12 replies to this topic

[astro]

Hi

 

Currently we do our metal detection hourly but sometimes this is not practical as production might be over by a few minutes from the hour. If I put a 10 mins window tolerance over the hour would BRC still accept this?

 

Thanks

[Slab]

IMEX a tolerance or variance in verification activities and critical limits isn't permissible, though I'm not 100% on BRC code.

 

Just so I understand your question, you mean that a run of product is only 45 minutes, and you return to the MD with no product to challenge?

 

It would depend on the exact verbiage of your policy or HACCP plan, however there should be a final challenge for that product by the operator or QC just to make sure thresholds and reject mechanisms haven't gone wonky since last checked.

 

 

[olenazh]

I used to work for the bakery with BRC (it was 2013-2014), and I remember we did metal detection every hour with a tolerance +/- 10 min - this wasn't objected by the auditor. However, BRC requirements might've changed since then.

[pHruit]

BRC is its usual vague self on this. The standard itself doesn't specify a tolerance, merely that "the site shall establish.. the methods and frequency".

The IG suggests that "a minimum of hourly testing is typically expected", but in the context that it relates to the site's ability to recover and retest product if a failure occurs. A 1hr +/- 10min tolerance is therefore really saying that 70 mins is reasonable for your site to achieve this, so perhaps the question is: can you demonstrate this to be the case?

[Ehab Nassar]

Hi Astro ,

 

The BRC Interpretation guide stated (4.10.3.3):

The frequency of testing must be based on the quantity and type(s) of product. However, the following should be considered:

• start-up and finish of shifts

• product changeovers

• change in machine settings following downtime for repairs

• customer requirements

• regular checks throughout production (these should consider the site’s ability to recover and retest product in the event of a failure; therefore a minimum of hourly testing is typically expected).

 

Accordingly you can set your frequency ( for example hourly or whatever is appropriate for you machine and lines)  as well as the above mentioned  , so  if the production  finished you can document the time and indicate it as end of production simply.

 

BR,

Ehab

[Charles.C]

Hi

 

Currently we do our metal detection hourly but sometimes this is not practical as production might be over by a few minutes from the hour. If I put a 10 mins window tolerance over the hour would BRC still accept this?

 

Thanks

In addition to auditorial necessities, there is a (painfully) simple financial/logistical  criterion -

 

How much time and money are you prepared to "suffer" in the event of a re-inspection ?

[Ryan M.]

You put in your SOP and policy the required frequency.  In practice you do it more often so that if you miss one it doesn't go against your policy or SOP.  This is where a lot of people screw up their food safety and quality programs.  In their SOP's and Policy they put in the frequency they actual do things (in theory), and fail to do them that often.  I've always hedged it where the frequency is less in the SOP and Policy than it is in operational practice.  This way if something is missed it doesn't affect our food safety or quality programs.

[AmeliaJacobs]

We currently hold a BRC Grade A (we had an AA but quality culture got a lot of us...) and our metal detector frequency has a +/- 15 minutes every hour. For the same reason. People are human and not always available at that exact moment. Would we want to risk safety to run down the hall for that exact moment? No.

Essentially, setting a MD test frequency is a risk tolerance... how much product are you willing to re-open/re-run if there's an issue? Once you decide that issue (in a way that you will happily stick to it going forward), you're fine. You can say, we'll test the MD once a day. That's fine and I've known some places that do that. Just know that if anything goes wrong, that's a whole day's production. 

The auditor is totally fine with our +/- 15 minutes (5 years running) as long as it's documented AS SUCH in our HACCP and our procedures, verified and adhered to. 

 

[Charles.C]

We currently hold a BRC Grade A (we had an AA but quality culture got a lot of us...) and our metal detector frequency has a +/- 15 minutes every hour. For the same reason. People are human and not always available at that exact moment. Would we want to risk safety to run down the hall for that exact moment? No.

Essentially, setting a MD test frequency is a risk tolerance... how much product are you willing to re-open/re-run if there's an issue? Once you decide that issue (in a way that you will happily stick to it going forward), you're fine. You can say, we'll test the MD once a day. That's fine and I've known some places that do that. Just know that if anything goes wrong, that's a whole day's production. 

The auditor is totally fine with our +/- 15 minutes (5 years running) as long as it's documented AS SUCH in our HACCP and our procedures, verified and adhered to. 

 

 

Hi Amelia,

 

Frankly I would have said that 1/day is not "fine" and "+/-15 min" indicates a shortage of QA personnel. Regardless I agree that it's Ok to compromise.

