Yes, I agree one can understand the FDA’s dissatisfaction with Allergen Control as a PRP as per the recalls. FSIS, I believe, have a vigorous corrective expectation where a PRP is visibly failing to be “verified”.
Perhaps the "commonplace" occurrence of allergen-related/labelling recalls has also created an impression of non-seriousness. I would have thought that the financial losses in having to relabel, possibly repackage, entire shipments would be large enough to cause a major manufacturer concern but seemingly not so.
There seems to be several sides to the issue of allergen labeling recalls– (a) manufacturing facilities’ scientific limitations regarding which products actually contain allergens, (b) disinterest or incompetence regarding allergen labeling requirements / necessity, (c) precautionary labeling. Whether elevating the allergen issue (ie harpc) will simply maximize manouevring via option (c) is unclear to me.
It is also a fact that the original trend ca 2000 to handle potential CCPs like Allergens within Prerequisite Programs was not totally unrelated to a distinct benefit in reduced documentation and auditorial probing. The launch of standards like iso22002 has added another stamp of approval to the (currently) wide scope of PRPs.
One objection to HARPC is that (perhaps similarly to iso’s use of oprp) a primary objective of haccp in focusing attention on “significant hazards” is damaged by the introduction of the Process-System-wide Preventive Control (PC) whose interpretation now contains the wonderfully subjective “significantly minimize the hazard” term. Traditional haccp of course tended to receive the opposite criticism - if a hazard was not a CCP, the hazard was "uninteresting" / relegatable to PRP handling. (iso22000 tried to bridge this gap but IMO has failed to achieve this objective.)
Harpc seems to have another notable blind spot (similar to very early haccp) due its curious (and afaik unexplained) inclusion of the "absence" of PCs while performing the hazard analysis (precise meaning of absence unclear to me). The result looks to be the likely approval of a truckload of PCs for raw finished products which are destined “to be cooked” by the consumer/onward user. Have so far seen no published refutation of this conclusion, eg within the downloadable FSPCA manual.
Time will (soon?) tell. I wonder why FDA could not have simply defined an intrinsic / significant cross-contamination, allergenic hazard as being associated with a CCP. I believe historical precedents exist.