The Qualification Process: Integrity of Samples for Qualification
Facing a highly competitive market, many pharmaceuticals, biomanufacturing and microelectronics companies are continuously looking to gain cost savings on the critical items utilized in their production facilities. This usually results in critical products undergoing a vigorous qualification process. While SOPs and specifications may need revision, the crux of the process is the actual vetting of production samples in order to insure product quality and other testing parameters (i.e. particles, extractables, form, fit, function, etc.) deemed important to the processes.
The integrity of the samples that are used to qualify the product is an extremely important component of any qualification and should be sourced indirectly (not directly from the manufacturer). This of course, would be more representative of the product that would be received in future shipments as it would more likely experience comparable manufacturing & shipping conditions, temperature, humidity, handling and much more.
In our experience, we’ve discovered that given the financial impact as well as the inherent risk associated with any technical qualification, some manufacturers may send out samples of lots with results that they are more comfortable with which may not be an actual representation of product that would be received in future shipments. In some cases, companies have written their specification to account for the qualification lots, only to find subsequent lots not meeting the sample lots testing parameters. Why introduce a variable which could potentially yield non-reproducible results?
Indirect sampling (off the shelf samples) allows you to reduce variability and provide assurance that you will be testing a lot representative of the product that you’ll receive in the future. There may always be an element of variability from lot to lot, however actual representative production sampling will more likely deliver more reproducible results than a “preferential” sample (or samples).