Process Capability in Sealer Validation
As part of ISO 11607 Part 2, medical device companies must have a validation program to demonstrate the repeatability and reproducibility of their sealing process. The most common statistical tool used in the industry to demonstrate this is the process capability index (Cpk).
I've seen sealing process validations with extremely high Cpk indexes. I've also seen validations that had visual integrity compromised in order to achieve a minimum Cpk value. In both cases, it was shown that the process was not only capable but also in a state of control.
Why are you having sealing issues if the validation has an extremely high Cpk value? Why did the Cpk fail to catch this problem? The easiest thing to say is that the minimum seal strength specifications and/or the test methods are wrong for the package. But what if they are OK for your application? The problem may be coming from the way the data is collected and entered into the formula. Sometimes we may use the average peel among X amount of pulls per package to calculate the Cpk. This is what I called the average of the average Cpk calculation. This practice normalizes the data by bringing the standard deviation closer to 0, thus getting higher Cpk values. This practice also does not evaluate the variation that can be caused within the sealing area by uneven temperature distribution on the platen, uneven pressure distribution, and other factors.
Another variable to consider is the test method itself. Do you understand how much variation you are getting from the test equipment? Has it been validated? Sometimes we take it for granted that we are using an approved ASTM test method, and as long we stay in accordance with it, we should be fine. Well, this can also be contributing to your lower Cpk values. What is the first thing that comes to people's minds when Cpk values are low? The answer in the majority of the cases is increasing the seal strength. If you take a close look at the Cpk formula, the biggest contributor to the Cpk value is the denominator, because it is multiplied by 3, versus the numerator, which is a simple subtraction. In addition, higher seal strength values typically bring higher standard deviations.
Remember that reducing the variation going into the test results will get you better controlled processes and thus better Cpk values. This variation can be coming from your sealer equipment as well as from your test method. Before you start compromising visual integrity to get higher seal strength values, stop and think about what you can do to reduce the variation you are seeing in your test results.
Charlie Rivera, CPP