As of Thursday, August 20, 2026, CDC counts 10,930 people sick in the federal Cyclospora outbreak traced to iceberg lettuce from Taylor Farms de Mexico, at least 454 of them hospitalized, two dead, across seventeen states, with illness onsets running from June 14 through August 11. On its separate surveillance page the agency counts 15,716 laboratory-confirmed cases nationwide since May 1 and at least 11,841 more awaiting analysis. My own fifty-state tally, built from what each state health department publishes, stood at 31,889 that day. Those are the numbers this argument is supposed to be about.

Over roughly thirty-six hours this week, the senior food adviser at the Department of Health and Human Services^[1] answered people asking about those numbers by going after the people instead. Renee Guilbault catalogued the replies with screenshots. I am not going to answer any of it in kind. I am going to set it out plainly, once, and then explain why it is a bad trade for the man making it.

Every one of those answers went to the person asking rather than to the question asked

Here is the record of the past two weeks in summary, as documented in Guilbault’s post and in his own public account.

Source: Renee Guilbault, The Informed Eater, August 22, 2026, which reproduces each reply as a screenshot, together with contemporaneous reporting. I have not independently retrieved every post; the screenshots and the accounts they came from are public and can be checked.

Six entries, and not one of them contains a case count, a hospitalization figure, an inspection date or a plan. That is a choice, and it is an old one. Arthur Schopenhauer wrote out thirty-eight tricks for winning an argument you are losing, and he put this one last, as the final resort: the moment you see you are being beaten, become personal and insulting. He was not recommending it. He called it the point at which a man abandons the contest of intellect and appeals to something else entirely.

It is not new, either, and the clearest case is the one that ought to embarrass him most. In July, speaking at a MAHA event at the Great American State Fair, he told an audience that Marion Nestle’s research had been bankrolled by the sugar and tobacco industries, which he said had funded her to argue that the base of the American diet should be carbohydrates. He posted separately that she was an architect of the 1990s Dietary Guidelines that produced an explosion of refined carbohydrates and ultraprocessed food. Tim Carman of the Washington Post took the charge to her, and she answered it in print. She sat on the 1995 Dietary Guidelines advisory committee as one member among many, and was not its architect. The rest of it is inverted. Marion Nestle has spent her career documenting who pays for nutrition research; she wrote the book that made industry funding of food science a public subject, and she still runs a weekly item on industry-funded studies. She also spent 1986 to 1988 as senior nutrition policy adviser at the department he now advises, where she edited the Surgeon General’s Report on Nutrition and Health. Her answer included the part he should have read first: she is not MAHA’s enemy, she supports pieces of its agenda, and she has been on record for years asking for coalitions around the objectives everybody shares. That is who he chose to accuse of being bought. A man who cannot tell an ally from an opponent is going to have a hard time finding either one when he needs them.

That is the cost, and it is not a cost to the people being insulted. It is a cost to the man posting. A government spokesman has exactly one asset, which is that when he states a fact, people provisionally believe him. Every insult spends a little of it, and every insult attached to a number invites the reader to check the number. People have now checked, and the number did not hold up.

The recall count does not measure whether food is safe, and the group that compiled the numbers says so in writing

The chart being posted says food recalls are at a multi-year low. The updated version of it now carries a citation, to the Get the Facts data team at Hearst, which used figures compiled by the U.S. PIRG Education Fund. That citation is the problem with the chart.

On August 16, updated August 21, PIRG’s consumer watchdog director Teresa Murray published a piece asking exactly this question: are we seeing more food recalls? Her answer is no, and her next sentence is that even if we were, it would not automatically mean our food is less safe. She writes that the annual number of recalls is not an indicator of whether food is safe.

The people who assembled the data say it does not measure what it is being used to measure. That is enough. I do not have to argue that recalls are up, and I am not going to, because the count would not prove my case either. It has been the wrong instrument in every administration I have watched, which is now thirty-three years of them.

Three datasets say 2026 is running flat, and they are counting three different things

There is no single national recall number. There are several, they are built on different universes of events, and they cannot be plotted against each other. Here is what each one actually counts.

