People I respect have told me I am unnecessarily critical of industry and of the government, and of FDA and CDC in particular. Some of them are friends. A few of them spent careers inside the agencies I write about. I have thought about it for a while now, and part of it is fair.
Here is the part that is fair. I have been representing people sickened by food since 1993, and thirty-three years of sitting with families who lost a kidney, a colon or a child does not make a person gentle. It is easy for the volume of my criticism to read as a judgment about the people doing the work, and that is not what I think. The career staff at both agencies are, in my experience, careful, serious and badly outnumbered. The epidemiologists who worked this outbreak did the interviews, built the traceback and found the supplier. They did it with fewer colleagues than they had two years ago. When I am wrong about them, I would rather be told than left alone.
Here is the part I do not accept. The implication underneath the criticism of my criticism is that complaining is all I do. That is checkable, and it is not true. Below are fifteen things the federal government could do about this parasite, sorted by how long each one takes. Four of them require no rulemaking, no appropriation and no act of Congress. Most of the rest were federal policy or federal practice inside the last fifteen years. I would rather be judged on this list than on anything else I have written this summer.
The reason for the urgency is the arithmetic. As of tonight, the federal government counts 6,358 illnesses, 278 hospitalizations and two deaths in fifteen states. CDC’s national surveillance page counts 10,468 laboratory-confirmed cases across 47 states plus more than 12,255 that are not confirmed. The fifty state health departments, added up one by one, report 25,836. Michigan alone reports 12,485 people and 279 hospitalized. The season CDC defines closes on August 31, which is twenty-five days from now.
This week, with a memorandum. No rulemaking, no appropriation, no act of Congress.
1. Grant the citizen petition that has been sitting at FDA since February. Stop Foodborne Illness asked the agency to adopt a policy of publicly naming every company associated with a foodborne illness outbreak, recall or no recall. The legal argument is clean: the Trade Secrets Act protects sales figures and proprietary processes, not the factual identity of a company that sells food to the public. It requires no rulemaking. The Commissioner could grant it this week with a memorandum. My one amendment stands, and this outbreak proves it: disclosure has to reach the processor and the grower, because in a produce outbreak there is very often no brand on the bag at all. And there is now a reason to think the agency has already answered in practice rather than on the docket. In April, FDA withheld the company names from its own report on a mango outbreak and said the information was covered by the trade secrets act. The petition is still pending.
2. Publish one national number, define exactly what it counts, and refresh it on a schedule. There are four numbers in circulation for the same disease in the same week. The outbreak page says 6,358. The surveillance page says 10,468 confirmed. Add the unconfirmed and it is 22,723. The states themselves say 25,836. Every one of those figures is defensible and no two of them describe the same thing, which is why the public hears a different number depending on which reporter called which office. Pick a figure, say in one sentence what is inside it and what is not, include probable cases or explain why not, and update it on a published cadence. An agency that will not state the size of a problem cannot expect to be believed about the solution.
3. Publish the case definition, and date every change to it. On August 5 CDC widened the definition of a case in this outbreak from people who reported eating at one restaurant chain to people who reported that chain or the recalled Taylor Farms de Mexico lettuce. The count went from 1,947 to 6,358 that evening. Nobody announced the change. It surfaced in a single sentence near the top of an FDA advisory, and anyone charting the outbreak learned about it by noticing that the line had jumped. A case count is not a measurement if the thing being counted changes without notice. Post the definition, date every revision, and republish the earlier series on the new definition so the trend line means something.
4. Publish the genotypes. CDC has had targeted amplicon deep sequencing for this parasite since 2018. It is the only tool in existence that can tell whether the fifteen-state outbreak, the six other federal Cyclospora files and North Carolina’s separate herb signal are one event or four. It is already built and already paid for, and it answers a question no state health department can answer on its own. Nothing has been published. This is the single cheapest thing on this list and probably the most valuable.
This season. Weeks, not years, and the season is not over.
5. Restore required Cyclospora reporting in FoodNet. I know the objection and I agree with half of it. FoodNet measures trends; it does not detect outbreaks and restoring it would not have found this one. What it would have done is tell us in the spring that something was different. Required reporting for this parasite came off effective July 1, 2025, ten months before the first illness in this outbreak. Reinstating it costs almost nothing, and nearly every letter Congress has sent this summer asks for it.
6. Make cyclosporiasis reportable in every state, not merely nationally notifiable. Those are two different things. Notifiable means CDC asks for the case. Reportable means state law requires a physician or a laboratory to send it. Pennsylvania is an outbreak state where reporting is voluntary, and the federal government has now counted more sick Pennsylvanians than Pennsylvania has. The parasite does not observe the distinction, and neither should the count.
7. Say what is in the other six files. FDA’s outbreak investigation table currently carries seven Cyclosporainvestigations. One of them is the lettuce. The other six say only that the product has not yet been identified, and together they hold 178 people. One of the six was opened on August 5, the same day the government announced the lettuce outbreak had reached 6,358. Another grew from ten cases to twenty-five, which FDA explains by saying it now includes additional illnesses based on similarities in reported exposures. That is the agency telling us it has found a common exposure pattern in a file whose food it will not name. Publish what is known about each of the six or say plainly why it cannot be published yet.
