I get audited several times a year, either internally or externally by the FDA, BSI, KEMA, etc. Recently, the FDA said that our calibration certificates need to be reviewed and signed. There is nothing in any regulation that states calibration certificates need to be reviewed 100%. I do perform a 10% monthly sample of each technician. I look for the following:
1. Following written calibration procedure for the instrument
2. Does the tech have a training record on file for the instrument he calibrated
3. Is the calibration certificate filled out correctly?
Any issues discovered are communicated directly to the technician, and I also use it as a training tool in my quarterly staff meetings with all the calibration staff. The main purpose of the record review is to assess the quality of work the technician is performing to, and also use it as a merit tool at the end of the year. The goal is to have less than 5% of paperwork related errors, and no leeway on following procedures.

So my question is: are there any docs that state this as a requirement for calibration records?

Thoughts?

Hi Charles,

I did a Google search and found the government requirement for you: 21 C.F.R. § 820.72(a) (http://a257.g.akamaitech.net/7/257/2422/10apr20061500/edocket.access.gpo.gov/cfr_2006/aprqtr/pdf/21cfr820.75.pdf)

Do a Google search of 21 C.F.R. § 820.72(a) and you might find the same interesting list of warning letters as I did. I hope it helps.

If your organization is ISO 9001:2000 or the medical device equivalent you are required by the standard under clause 7.4.3 (ISO 9001) to verify the purchased product. ISO 9000:2005 defines a product as also being a service so in this case you are verifying the calibration certifications issued by a service provider. These verifications are to assure that the services procured from the calibration service meet your contractual requirements, those imposed by the customer and of course the requirements contained in clause 7.4 and 7.6 of the standard and that you aware of and have taken action related to M&MDs found to be nonconforming inclusive of product recall when warranted.

In other words they are looking for objective evidence you have reviewed the certifications and accepted the calibration results before using the M&MD.

Hi Charles,

I did a Google search and found the government requirement for you: 21 C.F.R. § 820.72(a) (http://a257.g.akamaitech.net/7/257/2422/10apr20061500/edocket.access.gpo.gov/cfr_2006/aprqtr/pdf/21cfr820.75.pdf)

Do a Google search of 21 C.F.R. § 820.72(a) and you might find the same interesting list of warning letters as I did. I hope it helps.

Here is the text of the relevant bit about calibration records:
(2) Calibration records. The equipment identification, calibration dates, the individual performing each calibration, and the next calibration date shall be documented. These records shall be displayed on or near each piece of equipment or shall be readily available to the personnel using such equipment and to the individuals responsible for calibrating the equipment.
So there's nothing there about records having to be signed by anyone. I did the suggested Google search and encountered a few warning letters that did address deficiencies in calibration systems, but nothing about certificates not being signed.

I think the best advice is to ask the person who made the finding to provide the documentary basis for it. "Where's the shall?" in other words.

Is this about internal calibration or calibration performed by vendors?

Don't know about FDA/CFR, but ISO 17025 contains verbiage about approval of results. Perhaps you have run into something similar?

The best advice has already been given, of course: ask the originator of the finding to show you the shall.

I get audited several times a year, either internally or externally by the FDA, BSI, KEMA, etc. Recently, the FDA said that our calibration certificates need to be reviewed and signed. There is nothing in any regulation that states calibration certificates need to be reviewed 100%. I do perform a 10% monthly sample of each technician. I look for the following:
1. Following written calibration procedure for the instrument
2. Does the tech have a training record on file for the instrument he calibrated
3. Is the calibration certificate filled out correctly?
Any issues discovered are communicated directly to the technician, and I also use it as a training tool in my quarterly staff meetings with all the calibration staff. The main purpose of the record review is to assess the quality of work the technician is performing to, and also use it as a merit tool at the end of the year. The goal is to have less than 5% of paperwork related errors, and no leeway on following procedures.

So my question is: are there any docs that state this as a requirement for calibration records?

Thoughts?

Here is the text of the relevant bit about calibration records:

Quote:
(2) Calibration records. The equipment identification, calibration dates, the individual performing each calibration, and the next calibration date shall be documented. These records shall be displayed on or near each piece of equipment or shall be readily available to the personnel using such equipment and to the individuals responsible for calibrating the equipment.

