Ee Bin Liew Hi, if one value-stream-maps a complaint, it's amazing how much time, effort and cash is need to just even process a NFA (no further action), not to mention one that requires field action! This is irrespective of handling complaints manually or via software. perhaps companies are more comfortable shifting the initial goalposts in order to manage the enormous workload, and we're probably only talking about a small fraction of actual complaints because most don't get reported back to the manufacturer.

In addition, fixing problems are never going to be a sexy as new product development.. or perhaps the company feels a safety-related fix on the device can wait for the next iteration due to some skewed risk vs. benefit analysis?

the regs didn't change much, if at all, with regards to complaint handling, all around the world. But the same issue remains. Smaller companies or startups actually have a better opportunity to get it right, as there's no legacy issues to worry about. It's a matter of committing to it. Until the industry really commits to it, the WLs and 483s or the equivalent in other jurisdictions will keep coming. That actually shows that the system is somewhat working in my opinion - you get punished until you actually listen and learn.

hope it helps.

Cheers,
Ee Bin

James (Jim) Dent, LSSBB, DTMx2 What's the ratios of the complaints compared to the volume of the product or number of products on the market. If the shear volume of products has increased, and the widespread usage of the products have spread across the globe - then theoretically, the number of complaints will statistically increase proportionately. The number of complaints may also increase if the number of new advanced technologies are also introduced in a during the past 4 years as compared to the previous 4 years. Also examine whether the number of start-up medical companies, placing product into the market, has significantly increased over the past several years as compared to the previous period.

Even if the req's did not change, nor the frequency, nor the diligence of the audits changed - there are most likely these other factors that are resulting in the appearance of increased WL's and 483's. The number per individual company could be declining; but the appearance of increase is due to the newer companies which weren't in the statistics 4 years ago.

need for > greater profit + new hires - education and adequate training transition = ...

complaints+capa are the heartbeat of the quality system

Is it at all related to significant issues in the field and the FDA enforcing its role to protect consumers? Also, I know the FDA is moving toward having auditors who are specialists in product areas (not sure if this has started yet). Maybe auditors are more aware of failure modes and can thus better detect deficiencies in the complaint resolution process?

Dan O'Leary I do a lot of work in complaints, so I was intrigued by your question. I decided to look at the percentage of times that a Warning Letter that cites any QSR section also cites complaints, 820.198. I found the following: 2008 65.8%; 2009 59.3%; 2010 61.1%; 2011 56.5%; 2012 66.3%; 2013 61.7%; 2014 60.8%; 2015 68.8%;


I didn’t see a major change in the rates over time, but I did see numbers that are significantly higher than I had expected. I would have guessed 25% to 30%.

I don’t conclude there is change in the way companies have been interpreting the regs, but rather conclude it has been a uniformly bad job over time. I don’t have a sense of the kinds of problems cited. I would have guessed failure to have procedures, failure to designate a unit, and failure to evaluate for MDR reportability would be the top three, but based on the counts, I think that issue requires more analysis.

Brian Dawson Dan, thanks for generating the Warning Letter data. The high percentage was not a surprise to me as it seems complaints and CAPA are typically the most cited QSR elements in 483 inspection reports. I was also interested to see there was not a substantial change in how often complaints was cited. In my experience the problem areas are usually inadequate investigation, failure to implement corrective action and not evaluating MDR reportability correctly.

The failure to investigate and failure to implement corrective action go hand-in-hand, and seem to be one of the easiest parts of the QSR for an investigator locate a nonconformance. Almost all companies have some complaint types which recur and for which they have not implemented corrective actions. If the investigator feels the reported failure represents a failure to meet a specification and the company has not implemented corrective action it can easily become an observation.

Regarding MDR reporting, there are two predominant errors made. First, is the failure to report a serious injury when "medical intervention" is made to prevent a serious injury from occurring. Investigators tend to take a conservative view of actions taken by medical personnel as being actions to prevent serious injury, while companies seem to not recognize actions as interventions. Second is the failure to report a malfunction which is likely to result in a serious injury. Companies will sometimes point to a risk analysis performed during product development which says the malfunction is not likely to cause injury. That rationale is generally OK if there have been no reports of injuries occurring from the malfunction. Once an injury is linked to that malfunction, FDA considers that malfunction to be likely to cause injury upon recurrence and the risk analysis should be updated.

David Lim, Ph.D. RAC, CQA Brian, as for your MDR and two predominant errors being ("medical intervention" and "recurrence" for a malfunction), there is a BIG Gap between industry's understanding and FDA's expectation. Industry tends to look at the facts in a narrow sense (just the incident itself), but FDA looks at the facts "under the surrounding circumstances" at the time - thus, there are different outcomes depending upon how we view the situation based on the facts. Regulations presume that it will recur once it occurred unless it can be shown that no similar incident occurred in the past two years. Most of decision makers are not aware of these two BIG GAPs, rather making frequent, costly and sometimes deadly mistakes.

Dan O'Leary My analysis would not have found MDR errors, such as the type mentioned by Brian, because they would be under Part 803 in a Warning Letter. I looked at Part 820 exclusively. 820.198(a)(3) requires complaint evaluation for reportability. I see failure to evaluate as a common problem. The evaluation could, however, result in a conclusion that the complaint is not reportable.

Alvin Tai I don't think it's necessarily "errors" being made, I think that it's sometimes difficult to balance out the maturity of a CAPA/Complaint handling process. If you make it too complicated, you may fall into a resource trap and won't be able to maintain compliance to your own procedures. On the other hand, if you make it too simplistic, you run the risk of not even having a compliant process. This is such a great topic though, and I wrote a blog post with some data behind FDA observations: https://medium.com/@coreinput/quality-system-focus-for-medical-device-startups-8094e6ad7563

Alicia Racelis Dan,
Could you see the distribution of the company types being targeted over the years? Just wondering if specific types of medical device companies are feeling particularly susceptible to WL's and 483s at this time.
Warm regards,
Alicia

Dan O'Leary Alicia,

Good to hear from you after all these years.

I don’t have good data on company type. The only classification I’ve tried is the phrase at the beginning of the Warning Letter, that says some like “your company is a manufacturer of …” or “your company is a specification developer of …” etc.

I don’t think FDA is always consistent and I know I have not been consistent. For example, a company could identify a company as both a manufacturer and a specification developer, but I’m not sure I entered them consistently.

Alicia Racelis Hi Dan,

Great to talk to you again as well! I guess my question turned into the Heisenberg principle:) , but your data on the consistent trend overall remains quite interesting regardless. Definitely worth a further look at some time.

Kind regards!
Alicia