Who Really Designed the Design Space?

Quality by Design is a great idea for Pharma, but some dispute its industry origins

By Steven Kuehn, Editor-in-Chief

Share Print Related RSS

I do enjoy a little controversy, especially when it involves the origins of trusted engineering methods and practices. No matter the industry, debates of this character often swim below the surface of technical discourse only to breach for some air when advances or claims of innovation receive fresh scrutiny and perhaps some revisionist hindsight by the engineering/technical community.

Apparently, a bit of a debate is occurring between the Pharma and Chemical process engineering communities as to the origins of the design space concept and its application to pharmaceutical manufacturing process in the context of Quality by Design (QbD). I became aware of the “controversy” after publishing “Who Designed the Design Space?” — a column written by Ronald Snee, reprising a blog he posted on a respected Pharma industry website last year.

Snee’s take on QbD history walked one through R.A. Fishers’s pioneering work on statistical design of experiments (DOE) in the ’20s and the discipline’s further development, including seminal commercial application of its fundamentals by two engineers at Imperial Chemical Industries looking to optimize chemical production process through response surface methodology (RSM). Snee credits Merck scientist Joe Schwartz with introducing RSM to Pharma in the early 1970s as a means to optimize formulation process. According to Snee, FDA’s Ajaz Hussein picked up the ball about 12 years ago, introducing the discipline (coined as QbD by Juran in 1992) as the overarching system to pursue and assure pharmaceutical process quality.

Snee makes a subtle point, he said RSM never gained any real traction with Pharma in that era because “Apparently the process improvement need wasn’t yet identified in Pharma to improve its processes …” In hindsight, that’s quite an understatement. It’s at that point in time I see where the seeds of today’s controversy might have been planted.

One of my contributors told me, “The irony is, that before Ajaz Hussein suggested QbD to FDA — which is an application of chemistry and chemical engineering fundamentals — those fundamentals were well known and practiced in chemical engineering in the ’70s.”

At the crux of his critique was his impression that Snee’s column perpetuates the myth that QbD, statistical process optimization and its supporting design space framework are fresh ideas when in reality (and any chemical engineer worth his or her salt knows) they are just jargonistic new labels for Fisher’s DOE principles, codified by DuPont in “Design of Experiments” in 1974. “We were taught that every process you develop and design has to be such that you have complete command and control [of a process] to produce the same quality,” whether continuously or from batch to batch, said my colleague. “Please do not call fundamentals of chemical engineering new. We all know Pharma is a few generations behind and needs to catch up … giving fancy names [to things] has become Pharma’s habit as it considers itself well above its roots.”

To a degree I understand the frustration, because it appears that in the 1970s the Pharma process engineering community had a chance to embrace the chemical industry’s well defined and commercially applied process optimization principals, and for reasons I can yet fathom fully, chose to either ignore or reject them. Snee wrote, “Sometimes a new idea can take a generation or more before it becomes standard practice.” Well, that’s true, but I don’t think he meant that the idea or concept was actually “new” — more likely it was new to Pharma, back in the day, and only now being accepted doctrinairily by the industry. But I understand that for those ChemE’s who cut their teeth on DOE, listening to Pharma’s process engineering community prattle on about QbD and design space like it was some home-grown concept after rejecting its core principals for so long, could be a trifle annoying.

Share Print Reprints Permissions

What are your comments?

You cannot post comments until you have logged in. Login Here.

Comments

  • The annoyance is totally justified as Pharma has always often presented a superiority complex suggested their work was better or more important than others. If fact as noted Pharma were late comers to design and statistical control in processes. I see more as ignoring than out right rejections where I can think of several reasons for this delayed adoption, a few which you have already touch on: There is not the same CoG sensitivity in Drugs as most materials and therefore did not have to squeeze as hard to sustain profits and may have known about DOE etc but frequently did not bother to apply. Unfortunately the scales of many drugs can be limited by comparison to commodities where a "non-ideal" process may have been developed during clinical support and becomes implemented as easy path forward if is proved as "adequate" hence adds further resistance unless a major need identified to replace or alter the manufacturing. Part of that is there were/are inherent Regulatory burdens which inhibited ability and desire to make changes, even if shown to improve product control and economics. Until there was some recognition and evidence that FDA and other agencies would allow or create mechanisms for introduction of changes there was a big risk barrier to overcome. Another possible cause I see where I can not verify my observations that most Pharma Process groups that do the early development are heavy on Chemists and light on ChemEs so even though is correct DOE is standard part of the latter's training its rare for Chemists to get educated in such so is not part of their tool box (until perhaps added by exposure and interactions to ChemEs) and has taken longer to incorporate in fundamental activities for Pharma Process R&D.

    Reply

  • More on Who Designed the Design Space

    Snee Associates, LLC Ron@SneeAssociates.Com June 4, 2014

    Discussion of the history of a methodology is a good thing. It can help us deepen our understanding of what it is and why it has been found to be useful. Here are a couple of more pieces of the history of the design space and Quality by Design (QbD).

    First we need to keep in mind that QbD is more than the design space.

    As introduced by the FDA and I have described (Snee, Pharmaceutical Processing, February 2009) QbD contains at least 10 critical building blocks. The design space is a necessary, but not sufficient building block of QbD. Unfortunately some have been led to equate QbD with the design space. Following the work of Box and Wilson at ICI, response surface methodology and design of experiments (DOE) became over time widely practiced in the chemical, plastics, paper, electronics and other process industries. The methodology was widely used at DuPont beginning in the 1960’s formulated into the “Strategy of Experimentation” (SOE). DuPont began marketing the SOE approach in 1974.

    SOE was a major step forward as it defined what statistical designs should be used depending on the experimental environment encountered. This enabled scientists and engineers to quickly and effectively learn and use the methodology, and in the process learn design of experiments. I played a role in the development and use of SOE at DuPont where I worked for 24 years and use it today in my product and process development consulting practice for Pharma, Biotech and other process industries. The Pharma and Biotech literature contains numerous examples of the use of SOE.

    Another important piece to the QbD puzzle occurred in the early 1980’s when Taguchi’s DOE-based robust product and process design methodology got the attention of American industry. A leading user of the methodology was AT&T Bell Labs which did considerable work on the concept and use of QbD (Mayo, Quality Progress, April 1986) complete with educational programs and materials. Robustness is today one of the critical building blocks of QbD as practiced in Pharma and Biotech.

    Discussion of the history of QbD is enlightening, but let us not lose sight of the goal of effectively using the methodology to improve the performance of Pharma and Biotech products and processes.

    Ronald D. Snee, PhD Snee Associates, LLC Ron@SneeAssociates.Com June 4, 2014

    Reply

RSS feed for comments on this page | RSS feed for all comments