By Paul Thomas, Managing EditorYou cant improve or control what you cant measure, goes the familiar axiom. In a focused discussion at Septembers IFPAC Summit in Puerto Rico, discussion leader David Radspinner, Thermo Electrons PAT expert, offered two corollaries:
- Just because you can measure it doesnt mean you can control it.
- Just because you can measure it doesnt mean you should control it.
Drugmakers looking for success in PAT implementations must not only upgrade their quality systems, says Radspinner, who was involved an FDA PAT evaluation as one of Sanofi-Aventis PAT leaders. They must also adopt a new mindset and approach to quality, one that has risk management as a key component.It is important to focus not on mere procedural steps toward quality, he said, but on the concept of quality and its ultimate link to the consumer. Quality systems must be infused with risk-based approaches that consider the impact that a manufacturing failure might have on the performance of the product, if it were to happen. They should also consider the likelihood of such an event happening, and the ability to detect or notice that event.An iterative, continuous improvement approach to learning will also be critical, Radspinner adds. Such an approach allows for new details to be discovered well into the implementation and accounts for external forces that may shift conditions or requirements for the implementation.Finally, manufacturers must rethink qualification/validation approaches that:
- assume users know exactly what they want;
- assume all requirements are of equal criticality;
- view success as determined only by the completion of documentation;
- neglect or downplay risk assessment.
Give FDA a warm welcomeAnyone undertaking a PAT implementation should expect a visit in the early stages from PAT-trained FDA reviewers, Radspinner emphasizes. Because FDA neither endorses nor encourages any specific technologies or applications, manufacturers should view the visit as an opportunity to build trust with the Agency. They can do this by exhibiting a commitment to a new way of operatingshowing learning and rationales that support enhanced process understanding and decision-making that is focused on quality and based upon science and technology.Radspinner ended the discussion by answering questions addressed by members of the audience.
Below, we present answers to FAQs that he presented in August, at the 45th Annual Conference on Pharmaceutical Analysis in Wisconsin.Q: When is the best time to contact FDA?
A: As early as possible. Whenever the manufacturer has something to say about an idea, plan or results, is a good time to pick up the phone. Its a myth that manufacturers should wait until they have hard data to contact FDA. The PATriot team has made it clear that it is always willing to talk.
Q: Is PAT mandatory?
A: No, PAT is not mandatory. However, your business model should be driving it.
Q: Will it ever become mandatory?
A: No, but it will become a part of what you do. The expectation will be that you understand, monitor and control whats critical.
Q: Which Technologies or Applications are already approved by the FDA?
A: The PATRIOT team at the FDA neither endorses nor encourages specific technologies or applications. They have been very careful to avoid any sense of recommended technologies. Instead, they expect continuous innovation.
Q: If I move my analysis of X from the laboratory to the production floor, is that PAT?
A: No. This may improve your business (so by all means do it), but dont bother the FDA with a request for a PAT filing. Remember . . . design, analyze and control.
Q: If I look more carefully at my process, I may find things I dont want to know what should I do?
A: This question is frighteningly repeated more times than you may think. The process is what the process is. If there are issues, they are there whether you look for them or not.If your focus is on risk management, then only those things you find which indicate a higher level of risk would be mitigatedthat is what quality management is all about. If it is of limited risk, then prioritize your efforts and move on. If it is of a higher risk and you choose not to mitigate it, then let me know which drug you make so that I may avoid it!
Q: What if I look at a higher sample size and find out-of-specification results?
A: Consider difference between a specification and a standard. Most specifications are sample-size specifice.g., content uniformity is based upon a known, fixed and small sample size.Also, evaluate with risk assessment. Consider the difference between expected rate of occurrence and actual.
Q: Does real-time release = parametric release?
A: Real-time release is not parametric release. Rather, it builds upon this concept, but instead involves:
- Replacing laborious methods with rapid off-line methods (e.g., HPLC replaced with NIR);
- Moving traditional testing at-line or on-line, without linking to quality control;
- Measuring everything you can with unconnected specifications.
