The U.S. Food and Drug Administration’s outbreak investigation table currently lists three open foodborne investigations in which the product is not identified. The case counts are public. The food is not. That combination is the story, and it is not an invitation to guess a brand.
Three investigations, not one outbreak
FDA investigation 1415 is a Listeria outbreak with 11 illnesses. A traceback has been initiated. FDA first posted it on September 30, 2026. The product is not identified, and the investigation is active.
Investigation 1421 is an E. coli O157:H7 outbreak. FDA first posted it on September 23 with 25 illnesses. A September 30 update listed 38. Traceback and sample collection have been initiated. The product is not identified.
Investigation 1417 is a separate Listeria outbreak. It was first posted on September 23 with 40 illnesses and listed at 44 on September 30. A traceback has been initiated. The product is not identified. It is not investigation 1415, and the two Listeria counts should not be added together.
Today in Public Health added all three to the signal desk on October 1 from that FDA table. This article does not have a newer count. If FDA has revised any of the three since September 30, that revision is not claimed here.
Why a count can be public before a product is
This next section is an explanation of how these investigations are structured, not a quotation from FDA. A case count can be built from laboratory results and from reports by state and local health departments before investigators know what food the patients share. People get sick, specimens are tested, and a cluster becomes visible. Naming a product requires a further step: showing that the foods those people ate converge on a common item, supplier or production setting.
That is why FDA can disclose 11, 38 or 44 illnesses without naming a brand. The count is evidence that an investigation exists. It is not evidence that a particular company’s food caused it. Treating the count as a product identification is the error this table is easiest to misuse for.
What traceback and sample collection mean
Traceback means following the food people reported eating back through the places that sold, distributed or supplied it, to see whether those paths meet. It can start at a restaurant, a grocery store or a household and move upstream. “Initiated” means that work has started. It does not mean it has found a match, and it does not mean a recall exists.
Sample collection means testing food or the environment where food was handled. A sample can support or weaken a hypothesis. Collecting samples is not the same as having a positive result, and a positive result in one place is not, by itself, a brand name for every illness in the cluster. FDA’s table says sample collection has been initiated for the E. coli investigation. It does not publish a product result for any of the three.
Why guessing a brand is not a safety step
If the product is unidentified, a reader who picks a likely brand is doing something the investigation has not done. The guess can be wrong in both directions. It can accuse a company with no finding attached to it. It can also send a household toward the wrong food while the actual source, once found, is a different item entirely. Neither outcome is a precaution. It is a substitution of rumor for the traceback FDA says is still underway.
There is a contrast on the same FDA site. When an investigation has a recalled product, FDA says so and tells people not to eat, sell or serve that product. The alfalfa seed-lot outbreak is that kind of investigation. These three are not. For refs 1415, 1421 and 1417, FDA has not identified a product, so there is no product-specific consumer action in the source. This article will not invent one.
What a reader can do is narrower. Do not combine the two Listeria investigations. Do not treat a rising count — 25 to 38, or 40 to 44 — as proof of a named food. A rising count with no product means the exposure is still being traced, and the total may change again. Confidence is high that FDA posted these investigations and these counts. Confidence is low on source attribution, because the source itself says the product is not identified.