19 replies to this topic

[mgourley]

I just got back from a 2.5 day training seminar.

Unlike many training seminars, this one was full of applicable info and answered many of my questions as to what, exactly, the FDA will be requiring from a "qualified individual" and what the "food safety plan", is supposed to look like and what it is supposed to contain.

Quite honestly, it's very much like traditional HACCP, with some specific verbiage. 

Hazards are controlled by four Preventive Controls. Process (usual HACCP things like GMP's), Sanitation, Allergen and Supply Chain.

 

I have come to the conclusion that the vast majority of your current HACCP Plan ingredient and process hazard analysis can be simply copied and pasted to the new "format" that the FDA is looking for in a food safety plan.

Preventive controls need to be summarized for each type of control measure.

 

That being said, there is still quite a bit of detail that is not yet available. FDA is apparently working on a comprehensive Food Hazard Guide (similar, I guess to the one that Canada already has) that will be, supposedly, published by the end of the year.

 

Allergen cleaning does not have to be verified, however a "visually clean" check should be done.

 

I won't specify who (company) provided the training, but they are certified by the FSPCA and more of these training opportunities are becoming available. I would highly recommend this training for companies that will be required to have a "Preventive Control Qualified Individual" on staff to create and verify the Food Safety plan required by "FSMA"

 

Marshall

 

[Charles.C]

Hi Marshall,

 

Appreciate the Update.

 

Next comments are purely speculative since I haven’t been following the FSMA pronouncements other than being impressed but bemused by the proposed scheme to define Products as  “High/Low Risk”. It does seem to be taking a long time to see a draft list.

 

I get  the (conceptual) impression that FDA are re-inventing their Grand Plan for the Fishery Industry circa mid-1990s. The Associated Guide and Final Manifesto (= Rule) developed for the Fishery Scheme were indeed minor HACCP masterpieces IMO.

 

“Preventive Controls” sounds strangely affiliated to Preventive Control Measures (aka Preventative Control Measures). I haven’t spotted the difference between a “Control” and a “Control Measure” as yet. It all sounds a bit like  “6 characters in search of an Author”.

 

One comment  I found a little disturbing  in yr post – I deduce that FDA have proposed some kind of split between “GMP” content and “Sanitation”, that sounds like a head-on collision with 21 CFR Part 110  (and nearly everybody else).

 

The Fishery Plan was, i think, devolved onto the National Seafood Alliance to do the 'heavy lifting" regarding certification of qualified implementers.

http://flseagrant.ifas.ufl.edu/haccp/

[mgourley]

Charles,

 

There really is no split between GMP's and Sanitation.

Process controls can include GMP's, but generally are going to be things like metal detection, time and temp etc. Whereas Sanitation controls are going to be things to assure cleanliness of food-contact surfaces and prevention of allergen cross-contact and biological cross-contamination from objects and personnel.

 

Marshall

[Charles.C]

Charles,

 

There really is no split between GMP's and Sanitation.

Process controls can include GMP's, but generally are going to be things like metal detection, time and temp etc. Whereas Sanitation controls are going to be things to assure cleanliness of food-contact surfaces and prevention of allergen cross-contact and biological cross-contamination from objects and personnel.

 

Marshall

 

Hi Marshall,

 

Thks. Interesting stuff.

 

Does this mean that PRPs are going to disappear as a separate category and be absorbed within "Preventive Controls".? That would represent a major innovation.

[mgourley]

As I understand it the short answer would be yes.

It will also require that people take a more scientific approach to hazard analysis.

 

See the attached for the following example.

 

Currently, under a Codex HACCP model, you might say that a chemical hazard exists at a processing step because it is "possible" that chemical residue might be left on a food contact surface after cleaning. Your method of controlling that "possible" hazard might be your PRP of the Cleaning Program, which states that equipment is cleaned and rinsed and then tested to ensure there is no excess chemical residue.

 

The logic, as I understand it goes something like this:

If your Cleaning Program says that you need to rinse and test to make sure there is no excess residue, AND you are using chemicals at the label direction,AND you have verified that your verification of your SSOP shows that there is no chemical residue after the procedure is completed, then there would be no "potential" hazard at that step, thus there would be no need for a preventive control.

