By Bernard Tulsi
A cursory look at the current status of the FDA’s quality-by-design (QbD) initiative reveals a mixed picture. By mid-2009, early adopters had successfully employed QbD tools to file FDA applications. Other companies cherry-picked the tools while contemplating if and how to expand their piecemeal approaches.
Still, skeptics abound. A number of companies view QbD as the latest quality management fad or as yet another FDA information hurdle. Predictably, several companies want assurances about payback for the extra effort. Others extol the virtues of their current approaches and see no need to meddle with what they already consider successful.
Furthermore, QbD has long been dogged by the notable absence of authoritative descriptions, clear directives, or definite parameters. “QbD is an umbrella concept right now. It lays out what the FDA wants industry to do, but implementation is really up to the companies,” says Paul Skultety, director of pharmaceutical development services at Xcelience, a formulation development and contract manufacturer for the pharmaceutical industry.
Skultety says that the interpretation of QbD varies. “The FDA does not have a precise definition of what it wants yet, and it is still evaluating what the concept means.”
Not unexpectedly, this has led to uncertainty. “Generally, the establishment of parameters generates even more questions,” says Ronald Branning, VP of corporate quality assurance and compliance officer at Gilead Sciences, who is also a noted writer and instructor on QbD.
Concurring, Skultety believes that better definitions will emerge gradually over the next five to ten years from the FDA and industry and added, “Right now it is a moving target, and no one has a good handle on what it means.”
The History of QbD
At its most fundamental, QbD begins to sketch a framework aimed at linking drug treatments and their production methods to required quality and process controls. An FDA notice in mid-2008 to include biologics under the QbD umbrella stated that manufacturing would “depend on a risk-based approach linking attributes and processes to product performance, safety, and efficacy.”
The roots of QbD may be traced back to at least two International Conference on Harmonization (ICH) documents — Q8 Pharmaceutical Development (May 2006) and Q9 Quality Risk Management (June 2006) — and FDA’s industry guidance documents — PAT Framework (September 2004) and its quality systems approach to pharmaceutical cGMP (current good manufacturing practice) regulations (September 2006).
Even earlier (mid-July 2005), the FDA’s Office of New Drug Quality Assessment (ONDQA) spearheaded a pilot plan aimed at acquiring the experience to evaluate chemistry, manufacturing, and control (CMC) sections of New Drug Applications (NDAs). The 2005 pilot was deemed highly useful in identifying information to be shared with industry with respect to QbD and small molecule manufacture.
QbD principles are applicable to both small molecule drugs and complex biotech products, but the evaluation challenges are much greater in the latter. In mid-2008, the FDA initiated a QbD pilot program — and invited volunteer companies to participate — to gain experience for biotechnology submissions. The goal was to define clinically relevant characteristics of biologics and link them to the manufacturing process, as well as the desired performance, safety, and efficacy expectations.
The scientific portion of QbD has been in use for some time now, dating back to the early adoption of Design of Experiments (DoE). The DoE approach incorporates statistically valid methods for data collection, which are far more meaningful than replicating experiments one at a time, according to Branning. “DoE showed that statistically valid, designed experimentation provided the only way to detect interaction effects in processes that impact product quality,” he says.
The Concerns Of Adoption
Branning concedes that getting some industry scientists to see the importance of DoE could be difficult. He attributes this partly to the absence of DoE on universities’ curricula but adds that once scientists grasp the underlying concepts and the power of this tool, they often become strong advocates.
Because the QbD concept is still evolving, the FDA does not impose firm directives but it does want to see processes and procedures executed in a scientific manner that makes sense, according to Skultety. “If processes and the reasons behind them are on sound scientific footing, the FDA will most likely approve them even though different companies may do them differently,” he says. “That is how this will evolve. There is not a lot of black and white here, but also a lot of grey — somewhat like good manufacturing practice [GMP] where there are not many hardcore rules. Companies go about it differently but in a scientific manner.”
To be sure, there are other impediments, not the least of which is the perennial question of costs. Branning explains that costs will depend on whether the starting point is confined to new data and filings or will include all products already on the market.
To subject all existing products to QbD principles will pose a major hurdle to companies, and many will balk at the substantial financial outlay. Even if only new product candidates are included, companies that have not done QbD before will have to expend start-up dollars on staff training and to establish the mechanisms and operations to meet the requirements of different parameters.
Branning says that companies can address the cost questions by applying QbD slowly to work that is ongoing and to projects that have already been planned. A great starting point could be to use QbD tools to find solutions where they have experienced difficulties in setting up processes or where there are product development problems. “That represents one of the real benefits — the ability to surmount real problems by applying QbD techniques and then slowly building them into normal operations,” says Branning.
Despite its formidable teething problems, the experts believe that QbD could deliver important benefits. Branning says it is important to understand what QbD represents: “It is a methodology, and the first part of any methodology is to understand the goal, which in this case is delivering safe and effective products to patients. Next is knowing the starting point and who is responsible for what.”
Branning says that QbD should not be seen as an exercise where one department does an aspect of product development and passes it along — rather, it must be viewed as a continuum. “Unless there is this kind of overview and solid awareness of the QBD principles, implementation could become isolated pieces instead of a continuous whole.”
Such mapping of steps and functions will also provide the opportunity to ascertain whether handoffs among departments are as good as they should be, and help to bring cross-functional parts of organizations closer together on key issues: process development, understanding the impact of processes on products, how products perform in the clinic, and how to continually improve processes and products.
“QbD principles will help drug developers delve into the processes and examine how they affect product quality. This knowledge will enable them to tweak process parameters or change the order of processing and see how such changes affect product quality, while providing an overall picture about the potential impact on patients,” says Branning.
Asked about the uptake of QbD within the industry, Branning says, “Unfortunately, it is appallingly low.” He attributes the low adoption rate to inadequate understanding of QbD principles, several unanswered questions resulting in a lot of confusion, lingering controversy, and the usual resistance to change — in Europe, they bristle at the suggestion that QbD could be beneficial and strongly hold on to the view that science is best left unfettered.
“It will take some sort of change agent either within companies or the industry, and that is what FDA is trying to do to make a reality of QbD,” he says.
Another big impetus will come from the compelling success stories. The literature is replete with anecdotal data suggesting that product developers have been able to reduce development time by as much as six months to three years.
“Additional uptake,” says Branning, “will come from the understanding that it is easier and less expensive to attain the desired results using QbD tools, and therefore it must be more powerful and take less time than just randomly trying to figure out processes, the products they produce, and the impact on patients.”
One attraction of QbD is that with limited time and budget, it is possible to use DoE to optimize operations and generate the most data per unit time and cost. “Anyone who does not look at and assess QbD in an appropriate way may well be passing up an important opportunity,” Branning concludes.