On the other hand I looked at this part under 21 CFR 820.72 (b)

"When accuracy and precision limits are not met, there shall be provisions for
remedial action to reestablish the limits and to evaluate whether there was
any adverse effect on the device’s quality. These activities shall be documented."

Although I am not affected by 21 CFR and I have not seen this before, I do recognise this principle in ISO 17025 clause 5.5.2. We have had a non-conformity from our AB by not documenting that an external calibration met our stated requirements. We do this by confirming (and signing/dating) on the certificate that it complies with a requirement/limit etc. documented in a procedure.

I think our AB was trying to stop the following scenario: Certificate arrives in post; technician takes one look, seems to be Ok, puts away in drawer, forgets it and does nothing.

Interesting question. I do know that all calibration certificates (within my experience with FDA) are reviewed, and cannot go without being reviewed. However, to state the requirement for that is out of my realm.

I do think that calibration certificates (CoC) are considered raw data, and I do believe there are requirements for reviewing raw data. Also, CoC need to be reviewed for impact.

I certainly did not give you any meat for your hunger:). However, maybe there is some solace in the fact that my reality is unlike yours.

On the other hand I looked at this part under 21 CFR 820.72 (b)

"When accuracy and precision limits are not met, there shall be provisions for
remedial action to reestablish the limits and to evaluate whether there was
any adverse effect on the device’s quality. These activities shall be documented."

Although I am not affected by 21 CFR and I have not seen this before, I do recognise this principle in ISO 17025 clause 5.5.2. We have had a non-conformity from our AB by not documenting that an external calibration met our stated requirements. We do this by confirming (and signing/dating) on the certificate that it complies with a requirement/limit etc. documented in a procedure.

It's possible to document an activity without anyone signing anything.

Before going into any response to the great posts here, I'd like some points cleared up.

I read that Charles is a calibration lab supervisor. Is this lab doing calibrations for customers? The term "certificate" imples there's some kind of formal, traceable record the cal has been done--and a signed certificate is way beyond what's usually needed in internal calibration.

Are these calibration certificates the only in-house evidence validating accuracy with traceability to NIST? In such a case I suppose I'd expect a certificate too, the files of which ought to be reviewed for requirements--but not 100% of them unless their number is small.

Like Jim, I'd like to hear the "shall". Did the auditor name the requirement?

Thanks, Mr. Niceguy! Good stuff. Where did the below quote come from?



Quote:
(2) Calibration records. The equipment identification, calibration dates, the individual performing each calibration, and the next calibration date shall be documented. These records shall be displayed on or near each piece of equipment or shall be readily available to the personnel using such equipment and to the individuals responsible for calibrating the equipment.

"When accuracy and precision limits are not met, there shall be provisions for remedial action to reestablish the limits and to evaluate whether there was any adverse effect on the device’s quality. These activities shall be documented."

Although I am not affected by 21 CFR and I have not seen this before, I do recognise this principle in ISO 17025 clause 5.5.2. We have had a non-conformity from our AB by not documenting that an external calibration met our stated requirements. We do this by confirming (and signing/dating) on the certificate that it complies with a requirement/limit etc. documented in a procedure.I find the operative term is "When accuracy and precision limits are not met"... That tells me the clause is addressing the documentation involving the response to not meeting limits and not when they are met.

I read that Charles is a calibration lab supervisor. Is this lab doing calibrations for customers? The term "certificate" imples there's some kind of formal, traceable record the cal has been done--and a signed certificate is way beyond what's usually needed in internal calibration.

Are these calibration certificates the only in-house evidence validating accuracy with traceability to NIST? In such a case I suppose I'd expect a certificate too, the files of which ought to be reviewed for requirements--but not 100% of them unless their number is small.

Like Jim, I'd like to hear the "shall". Did the auditor name the requirement?

Good point, Jennifer. This is a great question (IMHO) and am glad we're discussing. Exactly who is being asked to review and sign the documents is a good question.

Charles, are you also under 17025? This is a dumb question... but I felt behooved to bring it up: What do your internal procedures say? Is there a higher level organizational procedure that requires this?

It's possible to document an activity without anyone signing anything.

I totally agree. It was the interpretation of our AB that somehow the external certificate must show evidence of our review and acceptance.

I find the operative term is "When accuracy and precision limits are not met"... That tells me the clause is addressing the documentation involving the response to not meeting limits and not when they are met.