Sources: U.S. PIRG Education Fund, Food for Thought 2026; RecallBench 2026 statistics, retrieved August 22, 2026; Sedgwick Product Safety and Recall Index for the first half of 2026, as reported August 20, 2026. Arithmetic check on the annualization: August 12 is day 224 of 365, or 61.4 percent of the year; 324 divided by 0.614 is 528.

None of those rows says multi-year low. Two say flat. The gap between 320 and 528 is the difference between counting the recalls that got a press release and counting every recall that got classified, and a year-to-date bar cannot sit beside full-year bars, which is where the decline in that chart comes from. Guilbault built the annualized version and posted it in reply. The arithmetic took her an afternoon.

The last row of the table above is the one that ought to worry the department most, because it did not come from a staffer’s account. The claim in it, that recalls are at their lowest in over a decade, went out over an official one. Susan Mayne, who ran FDA’s Center for Food Safety and Applied Nutrition for eight years and has been writing about this outbreak since July, checked it against the agency’s own public dashboard and found the number had been drawn across every category FDA regulates, drugs and devices and tobacco and veterinary products together with food, rather than food alone. A recall warning is not the place to run a numbers argument. People were being told that a product in their freezer might hurt them.

A recall is a notice sent after people are already sick, and most outbreaks never produce one

PIRG looked at all 28 foodborne illness outbreak investigations announced in 2025. Thirteen of them never identified even a food type. In four more, investigators named a food but never a brand. Recalls were announced in 11 of the 28. Whatever the total count of recalls is in a given year, the illnesses that never generate one are invisible in it.

Recalls are also, overwhelmingly, voluntary and company-initiated. Crediting an agency for a high recall count credits it for what companies did after the fact and crediting it for a low one credits it for silence. A recall is a lagging indicator triggered by illnesses that have already been counted. You cannot recall what nobody found.

The gap that produced this outbreak is an inspection gap, and it can be measured

The lettuce came from a plant in Guanajuato. GAO reported in 2025 that FDA averaged 917 foreign food facility inspections a year from fiscal 2018 through fiscal 2023, against an average of 8,353 domestic ones, and Congress wrote a schedule into the Food Safety Modernization Act that would have reached more than nineteen thousand foreign inspections a year by 2016. ProPublica reported last November that the foreign number has since fallen below two hundred, and that the cause was not a shortage of inspectors but the loss of most of the staff who booked their travel and processed their reimbursements.

Nobody at FDA decided to stop looking at that plant. No one was assigned to go. The investigators, the state epidemiologists and the traceback analysts who have worked this outbreak all summer have done it with what they were left, and they are not who I am arguing with. The plant was last inspected seven years before this outbreak, and the visit before that was in 2013, the year of the last Cyclospora outbreak traced to the same company and the same Mexican state.

Four things inside this administration’s own control would help more than any argument on X

First, put the Food Traceability Rule back on the schedule it had. It was published November 21, 2022, took effect January 20, 2023, and carried a compliance date of January 20, 2026, which was moved to July 20, 2028. Iceberg lettuce is on the food traceability list. The jalapeño peppers behind the Salmonella outbreak running alongside this one are on it too.

Second, restore the surveillance. FoodNet went from eight pathogens to two, and Cyclospora was among the six dropped. A parasite nobody is watching for is a parasite nobody finds until thirty thousand people are sick.

Third, rebuild the foreign inspection program, understanding that it is a travel and reimbursement problem as much as a headcount problem, and fixable this year.

Fourth, name Cyclospora in the pre-harvest agricultural water rule. The indicator organisms growers test for do not detect it, and the antimicrobials they use do not kill it. Testing for something else and calling the result clean is not a control.

Not one of those four requires Congress to act first, and not one of them requires anybody to lose an argument in public.

Nobody has to lose this argument for people to stop getting sick

I have spent a good part of this summer disagreeing with this administration’s food safety record, and I have written about this particular adviser before. None of that changes the arithmetic above, and none of it is a reason to answer an insult with one. The country is not served by a fight between a plaintiffs’ lawyer and a political appointee about who is arguing in bad faith.