 

It seems to me that these regulations are more geared towards processes that have a propensity towards biologic hazards. In my case, that's not really an issue.

 

The attached is an example of the template that was shown for hazard analysis.

 

Marshall

 

Attached Files

•  Hazard Analysis.xlsx   12.93KB   147 downloads

[mgourley]

I'll post a more complete example when I get back to the office

 

Marshall

[mgourley]

And as an aside, 21 CFR 110 is being replaced with 21 CFR 117. The long awaited change to cGMP's

 

Marshall

[Charles.C]

Hi Marshall,

 

Thanks for the above responses.

 

I just did a brief look-around and noticed this 2016 publication which, if still current, clarifies some of the aspects I was querying. Perhaps "muddies" is equally applicable.

 

http://www.foodsafet...od-safety-plan/

 

I deduce a lot of discussion is still going on into the nitty-gritty of the associated HARPC. A lot of the concepts, eg using PRPs to control significant hazards, are not new of course and have been implemented in various guises for many years already. Think PAS220.

 

Frankly I was horrified to note that the [to me] dreaded oprp function has seemingly found a new home outside of ISO. A Harpinger of Doom IMHO.

 

This Project already has the whiff of another Consultancy Goldmine for the next decade or so. Looks a worthy contender to follow the iso22000 windfall.

 

I suspect USDA may not regret having watched from the sidelines although I daresay the jumble of "Preventive Controls"  was not  the primary reason.

[mgourley]

Here is a sample of the Hazard Analysis that was provided in the training.

 

Marshall

Attached Files

•  HA Sample.xlsx   11.13KB   138 downloads

[Charles.C]

Hi Marshall,

 

Thanks for the file. I accidentally missed the first one for which thanks also.

Frankly, although my regretfully limited harpc understanding is solely based on IT material, there appears to me to be a disconnect between yr template hazard analysis examples and currently published documents.

 

The examples generated several basic queries for me.

 

(1) I note the presence of  “significantly minimized” which  appears fundamental to the def. of PC in harpc.

Nowhere have I so far seen an attempt to explain/give an example of what it actually means in harpc.

I imagine it will somehow be connected with PRPs and the like. I can appreciate the phrase is in line with the FDA’s existing “non-haccp” presentations for Hazard Risk Minimization for Fresh Produce. IMEX this phrase is not new to haccp but was quietly phased out of the early haccp CCP1/CCP2 approach due its concluded redundancy.  It also has some potential linkage to the equally ambiguous iso22000 def. of oprp.

 

(2) I have not seen any recent direct references as to the potential uses of  (iso) “OPRP” within a  PC context. Maybe, just like CCPs, PRPs,  FDA simply don’t care how you label yr PCs as long as the “requisite”  info. is provided. So maybe there will soon be an  OPRPH in action.

 

BTW, as a harpc newbie, I quite liked this introductory AIB document which was issued prior to the final rule –

 

 FSMA-HARPC, AIB update, Jan 2015 - FDARules.pdf   204.78KB   82 downloads

 

(3) Several publications (see ps 1-2 below) have described harpc as a paradigm shift. And some not (eg ps3). Some promotional aspects no doubt exist but I had initially rather sided  with the former based on IT info.  Yr template examples do suggest the opposite. But see (4) below.

 

 ps1 - FSMA-HARPC,Paradigm shift,Sep 2015.pdf   241.99KB   57 downloads

 ps2 - FSMA-HARPC,Paradigm shift.May 2015.pdf   511.04KB   51 downloads

 ps3 - Does an existing haccp plan fit into HARPC,Dec 2015.pdf   222.89KB   69 downloads

 

(4) These 3 (latest I could find) references below seemed  variously at odds with yr template examples, eg (a) with respect to "including"  consideration  of  likelihood/severity,  although I have not seen any specific related harpc documentation requirement, (b) regarding statements like this in files fin2/3  (eg for fin2, I assume quoted from somewhere)  –

“The [harpc] hazard analysis must include severity and likelihood in the absence of preventive controls” . Further comment in respect to BRC compatibility is also included in fin2.

 fin1 - HARPC - Final Rule Summary,Sep 2015.pdf   51.76KB   55 downloads

 fin2 - BRC Sites and-HARPC,March 2016.pdf   782.61KB   63 downloads

 fin3 - FSMA preventive control rules,Dec 2015.pdf   1.29MB   60 downloads

 

I suspect the situation “on the ground” may still be fluid / open to interpretation so perhaps yr examples will shortly prove to be “the real McCoy”.