I thought about that as well when I posted before, but our AB would get you every time if this was ISO 17025. If you do nothing where is the evidence that it was OK ? Maybe if you documented elswhere - not on the certificate -that no action was necessary, providing it was traceable to the certificate, that would be OK.

It is a constant irritation for us in a testing lab because it also applies to instrument service records given by manufacturers' service engineers.

Let's think about this for a minute. 820.72 b states "Calibration procedures shall include specific directions and limits for accuracy and precision. When accuracy and precision limits are not met, there shall be provisions for remedial action to reestablish the limits and to evaluate whether there was an adverse effect on the device's quality. These activities shall be documented"

How would you know the equipment was out of calibration unless you reviewed the certificates you receive back from your cal service? Do they automatically notify you if a piece of equipment if found OOT or do they just bring it back within tolerance?

Thanks, Mr. Niceguy! Good stuff. Where did the below quote come from?

It was Jim Wynne's quote from CFR 21.

Sorry about that - I was not using the proper Quote button :o

Let's think about this for a minute. 820.72 b states "Calibration procedures shall include specific directions and limits for accuracy and precision. When accuracy and precision limits are not met, there shall be provisions for remedial action to reestablish the limits and to evaluate whether there was an adverse effect on the device's quality. These activities shall be documented"

How would you know the equipment was out of calibration unless you reviewed the certificates you receive back from your cal service? Do they automatically notify you if a piece of equipment if found OOT or do they just bring it back within tolerance?

I have been asking around about this, and this is the general response I am getting too. Signing the CoC is assurance that review has been made to impact, following of procedures, acceptance of findings, etc.

Thanks for all the replies. We are an internal calibration lab for the company; therefore, we perform internal calibrations only. We record our results on a "calibration certificate" showing the before and after data (if an adjustment was made). The calibration certifcate is then signed and stamped by the calibration technician and entered into our database for scheduling the next cal date. The calibration certificate is then filed.

The FDA auditor was basing his remarks that the calibrations were for "equiment that impacts drug processing".

We are not under 17025.

"When accuracy and precision limits are not met, there shall be provisions for remedial action to reestablish the limits and to evaluate whether there was any adverse effect on the device’s quality. These activities shall be documented."
They are refering to activites when you find out of tolerance conditions with the equipment. We have always done this and it's documented.

To be unusually blunt, it's not very difficult to review a cert of cal and sign the bottom to show approval and document the review. We do this for all calibrations and it works just fine for us. Would it save us much time or money to not sign these, not really...would it be unpleasant if we found out that a majority of our certs were filed with errors in the document, you betcha. To me, reviewing the cert of cal is about the same as rereading an email to check for errors before sending it out, just good practice.

Well to be honest, I would not want to sign one unless I knew the information recorded was correct. This means several things:
1. Was the calibration procedure followed (i.e. using the correct test points)
2. Did the tech accept the calibration, but it was actually out of tolerance (math error)?
3. Does the accuracy statement on the calibration certifcate match the statement in the calibration procedure?
4. Is there a training record on file, and did they use the current rev of the calibration procedure?
5. Typo's (e.g. forgot to record the standards used, forgot to sign, etc)

This is a very intensive review if you wanted to do one to that level.

How would you know the equipment was out of calibration unless you reviewed the certificates you receive back from your cal service? Do they automatically notify you if a piece of equipment if found OOT or do they just bring it back within tolerance?

Good point; however. requiring a signature, initials or other "objective evidence" of the review seems to be going a little too far in the direction of making sure all the i's are crossed and the t's are dotted.

How about another set of initials to confirm that the first party actually reviewed the documents, as opposed to just marking them?:tg:

To be unusually blunt, it's not very difficult to review a cert of cal and sign the bottom to show approval and document the review. We do this for all calibrations and it works just fine for us. Would it save us much time or money to not sign these, not really...would it be unpleasant if we found out that a majority of our certs were filed with errors in the document, you betcha. To me, reviewing the cert of cal is about the same as rereading an email to check for errors before sending it out, just good practice.

All a signature proves is that somebody signed a piece of paper.

As I said before, I don't do FDA. That said, I would advise my ISO (9001 and 17025) clients to push back if an auditor asked for signatures to prove "review" of CoC's.

