I need some advice from you seasoned medical device manufacturers:

We were audited by the FDA a few weeks ago. One of the trails they pursued was a medical device unit we had in for repair last year. When we went to pull the folder for the repair information (i.e. travelers, incoming inspection report, repair evaluation), the folder wasn't there. We searched during the audit and even after the audit with no luck in finding the folder. We told the auditors we couldn't find the folder and luckily they backed off on that matter. (I still don't know if the folder was misplaced or if it even existed.)

Now, a high-ranking person here is pushing to replace the documentation by having new records generated. The information is not something readily available from other sources so the names and dates being signed are best estimates. Personally, I am refusing to sign a record that has a high probabilty of being innacurate. This just to have a record on file.

My question is, what are the consequences of not being able to produce a repair record for a medical device? I am looking for some support to justify not signing the new repair record.

I need some advice from you seasoned medical device manufacturers:

We were audited by the FDA a few weeks ago. One of the trails they pursued was a medical device unit we had in for repair last year. When we went to pull the folder for the repair information (i.e. travelers, incoming inspection report, repair evaluation), the folder wasn't there. We searched during the audit and even after the audit with no luck in finding the folder. We told the auditors we couldn't find the folder and luckily they backed off on that matter. (I still don't know if the folder was misplaced or if it even existed.)

Now, a high-ranking person here is pushing to replace the documentation by having new records generated. The information is not something readily available from other sources so the names and dates being signed are best estimates. Personally, I am refusing to sign a record that has a high probabilty of being innacurate. This just to have a record on file.

My question is, what are the consequences of not being able to produce a repair record for a medical device? I am looking for some support to justify not signing the new repair record.

I am not an FDA-related requirements expert. But personally, I would be very afraid to sign re-created documents...:mg:, even with the best intentions..

Stijloor.

I am not an FDA-related requirements expert. But personally, I would be very afraid to sign re-created documents...:mg:, even with the best intentions..

Agreed....it is not a standard, it is the law. Bad things can happen.:cool:

I agree with the posts above: Signing off on 'reconstructed' records is dangerous. If you do sign off on 'reconstructed' records, make a list (Ah, another record) of each one you sign off on (and date you did so, etc.) with a disclaimer of some kind preferably signed by your boss so that if anything does come of it your :ca: is covered.

You don't mention it, but I *assume* you're designing / implementing a new system ensure there aren't problems in the future.

You don't mention it, but I *assume* you're designing / implementing a new system ensure there aren't problems in the future.

Eye, there's the rub!

I wish I knew exactly what happened. One account I've been told is that the folder is simply lost. Another has been that other information for this serial number is missing from contracts; like it got pulled out for another audit and never returned. Yet another account has it that the information/folder was never generated because it was "a quick, turn-around repair" :mg:.

The key to the solution is defining the problem.......and/or modifying the system so that it is robust enought to withstand (almost) any problem.

I am not an FDA-related requirements expert. But personally, I would be very afraid to sign re-created documents...:mg:, even with the best intentions..

Stijloor.

Agreed....it is not a standard, it is the law. Bad things can happen.:cool:

I agree with the posts above: Signing off on 'reconstructed' records is dangerous. If you do sign off on 'reconstructed' records, make a list (Ah, another record) of each one you sign off on (and date you did so, etc.) with a disclaimer of some kind preferably signed by your boss so that if anything does come of it your :ca: is covered.

You don't mention it, but I *assume* you're designing / implementing a new system ensure there aren't problems in the future.

I agree 100% with these guys. I'm no expert, but I've been in medical devices long enough to know you don't do this kind of thing...or as Bob has stated, "bad things can happen"...:agree1:

You don't mention it, but I *assume* you're designing / implementing a new system ensure there aren't problems in the future.

One word: scan!

Scan soon and scan often. Back up scans. That way, something falls off the desk into the wastebasket, there is a scan to back it up.

Cheap insurance - even with the cost of a high end copier that scans (probably the fastest scanning device you will find.) :cool:

Eye, there's the rub!

I wish I knew exactly what happened. Have you audited other repair orders/records to ensure this one was a 'fluke'? In short, have you verified the system? What have internal audits shown?

Have you audited other repair orders/records to ensure this one was a 'fluke'? In short, have you verified the system? What have internal audits shown?In addition to that, I would recall the unit in question and perform a thorough evaluation, (keeping records this time) as a way to mitigate the risks involved. I am surprised that FDA inspectors would be so "understanding" and did not write the organization up.

Since I am not sure how critical the device is, it is hard to offer additional advice.