Mr. Means, this part is for you. You have a job most people in food safety would take, with a Secretary who listens to you and a department that can move things I have spent thirty-three years asking for. The four items above are inside your reach. Nobody who works on food safety for a living is going to be talked out of the numbers by being called a name, and every time you post one you make it easier for the next reader to skip past a true thing you say. If you want to know what this outbreak actually did, I will introduce you to some of the families. An hour, no press, no reporters, nothing on the record. That offer has been open to every administration since 1993, and it is open to you.

Every correction that has to be issued costs more credibility than the original claim was worth. And the 10,930 people in the federal count, the 454 who were hospitalized and the two families in Michigan who buried someone are not a talking point on either side of this. They are the reason there is anything to talk about.

^[1] Calley Means has been a senior adviser in the Office of the Secretary at HHS since November 2025, where he supports food and nutrition policy and works directly with Secretary Kennedy; HHS confirmed the appointment to The Hill. He came into the administration in March 2025 as a White House special government employee, a designation capped at 130 days, and left when that term expired in the fall. He co-authored the May 2025 MAHA report and the Make Our Children Healthy Again strategy that followed it, and he co-wrote the book Good Energy with his sister. Before government he co-founded the health-benefits company Truemed and worked as a consultant to food and pharmaceutical companies; Reuters has described him as a former food industry lobbyist. He has not held a food safety or public health position.

Saturday, August 22, 2026

This is my seventh Weekend Wrap-Up. Last Saturday this column was called On Tuesday They Said the Food Supply Was the Safest in the World and the Outbreak Was Contained. By Thursday Night the Government’s Own Outbreak Had Grown by Half. This week the message stayed the same and the arithmetic went the other way again. Three states cut how often they publish. Officials in two more said the worst had passed. On Friday two national outlets ran pieces reporting the outbreak may be tapering off. And then every count that moved this week moved up, one of my own readings turned out to be wrong, CDC added fourteen hundred people to the federal outbreak without adding a single state, and a sprout outbreak that was forty-one people in two states on Thursday was fifty-five people in fifteen states by Friday afternoon.

Tuesday’s confirmed count reached 15,716, and the backlog grew instead of clearing. CDC’s surveillance page now carries 15,716 laboratory-confirmed cases of Cyclospora acquired in this country since May 1, up from 13,895 a week earlier, with 828 hospitalizations and the same two deaths in Michigan. Forty-seven states plus the District of Columbia and Puerto Rico. The number underneath it is the one I would watch: the pile of cases still under investigation went from 10,455 to 11,841. Confirmed rose by 1,821 while unconfirmed rose by 1,386, which means new arrivals outran the laboratories again. Combined, the federal count is 27,557 people.

A correction to something I published, and it is mine rather than anyone else’s. A week ago the unconfirmed pile had fallen for the first time all season, and I read that as the reporting lag finally clearing. I flagged at the time that one week is not a trend. One week later it grew by 1,386. It was not a trend. The drop was an artifact, the queue is deepening rather than emptying, and anybody who took comfort from my paragraph should take it back out.

The federal government has now told the public, in writing, that six clusters of this parasite have no explanation attached to them. CDC added a section to that same page headed Clusters of illness under investigation and pointed readers to FDA’s active investigation table, where six files carry 236 people between them and name no food at all. None of it is new to the table. What is new is that it now sits on a page a worried person might actually read.

Then on Thursday CDC added 1,449 people to the outbreak and did not add a single state. The Taylor Farms de Mexico lettuce outbreak now stands at 10,930 illnesses, at least 454 hospitalizations and two deaths across seventeen states, with onsets running June 14 through August 11, and the Fast Facts block says seventeen states, none new. That has not happened before in this investigation. Every prior update grew the map as well as the count, from five states to nine to fifteen to seventeen. FDA posted the same figures the same day and changed one sentence worth reading twice. On August 13 it had begun an onsite inspection at Taylor Farms de Mexico, the processor. It now says onsite inspections and sampling at iceberg lettuce growers in Mexico, plural, with international partners. The investigation has moved upstream from the plant to the fields, which is where a waterborne parasite reaches lettuce in the first place.

