Fit-for-purpose Quality by Design (QbD) implementation, part 1: Risk analysis and statistical Design of Experiments (DoE)
Even a lot of luck, and a forgiving process, cannot justify the “failed-the-first-time” technology transfer cases where trial-and-error univariate experimentation is used in the name of perceived higher speed. Large pharmaceutical companies have practiced Quality by Design (QbD) for many years before the FDA QbD initiative was published in 2002. Significant success has been achieved while refining the QbD methodology, and this has contributed to the clarification of QbD terminology. The road to these accomplishments has sometimes been convoluted, and some old challenges still persist, while during QbD practice, new ones have been identified. QbD adoption by mid-size and small companies has been slow, however there has been a noticeable increase in the use of risk analysis, statistical design of experiments (DoE), and bench Process Analytical Technology (PAT) recently. The use of chemical engineering principles is more often present in the development of robust processes, and this QbD element is slowly practiced as well. QbD objectives of large companies are different than those of mid-size companies, with the former executing “complete” QbD process development, from design, to regulatory submission, and smaller companies implementing key QbD elements only. Herein we briefly highlight aspects of fit-for-purpose risk analysis and statistical design of experiments (DoE) in the practice of QbD. What we claim is that meaningful, systematic process development is always good development, whether for a large, or for a small company. Good QbD process development is fit-for-purpose, and fit-for-resources; it is the combined creativity of chemists and of chemical engineers that is key to delivering robust processes, while minimizing the cost of poor quality.
Quality by Design (QbD) implementation started, as expected, on unequal footing; many large organizations were practicing QbD elements before the 2002 FDA announcement regarding “Pharmaceutical Current Good Manufacturing Practices (cGMP’s) for the 21st century”(1). Impressive accomplishments were obtained, and QbD methodology improved as a result of such extensive QbD work. Many small and mid-size companies followed suit, and several QbD methodology elements, such as risk analysis, statistical design of experiments (DoE) and bench Process Analytical Technology (PAT) have been extensively employed in the development of robust chemical processes. The early use of chemical engineering principles for low risk process scale-up is also practiced, even though more slowly than the other QbD elements. As expected, because the QbD guidelines are not prescriptive (2), companies tend to develop implementation strategies fitting their organizational culture and resources. Unnecessary challenges emerge when different groups in an organization, or different companies, partner for the purpose of process development using QbD, attempt to rea ...