Validation: Does the Supplier Share Your Interpretation?
Validation is a critical element in medical device manufacturing success, but what if a supply partner doesn't share the same interpretation? How will this impact the bottom line and the perception of the product or even the company? This article reviews several important aspects of the validation process that must be discussed between a device manufacturer and their supply partner before production begins.
Dana Schramm is the VP of engineering at Donatelle. He has over 15 years experience in plastics, more than 10 of which is exclusive to the medical industry. Schramm can be reached at 651-633-4200 or firstname.lastname@example.org.
The term validation has become a frequently used term in medical device manufacturing. This does not mean that there is a uniform application of this term as there are still many different interpretations on when it is necessary, what is included, and what success looks like. With the increased scrutiny by the FDA on supplier controls, can medical device companies afford to take the risk that the manufacturer of its product is not effectively conducting process validations? Or worse yet, that it is not conducting them at all?
The FDA requirements, as defined in QSR 21 CFR Part 820, are very clear; they state, " . . . where the results of a process cannot be fully verified by subsequent inspection and test, the process shall be validated . . . " There are no exceptions for low volume or low risk products; this applies unconditionally to all products covered under the QSR. This surely sets the minimum requirements, but should the process be used to satisfy a regulatory requirement or should it really be an essential step in developing the overall manufacturing plan for a product? When ISO 9000 was first introduced, many people saw this as a burden on their business instead of realizing that a sound, quality system is actually a benefit to the business. Process validation is similar in that if used effectively, it results in significant benefits to the business–both the suppliers and the OEMs.
Like many standards and regulations, there is significant room for interpretation within the QSR. To provide further guidance, the Global Harmonization Task Force (a group of representatives from the medical device regulatory agency and industry) created a document which provides further definition on process validation. This document, GHTF/SG3/N99-10:2004, helps define the purpose of process validation, when it's required, and general guidelines on how to perform a process validation. There are some processes defined, such as injection molding, that can't be effectively verified and require a validation in all cases. Other processes, which may be verified, are left for the manufacturer to determine whether the validation route is the proper path. For the validation to be valid, two high level requirements must be met. Objective evidence must be collected and then that evidence needs to be assessed against a set of predetermined requirements.
Objective evidence means that there is rationale behind the data being collected, how it's being collected, and how it is analyzed. For a complex device, the cost of the validation is greatly impacted by this portion of the validation. The number of parts produced during a validation is directly tied to the confidence statement that can be made about the process following the validation. The larger the sample size, the greater the confidence and/or reliability statement. For example, a sample size of "X" allows a statement that with 95% confidence, 99% of the product meets specifications. The confidence statement that companies intend to make should drive the sample size. Conversely, they do not want to perform a validation without thought given to the sample size. If the sample size was arbitrarily chosen at a low number due to cost or timing, it would be unfortunate to pass a validation, only to find that on a critical feature, the confidence statement results in a 90% confidence that only 80% of the product meets the specification. If this is the result, is it realistic to move into manufacturing with this process and not expect defects? To make decisions based on statistical data, there needs to be effective means of collecting this data. Establishing pre-determined requirements for gage reproducibility and repeatability (GR&R) is a necessary step for any feature being analyzed in the validation. Most companies have company-wide established requirements, such as 20% or less for a GR&R. This means that it's possible for up to 20% of the specified tolerance to be consumed by the measurement method. It is foolhardy to make statistical statements if there is not confidence in the data being collected. Developing a suitable measurement method as part of the validation plan also is necessary as part of an ongoing control plan once in production.
Pre-determined requirements mean there should be forethought put into defining the product requirements. A common practice is to apply standard tolerances to most dimensions on a print specification. If product requirements are unnecessarily constrained (e.g., tight dimensional tolerances), then there are unnecessary costs related to validating a process as capable of meeting these requirements. These costs are in the form of additional development time, increased inspection activities, data analysis, report writing, etc. In addition, if the validation fails as a result of these requirements, the product designer is faced with the potential of changing requirements to match the output of the process. To any auditor, this could potentially raise a red flag. It really raises two questions:
• How much thought was really put into the design stage if the design requirements can be changed to match a manufacturing process output?
• Are the changes to the product requirements really just a step to formally close-out the validation phase?
To preclude such situations, a well laid-out audit plan is necessary with input from both design and manufacturing personnel.
Once the product requirements are well established and the evidence that will be collected to demonstrate the manufacturing process is consistently capable of meeting these requirements, the validation plan can be created and executed. It isn't necessary to toil over the format of a protocol. In addition to an example in the GHTF document, templates and examples are widely available on the internet. The importance of the protocol is really the thought that was put into the contents. In addition to the previously mentioned elements, there are several other critical portions of the validation. An essential element to any validation is a well understood manufacturing process. There are two reasons for this. First, the validation should not be used as a process development tool. The validation should be a confirmation tool that proves that the process is capable of meeting the product requirements. Second, without a well understood manufacturing process, it would be difficult to establish the critical process parameters that will be challenged in the validation. Another critical element is establishing a process control plan that ensures the manufacturing process is within the scope of what was proven during the validation. Finally, the validation should provide scope around when a re-validation is necessary. This portion should identify what manufacturing process variables that, if changed, would have a significant impact on the process output and consequently, would no longer be consistent with the process proven effective during the validation.
Many of the activities that occur as a result of developing and executing a validation should occur as part of any good practice to developing a manufacturing process. The validation process formalizes these activities into a commonly accepted framework within the medical device industry. The validation is not a substitute for process development but as a confirmation that the process development is complete and ready to move into production. The thought process and methodology that goes into a validation will be directly proportional to the confidence in the manufacturing process once complete. Choosing the wrong sample size or process parameters will not only increase the risk that product will be produced outside of product specifications but will quite likely require process and/or product specification changes at a later date. The cost to implement changes at a later date becomes increasingly expensive and may impact product design validations and/or FDA submittals. A well thought out and executed validation plan is inexpensive insurance against these risks.