My definition of “specification” is rather simple: it’s a promise.
Just like any other promise—you’d better be sincere when you make it and able to keep it. Failure to keep a promise brings disappointment. Frequent failure leads to distrust. And consciously breaking a promise is nothing less than deceitful.
Just as in our personal lives, the pharma industry makes a promise to its healthcare professionals and patients every time it establishes a product specification.
Whether it is a raw material or component from a supplier, in-process material, or final product, a specification is your promise to provide a product that possesses the attributes known to make it work.
Anything outside of the specification range is either unknown because it has not been studied, or known to have some probability of a negative effect. Neither is acceptable.
The same could be said for process control ranges. Although they are applied to the manufacturing process and facilities, they nonetheless are “promises” based on a scientific field of study with respect to product quality.
Thus it seems to be particularly egregious when specifications and process controls are capriciously established or changed.
Take for example River’s Edge. The Warning Letter1 stated that Lidocaine HCL 3%/ Hydrocortisone 0.5% lots were released and distributed even though they failed the initial release and stability viscosity testing at 3, 6 and 9-month time points. The firm responded by revising the specification range.
Additionally, after failing to investigate over 30 complaints of discoloration of Hydroquinone 4% Cream, River’s Edge responded that it would address the issue by revising the labeling to describe the cream as “tan to slight brown on storage.”
Another example is Gilead. The Warning Letter2 stated that their aseptic processing room was not adequately constructed to meet design specifications. The room, in which partially-opened sterile drug is transported, failed to meet ISO design criteria. The firm responded by reclassifying the room to a lower standard.
FDA was not amused in either case.
Quality Assurance is responsible for being the guardian of specifications and making sure that systems are in place and followed for developing and changing specifications.
It is a system based on statistical analysis of experimental studies, which is part of the process and product knowledge base. You know—the institutional knowledge of the scientific and technical aspects of what you make, how to make it and why your specifications are what they are.
I am aware of some shoddy specification-setting practices, like—taking the highest and lowest values on record, plus or minus your shoe size—just to file the NDA quickly to beat the competition.
But this always causes problems. Manufacturing processes are not efficient and varied product attribute results are obtained. It all adds to a high nonconformance rate and adds cost to the operation.
So a specification is not only a promise, but established properly you get something in return. It’s the “sweet spot” for economical operations.
1 River's Edge Pharmaceuticals, LLC, Warning Letter 10-ATL-05, May 20, 2010. http://www.fda.gov/ICECI/EnforcementActions/WarningLetters/ucm220315.htm
2 Gilead Sciences
, Warning Letter 44-10, September 21, 2010. http://www.fda.gov/ICECI/EnforcementActions/WarningLetters/ucm227649.htm
I would make one comment-QA is indeed the guardian of specifications BUT QA does not set specifications! If R&D sets a specification it is their job to justify it and not QA's responsibility. Remember that when one is talking about quality risk management in terms of lifecycle it is not unreasonable to have a change in specifications. What went wrong in the cases you quote is that the companies incorporated these changes-not via change controls but as a response to an audit. That is the real issue here
ReplyDeleteYou are absolutely correct, Malcolm.
ReplyDeleteQA ensures that an appropriate system exists whereby specifications are not just sucked out of R&D's thumb. But R&D typically sets specifications. In fact, R&D can even write the procedure(s) that describe the system for establishing specifications--but QA approves.
And---correct again. It is not unreasonable for specifications to evolve over time as more data accrues with operational and experimental experience. The change control system gives QA the opportunity to review the justification.
In these cases, one wonders what QA's role was in reviewing the 483 response. The FDA response was certainly a predictable.
Thank you, Malcolm.