Welcome to this week’s installation of Warning Letter of the Week (WLOTW). Each week we will review a different FDA Warning Letter to gather as much information as possible to avoid repeating those mistakes. The Warning Letters will range from recently posted letters to letters covering specific markets, and anything else we find interesting.
If you want to learn more about Warning Letters, check out FDA Warning Letters: Everything You Need to Know.
This week we will review a Warning Letter given to Olympus Medical Systems (full Warning Letter). As the Warning Letter is extensive and contains a lot of valuable information, this will be part one of a multi-part series diving in depth into the Warning Letter.
The Olympus Medical Systems site in question manufactures endoscope accessories and had major violations related to complaint handling and corrective actions procedures.
Let’s look through the violations to understand the extent of the issues.
Table of Contents
Overview of First Violation
The first violation is that Olympus did not: “adequately establish and maintain procedures for implementing corrective and preventive action, including analyzing processes, work operations, concessions, quality audit reports, quality records, service records, complaints, returned product, and other sources of quality data to identify existing and potential causes of nonconforming product”.
This is not surprising for a finding, given that CAPA and complaint procedures are the two most cited areas for 483s and Warning Letters.
Also, based on the information provided in the Warning Letter, it sounds like this was a for-cause FDA inspection. In these situations, the FDA is going to review the complaint process with even more scrutiny than usual.
Following the overview of the violations, the FDA lists specific instances that lead them to the conclusion that the procedures were not adequate.
Not Taking Action on Trending Data
The Warning Letter states that Olympus “received approximately 160 complaints describing the “distal end cover” model number MAJ-2315 has ‘dropped out.’”.
Next, it says that the company did identify that the number of complaints received was higher than expected, but did not consider this reason to initiate a CAPA.
I am very curious about the total number of complaints the company receives, to help understand what the decision-making process was that led to a CAPA not being generated.
If the company receives thousands of complaints for various devices, it would make some sense that one increasing complaint trend would not trigger a CAPA (although, there would of course be other concerns about product quality).
However, if there is a smaller number of complaints, it would be extremely interesting to learn how the spike in complaints did not trigger a CAPA.
What I believe was the downfall in this situation is that one device had so many complaints for the same issue. It would be different if one device received a significant amount of unrelated complaints. But the sheer number of complaints for the device all focused on the same failure should generate a CAPA.
This is a good reminder to check your CAPA triggers and quality data. Just because your procedure does not force you to create a CAPA for a negative trend in complaints, does not mean that will be satisfactory for the FDA.
Reportable Complaints not Generating CAPAs
The next example of not having adequate procedures is truly astounding. The Warning Letter reveals that Olypmus’ complaint handling flowchart “describes that complaint investigations that would require evaluation for reportability would not be considered for Corrective and Preventive Action”.
Anyone with medical device complaint experience could tell you what an enormous oversight had been made.
Not only do all complaints need to be evaluated for reporting and inclusion in the CAPA program. But as the FDA says “complaints that are deemed reportable are of the highest risk complaints and therefore especially important to be considered for corrective and preventive action implementation.”
I genuinely wonder if the company made a typo-style mistake when creating its’ complaint flowchart. It does not seem feasible that they consciously determined that reportable complaints would not generate a CAPA. Either way, it is a tremendously bad look for the organization during an FDA inspection.
While there are no hard and fast rules, as a general principle any complaint which must be reported should generate a CAPA. This is what the FDA expects, and it will not go well if you choose to not even consider taking action on reportable complaints.
Now would be a good time to review your complaint handling documents to make sure you are not making any similar mistakes. You can also review our complaint handling guide, for further information to help you create a successful ISO 13485 and FDA complaint handling procedure.
Yield Control for Packaging
The third example we will be reviewing today has to do with packaging nonconformances found in the same medical device that had a negative trend in complaints.
The issue was that the company was using yield control to justify not following their Nonconforming Product Procedure for packaging concerns.
“Non-Conformances (b)(4) and (b)(4) were opened for observed wrinkles in the package seal of Distal End Cap/MAJ-2315…
Your firm stated that yield control may be appropriate for some low risk nonconformances instead of using its Nonconforming Product procedure. Yield control allows for an acceptable level of package seal defects, even though there is no assurance that visually observed seal defects would not compromise the sterile barrier.
Package sealing is a validated process, which results cannot be verified when the validated requirements of the seal are not met.”
