7 replies to this topic

[porky]

Hello All...

First I have to say that I consider this foruns very useful for people who works in the agro-food sector.

My main question is:
While conducting a hazard analysis, do you consider, on the risk assessment step (to classify the probability...) that the existing control measures in place can fail?

I already read a few things on HACCP but I never get a clear anwer.

[yorkshire]

Dear Porky,

We always do a risk assessment on potential hazards.

It is very basic and we score severity and risk between 1 and 3.

The first question is "What would the severity of illness / injury be if this hazard was not controlled?"

Score 1 for mild, 2 moderate, 3 severe.

The second question is " What is the likelihood of this hazard occuring?"

Score 1 unlikely, 2 may happen, 3 will happen.

You then multiply both numbers together.

At our site we filter out hazards that score a total below 3 which means we only consider "significant" hazards.

This can mean that real hazards get ignored, but as HACCP is a system that is always developing all real hazards should get picked up during your reviews.

[porky]

The second question is " What is the likelihood of this hazard occuring?"
----------------------------------------------------------------------------------

Is in this type of question that I have my issues

For the likelihood do you consider that control measures in place (pre-requisite program) can fail or not?

Can I make my self understandable or not

Thanx

[Gaskit]

Dear Porky,

I am in the same industry as you, please consider the under mentioned as an example of considering a pre-requisite programme has failed(ish).

Process step: Storage
Hazards: Microbiological / Physical hazards from pests
Control Measure: Pest control contract
Critical Limits: No pest infestation
Monitoring procedure: As defined in pest control contract
Corrective action: As recommended by pest control contractor
Responsibility: SG

Hope it gives you a point in the right direction.

Regards,

Steve

[yorkshire]

Dear Porky,

You need to decide if you want to consider the risk with or without the pre-requisites. If you don't consider the pre-requisites then many of them could become CCPs. However if you consider the pre-requisites are under control, and regularly check that they are under control, then you should be OK.

On the subject of likelihood you have to go by your experience. A machine with blades that routinely break would be very likely to contaminate the product with metal (Risk 3). With a vat that has a stirrer near to the edge it may be foreseeable that it could move and scrape the edge, thus contaminating the product with metal (Risk 2). If the stirrer was well away from the edge you may decide it is very unlikely to contaminate the product at all (Risk 1).

You need to use the experiences within the business and include engineers and operators when looking at likelihood.

[frawat]

Hi all,

ISO 22000 says that you may consider the likelihood of failure in the functioning of a control measure to categorize if it should be managed through an oPRP or a CCP. Therefore I think the answer to your question is yes.

The example of Gaskit is fine, but makes me think the following:
Why is he establishing a critical limit? I think oPRPs do not have critical limits because you cannot establish a dividing line between a safe and a potentially unsafe product, and also because in pest control there is a synergy of control measures, so how could you establish a specific critical limit? If the critical limit is 'no pest infestation', then you are not really controlling anything, just checking if the hazard is present or not (in other words, inspection may not be a good control measure in this case). Sorry Gaskit if I am misunderstanding you.

I think that generally PRPs should have a low probability of failure, otherwise they could be considered as CCPs. Also, you must consider them before the hazard analysis.

thanks and best regards,
Francis

[jamesgibb]

The example of Gaskit is fine, but makes me think the following:
Why is he establishing a critical limit? I think oPRPs do not have critical limits because you cannot establish a dividing line between a safe and a potentially unsafe product, and also because in pest control there is a synergy of control measures, so how could you establish a specific critical limit? If the critical limit is 'no pest infestation', then you are not really controlling anything, just checking if the hazard is present or not (in other words, inspection may not be a good control measure in this case). Sorry Gaskit if I am misunderstanding you.

Francis,

I agree entirely with your logic, This is precisley the type of situation that causes problems with ISO22000,

As I understand it an uncontrolled hazard identified as part or the analysis is controlled as follows:
critical limit = HACCP Plan
no critical limit = oPRP

Anyone else have a theory on this?

James

[frawat]

Hello James,

I think you are right.
I have just finished attending an ISO 22000 seminar in Perú, with an expert (and outstanding teacher) from Colombia. Overall, one of the things we concluded was that the standard was not too didactic, but that it was a somewhat unique standard (much more rigorous in many aspects, like validation for example) and that it improves traditional HACCP approaches.

We had much discussion, specially about CCPs and oPRPs.
I have never really used the usual CCP decision trees because I did not find them useful and thought they did not discriminate well between CCPs and oPRPs (in fact that was the reason I used to have, at the beginning of our HACCP program, too many CCPs!).


In the seminar, we also agreed that the significance of the hazard was a key concept (check Yorkshire comments), and this has to do with many factors, like (see 7.4.3):

- the nature of the hazard itself (severity)
- the probability of failure of the control measure or process variation
- if the hazard could be controlled at a further step,
- etc., etc.

This leads to the question whether it makes sense to establish a CL or if the hazard can be controlled just with oPRPs. So it kind of confirmed to me that the traditional CCP decision tree is incomplete as tool or method to identify a CCP. It is interesting that ISO22000 requires you to record and document your method of CCP identification.

Nevertheless, I also think "Yorkshire" comments on risk assessment are very useful and appropriate.

regards,
Francis