Guidelines on determining the frequency of finished product testing?
Other than annual testing for micro identified in our HACCP plans (which always comes out in favor of our HACCP), we do not test our products for micro (APC, Coliform, Yeast & Mold) and we received a minor for not having such procedure. Now we are preparing a procedure for testing but seem to hit a wall when it comes to frequency of testing.
We are a small scale company and cannot afford to test every lot or do a skip lot test. If there any guidelines on determining the frequency of finished product testing? I would like to have document as a back up for the frequency we determine.
Thats a tough one, frequency of testing from a statistical standpoint is complicated, and typically has the end conclusion of "well, none of it has significance anyway....".
I think the trick to justifying end item testing for pathogens is to tie it to your trace/recall program. You need to determine what you will do with a positive test first, see what your scope of recall would be, and come up with a frequency that matches well with the amount of risk you want to assume.
Questions to think about:
1. If I get a micro positive, what products would that implicate?
>anything produced on shared equipment?
>anything since last sanitation event?
>anything since last micro negative?
2. If that micro positive is traced back to a raw material, what would be implicated?
>all products that used that same lot of material? Any other products produced on shared equipment like above?
Come up with what your response procedure would be. If you decide that your testing interval would create too large a scope for you to accept if faced with recall, then shorten the interval to a more manageable scope. If you think you could have control of all affected materials in the event of a positive, then maybe create a light interval for "verification" and a heavy interval for "investigation" that would let you reduce the scope of affected material. This will affect other things like raw material tracability and sanitation intervals as well as you learn what the "stop" points would be as your recall scope expands.
Other than annual testing for micro identified in our HACCP plans (which always comes out in favor of our HACCP), we do not test our products for micro (APC, Coliform, Yeast & Mold) and we received a minor for not having such procedure. Now we are preparing a procedure for testing but seem to hit a wall when it comes to frequency of testing.
We are a small scale company and cannot afford to test every lot or do a skip lot test. If there any guidelines on determining the frequency of finished product testing? I would like to have document as a back up for the frequency we determine.
Hi chavalik,
So what type of products are involved?
I assume the reason for mentioned micro testing of finished products is to verify the haccp plan.
JFI, Other than possibly "mould", the micro. items you refer are not related to Food Safety in a product context.
So i am rather puzzled as to the reason for yr receiving a (I presume) safety-related NC ? (If you had included pathogens it might make more sense.)
The sampling/testing frequency is typically required to be based on risk assessment.
If yr unknown products are all haccp low risk and yr analytical results are 100% non-argumentative from a safety POV, the actual requirements are probably minimal from an auditorial POV.
A quite elegant, albeit simple, risk-based answer is offered in the first link within this post and an example of its forum use in 2nd link -
http://www.ifsqn.com...611/#entry90991
Other links in same post discuss various related frequency situations
Assuming all products are low risk as above, you could simply define a number of "rotated representative" items to be tested on a weekly/monthly/bimonthly/etc basis (as logistically/financially realistic) and use the above linked risk procedure to show satisfactory overall adequacy.
Alternatively, you might (for "show") expand above by initially slightly exceeding yr intended frequency then reduce it based on "no problems".
Obviously if yr products are not all equal low risk or there are some existing micro difficulties, the above method(s)may not work without some modification.