 

When auditing (not GFSI), as long as no significant quantities of rejects, etc,  I have always (mentally) "disregarded" a wad of tables showing everything checked right on the hour.

[Ryan M.]

Well...to Amelia's point it is up to the tolerance for risk in case of metal is introduced into the products.  At a previous facility it was once daily and never ran into an issue during our SQF audit or our multitude of customer audits.  Some customers tried to argue the point they wanted it more frequently, but we didn't bend on it.  The way our process was setup it made it very challenging to conduct a verification on the metal detectors and cause significant product loss.  Our CEO wanted to pull them out altogether, but we agreed on a daily check.

 

Why does a "+ 15 min" indicate a shortage of QA personnel?  Every plant I've been in doesn't have people always standing at the ready to do checks or rounds.  I suppose it is a different QA setup, but typically don't have a lot of free time to just do checks.  As Amelia has said, things come up.  I'm just trying to understand why this would even be an issue?  So long as the policy /SOP reflects it, risk assessment completed, and the checks are within that time frame.  Is there a reasonable alternative?

 

 

Hi Amelia,

 

Frankly I would have said that 1/day is not "fine" and "+/-15 min" indicates a shortage of QA personnel. Regardless I agree that it's Ok to compromise.

 

When auditing (not GFSI), as long as no significant quantities of rejects, etc,  I have always (mentally) "disregarded" a wad of tables showing everything checked right on the hour.

[Charles.C]

Well...to Amelia's point it is up to the tolerance for risk in case of metal is introduced into the products.  At a previous facility it was once daily and never ran into an issue during our SQF audit or our multitude of customer audits.  Some customers tried to argue the point they wanted it more frequently, but we didn't bend on it.  The way our process was setup it made it very challenging to conduct a verification on the metal detectors and cause significant product loss.  Our CEO wanted to pull them out altogether, but we agreed on a daily check.

 

Why does a "+ 15 min" indicate a shortage of QA personnel?  Every plant I've been in doesn't have people always standing at the ready to do checks or rounds.  I suppose it is a different QA setup, but typically don't have a lot of free time to just do checks.  As Amelia has said, things come up.  I'm just trying to understand why this would even be an issue?  So long as the policy /SOP reflects it, risk assessment completed, and the checks are within that time frame.  Is there a reasonable alternative?

Hi Ryan,

 

As you say, it logically relates to (a) risk assessment + (b) consequences of a failure.

 

TBH, I have never encountered a situation where a Company has evaluated (a,b) and concluded that a verification frequency of 1/day was appropriate..

 

IMEX QA functions involve a team, not just a 1-man/woman show.  I am surprised that, presumably based on experience, a +/-10 min tolerance  might be found necessary.  20 minutes = 33% of the stated Interval !.

 

 You are correct in that a ("likelihood of failure") review would probably justify an interval  >> 1 hr  but (b) has always been given priority. It's a Judgement call.

 

@ Amelia - I think you have been lucky with auditors .  IMEX an auditor would first query whether a  stated 1-15 minute tolerance band on the testing time included the start/end of shifts and machine failures. ++.

 

PS - I concur that people have less time than they used to for "Walking the Floor". And we all know where to lay the blame for that !!

[Ryan M.]

Thanks for the insight and that makes sense.  Any process / facility needs to do a thorough risk assessment and include upper management to determine verification frequency and ensure the resources are there for it.  Upper management does not always understand the resources needed to conduct an activity at the prescribed frequency and this definitely needs to be part of the discussion with risk assessment and frequency.

 

Hi Ryan,

 

As you say, it logically relates to (a) risk assessment + (b) consequences of a failure.

 

TBH, I have never encountered a situation where a Company has evaluated (a,b) and concluded that a verification frequency of 1/day was appropriate..

 

IMEX QA functions involve a team, not just a 1-man/woman show.  I am surprised that, presumably based on experience, a +/-10 min tolerance  might be found necessary.  20 minutes = 33% of the stated Interval !.

 

 You are correct in that a ("likelihood of failure") review would probably justify an interval  >> 1 hr  but (b) has always been given priority. It's a Judgement call.

 

@ Amelia - I think you have been lucky with auditors .  IMEX an auditor would first query whether a  stated 1-15 minute tolerance band on the testing time included the start/end of shifts and machine failures. ++.

 

PS - I concur that people have less time than they used to for "Walking the Floor". And we all know where to lay the blame for that !!

[astro]

Thanks all very helpful