Stuart R. Gallant, MD, PhD
In Pharmaceutical manufacturing, frequently process development uses designed plays—just like in American football. In football, the play caller looks at the situation (first down, short yardage, red zone, etc.) and picks from a small number of options. For instance, if there is less than a yard to go on a critical possession, it’s a pretty good bet that an NFL team will call a quarterback sneak.
Most of the time in pharmaceutical manufacturing, the development team considers the product—for example, an antibody—and selects from a small number of options. In the case of an antibody, the manufacturing process will almost certainly include cell debris removal, Protein A, anion exchange, viral inactivation, and viral filtration. Usually, there’s only about one decision to make in an antibody process: what’s the third chromatography going to be?
But, occasionally, there’s no play in the playbook that fits the situation. These are truly novel design problems. Some examples include:
- Proteins which require special processing conditions
- Molecules which are difficult to formulate
- Recurrent reactor contaminations which continue in spite of standard control techniques
Facing up to such novel problems can be a source of anxiety for the development team.
Today’s post addresses the design process. It includes a general approach to ground-up process design, as well as a few examples.
Picking Your Team
Picking your team is very important to a successful design process. Some things to consider include:
- Knowledge breadth: Make sure that all the relevant disciplines are represented (process development, analytics, Quality Assurance, Regulatory…).
- Knowledge depth: Here, there are some trade-offs to think about. A less experienced team member may bring enthusiasm to the team, as well as strong commitment to the project. A more experienced team member may bring lessons learned from past projects, but at the cost of less availability (due to other obligations).
| Be Outward Focused: Once, when I was working with a team addressing a reactor contamination problem, we decided that we needed an outside expert to support out team. This expert brought world-class skills in cell bank characterization, and they were worth every penny of their fee. |
The Design Process
1) Problem statement: The question you ask will dictate the answer you find. Time spent clearly defining the problem you are to solve will pay big dividends as the project proceeds. Make sure you consider:
- Project needs
- Project goals
- Project parameters
- Assumptions and limitations
- Budget and timeline
- May need to do research to adequately define the problem
- Discuss how the solution will be evaluated (minimum cost, ability to meet certain criteria, time to implement, etc.)
- After you think you have the problem defined, check back with the project sponsor to be sure you nailed it. You don’t want to build the solution to a problem you misunderstood.
2) Refinement of the problem statement: Now that your sponsor and your team agree what the problem is, you can begin thinking about the problem and its solution:
- Are there elements that need clarification?
- Conduct an initial analysis of the problem: Which aspects of the problem are the most difficult and most easy to solve? Does this problem fall into a class of similar problems? Has this or a similar problem been solved previously? Conduct a review of scientific and patent literature.
- Revise and refine the problem statement if necessary.
| The Internet—Your Best Friend: I conduct a thorough literature review of every problem that I confront—writing up the results of the research in report format. Looking back at the report later helps me see the evolution of my thinking. It’s also a great way to bring new team members up to speed on a problem. |
3) Brainstorming: Once you have the problem adequately defined, you can start thinking about novel solutions. You’ll want to involve your entire team in this process. Consider:
- What method(s) of brainstorming might be most effective?
- Defer judgement; initially, the goal should be to maximize the generation of possible solutions
- Resist the urge to jump on the solution that seems most attractive (cheapest, quickest, requires least resources, etc.). Note all possible solutions (the good, bad, and ugly)
- Consider different viewpoints
- Consider different methods of representing the options (visual, verbal, graphical…)
- Make sure that the discussion allows free thought—no bad ideas
- Many possible solutions should be captured—for some problems, this could be dozens
- Some team members are better in groups, and some are better one-on-one. Make sure you follow up one-on-one with all your team members to be sure that everyone is contributing their best ideas
| Be Flexible: Once, when I was working with a team designing a virus purification process, we came up with two solutions: 1) chromatography columns (slow) or 2) membrane absorbers (faster and cheaper). Once we got comfortable with the newer membrane technology, it was clearly the better choice. |
4) Organize and evaluate possible solutions: Now that you have a tentative list of solutions, you will want to categorize them.
- Categorize and arrange the ideas in logical ways
- Research the solutions which are poorly understood (refine your understanding of their benefits, costs, risks, and time required)
- Look at what they have in common and in contrast, their benefits and costs
- Risk rank the solutions
- Strike out the solutions that have too many costs, risks and too few benefits to be of practical utility
| Try Not to Kick the Can Down the Road: Once, when I was working with a team attempting to formulate a low-solubility injectable, we considered some exotic formulation technologies. Unfortunately, using those technologies would have bound us to a small number of possible manufacturing sites. We persisted and found a more manufacturing-friendly formulation which made finding a CMO much easier. |
5) Refine the leading candidate solutions:
- Build out and tighten up the best ideas
- Improve your analysis of their costs, benefits, and plan of implementation
- Consider whether more data must be gathered. Regarding data, use these guidelines:
- Whenever you conduct an experiment, ensure that the experiment is hypothesis driven and designed to help you: 1) decide between options, 2) optimize one or more of your options, or 3) define the risk(s) of one or more of your options
- Consider how long it will take to plan, execute, and analyze the experiment. Data which will be returned too late to affect your final decision may still be worth collecting—it can be used to buttress the final decision. But, no data is free—all experiments cost money
- Be aware that you will almost always have too little data to be 100% certain
6) Ratify twice: You will want to ratify your solution two times. First, you need to get buy-in within your team. This should be relatively easy because the team has been intimately involved in the design process—so they are aware of all the data. However, you need to be aware that sometimes individual team members can feel special ownership of their contributions—that can throw of their internal decision-making process. Once, you have the team’s buy-in, you need your sponsor’s buy-in. Hopefully, you have been communicating with your sponsor (and/or customer), so the solution should not come out of left field. But, you may need to sell the sponsor on the solution because they have less invested in your particular solution than the team does. As you seek ratification, consider:
- Is there one best solution or several acceptable solutions? (Be wary of mindsets which allow only one solution—usually there is more than one solution to a given problem.)
- Discuss the possible solutions with stake holders and sponsor and gather feedback. Poll stake holders and sponsor for both solutions they find acceptable and their preferred solution
- Discuss next steps with the sponsor/customers
- Hold a meeting, publish, or otherwise convey the results of the process within the company and outside the company
| Build Support: Once, when I was working with a team manufacturing a difficult protein, we decided that the protein required low temperature handling. To get management support, we needed to assemble a compelling data set to justify the expense and hassle of processing at 2°C to 8°C. |
7) Evaluate:
- Review the design process and how it was conducted
- Review the quality of the solution, both immediately and after some time to see if the solution has held up
- Document the solution and the learning that occurred for future use
Those are the key ideas in pharmaceutical design. Design is intimately tied to technology transfer. I would like to do a post on technology transfer some time soon. So, keep an eye out at PharmaTopoTM.
Disclaimer: PharmaTopoTM provides commentary on topics related to drugs. The content on this website does not constitute technical, medical, legal, or financial advice. Consult an appropriately skilled professional, such as an engineer, doctor, lawyer, or investment counselor, prior to undertaking any action related to the topics discussed on PharmaTopo.com.