[mgourley]

Here is a complete food plan example.

 

Marshall

Attached Files

•  Food Plan Example.pdf   836.62KB   167 downloads

[Charles.C]

Here is a complete food plan example.

 

Marshall

 

Hi Marshall,

 

Thanks for the aptly chosen example. Interesting stuff.

Rightly or wrongly, seems to not address my comment in item (4) in previous post.

It is an intriguing approach but, based on this document, looks to be a WIP.

Not so sure about the Sanitation PCs. Sort of Retro. After striving to reduce CCPs, this now puts them back.

 

 

 

(

[Charles.C]

Hi Marshall,

 

I forgot to add one comment about the Food Plan Example in Post 11 which puzzled me.

 

The use of the Supply Chain PC is marked at receiving  as being  "applied at receiving ". Surely this is operationally incorrect ?

 

One notably good thing about the example IMO is the absence of references to oprp. I have already noticed this entity has popped up in some recent harpc-oriented publications together with the usual variety of  "explanations"  interpreted from iso22000.

 

As a corollary to my comment in previous post I don't get the use of column 4 in the hazard analysis table from a risk assessment POV. Statements like "vegetative pathogens such as Salmonella, pathogenic E.coli ..." is not my idea of risk assessment. And appears mismatched to the column heading. Similarly P = None.

It appears that (ultimately) the terminology/concept of  "significant hazard" has been eliminated despite the fin1 file in my previous post.. If so, that would IMO constitute a major change.

[mgourley]

Charles,

 

If you require COA's for materials that show absence of pathogens, and you check those COA's at receiving, then It would be operationally correct.

 

FDA wants hazard analysis to be more scientific. They define a hazard requiring a preventive control as "a known or reasonably forseeable hazard for which a person knowledgeable about the safe manufacturing, processing, packing or holding of food would, based on the outcome of a hazard analysis (which includes an assessment of the severity of the illness or injury if the hazard were to occur and the probability that the hazard will occur in the absence of preventive controls), establish one or more preventive controls to significantly minimize or prevent the hazard in a food and components to manage those controls (such as monitoring, corrections or corrective actions, verification and records) as appropriate to the food, the facility and the nature of the preventive control and its role in the facility's food safety system."

 

They define known or reasonably forseeable hazard as: "A biological, chemical (including radiological), or physical hazard that is known to be, or has the potential to be, associated with the facility or the food."

 

So, we know that vegetative pathogens exist on raw vegetables and in the soil. (Scientific evidence and actual food borne illness outbreaks).

So column 4 is a justification for your decision for a preventive control at that step. Since we know that the hazard exists or has existed and we know that the severity would be really bad if the hazard were not controlled.

 

As far as P = None, it can be assumed that for the purposes of this example plan, physical hazards are not known  to exist at the step, or there has been no evidence of there being physical hazards at that step, or one makes the assumption (for this example plan), the suppliers of the ingredients take the steps to ensure that the product being received does not have physical hazards present when they sell it.

 

As far as "significant hazard" vs. "potential hazard". I would agree.

 

Marshall

[Charles.C]

Hi Marshall,

 

Thks above post.

 

My comments below may not relate to the most current FDA interpretation of the project so please correct  if my ref. quote below is out-of-date/wrongly interpreted.

 

If you require COA's for materials that show absence of pathogens, and you check those COA's at receiving, then It would be operationally correct

.

If the checking activity  is the PC, I do agree it is at the stated step. But then the PC as stated in the table is surely either incorrect or incomplete. Either way, its only semantics.

 

I have always been somewhat mystified over the intended meaning of “absent” where RAs are involved..

