of spin-offs from universities or other research organizations. Given that most NDAs are for biotechnology-derived products, this is undoubtedly an opportunity for CMOs to differentiate themselves in a fiercely competitive market environment.

Industry Trends

Many quality and compliance processes are still stuck in the paper and pen era, making it onerous, if not impossible, to garner the benefits from automated process controls, product and process knowledge and continuous improvement. The implementation of integrated computerized systems that provide an interface between operations and quality is a must, and is driven by the need for cost control and benefit optimization. The business case for such investments is easily made, whereas the same cannot be said for Quality by Design. Only companies with a long-term strategy, an agile operational and quality setup, a drive for excellence and the ability to provide adequate resourcing will be and are embracing QbD. They are indeed shaping of what will be seen as QbD best practices in the decades to come.

References

1. Siegfried Schmitt, editor, Quality by Design – Putting Theory into Practice, 2011, www.parexel.com/services-and-capabilities/consulting/strategic-compliance/siegfried-schmitt-quality-by-design-qbd-book/
2. Pharmaceutical Development Q8, http://www.ich.org/products/ guidelines/quality/article/quality-guidelines.html
3. FDA Guidance for Industry, Process Validation: General Principles and Practices, January 2011, www.fda.gov
4. Nick Taylor, First-to-file race slowing QbD adoption in generics, In-Pharma Technologist, 8 August 2011, www-t.wrbm.com/r/?id=h22346114,8953372,8955e34&p1=yHMpPsim%2FVugGvZmg8R7KQ%3D%3D

Dr. Siegfried Schmitt is principal consultant at PAREXEL. He can be reached at siegfried.schmitt@parexel.com.