The ramifications for missing repair files on a device are troublesome if there was an MDR/Serious Injury report associated with that device and there were no records to show that the device did not contribute to the injury. Missing repair records on a device that could have contributed to a death or serious injury could look like you are trying to hide something from the FDA.

I guess luckily there was no MDR associated with this device, or you would have gotten a write up (483) for this.

As far as "creating" records, don't do it. It constitutes falsification of records that can be used against you in a court of law. Think of anything you sign as a document that can be used in a court of law that will hold you personnally accountable and subject to jail time and fines. Take it from an ex-fda inspector, you don't want to be defending "created" records.

I assume they didn't write it up because it was only one record. Don't create a record, issue a CAPA instead.

I assume they didn't write it up because it was only one record. Don't create a record, issue a CAPA instead. I agree, but Sidney has a very good point considering the critical nature of what you have described.

In addition to that, I would recall the unit in question and perform a thorough evaluation, (keeping records this time) as a way to mitigate the risks involved. My criteria was simple: "Do I want to spend time in prison for this?" Recall the unit to mitigate risk.

I agree, but Sidney has a very good point considering the critical nature of what you have described.

My criteria was simple: "Do I want to spend time in prison for this?" Recall the unit to mitigate risk.

Well, with the FDA, you have to tread very carefully. Even recalling that one unit would require reporting it to the FDA if it was recalled for inspection, repair, etc and there may be a remote risk to health (report it publicly as a recall). I would recommend that you write a memo to file to place in that specific repair file stating something to this effect "It was discovered that repair records could not be located during the FDA audit that occurred during (time frame). The incident has been reviewed by FDA and no follow-up action has been required for this issue".

As other's have stated, I would recommend sampling the other repair files you have to determine if it was an isolated incident or systemic. If systemic, I would recommend opening a CAPA.

In addition to the verbiage for the memo to file, I would add that you investigated and have determined that there were no complaints or adverse events (injuries/death) associated with the device in question.

Total cost to recall one unit vs. Total cost of unit failure (including possible publicity and legal fees).

Many companies are reluctant to recall units and reporting to the FDA. Like mentioned above, there is the cost of negative publicity. I recommend scouring your complaint database to see if there was a complaint for this device. If so, and if it may have caused or contributed to an injury or death, definitely recall.

Otherwise, FDA is trained to be on the look out for unscheduled maintenance/repair, as opposed to scheduled repair (due to component predicted mean time to failure). Repairs due to expected component failure are okay, repairs due to unexpected component failure is not. Unexpected failure indicates that there are either design or manufacturing/supplier issues.

Is there any RMA data you can dig into to discover the root cause for the repair?

You could create a new record--dated now--that explicitly documents that you are replacing a missing record, believed to have once existed and to have been created contemporaneously with the repair of a particular unit, with a new record of people's recollections of that repair. If what everyone is attesting to is true and accurate, as opposed to fraudulent, it's signable.

As to recall...recalls are not based on missing records. They're governed by clear rules in 21CFR7, and are based on either evidence of injury or death, or a conclusion that excessive risk exists of a future significant injury, death, or malfunction that could result in significant harm. What class is the device? Is the nature of your repair process, and your staff's recollections of what was wrong with it and what they did to fix it, and any other circumstantial evidence you have to indicate the condition of the device-submitted-for-repair when it left your firm's possession after that repair visit, sufficient for you to conclude that there is no excessive risk of significant injury, death or a malfunction that could result in significant harm?

If you conclude that there's a remote risk but not enough that the device would be regarded by FDA as violative of the law, or you just want to bring the unit back to verify that it was repaired correctly, and if there's no categorical basis for a 21CFR7 recall, you could do a 21CFR806 market withdrawal of the particular unit in question. A market withdrawal is used when what's needed is "physical removal of a device from its point of use to some other location for repair, modification, adjustment, relabeling, destruction, or inspection", and the risk analysis does not call for a higher-level action.

Regarding other comments about recall in this thread: perhaps it's inappropriate to offer advice on FDA recall issues if you're not up-to-speed on 21CFR7, 21CFR803 and 21CFR806.

You could create a new record--dated now--that explicitly documents that you are replacing a missing record, believed to have once existed and to have been created contemporaneously with the repair of a particular unit, with a new record of people's recollections of that repair. If what everyone is attesting to is true and accurate, as opposed to fraudulent, it's signable.

