Q: I have a situation where we can’t replicate a customer’s problem when we test the returned product samples. What must we do to close such complaints? Furthermore, in cases where we suspect user error, how do we handle those complaints and still meet FDA’s expectations for proper closure?
A: If you can’t replicate the event or failure mode, you can close the complaint for insufficient data. That’s assuming this is a first-time complaint — or the complaint has occurred before and is recurring with the expected severity and frequency defined in your product risk assessment.
[pullquote align=”right” cite=”” link=”” color=”” class=”” size=””]When you try to replicate the problem, make sure you do exactly what the complainant claims he/she was doing at the time.[/pullquote]
But this step is important: close the complaint with a commitment to continue to track and trend the issue in case you get similar complaints that help you better understand what’s happening. Also, when you try to replicate the problem, make sure you do exactly what the complainant claims he/she was doing at the time. It’s not sufficient just to use your normal finished-product test, unless you can prove the test will detect the kind of failure described by the complainant.
In cases where you suspect user error, you should have on file investigations that show the results of poor technique or faulty use (or other known problems such as improper product storage, using expired reagents, etc.). If you don’t have these types of investigations, open a CAPA and investigate what user-dependent factors can lead to poor product performance. For each of your products, you should have a documented investigation that allows you to say, “What the customer says happened, we know will happen when the following poor technique (or whatever) is used – see Investigation #123.”
In this way, you need only to show the product is testing OK when proper technique is followed. Because you can’t replicate the customer’s problem, you believe the most probable cause is poor technique, and the number of complaints for this lot is occurring within the expected frequency and severity defined in your product risk assessment. Therefore, no further investigation is required, but the event or failure mode will continue to be tracked and trended.
Answered by Denise Dion, EduQuest Vice President of Regulatory and Quality Services (18 years as an FDA investigator and co-editor of FDA’s Investigations Operations Manual) and instructor of EduQuest’s The CAPA Clinic: FDA Rules for Effective CAPA Systems, Failure Investigations and Complaint Management.