For the first time all summer, laboratory work ties these illnesses to one another. At the end of Thursday’s update sits a paragraph that was not there a week ago. Investigators are genotyping samples from sick people, and CDC says the parasites that caused these illnesses are genetically related, which it calls additional evidence that everyone in this outbreak got sick from the same source. Read it precisely, because it cuts both ways. It ties the patients to each other. It does not tie them to a head of lettuce, because there is no product isolate to compare against and there never was. What it does do is close off the argument that this summer is a great deal of unrelated cyclosporiasis piling up at once — which is the best case against the lettuce, and I had set that case out as well as I know how the day before. San Francisco showed within a day what the tool does: one resident’s infection was genetically associated with this outbreak, in a state where the recalled lettuce was never distributed.

CDC also deleted the paragraph that told readers where the lettuce went. The passage naming the distribution — at least thirty-one states, including Taco Bell locations and Marketside products at select Walmart stores — is gone from the consumer page, and the link to the recall notice went with it. The caption now says the outbreak was linked to that lettuce, past tense, where a week ago it said is. The tense is defensible; the best-by dates really have passed. The distribution list is a different matter. FDA still publishes both lists, so the information has not vanished from the government. It has come off the page a worried person is likeliest to land on, on the day the count rose by 1,449.

Seven in ten confirmed cases are now assigned to the lettuce, and two of every three people counted in this country are assigned to nothing. The 10,930 set against 15,716 confirmed is roughly seven in ten, where a week ago it was six. Against the combined federal figure it is four in ten. Against my own fifty-state count, which stood at 31,889 on Thursday and has since taken corrections of 177 in Illinois and 31 in Washington, it is about a third. Attribution here means a sick person told an interviewer what they ate, and not one state publishes its own cases split that way. New Hampshire is the exception: thirty-two of its sixty-nine cases are tied to recalled product, and the cluster behind them began with nine people who ate from a hospital cafeteria salad bar in Keene. A hospital can see illness in its own workforce and report it the same week. A grocery store cannot. And in Lee County, Florida, a nurse practitioner at a free clinic told a reporter on Thursday that her clinic is not testing anyone, because her patients are uninsured and the test runs into the hundreds of dollars. She has seen cases anyway, in patients and in her own staff, and none of them are in Florida’s count.

The jalapeño outbreak added eighty-six people and five states, and its epidemic curve runs into August. CDC moved that count for the first time in fourteen days: 431 sick in 32 states, 57 hospitalized, nobody dead, with Florida, New Jersey, New Mexico, Pennsylvania and West Virginia added. Minnesota and Colorado hold 247 of them between them. CDC also confirmed what its own published curve showed before its text caught up: illnesses ran from June 19 to August 2, not to July 20. Twenty-four people got sick on or after the day Chipotle switched suppliers, and five after QDOBA stopped serving jalapeños. I am not going to tell you what that tail means, because the interviews behind it have not been published. Roughly 164 grocery products have now been pulled across eight companies, the importer has still never published a recall notice, and the grower in Sinaloa has still never been named.

Thursday’s new sprout outbreak was a fifteen-state outbreak by Friday. Minnesota’s health and agriculture departments named alfalfa sprouts grown by Everything Sprouts of Minneapolis, sold under the Everything Sprouts and Calco brands, behind twenty-three illnesses there, with Wisconsin reporting seventeen more. On Friday CDC and FDA both posted: fifty-five people in fifteen states, four hospitalized, nobody dead, onsets running May 31 through August 8, and twenty-six of the thirty-four people interviewed had eaten alfalfa sprouts. The outbreak did not grow overnight. It was always this size, and two state health departments found it before the federal system did.

Read the pathogen line twice, because it is the whole story. Those fifty-five people are carrying Salmonella Agona and three separate serogroups of Shiga toxin-producing E. coli — O26:H11, O103:H25 and O168:H8 — and two of them are carrying one of each at the same time. A kitchen does not do that. A seed lot does. Sprout seed is grown in open fields, and whatever lands on it goes into a warm, wet, four-day germination that is very nearly an incubator by design. There is no kill step afterward, because the entire point of a sprout is that you eat it raw. We have been here before, with the same two states and the same vegetable: in 2016 an E. coli O157 outbreak traced to alfalfa sprouts from Jack & The Green Sprouts of River Falls, Wisconsin sickened eleven people across Minnesota and Wisconsin. Minnesota’s epidemiologists found that one and found this one, which is worth saying out loud in a summer spent criticizing the architecture. Sprouts have been linked to more than fifty outbreaks in this country since the mid-1990s. I do not eat them and will not feed them to my family, and I have never had to revise that.