If it was an internal calibration, presumably the tech who did the work signed off on it.

In any system where I have influence, the tech is responsible for initiating the reaction if an instrument is out of cal.

If another party signs off, who then is responsible for the work?

All a signature proves is that somebody signed a piece of paper.


That statement sounds rather like something Bill Clinton would say.

A signature by an offficer of the company is a notation accepting responsibility for the task of the review, at which point the signing member can be held legally responsible for his/her actions, fraud, liability, etc. To say it doesn't prove anything is a bit naive.

Either way, it is our practice to have the technician review their work, sign to accept that the information is correct etc, it is then reviewed by their superior (signed by them) and filed. This is not an overly time consuming task and moreover it gives us certainty that our instrumentation and its associated documentation is in order, which for a laboratory is quite critical.

This is simply the way we do our calibration cert reviews here, it works well for us and isn't a difficult task.

Good point; however. requiring a signature, initials or other "objective evidence" of the review seems to be going a little too far in the direction of making sure all the i's are crossed and the t's are dotted.

How about another set of initials to confirm that the first party actually reviewed the documents, as opposed to just marking them?:tg:

Industry Standard is that the tech initials/signs and dates they did the work and it is accurate and quality (or another independent function) reviews the data and signs if any OOT conditions are found and the unit if brought back within tolerance an investigation is performed (documented) to determine if any adverse affects could have been caused. it also allows us to see if the interval needs to be changed due to drifting or stability.

Well to be honest, I would not want to sign one unless I knew the information recorded was correct. This means several things:
1. Was the calibration procedure followed (i.e. using the correct test points)
2. Did the tech accept the calibration, but it was actually out of tolerance (math error)?
3. Does the accuracy statement on the calibration certifcate match the statement in the calibration procedure?
4. Is there a training record on file, and did they use the current rev of the calibration procedure?
5. Typo's (e.g. forgot to record the standards used, forgot to sign, etc)

This is a very intensive review if you wanted to do one to that level.

That’s funny, is this not something you should do to assure the certificaiton does not contain these errors.

A signature by an offficer of the company
Surely you are not claiming that officers of corporations sign cal certs, are you? An officer of a company generally has a job title containing words such as "president," "vice president," "general counsel," "treasurer," "comptroller" or, helpfully, "officer." Now perhaps someone with a title like that has at some time signed a certificate of calibration, but I sort of doubt it. It is also remotely possible that those who routinely sign CoC's have been named as officers of their corporations, but I doubt that, too.

is a notation accepting responsibility for the task of the review,First of all, the signature proves nothing about the depth of the review or about the qualifications of the signer or about the quality of the signed-for work. At worst, it proves only that some clueless pointy-haired boss was hornswoggled into signing off.

Second, the meaning of a signature is open to debate unless the signed document clearly indicates what the signature means. Words like, "by signing here I affirm that I have assiduously reviewed the work in question and find it to be adequate. I hereby relieve the technician of all responsibility," would help to tie down the meaning. But how often do those kinds of words appear on QMS documents?

at which point the signing member can be held legally responsible for his/her actions, fraud, liability, etc. To say it doesn't prove anything is a bit naive.I'm not a lawyer and neither are you. But I think you have this bit wrong. Corporations are persons, and they generally bear the responsibility for the actions of their employees as they relate to the business of the corporation. That's why, for example, my class-action suit payout from Toshiba came from the person known as Toshiba, not from Toshiba and the chief of Toshiba engineering. This, even though Toshiba's chief engineer was certainly head of the function, and therefore responsible for the problem that caused the lawsuit.

Either way, it is our practice to have the technician review their work, sign to accept that the information is correct etc, it is then reviewed by their superior (signed by them) and filed. This is not an overly time consuming task and moreover it gives us certainty that our instrumentation and its associated documentation is in order, which for a laboratory is quite critical.That's YOUR practice. If it helps you sleep better at night, great. Personally I would rather rely on the knowledge, integrity and accountability of the technician doing the cal.

And what if there's a foul up? If there's only one signature on the cert, you only have one problem: how did it happen? If there are two signatures, you have two problems: how did it happen; and, how did the reviewer miss it.

It all comes down to the old question about who watches the watchers. How many layers of lookers-over-the-shoulder do you need? Are you really sure just one is enough?

