Recall vs. Field Correction - Field Corrections instead of a Recall

[robert.beck]

Hello Robert,
Interesting situation.

The actions you describe would not be considered a "Recall" per FDA regulations 21 CFR Part 7 or Part 810, since only FDA can determine that a Correction or Removal is a FDA Recall.

Regarding reportability of corrections or removals to FDA per 21 CFR Part 806....reports to FDA are required if the Correction/Removal is initiated to reduce a risk to health. You need to document a Health Hazard Evaluation (HHE) to make that determination.

When you replace the existing device with a new device, do you ask the customer to return the existing device?
If Yes, that would be a removal.
If No, your replacement action does not directly meet the definition for a Correction or Removal.

§806.10 Reports of corrections and removals.
(a) Each device manufacturer or importer shall submit a written report to FDA of any correction or removal of a device initiated by such manufacturer or importer if the correction or removal was initiated:

(1) To reduce a risk to health posed by the device; or

(2) To remedy a violation of the act caused by the device which may present a risk to health unless the information has already been provided as set forth in paragraph (f) of this section or the corrective or removal action is exempt from the reporting requirements under §806.1(b).

Regarding the "hotfix", the same requirements apply. You should document a HHE for the hotfix Correction, and if it meets FDA reportability criteria then you should report it to FDA.

You may find the attached presentation useful.

Thanks for the presentation. Slide #5 implies that there are corrections and removals that do not have to be reported and later slides go into more detail, identifying these situations as "exemptions." The company keeps detailed records on this and is planning to implement an improved tracking and record keeping system in the near future. I think we are in a grey area because most of the device replacements are done as the most time-efficient method of getting a working device into the patient's hands. A small percentage lead to device improvements, and an even smaller percentage could by described as "user error." I am trying to get them to stop doing hotfixes because they should be able to include the vast majority of these into a scheduled maintenance release. Because the device is so easy to use, it is prescribed about 100,000 times/year. Sometimes it is prescribed inappropriately, for instance to a patient with mental health issues, who took it off and threw it across the room.

I have another question along the lines of the same issue. This particular device is very safe and does nothing harmful. It is supposed to monitor some physiological parameters which are non-life threatening. It does nothing more than monitor and report. There is another set of similar physiological parameters, which indicate a serious potentially life-threatening condition, but this device is not intended for or indicated for monitoring of those parameters. The company certainly does not market it for this situation, but sometimes physicians use it that way because they cannot distinguish between the two situations because of their own deficiency as a physician or for convenience, and rarely this leads to a patient death. The company insists their device is not at fault and they cannot control physicians, and a certain amount of misuse is inevitable. The frequency is very low; a high estimate is 0.04% per year. What should be done about these events? My opinion is that they should at least be reported to MAUDE, but the management feels this will seriously and adversely affect their business and argues against it as I noted above. I'm considering leaving them because of this. Am I overreacting or not?

 

[Ronen E]

I have another question along the lines of the same issue. This particular device is very safe and does nothing harmful. It is supposed to monitor some physiological parameters which are non-life threatening. It does nothing more than monitor and report. There is another set of similar physiological parameters, which indicate a serious potentially life-threatening condition, but this device is not intended for or indicated for monitoring of those parameters. The company certainly does not market it for this situation, but sometimes physicians use it that way because they cannot distinguish between the two situations because of their own deficiency as a physician or for convenience, and rarely this leads to a patient death. The company insists their device is not at fault and they cannot control physicians, and a certain amount of misuse is inevitable. The frequency is very low; a high estimate is 0.04% per year. What should be done about these events? My opinion is that they should at least be reported to MAUDE, but the management feels this will seriously and adversely affect their business and argues against it as I noted above. I'm considering leaving them because of this. Am I overreacting or not?

If a manufacturer is well aware of an off-label use, and does nothing official to declare it a contraindication, it becomes part of the intended use even if the manufacturer did not intentionally endorse it. IMO they should clearly add it as a labelling contraindication; then if physicians still ignore it becomes abnormal use rather than misuse (provided that the warnings are prominent and no other practical means can prevent the unwanted use). 0.04% death rate for a device prescribed 100,000 times a year means 40 people dying this way each year! Not acceptable in my world.

 

[robert.beck]

There are no practical means to prevent this unwanted use, because physicians will make mistakes. This particular situation is clearly identified in the IFU as a contraindication, and in the approved intended use as "Not for ..." but it happens anyways and the company is ignoring it for business reasons which IMO are weak. The company is well-aware of this and other off-label uses. We are in the process of getting a new 510(k) to deal with one of the other less-critical off-label uses but not this one. So given that, what should be reported?

 

[indubioush]

1. Report the adverse events now (read this) or
2. If you are not allowed to report the events, never ever sign any justification for not reporting.
3. Ensure your risk management file appropriately reflects this "reasonably foreseeable misuse."

 

[robert.beck]

This is good advice for a difficult situation, and I thank you. How does one mitigate or justify reasonably foreseeable misuse that is not due to the device directly but is a human mistake. I don't think it can be mitigated except by strenuous education and it will still happen to some extent. Do you see a justification based on a low incidence relative to a large number of patient benefiting from the device as valid?

 

[Ronen E]

There are no practical means to prevent this unwanted use, because physicians will make mistakes.

There are almost always ways to REDUCE (not necessarily completely eliminate) mistake rates - research "error proofing". Has any usability study been done on this? If I was auditing this situation the first thing I'd ask to see is a project/activity to methodically look into the issue and implement what's possible. Once that's done, you have grounds to claim "we did everything reasonable".

the company is ignoring it for business reasons which IMO are weak.

(emphasis added)
THAT is the issue. To me it doesn't sound acceptable, and I would seriously look at employment alternatives if I couldn't change that attitude.

So given that, what should be reported?

Reporting is important, but in cases involving death reporting is IMO secondary to attempting prevention. What good is it to the dead person, their family and friends, that the "incident" is reported?...

How does one mitigate or justify reasonably foreseeable misuse that is not due to the device directly but is a human mistake.

First, is it actually a "reasonably foreseeable misuse" or an abnormal use? I already listed the criteria for abnormal use, and I'm quite sure that if you'll look into them seriously you'll already be better positioned, both in terms of mitigating the risk and in terms of having a more solid justification for you actions (or lack thereof).

Second, it's almost always DUE TO THE DEVICE DIRECTLY. In usability terms, what matters is the User Interface, and that includes the labelling too. "Human mistakes" are encouraged by lacking UIs, and can be reduced through better ones. Once you have a serious usability study of the case done and dusted, you'd not only have a better chance of improving, but also be better positioned in claiming "everything reasonable has been done to prevent the abnormal use (=going against a prominent warning/contraindication)".

I don't think it can be mitigated except by strenuous education and it will still happen to some extent.

Do you have a methodical, team-based, documented analysis on file to support that statement, or is it only your own gut feeling? And if that's really the only option, has it been tried, to any extent? In terms of managing abnormal use (you can call it abuse), it makes a difference whether you've actually tried all reasonable means even if they ended up not solving the problem.

"Still happen to some extent" might be acceptable - depending on what that extent is, what's been done to achieve it, what needs to be done to further reduce it and by how much, etc.