The company recalled on Friday afternoon, and the notice lists no lot numbers. FDA said it had recommended a recall and that none had issued, and Food Safety News ran the story under a headline saying the company declined. Later that day Everything Sprouts posted a notice on its own website recalling alfalfa sprouts distributed since May 27. Read the document. It says the products subject to the recall include the following lot numbers, and then lists none. It says no illnesses have been reported to the company, in an outbreak of fifty-five people. It calls the company Everyday Sprouts, twice. It reaches Minnesota and Wisconsin, in a fifteen-state outbreak. And it arrives with FDA’s sample press release still stapled to the bottom of it, the specimen one about XYZ Inc. of Anywhere, Mississippi recalling five-ounce packages of Snackies. FDA had inspectors in that plant on August 19, and it had warned this company in March 2025 about a roof leaking onto exposed sprouts and promised to check the corrections at its next inspection. There was no next inspection. A recall here was never going to clear a shelf, because the product lasts days. What a recall produces is lot codes, a customer list, and notice down the chain to everyone who received it. None of that is in this document.

The produce industry’s own chief scientist told Congress this week to take an independent outbreak board seriously. De Ann Davis, senior vice president for science at Western Growers, published a piece on August 19 supporting Frank Yiannas’s proposed National Foodborne Outbreak Investigation Board, conditionally, and her diagnosis is the best sentence written on this subject all year: responsibility in the American food safety system is distributed and accountability is not. Her seven design questions mostly have no answers yet, because there is no bill. Thirteen congressional letters have gone out about this outbreak and not one of them asks for this. I answered her on Friday and named six people who could sit on such a board. I am not one of them, and neither is any other lawyer who bills an hour on these cases.

The lobbying record now has names in it. I spent Sunday and Monday in the disclosure database after the Guardian pulled the first filings. One association reported its work as support for the thirty-month compliance extension in its own words. A second named the shutdown bill by number in the quarter it passed. And a 2024 bill to strip the lot-code requirement out of the traceability rule was introduced by a member of the appropriations subcommittee that writes FDA’s money and cosponsored by that subcommittee’s ranking member. The bill died. Its language turned up in three appropriations cycles and became Section 780. The Senate’s own appropriators wrote the opposite instruction, and their language did not survive conference. Nobody has asked in public why.

FDA held its two-day Listeria meeting in College Park on Tuesday and Wednesday, and the docket is still open. I did not make it and filed written comments instead. The agency’s own background note says listeriosis incidence has not moved in more than a decade, against roughly 1,250 cases and 170 deaths a year, and not one of the seven questions it published asks what regulatory or enforcement change would move that line. Docket FDA-2026-N-7914 does not care whether you were in the auditorium, and a paragraph from somebody who has chased a persistent strain through a drain is worth more than another page from me.

The country has a nominee to run the FDA, and she will inherit all of this on her first day. The President announced the nomination of Heidi Overton on Wednesday. She is nominated and not confirmed — the Senate health committee still has to act, and Kyle Diamantas remains acting until it does. My prediction is a three-hour confirmation hearing with no question about lettuce. If I get one thing into her in-box it is not on any of my policy lists. Spend an hour with somebody it happened to. Every other item I have asked for costs money, rulemaking, or an act of Congress. That one costs a calendar entry.

We filed twice this week, in California and in Illinois. Whitney and Casey Hight sued Ramona Family Naturals Market and the Eben-Haezer’s Happy Hens egg ranch in San Diego County Superior Court. Twenty-one people are sick and seven have been hospitalized in an outbreak the county traced to house-made mayonnaise built on raw, unpasteurized shell eggs — which the California Retail Food Code forbids by name, mayonnaise included. The county has since named the farm and says the same Salmonella strain is turning up in people who never set foot in that market, and there is still no recall and no published list of the stores that received those eggs. Then on Friday we filed in the Northern District of Illinois for a Chicago librarian and his wife, the first case of this outbreak filed in that state and the sixth we have filed. He ate a Chalupa Supreme box meal on July 11 and ordered it through the app, so the purchase is in his bank records rather than his memory. CDC issued its health alert on July 14. He was in an emergency room on July 15. FDA named the lettuce on July 16 and named Taylor Farms de Mexico on July 17. Nobody tested him for the parasite for another nine days, and it came back positive on August 4, twenty-three days after he got sick.

