FSIS has a new pilot. The agency is inviting poultry plants to volunteer to measure Salmonella more precisely as birds move through slaughter and processing — either by biomapping the plant top to bottom, or by building a validated critical control point for Salmonella into their HACCP plan, quantifying and serotyping as they go. In return, a plant that signs up gets a shorter compliance window — as little as 13 weeks instead of the current 52 — and can ask FSIS for waivers to test less often. The agency says the data it collects “could inform potential future regulatory changes.”
I have read a lot of these constituent updates over the last thirty years. This one measures. It does not decide.
It is the sequel to the Salmonella Framework FSIS proposed in August 2024 and then quietly withdrew in April 2025, after more than 7,000 comments and a lot of industry heartburn about legal authority, the science, and the cost. Then came a public meeting in January. Now a voluntary pilot. The direction of travel could not be clearer: more study, more meetings, more measurement — and still no line in the sand.
Here is what gets me. The science in this pilot is pointed in exactly the right direction. Quantifying the organism instead of a yes/no presence test. Serotyping to find the Salmonella that actually puts people in the hospital. That is precisely the machinery you would need to declare pathogenic Salmonella an adulterant. The old excuse — “we can’t tell which Salmonella is dangerous” — disappears the moment you can. And yet FSIS keeps treating that emerging science as a reason to keep studying rather than a reason to act. The agency’s new head of food safety went on camera and said she is even more convinced Salmonella should not be an adulterant, that first we have to go find the “most pathogenic Salmonella.”
Fine. Go find it. We handed you the list six years ago.
In 2020, my firm petitioned FSIS on behalf of Rick Schiller, Steven Romes, the Porter family, Food & Water Watch, the Consumer Federation of America, and Consumer Reports, asking the agency to declare a defined set of outbreak serotypes adulterants in meat and poultry:
Agona, Anatum, Berta, Blockely, Braenderup, Derby, Dublin, Enteritidis, Hadar, Heidelberg, I 4,[5],12:i:-, Infantis, Javiana, Litchfield, Mbandaka, Mississippi, Montevideo, Muenchen, Newport, Oranienburg, Panama, Poona, Reading, Saintpaul, Sandiego, Schwarzengrund, Senftenberg, Stanley, Thompson, Typhi, and Typhimurium.
Thirty of the thirty-one come straight out of the CDC’s own Salmonella Atlas — forty-two years of laboratory-confirmed illness. Those are the serotypes making people sick. That is the “most pathogenic Salmonella” everyone keeps saying we still need to identify. It has been sitting on the agency’s desk this whole time.
Meanwhile, the numbers do not move. Roughly 1.35 million Salmonella infections a year in this country. About a quarter of the foodborne cases trace back to poultry. Between 2017 and 2021, FSIS watched the share of chicken samples testing positive for Salmonella fall by more than half — and the human illness rate did not budge. That is the whole point, and it is the thing the agency will not say out loud: measuring contamination more precisely is not the same as making fewer people sick. We have twenty years of flat illness curves proving it.
And read the fine print on this pilot. A plant that joins gets a shortened compliance window and can apply to sample less frequently, and FSIS may even stop posting that plant’s Salmonella category where the public can see it. The “reform” hands industry a lighter regulatory touch today in exchange for data the agency may or may not ever use to regulate tomorrow. The data gets collected. The rule never comes.
None of this required a pilot to figure out, because the law has not changed since I wrote about it in March. The Federal Meat Inspection Act and the Poultry Products Inspection Act both say a product is adulterated if it: “bears or contains any poisonous or deleterious substance which may render it injurious to health.”
Salmonella that hospitalizes tens of thousands of Americans and kills hundreds every year is a deleterious substance. It is not a close call. The Fifth Circuit in Supreme Beef leaned on the comfortable fiction that cooking cures everything and the consumer will sort it out — the same 1974 logic that once told American housewives to handle raw poultry carefully and not bother the meat industry about it. That reasoning was wrong then, and it is wrong now.
So here is my take on the pilot, same as my take on the framework, same as my take on the petition. We do not need FSIS to measure Salmonella more precisely to know it belongs on the list of things in our food that can poison or kill a person. The agency has the authority. It has the serotypes. And now it has the very quantification tools it always claimed it was waiting for.
Stop measuring the problem and start calling it what it is. You have the authority — you just need to use it.