Now I understand that ISO 17025 requires review of test and calibration results. That's why, in my 17025 work, I make darn sure that the system documentation indicates that the technicians doing the work are competent to review their results and have the responsibility and authority to do so. Then I work with the client to make darn sure that the technicians are, in fact, so competent.

Great comments today on this one. It seems there has become a moderately passionate disagreement on this one. Passion is excellent, as long as we keep it towards the topic, and not each other!:D

Here is what I see so far (please feel free to edit):

1. The original question is do calibration certificates (CoC) need to be reviewed?
2. And if "yes", what are they being reviewed for?
3. Is there a regulatory requirement for reviewing CoC?
4. Do CoC need to be signed?

In my years of being involved with calibration, there have always been reviews of the CoC. Either a supervisor of a cal lab (internal calibration) or a customer. I think that is a good approach to keep for a few reasons.

#1, you need to review what you paid for. Certificates are meaningless if you cannot extract some value from it. You may need to determine if you calibration interval is too long, too short, etc.

#2, people who generate calibration certificates are not perfect. There will be math mistakes, typos, errors, training issues, etc. Depending on your situation and regulatory exposure, that can range from no big deal to big deal.

#3 Calibration information is observational data; an experiment of sorts. All observational/ experimental data I've seen has always been reviewed/ signed by someone. Your scenario may be different.

If I understand the OP as a supervisor of a lab doing FDA work, I cannot imagine being OK without reviewing the documentation. There are just too many things that need to be checked on there. If you have other processes that eliminates such errors, then I guess you would not have to review. But I would rather catch an error than my internal customer, which is not as bad as the internal auditor, that saves us from the external auditor, who hopefully has assured a proper process that the FDA auditor does not find it! :)

Now, leaping into having to sign/initial the CoC is another matter. I might think it's good to make some mark on the CoC, equipment log, database entry, something to show that "yes, someone acknowledges the review, and that it is legit".

Great comments today on this one. It seems there has become a moderately passionate disagreement on this one. Passion is excellent, as long as we keep it towards the topic, and not each other!:D

Here is what I see so far (please feel free to edit):

1. The original question is do calibration certificates (CoC) need to be reviewed?
2. And if "yes", what are they being reviewed for?
3. Is there a regulatory requirement for reviewing CoC?
4. Do CoC need to be signed?

In my years of being involved with calibration, there have always been reviews of the CoC. Either a supervisor of a cal lab (internal calibration) or a customer. I think that is a good approach to keep for a few reasons.

#1, you need to review what you paid for. Certificates are meaningless if you cannot extract some value from it. You may need to determine if you calibration interval is too long, too short, etc.

#2, people who generate calibration certificates are not perfect. There will be math mistakes, typos, errors, training issues, etc. Depending on your situation and regulatory exposure, that can range from no big deal to big deal.

#3 Calibration information is observational data; an experiment of sorts. All observational/ experimental data I've seen has always been reviewed/ signed by someone. Your scenario may be different.

If I understand the OP as a supervisor of a lab doing FDA work, I cannot imagine being OK without reviewing the documentation. There are just too many things that need to be checked on there. If you have other processes that eliminates such errors, then I guess you would not have to review. But I would rather catch an error than my internal customer, which is not as bad as the internal auditor, that saves us from the external auditor, who hopefully has assured a proper process that the FDA auditor does not find it! :)

Now, leaping into having to sign/initial the CoC is another matter. I might think it's good to make some mark on the CoC, equipment log, database entry, something to show that "yes, someone acknowledges the review, and that it is legit".

The point of signing and dating a docuement, in the FDA's view, if a record is not signed and dated, then it did not happen. Several week ago I attend a AMMI training seminar for 21 CFR PArt 820 Quality System Regulations. The message was very clear by the FDA that records needed to be signed and dated.

The message was very clear by the FDA that records needed to be signed and dated.

Did they tell you where the written requirements say that? Have we established that there's a documented requirement for signatures on calibration certificates? I know this is the government we're dealing with here, and I have no experience in this area, but it seems to me that unless there's an explicit "shall," all bets are off.