8. Publish a death-ascertainment method. The federal death count for this outbreak went from zero to two on August 3. Both were in Michigan and FDA has since said on its own page that both people had illness onset before the July 17 recall. What nobody has described is how a Cyclospora death is determined and counted in the first place — who reviews the record, what the criteria are, and whether other states are applying the same ones. Two may well be right. It is impossible to argue about it until somebody says how a death gets onto the list.
9. Go look at the water and answer the question that is thirteen years old. Everything above this line is about counting sick Americans. None of it sends a single person to the field. In August 2013, after the last large outbreak traced to the same company and the same region, Taylor Farms de Mexico suspended production and resumed with FDA concurrence on a stated commitment to a comprehensive sampling program covering its products and its water. Nobody has ever published whether that program ran, for how long, or what it found. FDA’s own 2013 environmental assessment told the firm to determine whether this parasite is a reasonably likely hazard and to re-evaluate the wash step. Ask what happened to both. Then send someone to sample the irrigation water and publish the result whichever way it comes out.
Before next summer. These take rulemaking or money, which is exactly why they should start now.
10. Bring back the Microbiological Data Program, and this time give it a parasite target. The MDP was created under President Bush’s 2001 food safety initiative, tested roughly fifteen thousand produce samples a year, and accounted for the large majority of all federal produce pathogen testing until it was zeroed out and shut down in December 2012. I have been asking for it back since 2024. The honest caveat, which I would rather state myself than have stated at me: MDP tested for bacteria, not parasites, so restoring it exactly as it was would not have caught this outbreak. Restore it with Cyclospora on the panel and the sampling pushed upstream toward the field.
11. Add the parasite to national wastewater surveillance. CDC’s National Wastewater Surveillance System draws from roughly fifteen hundred sites covering about half the country, and it reports SARS-CoV-2, influenza A, RSV, mpox and measles. Cyclospora is not on that panel, and it is not on the main academic panel either. The infrastructure exists, the samples are already being collected, and this organism is shed in enormous numbers by people who never see a doctor. Adding a target to an existing network is a small ask with an unusually good return.
12. Put the FSMA 204 traceability rule back on its original schedule instead of 2028. Susan Mayne, who ran FDA’s Center for Food Safety and Applied Nutrition for eight years and has actually supervised lettuce tracebacks, told CBS Sunday Morning that the rule would have been enormously helpful here and that without it the same thing happens again next summer. She is right, and every week the delay stands is a week of the next outbreak being investigated with the tools we used for this one.
13. Name Cyclospora in the pre-harvest agricultural water rule. The only numeric microbial criterion in Subpart E of the final water rule is generic E. coli, and FDA’s own guidance on this parasite says that traditional fecal-indicator testing will not identify it. We are requiring growers to test water for the wrong organism, and to treat that water with chlorine chemistry the agency says does not reliably kill this one and calling the result a safeguard.
14. Fix the import side, because that is where every one of these outbreaks has come from. CDC has put it in print that for every United States Cyclospora outbreak from the mid-1990s through 2015 in which both the vehicle and its source were identified, the produce had been imported — Guatemala, then Peru, then Mexico. For a foreign farm the Produce Safety Rule is enforced in practice through verification obligations that sit on the American importer, which is paperwork about a supplier, plus rare foreign inspections. You could fully staff every domestic position on this list, and nobody would walk that field.
15. Stand up the independent board Frank Yiannas has been asking for. An NTSB for foodborne outbreaks: investigating across agency lines, publishing root cause and lessons learned, with no stake in the reputation of the agency whose work it is reviewing. Yiannas is a former deputy commissioner of FDA and a past president of the profession’s own society, and he has said this summer that the federal response here is approaching a catastrophic level of mismanagement. His framing is the right one and it is the reason this belongs on a solutions list rather than a grievance list. The point is not to find fault. It is to find facts.
Two things about this list that I would rather say myself. None of the fifteen would have prevented this outbreak. The contamination happened in a field in central Mexico in the spring, weeks before the first American got sick, and no surveillance system counts a case that has not happened yet. What these fifteen would have done is make the outbreak shorter, smaller and legible — found sooner, described honestly while it was running, and understood afterward. That is a real thing to want, and it is not the same as prevention. And not one of them is a partisan measure. The Microbiological Data Program was created under a Republican president and killed under a Democratic one. The traceability rule was written under one administration and delayed under two. Most of these were federal policy or federal practice inside the last fifteen years.
I have watched administrations of both parties handle outbreaks badly. What is different this time is not the mistakes. Mistakes are ordinary. The false positive was ordinary. A parasite with a two-week incubation period, a six-week reporting lag and no whole genome sequencing route is ordinary bad luck, and I have said so in every piece I have written about it. What is different is that three weeks into the largest Cyclospora outbreak this country has ever recorded, nobody at the top has said the number out loud.
Twenty-five thousand eight hundred thirty-six. Say it, and then say which of these fifteen you are going to do.