On the recall ledger, the jalapeño chain kept unspooling a week after everybody stopped watching it. FDA posted Hardie’s Fresh Foods on Monday — five-pound bags of whole jalapeños sold at two Texas Costco warehouses, seven lot codes, a notice that never once mentions the outbreak that caused it. The company had announced it five days earlier. Salata Dressings of Houston pulled a jalapeño avocado dressing sold at Texas H-E-B stores. Neither appears on FDA’s own outbreak advisory, which is now several recalls and eighty-six illnesses behind FDA’s own recall page. As always, Phyllis Entis at eFoodAlert is catching the ones the rest of us miss, and her weekly roundup is where I check my own.

One piece of housekeeping, and it is the most useful thing in this column if it reaches the right person. We are hiring. Marler Clark is looking for two people: an associate attorney with at least five years of litigation experience, and a paralegal with at least ten. There is more work than there are lawyers to do it. Everything above is the docket — six cases filed against Taco Bell, four franchise operators and Taylor Farms, the first jalapeño case, infant botulism against ByHeart and Nara Organics, Listeria out of Boar’s Head, E. coliO157:H7 from beef kofta and romaine, Salmonella from a Ramona egg ranch and a Lexington pie shop, and whatever those sprouts turn into.

Here is the part of the job nobody advertises. Clients find us; that has never been the hard part. Finding the defendant is. You will read FDA inspection reports and environmental assessments, USDA establishment records, whole genome sequencing and PulseNet clusters, epidemic curves built by state health departments and, this summer, the sanitation records of a Mexican municipality. That work happens before the complaint is filed, not in discovery. There is no billable hour requirement, no origination credit, and the firm carries the case costs. The other half of the practice is petitions, rulemaking comments, legislation and writing, and people here put their own names on that work. We would like you on Bainbridge Island rather than commuting to it — a few minutes up from the ferry terminal, thirty-five minutes across the water from Seattle — and we will help with the move. Having the right person matters more than the zip code. Competitive salary depending on experience. Write to me directly at bmarler@marlerclark.com and tell me why this work rather than any other.

That’s the week — a backlog that grew, fourteen hundred people added to a federal outbreak that did not add a state, the first laboratory evidence anyone has produced in three months, a paragraph about where the lettuce went quietly deleted, a jalapeño curve that ran two weeks past where it was supposed to stop, a sprout outbreak that went from two states to fifteen in a day and produced a recall notice with no lot numbers in it, and a nominee who gets all of it handed to her. Michigan had its smallest week of the outbreak on the same day CDC added 1,449 people to it. Both of those things are true at once. The season CDC defines closes on August 31. That is nine days. Check back next Saturday.

About Bill Marler

William “Bill” Marler has spent more than thirty years as a food safety lawyer and advocate—work that began with the 1993 Jack in the Box E. coli outbreak and has never really stopped since. In the years that followed, he has represented victims in nearly every major foodborne illness outbreak in the United States. That case, and the movement it launched, is the subject of the book “Poisoned” and the Emmy Award–winning Netflix documentary of the same name. Bill’s work has been profiled in The New Yorker (“A Bug in the System”), the Seattle Times (“30 years after the deadly E. coli outbreak, a Seattle attorney still fights for food safety”), the Washington Post (“He helped make burgers safer. Now he’s fighting food poisoning again”), and many others.

Dozens of times a year, Bill speaks to industry, regulators, and universities across the United States, Canada, Europe, Africa, China, and Australia about a simple idea: outbreaks are preventable. He has testified before Congress on the Food Safety Modernization Act and teaches food safety at institutions including the Harvard T.H. Chan School of Public Health. He writes regularly about food litigation and food safety at Marler Blog, and in 2009 he founded Food Safety News, which he continues to publish.