The issue here is that the packaging was a validated process, and the validation did not include visual seal defects such as wrinkles. Therefore, there is no reliable way to know if the packaging maintained a sterile barrier.
The FDA considers maintaining a sterile barrier high risk for this type of device. Therefore any issue which could lead to a compromised barrier should likely result in the generation of a non-conformance (and potentially CAPA).
And because the parts were not considered non-conforming, the “nonconforming product (was) not included as quality data for the corrective and preventive actions.”
Besides good processes not being maintained, the FDA likely included this example for two reasons. The first is that this is the same product as the first example. When the FDA reviews a product for complaints, there is a good chance they are going to look at how non-conformances were handled for the product.
The other reason is that this was another instance of the company not using trending quality data to generate a CAPA. CAPAs are the most important element of a medical device quality management system in the eyes of the FDA. Because they did not generate a CAPA from a trend in high-risk non-conformances or complaints, it showed the company did not have a good enough CAPA system.
Follow-up from Packaging Non-Conformances
Last but not least, let’s review Olympus Medical Systems’ response and the subsequent FDA response to the above findings.
After receiving the initial 483, Olympus Health Systems “initiated CAPA-200735 to investigate complaints of MAJ-2315 “dropped out” and after conducting a Health Hazard Analysis (HHA), concluded that no additional action is required and they will continue to monitor complaints for the distal cap falling off in the patient.“
However, the FDA did not agree that the actions were satisfactory given the risk level of the device for several reasons:
- “The HHA conducted does not discuss how/why the risk to the patient is not increased by exposure to the uncovered endoscope during the procedure as it is removed.“
- “The HHA does not consider the risk of tissue injury from navigating the gastrointestinal tract with an uncovered scope tip.”
- “your firm’s documentation concludes that the distal cap is unlikely to fall off if the user has applied it correctly. CAPA-200735 does not consider whether the device actually meets user needs as required by 21 CFR 820.30(g), including the ability to recognize imperfections in the cap and attach it to the scope correctly without damaging it.”
These are helpful examples if you are ever in a situation where you need to provide a response to the FDA.
The main focus of the FDA when it comes to medical devices is the risk to the patient. If you are conducting an analysis on an issue, it should focus on the risk to the patient. This includes why the patient was put at a lower or greater risk, and what the possible injuries could be in the result of a device failure. (That information should also be included if you are performing a Failure Mode Effect Analysis (FMEA), and highlights what should be included in the FMEA.)
The missing elements of the HHA are proven to be a significant oversight later given information presented later on in the Warning Letter, which we will review in a subsequent section of this multi-part series.
The last referenced example focuses on design flaws in the device. If you are going to state that there will not be issues if the device is used correctly, you should also prove that the device can easily be used correctly. This information should also be captured in the initial testing of the device and include real-world use cases.
Seeing as the company received 160 complaints about the cover dropping out, it is clear that this was not just a one-time event of someone not using the device correctly.
Lessons Learned
There is a ton to learn from this Warning Letter, which is one of the reasons the Warning Letter will be covered in multiple parts. Let’s look at some of the actionable items that come out of this section of the Warning Letter.
- Ensure you have a sufficient CAPA procedure, with clearly defined CAPA triggers. It cannot be stressed enough how important a well-running CAPA program is to the success of a quality management system.
- Review how trending complaint and non-conformance data are handled. If there is an increasing amount of complaints or non-conformances for the same part, and the same issue, this may require a CAPA to be generated.
- If a complaint needs to be reported to the FDA, you should almost always perform a CAPA for that complaint. There are some exceptions, but this should be the go-to when setting up your complaint procedure.
- Do not allow parts or packaging with problems to move through without a non-conformance being generated for validated processes. Unless the problems are included in the validation, you need to do some type of investigation before allowing the parts or packages to be used.
- If you are performing a Health Analysis as the result of a complaint, it should heavily focus on the risk to the patient. Your goal as a medical device manufacturer should be to fully understand the potential risk to the patient that is present with your device.
Wrapping Up
We hope you enjoyed this week’s WLOTW article. The Warning Letter we discussed has a massive amount of detail and provides some useful insights for QA professionals. Next week, we will continue in the multi-part series, by reviewing more findings related to validation and complaints.
As always, if you have any questions or thoughts on this Warning Letter, feel free to leave them in the comments below.