For example, in the veg.salad  washing step, if the antimicrobial component  is  typically present in the process, does one now have to imagine another  process where there is no antimicrobial component for the purposes of the RA  etc, ad infinitum ? I have assumed Yes, eg some untrainable person is forgetting the chlorine addition.

(The traditional haccp approach whereby one hypothesises/analyses  the consequence of something going wrong/its likelihood has always seemed much more intelligible to myself).

 

Ignoring my interpretive limitations, file Fin1 has -

 

Delegation of Hazard Control Within Supply Chain:  In recognition of the reality of modern distribution chains, FDA does not require a manufacturing/processing facility to implement a preventive control in certain circumstances when the hazard requiring a preventive control will be controlled by another entity in the chain.  For example, if a facility’s customer (or another entity in the chain) will control the hazard, then that facility can rely on its customer to provide annual written assurance that the identified hazard will be controlled by a subsequent entity.

 

As I interpret the above, this  “addendum”  neutralizes  the “absent”  in the harpc RA as compared to yr text/file fin1 for, for example, "to be customer cooked" products. So the result  of a harpc RA procedure based solely on “absent” would, presumably be invalid for such cases. The consequence of applying the addendum is presumably that in this case harpc = haccp as far as (veg.micro) process factors are concerned.

 

For veg.salad example which is RTE,  I  hypothesise that the "absent" requirement will effectively auto-justify  the necessity of implementing a PC at any haccp, process CCP-questionable steps (sanitation factors, etc will  presumably have similar harpc/PRP results). Based on the veg.salad listed results harpc appears to yield 4 Process PCs whch would maybe = 2CCP/1-2PRPs from a haccp RA. So "score" is approx equal  and perhaps some increased monitoring/documentation for harpc due the increased  number of critical limits - as design intended(?)/ Yippee.

 

As far as P = None, it can be assumed that for the purposes of this example plan, physical hazards are not known  to exist at the step, or there has been no evidence of there being physical hazards at that step, or one makes the assumption (for this example plan), the suppliers of the ingredients take the steps to ensure that the product being received does not have physical hazards present when they sell it

No problem but perhaps a trifle disingenuous.(just kidding )

[mgourley]

Charles,

Good points and I will try to address them, although I am not an expert, as i just got the training.

 

As I understand it, if the untrainable person is in charge of the chlorine addition, either you are too stupid to have such a person in charge, or you are not doing proper monitoring to ensure the level is correct.

Both of which would be a bad thing. If your PC says that the chlorine solution should be  xppm, and that is your PC, then documentation should prove that.

 

I am not sure what you mean by "absent" from a RA. I see that all of the time on a product spec though. I find it difficult to understand why a supplier would send a COA along with a particular lot of ingredient if the COA actually stated a "present" for something that the spec says should be "absent". Would that not trigger a control at the manufacturers site?

 

I find that with many HACCP HA there is the tendency to include "what if" as a possible hazard. For example, in a large mixer, if the actual operation of the mixer does not include metal on metal contact, but IF PM was not performed correctly, there is the "possibility" that there "might" be metal on metal contact at that step. How to control that hazard is pretty much always metal detection.
But is there a "known" hazard and has there actually been proof of metal contamination from that step? This where I think the PC's differentiate from PRP's.

 

I may not have adequately stated it before, but as I understand it. the Process PC's are traditionally HACCP CCP's and those PRP's that do not exactly fit into the PC categories of Allergen, Sanitation or Supply Chain.

 

Traditional HACCP "Critical Limits" do not mean the same in the new rule. I'll have to refer to the course material to be more specific.

 

Marshall

[Charles.C]

Hi Marshall,

 

Sorry, maybe my post not clear enough (plus i have some uncertainty over some of this stuff myself). My "absent" is  same one as in the def. in yr own post and the refs. It's typical RA usage. Not related to the product as such.. Hence my  attempted "example" and the "auto-justify" later on.

This scenario was precisely what generated tons of CCPs in the early haccp plans, ie every hazard in sight was a CCP, including Sanitation and the rest.

And then came "Significant hazards" along with customer usage. (There is a comment in the 2nd/3rd edition of the FDA's Fishery Guide that everybody else had cottoned on to the customer usage option so they had decided to follow suit despite the logic being debatable ITO.)