Speaking of government and signatures, I know someone who, while in the service, had a desk job for a time that required him to review documents and then initial them. One day he got one back from some higher-up with a note attached to it that said, "You were not authorized to review this document. Erase your initials, and initial your erasure." :frust:

Jim,
This is cited in 21 CFR Part 820, section 820.40 Document Controls.
• 820.40(a) document approval and distribution states: The approval, including date and signature of individual(s) approving the document.
• 820.40(b) document changes states: the signature of the approving individual(s), the approval date, and when the change became effective

Phil

Jim,
This is cited in 21 CFR Part 820, section 820.40 Document Controls.
• 820.40(a) document approval and distribution states: The approval, including date and signature of individual(s) approving the document.
• 820.40(b) document changes states: the signature of the approving individual(s), the approval date, and when the change became effective

Phil

I'm seeing this more clearly now. I looked at 21 CFR Part 820 (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=820), and perhaps I can tie things together for clarity here. First, Subpart D (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=820&showFR=1&subpartNode=21:8.0.1.1.12.4) (Document Control) says,
Each manufacturer shall establish and maintain procedures to control all documents that are required by this part...Each manufacturer shall designate an individual(s) to review for adequacy and approve prior to issuance all documents established to meet the requirements of this part. The approval, including the date and signature of the individual(s) approving the document, shall be documented. Note the explicit requirement for all documents required by this part to be dated and signed.
Moving on to Subpart G, §820.72 (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=820.72), we find,
Each manufacturer shall establish and maintain procedures to ensure that equipment is routinely calibrated, inspected, checked, and maintained. The procedures shall include provisions for handling, preservation, and storage of equipment, so that its accuracy and fitness for use are maintained. These activities shall be documented.Thus there's a clear requirement for calibration activities to be documented, and another clear requirement for all documents to be signed and dated. That seems to answer the question, no?

Surely you are not claiming that officers of corporations sign cal certs, are you? An officer of a company generally has a job title containing words such as "president," "vice president," "general counsel," "treasurer," "comptroller" or, helpfully, "officer." Now perhaps someone with a title like that has at some time signed a certificate of calibration, but I sort of doubt it. It is also remotely possible that those who routinely sign CoC's have been named as officers of their corporations, but I doubt that, too.

Yes, actually our company is a small company, the technical manager is an officer of the company and is responsible for review of calibration as one of his tasks, you make a lot of assumptions in this post.

First of all, the signature proves nothing about the depth of the review or about the qualifications of the signer or about the quality of the signed-for work. At worst, it proves only that some clueless pointy-haired boss was hornswoggled into signing off.

Again, this is an assumption on your part, the review is outlined in a quality document and is a specific task, we deal in data and therefore we take great care in ensuring our documentation is pristine. Nor is our boss pointy haired.

Second, the meaning of a signature is open to debate unless the signed document clearly indicates what the signature means. Words like, "by signing here I affirm that I have assiduously reviewed the work in question and find it to be adequate. I hereby relieve the technician of all responsibility," would help to tie down the meaning. But how often do those kinds of words appear on QMS documents?

If you're trying to be a lawyer here you're doing a good job, in the realm of law everything is "open to debate". Again I'd refer to our review procedure for clarification to narrow down the "debate".

I'm not a lawyer and neither are you. But I think you have this bit wrong. Corporations are persons, and they generally bear the responsibility for the actions of their employees as they relate to the business of the corporation. That's why, for example, my class-action suit payout from Toshiba came from the person known as Toshiba, not from Toshiba and the chief of Toshiba engineering. This, even though Toshiba's chief engineer was certainly head of the function, and therefore responsible for the problem that caused the lawsuit.

In a class action law suit against a corporation sure. I'm talking personal liability here. For example we had a tech once who slept through his shift while pouring 25,000 gallons of JetA fuel onto the ground, he was arrested, his boss had a lien put against his house, and the corporation had to deal with the clean-up costs/emergency worker costs. By your reasoning only the corporation would be liable? So if I always have a boss I never will be held responsible?

That's YOUR practice. If it helps you sleep better at night, great. Personally I would rather rely on the knowledge, integrity and accountability of the technician doing the cal.

Isn't that what I said in my post...that this is our practice, it works for us, it's easy and, since we don't have a pointy haired boss, the review is actually a useful process which doesn't require a huge commitment of time and work. So why try to contend that? I wasn't saying you personally have to do things this way, only that we do so and find it useful and easy!