As to recall...recalls are not based on missing records. They're governed by clear rules in 21CFR7, and are based on either evidence of injury or death, or a conclusion that excessive risk exists of a future significant injury, death, or malfunction that could result in significant harm. What class is the device? Is the nature of your repair process, and your staff's recollections of what was wrong with it and what they did to fix it, and any other circumstantial evidence you have to indicate the condition of the device-submitted-for-repair when it left your firm's possession after that repair visit, sufficient for you to conclude that there is no excessive risk of significant injury, death or a malfunction that could result in significant harm?

If you conclude that there's a remote risk but not enough that the device would be regarded by FDA as violative of the law, or you just want to bring the unit back to verify that it was repaired correctly, and if there's no categorical basis for a 21CFR7 recall, you could do a 21CFR806 market withdrawal of the particular unit in question. A market withdrawal is used when what's needed is "physical removal of a device from its point of use to some other location for repair, modification, adjustment, relabeling, destruction, or inspection", and the risk analysis does not call for a higher-level action.

Regarding other comments about recall in this thread: perhaps it's inappropriate to offer advice on FDA recall issues if you're not up-to-speed on 21CFR7, 21CFR803 and 21CFR806.

What you posted is correct, to a point. However, without knowing the risk of the device, it's difficult to determine what needs to be done in regards to corrections/removals due to missing records. This was mentioned only to rule out any adverse events associated with the dedvice. From the original post, it appears this was an isolated incident. What also needs to be taken into account is the company business model, is it sale only or rental. If rental, it's very easy to justify checking the unit when in cycles back in after the rental contract is complete. Before stating that it is inappropriate to offer advice on the FDA recall process, you should perhaps consider the business model (rental vs sale) and what the company considers routine repair/servicing. Market withdrawal is a correction/removal, but it must have sound justification and be adequately documented. I didn't say that a recall was necessary unless the device was associated with an adverse event/death or remote possibility of the like.

And once again, you can't recreate a record from people's recollection of what happened. That will not stand up in court and is unsignable. Period.

Total cost to recall one unit vs. Total cost of unit failure (including possible publicity and legal fees).

Total cost of seeing a high-ranking employee lead away in handcuffs after falsifying records - PRICELESS! (Sorry - I couldn't resist that one.)

Now back to serious business....
I checked other repair records around that time frame and all documentation was found to be in place. It looks like this is an isolated incident, which are usually the ones auditors seem to find.

We are a subcontracted manufacturer and our customer tests and burns-in all units we send in. Therefore, the possibility of an end user receiving a defective device is extremely low.

I am heeding the advice given here, pulling in what FACTUAL information I have and adding a cover letter stating the actual record of repair has been misplaced but the unit has been tested by our customer.

As far as scanning records, that is a very good suggestion. Unfortunately, I have scanned other types of records only to find that they have been destroyed on our server (yes, another problem to look into). I will look into the matter though as I'm tired of pushing paper and we are running out of room to store paper too.

Regarding other comments about recall in this thread: perhaps it's inappropriate to offer advice on FDA recall issues if you're not up-to-speed on 21CFR7, 21CFR803 and 21CFR806.I take the hint and will confess that I am ignorant on the cited regulations. But I am not advocating a recall mandated by the FDA, but a VOLUNTARY recall by the organization that serviced the equipment.

The device history is compromised at this point. Should this unit fail or malfunction in the field and the organization knowingly did not act upon discovery of missing records which could prove the device was adequately serviced and tested, they could be in legal trouble.

So, why a voluntary recall of the unit would be such a bad idea? There are obvious risks at hand.

So, why a voluntary recall of the unit would be such a bad idea? There are obvious risks at hand.No such thing. Actually all recalls (http://www.fda.gov/consumer/updates/recalls123107.html#checks) are voluntary. "Recall" requires notification to FDA

Actually all recalls (http://www.fda.gov/consumer/updates/recalls123107.html#checks) are voluntary.Al, actually the text reads Recalls are almost always voluntary.Subpart B--Mandatory Medical Device Recall Procedures (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=810.15)

I think people are too hung up on the recall word. In this case, there is no concern with the product type. But one specific device which has it's history compromised. If a manufacturer is being proactive and bringing it back for a thorough check, they should be commended for it, not penalized, in my opinion.

Al, actually the text reads Subpart B--Mandatory Medical Device Recall Procedures (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=810.15)

I think people are too hung up on the recall word. In this case, there is no concern with the product type. But one specific device which has it's history compromised. If a manufacturer is being proactive and bringing it back for a thorough check, they should be commended for it, not penalized, in my opinion.It usually doesn't get to that point.