 

I hadn't realized critical limit had been re-defined. Interesting.

 

And yes, i sort of picked up a trend that the net result was PCs = old(CCPs + PRPs) + some "critical limits" etc. (ignoring the oprp camp).

I presume this is why some publications are claiming that an existing GFSI compatible FS should work within FSMA bar a few things.

 

in fact I saw some comments that use of iso22002 will surely exceed FSMA requirements but I think this was before the fraud stuff got into the act.

 

And yes again, yr "what if" equates to my "going wrong". IMO the advantage of this approach is that one can envisage/analyse a small defect in the critical limit parameter as compared to my understanding (rightly or wrongly) of "absent" being a gross change.

[Charles.C]

addendum -

 

I think the extract below from my previously attached file fin2  (which deals with presenting a BRC haccp plan also compliant to FSMA/HARPC) spells out the difficulty (more clearly than myself) which  I have encountered with the procedure for hazard analysis in the harpc example  -

 

 hazard analysis in harpc.png   54.14KB   0 downloads

("but be" looks like a typo, probably = "is")

 

This difference in mind-set between traditional, eg Codex, haccp and harpc is quite fundamental IMO. it may be discussed in the manual (Ch8) associated with  yr veg.salad example.

 

And to emphasize the def. in yr Post 14, here is a 2nd extract from fin2 -

 

 hazard analysis in harpc - 1.png   21.43KB   0 downloads

-------------------------------------------------------------------------------------------

The following is my own interpretation of the above 2 quotes but this is speculative -

 

Crudely stated, it appears that -

 

All potential hazards (as, i think(?), are more or less equally interpreted by Codex or harpc) will in HARPC require PCs (excluding cases such as related to the addendum in my previous post)

(the "All"  is clearly not 100% accurate by inspection of the veg.salad col.3 [15-5] but is not so far out IMO [plus i didn't agree with the carrot "None" !])

 

In comparison, Codex divides the necessary handling of its potential hazards 3 possible ways - CCPs, PRPs, no action.

The codex split is based on first defining PRPs for certain potential hazards.  Then determining whether a potential hazard is (a) handleable by a pre-defined PRP or (b) Not. If the result is (b) the handling then depends on whether the risk (LxS) corresponds to (i) a "Significant hazard" or (ii) Not. if (i) >> CCP, if (ii) >> no action required.

 

So, from a BRC/Codex haccp POV,  the net result of the harpc procedure (ie 2nd quote above)  is effectively a  flood of "significant hazards"

--------------------------------------------------------------------------------------------------------

 

PS - Regardless of all these theoretical meanderings, the most visibly obvious query seems to be  -  Where is all the L/S data ?

Perhaps the Rule(s) have changed again.

[mgourley]

Charles,

 

I would guess that the lack of the L/S data is intentional.

It would have been used to determine whether or not a potential hazard exists, but is not a "part" of the actual food safety plan.

You would certainly have it on hand along with other documentation, but again, it is not "required" to be a part of the actual Food Safety Plan.

 

Marshall

[Charles.C]

Hi Marshall,

 

From the FSPCA manual I deduce “potential hazard” as used in the veg.salad hazard analysis = “a known or reasonably forseeable hazard”. (eg Pgs 8.3-8.5).

It also appears various table formats are in use, w/wo explicit LS data, and w/wo explicit  “significant hazard”.

The transition from col2 to col 3 in the  veg.salad hazard analysis looks equivalent to  an identification procedure for a [ process “harpc - Significant Hazard” ] (PHSH).  The latter are seemingly identifiable either intuitively or via L/S columns (all  presumably validatably).

Any identified PHSHs require a PC.

So i think the HARPC (Process) hazard analysis for PCs is equivalent to haccp for CCPs  BUT  (major) differs  due  the change in interpretation of  “Significant” eg  harpc  requires an evaluation in  the “absence of preventive controls”.

The Sanitation / Allergen / Supply Chain controls etc seem intuitively allocated to PCs. Similar to PRPs.

 

It will be fascinating to see a harpc hazard analysis for a “ready-to-cook” finished product. Afai can see The FSPCA manual has almost totally (intentionally?) avoided this category. Google also no luck.