Our piece, Freshway’s Traceability System Worked Like A Charm: FDA And Buyers Don’t Care, focused on lessons the Freshway Foods recall might offer for the usefulness of highly specific traceability data. One very well informed attorney sent along this note:
I have a quick comment on the Freshway recall. While I certainly agree with your general views on the importance of limiting the scope of recalls when there is adequate traceability data, this may be a unique situation (although I am not representing any of the currently affected companies and don’t know the facts beyond those disclosed by FDA, the associations and the media).
Given that only ~5% of labs are set up to test for E. coli 0145, FDA, the CDC and the states don’t really know, at this point, whether other lots may be implicated (particularly if from the same geographical area but there also could be cross-contamination in a processing facility) and whether the cases to date represent the beginning, middle or end of the outbreak curve. Because the strain was not previously linked to foodborne illness but has resulted in HUS, in young, healthy people, the outbreak is potentially serious and would be considered an emergency situation at FDA that could justify (under their procedures) seeking a broader recall on a public health basis.
Our comments on the “disconnect” between the specificity of the traceback and the broadness of the attitudes of FDA and the buying community were not intended as a critique of FDA or the buyers. It was really raising the question of how specific industry traceability needs to be.
In truth, the FDA was, if anything, very circumspect in its recall. We would say circumspect to a point of illogic. And this points to an issue with case-specific traceability.
If Campbell’s Soup has to recall a lot, there is no logical basis for the FDA to demand that Progresso also recall its soup. They are made in different factories, in different parts of the country, with different work forces, water supplies, etc. They may share a few suppliers but the cooking process should handle any issues there, so the problem is clearly with the individual plant.
In contrast, in most produce outbreaks you are dealing with a farm that is structurally identical to many other farms — the farm may draw on the same river or aquifer for water, be physically proximate to other farms… so they all share the same intrusive animals, even the work crews might cross over.
For traceability to be meaningful, different lots have to be meaningfully distinct from other lots. This can be as simple as saying that the machinery was sanitized between lots. If you don’t do this, then you can assign all the lot numbers you want. It really is all one lot because cross-contamination can occur.
But if the FDA hears about a newly dangerous E. coli on food and thinks the problem may be river water contamination, the scientifically meaningful distinction is product that drew on the suspected river and product that did not.
If we assume that zero tolerance of illness possibility is the policy, and FDA suspects river water and nobody was testing for that pathogen, then FDA should, in other words, close down the whole sector because it draws on river water that could have been contaminated.
It is another topic whether zero tolerance makes sense as a public health policy. For now, our question was to what degree case-level traceability — as called for in PTI — is actually going to be useful to the FDA? If the answer is “not often, maybe close to never,” then maybe that is an investment not worth making?