And what if there's a foul up? If there's only one signature on the cert, you only have one problem: how did it happen? If there are two signatures, you have two problems: how did it happen; and, how did the reviewer miss it.

And what if there's a foul up? If the technician misses something, or makes a calculation error and the reviewer catches it and makes the correction prior to putting the instrument back into production use? Then you have prevented numerous measurement errors...There are a lot of what ifs in this world, I'd rather have better odds of having the data correct in the first place than worry about asking why two people missed something.


There are going to be varying viewpoints on this issue, I'm simply offering mine and you can take it or leave it. We utilize complex instruments and the calibration of those instruments is of utmost importance, because the calibration is not an automated process and is vulnerable to human error we find the review process to be beneficial. And yes, our technicians are extremely competent and we are confident in their abilities, but human error is a fact of life, reviewing their work has proven useful. If you're just calibrating micrometers by all means a review with a signature isn't going to be very useful, but if you're programming variables into a complex instrument based on calibration calculations I'd rather have a review and overcheck so I can ensure that instrument is indeed working correctly.

Thus there's a clear requirement for calibration activities to be documented, and another clear requirement for all documents to be signed and dated. That seems to answer the question, no?

At least for the FDA requirement, in my book, Yes. Thanks to Phil for the citations.

There have been some excellent posts on this one (both sides). I thank those for sharing their thoughts, as it makes for a great reference for others.

As stated earlier, in my world of "Brad is king for a day", all CoC should be signed/initialed by a responsible person assuring completion, accuracy, etc. Not because you're "told" to, but because it makes good sense.

Thanks to Phil for the citations.
Ditto.

As stated earlier, in my world of "Brad is king for a day", all CoC should be signed/initialed by a responsible person assuring completion, accuracy, etc. Not because you're "told" to, but because it makes good sense.
Define "responsible person." Does this assume that a person who drafts a calibration certificate isn't responsible? Aren't we supposed to be moving away from the idea that everyone's work should be inspected? I can understand (if not completely sympathize with) a requirement for checks and balances in things like medical devices, pharmaceuticals, and nuclear explosives, but in most instances, I think we need to trust trained and skilled people to do things correctly.

Ditto.


Define "responsible person." Does this assume that a person who drafts a calibration certificate isn't responsible? Aren't we supposed to be moving away from the idea that everyone's work should be inspected? I can understand (if not completely sympathize with) a requirement for checks and balances in things like medical devices, pharmaceuticals, and nuclear explosives, but in most instances, I think we need to trust trained and skilled people to do things correctly.

Thank you Jim for the clear explanation. This is exactly the point I was trying to make earlier.

Ditto.

Define "responsible person." Does this assume that a person who drafts a calibration certificate isn't responsible? Aren't we supposed to be moving away from the idea that everyone's work should be inspected? I can understand (if not completely sympathize with) a requirement for checks and balances in things like medical devices, pharmaceuticals, and nuclear explosives, but in most instances, I think we need to trust trained and skilled people to do things correctly.

Ok, good point. By stating responsible person, that does not apply that everyone else is irresponsible. It means operationally defining an individual who knows the appropriate requirements, appropriate standards to use, etc. Depending on the scope of the calibration activities, levels of requirements, etc., that may/ may not be within the realm of the person performing the calibrations.

To me, this has nothing to do with trust. If I did not trust people doing the work, I would find someone else. As I stated earlier, if you have established a process that ensures that all those potential errors cannot be made, then "no", I guess you don't have to inspect anything.

As I stated earlier, I don't think there is any slam-dunk answer for this one; just different people's experiences and realities.

I am going to admit that I read about the first three pages then started skipping the longer replies do to time constraints. (Lean/kaizan) :cool:

I handle it this way and it seems not to bring up any problems:

Internal calibrations: Done in Gage Trak and have no signatures. The person who does the calibration is associated with it thru the software.

External calibrations: The certificates are inspected to make sure the item was not found to be out of calibration-

1. If found good then "OK" and initials are placed on top of cert. along with gage ID (because its easier to see that way when looking thru the binders)
2. Once Activated then "Activated On xx/xx/xxxx" is also wrote on cert.

