2 replies to this topic

[MRios]

I have a real problem with the fact that all decision trees for determining CCPs say :

Q1 - Do control preventative measures exist?
A1 - Yes

Q2 - Is this step specifically designed to eliminate or reduce likely occurence of hazard to an acceptable level
A2 - Yes

I work at a flour mill, where we have a separator that takes out all particles that don´t have the same density, size or shape of wheat kernels. This in itself is a preventive measure, IMO. It, in fact is specifically designed to eliminate or reduce likely ocurrence of hazard to an acceptable level. So is it a CCP?
After this step there´s other equipment that takes out dust, magnets to pick up any metal that could´ve gone through the separator, etc. (Wheat comes with all kinds of impurities.) The magnets for example, and I would venture saying that the separator, are more for the protection of the rolls and other equipment, because anyways flour gets sifted till you have particles that are smaller than 180 microns. (Ok, I do have to say that the point of the separator is milling only wheat, so you don´t end up with milled corn cobs or other debris in your flour, that would affect quality.)
So what about the fact that further on down the line there´s other equipment that takes care of these issues? If you don´t take this into account, all magnets would be CCPs.

Please enlighten me on this topic!!!

[Charles.C]

Dear MRios,

Somewhere in this forum, there is another thread which goes into considerable detail on yr specific query.

I don’t use d-trees for reasons like those in the next paragraph but I thought to start the ball rolling. Other direct dtree users are welcome to expand on yr question.

It’s important to remember that the d-tree process is a methodology to “standardise / visualise / approximate ” a more fundamental (likelihood x severity) risk assessment (which itself is already a derivative of things like FMEA etc etc) . All these techniques involve subjective judgement (ie probabilities) so that in general, black-white conclusions don’t exist. This limiting aspect of d-trees is sadly often stated in very small print at the bottom of the flow pattern / associated table of d-trees. Awkward cases (perhaps like yr own) tend to be handled by specific tree clauses. This is not necessarily unacceptable but illustrates the subjectivity involved.
There are other versions to the codex tree which try to “improve” the implementation but things then tend to also become more complicated.

Very, very, briefly, for yr case, the last “specially introduced” detector in a chain is now traditionally “the” CCP but you can also find references which demand correction at the earliest possible point and thereby make this ‘their” CCP also. But then you get a lot of CCPs …..

My personal solution is to stick to a, knowingly, less than 100% perfect, risk matrix analysis while also adding the "last in a chain" get-out clause. IMEX, auditors happily accept this and one can move on to other HACCP issues

Rgds / Charles.C

added -

some other threads -

http://www.ifsqn.com...showtopic=10486

http://www.ifsqn.com...t...6&pid=23155

http://www.ifsqn.com...?showtopic=9753

IMO the "acceptable" level in the dtree text must require a separate risk analysis but it seems many people conveniently pass over this detail.

[MRios]

Wow Charles! I think I just got myself a complete education on sieving and CCP´s with those links.
The part about a wire that is less than 5mm wide but 25 mm long going through a sieve reminded me of a piece of equipment we have before our CCP. It´s a safety sifter (not sure of the translation). We check its integrity weekly (from the manual and experience we´ve determined this is frequent enough), to prevent any wires (metal or plastic) from the sieves from going further on down the process. When we find rejected material from the turbosifter, which is our CCP, it is usually lint from the material that protects the sieves.
thank you so much for your help!