Guest Column | December 23, 2019

When Should Formal Quality By Design Documentation Begin?

By Kevin Wall, principal consultant and owner, Cincero Consulting

iStock-663958080 (1)

When is the best time to start quality by design (QbD)? This question is asked most frequently among many small firms. A better way to phrase it is: When should I formally document my QbD activities? The short answer: Document the current state of knowledge and risk before the registration campaign begins. In reality, QbD activities begin before Phase 1. This article discusses three reasons why the registration campaign is the perfect time to formally document the quality risk assessments (QRA) for the drug product and manufacturing process. The principles of QbD have become embedded into industry thinking, as I discussed in my previous article, “Debunking the Top Three Myths About Quality by Design (QbD).”

Reason 1: Capture The Current State Of Knowledge

Small firms run lean and rely heavily on contract development and manufacturing organizations (CDMOs). Luckily, most CDMOs understand the basics of QbD. Early development activities are carried out with the critical quality attributes (CQAs) in mind, even if they are not formally documented. Knowledge of the product and process is usually contained in a fixed set of project management deliverables from the CDMO, such as campaign reports, weekly project updates, change controls, and deviation investigations. It is not uncommon in early phases to have process development runs executed just prior to a production campaign. In many instances, the parameters are tweaked between batches of the campaign. Products and processes are not locked and there is a heavy reliance on end-product testing to demonstrate fitness for use. At this point, most readers are probably thinking, “Tell me something I don’t know.”

The registration campaign is a natural point to formally assess the current state of knowledge and plan the next phase of development. Registration material is typically made for Phase 3 studies or for late Phase 2 studies for products on a fast track. Either way, the registration campaign means the product and the manufacturing process have defined materials and equipment. While further optimization and QbD studies will be executed in the following months, the amount of change from this point forward is greatly reduced. The registration campaign is the perfect time to determine what is known versus not known and is not the time for assumptions. The QRA reflects the current state of process understanding and serves as a gap analysis between current knowledge and what is still needed.

It is commonly understood that risk is inversely proportional to process knowledge. I use preliminary risk assessments that measure uncertainty as a risk component. Uncertainty is a measure of the quality and depth of data supporting the risk decision. The more uncertainty, the greater the risk because of the lower confidence in the final risk decision. Uncertainty is particularly useful with QRAs in development, which work with limited data sets. I ask SMEs to reference the data upon which their conclusions are based. Three main questions really gauge the level of knowledge:

  • What do you know?
  • How do you know it?
  • How good are you?

Table 1 is a subset of the uncertainty criteria showing a major difference in the level of knowledge. Category 3 reflects expert opinion and experience, while category 2 indicates experimental data supports the conclusion. When data supports the conclusion, one can go to the study, report, or other document where data is captured. Sometimes, the study is well written and executed. In these instances, the uncertainty is low, and the conclusions are fully supported. Other times, data is present and gives a good sense of direction but does not suffice for the final data. It can be for a variety of reasons, including a statistically poor sampling plan or unvalidated analytical methods. On rare occasions, the conclusion is based on data assumed to be available but that, in fact, is not. It is not the expectation that everything be final with fully characterized processes or perfect data sets. The important point is understanding the known from the unknown. The QRA process allows the SMEs to compile and present the knowledge in a logical format, which grounds the project team on a common understanding. In my experience, it has been one of the most value-added activities on a project.

Table 1: Uncertainty Criteria

Reason 2: Align With Your CDMO

Engaging the CDMO on the QRA prior to the registration campaign has multiple benefits. Your CDMO likely has been heavily involved in developing the process and it certainly knows its process flow and equipment. Formally documenting its knowledge moves it from the realm of tribal knowledge to institutional knowledge. It makes the QRA better and ensures the sponsor is on the same page as the CDMO.

A formal meeting with all functions present is essential. Many details are overlooked if sponsors do the risk assessment in isolation. It is not a good idea to simply have the CDMO review your output. Risk assessments work best with collaborative interaction, as the group chemistry, with back and forth discussion, teases out information that would otherwise go unnoticed. The discussion should be cross-functional. It is the time to understand the manufacturing process from different perspectives.

I particularly like to have these discussions at the CDMO’s manufacturing facility. Going to the production line to see the equipment and ask questions gives a better appreciation of the process. This QRA process moves from theoretical to fact-based. The main goal is documenting the current state of knowledge, and that happens best where the action takes place. The factory setting makes the discussions more concrete and specific and lays the foundation for better communication with the CDMO. Effective communication is the key intangible that ensures success.

Get The Best Results

I find it works best to have a facilitator lead the QRA meeting. Ideally, it would be someone with a technical background. It should definitely be someone who understands the risk assessment process. Having an independent facilitator accomplishes a few things:

  1. Facilitators allow the SMEs to focus on the details.
  2. Facilitators help the team ask the right questions.
  3. Facilitators keep the team from getting bogged down in weeds, knowing when to table some of the discussion.
  4. Facilitators help ensure a consistent output.

Reason 3: Prepare Development Plan For Commercialization

By far the best value of the preregistration campaign QRA is ensuring the campaign produces the best quality of information to close specific knowledge gaps. The registration campaign provides an opportunity to sample multiple locations and time points. Don’t miss the opportunity to further characterize the process. On many occasions, the process risk is high because the level of knowledge supporting the decision has high uncertainty. Table 2 represents a preliminary risk assessment for drying after wet granulation for an immediate release tablet. In this case, the formulation scientist has extensive experience with the excipients and the particular manufacturing equipment. The project has progressed quickly, with limited data collected. Expert opinion with some preliminary data is the basis for most knowledge. The elevated risk rating reflects the level of uncertainty. The QRA identified some key areas of study for the preregistration engineering batches and the registration campaign itself. In particular, the team decided to challenge the time and temperature impact on impurities during the engineering batch. They would collect data on impurities with the newly validated method. They also ensured the test plan included correlating moisture with potency and dissolution. The important lesson is the project team determined how to best gather data to fill in holes in the risk assessment. They went into the campaign with clear data-gathering objectives based on the initial risk evaluation.

Table 2: Preliminary Risk of Drying

The QRA should be updated after the registration campaign to incorporate the new learning. The updated QRA provides the basis for the development activities conducted between the registration campaign and process validation. The iterative nature of the QRA allows the project team to adjust experiments to focus resources where they can be most effective.


The registration campaign is a huge investment, and getting the most from it makes business sense. The best practice is to work with the CDMO to assess the risk. Planning data collection beforehand ensures the most valuable data is collected. Good risk management execution not only ensures a successful registration campaign, it lays down the path for development activities, culminating in process validation.

About The Author:

Kevin Wall is principal consultant and owner of Cincero Consulting. He has worked with five virtual pharmaceutical companies to move eight of their drug substance and drug product projects through development. He applies “fit for phase GMP” from Phase 1 to commercialization. Wall’s process for identifying critical quality attributes (CQAs) and critical process parameters (CPPs) forms the basis for the CMC section of the regulatory submission (ICH Q11). He led the Johnson & Johnson development and global deployment of quality by design (ICH Q8) and quality risk management (ICH Q9) in pharmaceutical development and operations. He helped define and deploy Janssen’s risk-based qualification and process validation systems (ICH Q7). You can contact Wall at 817-915-0822 or, or visit his YouTube channel, cGMP Made Easy.