Example: Activated On 11/30/2007
01-0001
OK DB

3. If found out of tolerance then we make sure that it did not affect any of production and the Quality Manager writes and signs a letter that states that product was not affected. (Just a note-this is e-mailed and I save in a folder just in case)

4. If found out of tolerance and it does affect the product well lets just say FUBAR :mg:

That is how I handle it and I may not have to but it is easy to do and I am never questioned.

Thanks
Dale

Thanks everyone for all the comments. I never expected this many comments on this subject. I only wish that the FDA would update the regs to state this requirement, as I still feel that it's not clearly stated as a requirement and left to interpretation. Also, you have to ask if this is really value added function. Having a 2nd reviewer does not assure 100% compliance, as I'm sure most reviewers make sure that the data recorded is complete not necessarily accurate without mistakes.

The approach that I'm proposing will require the reviewer to look for a calibration done to the current revision, and they used the correct test points, along with looking for typo's. I figure this will require ~100 man hours of additional work (value added?). I guess it's "better to lose the battle in order to win the war."

Charles,
Have you challenged your FDA auditor to state where in the regs it requires such a review? We do in house calibration for mics and calipers only, we do not require that a review be performed (I do one every 4 to 6 month anyway).

I would not add extra work if the auditor can not define the regulation. If the auditor can not give you a good answer, you can always go up the up to the next level, the FDA Center that you use and ask for clarification.

Did you post the exact finding/observation the FDA cited you for?

Phil

In our first surveillance audit I also received this same feedback. Verbatim the 'deviation' (or finding) states: 'Objective evidence to show review and approval of Calibration Certificates not available for all certificates sampled and test data from service provider not always present'. We have our equipment calibrated by an outside source, and the data is available per request. Initial data comes with the equipment, but subsequent calibration data is a pass/fail. Per our policy the certificates of calibration are checked into our electronic calibration system, therefore reviewed.

During the audit I challenged this, and the auditor said that they were going to check on it with some collegues. I was just curious if I had a leg to stand on or do I just accept it? BTW I am talking ISO 13485:2003, and the supplier of the calibration service has been approved. My argument was: if we approve the supplier for use, why would question that they haven't done their job? If we were questioning whether or not their calibration abilities are satisfactory, why would we have them as our approved supplier?

Thank-you in advance for any feedback!

In our first surveillance audit I also received this same feedback. Verbatim the 'deviation' (or finding) states: 'Objective evidence to show review and approval of Calibration Certificates not available for all certificates sampled and test data from service provider not always present'. We have our equipment calibrated by an outside source, and the data is available per request. Initial data comes with the equipment, but subsequent calibration data is a pass/fail. Per our policy the certificates of calibration are checked into our electronic calibration system, therefore reviewed.

During the audit I challenged this, and the auditor said that they were going to check on it with some collegues. I was just curious if I had a leg to stand on or do I just accept it? BTW I am talking ISO 13485:2003, and the supplier of the calibration service has been approved. My argument was: if we approve the supplier for use, why would question that they haven't done their job? If we were questioning whether or not their calibration abilities are satisfactory, why would we have them as our approved supplier?

Thank-you in advance for any feedback!

How timely the same subject pops up in such a short window.

After this thread came out, I began to ask around to people in my circle of influence (primarily related to FDA). Bottom line: You review the certificates, and sign them to show you reviewed them. Now, if your electronic system has electronic signature that shows you reviewed it, that should be acceptable.

Reviewing the calibration certificate is not about questioning who did the work. It's looking to assure no typos, was your requirements satisfied, is it OK, as-is, to go back into use, etc. The way I look at this is- I would rather find the error than the internal customer, QMS auditor, and lastly (and the most damaging) an FDA auditor.

FOLLOW UP TO THIS THREAD

Since we had to implement this review, now the FDA would like the supervision to review the reviewer (when will this ever end). According to the FDA, "calibration supervision shall additionally conduct a documented sampling review of calibration records per an accepted sampling plan monthly".

So we have now added a statement in our procedures discussing this sample. We will be using ANSI-ASQC Z1.4 General Inspection Level II, AQL 0.65 for critical items, and AQL 2.5 for all others.

The cirtical items will be training requirements, compliance to calibration procedures, and verification of data and results. Non-critical will be documentation practices.

I just wanted to bring this sore subject up again, as it seems the FDA is really pushing this type of review for those companies in the medical side, especially in the pharma side.

What will it be next???

Thank you for the follow-up, Charles. I suspected this.