Shoppers Are Walking Past the Produce Aisle. Telling Them We Have the Safest Food Supply in the World Will Not Bring Them Back.

By Bill Marler on September 2, 2026

Consumers are buying less fresh produce this summer, and they are doing it because of what they have read, not because they are confused. Salad-kit purchases, lettuce prices, restaurant traffic and produce shipments all fell after the Cyclospora outbreak went public in July, and the federal response was a slogan. If government and industry want people back in the produce section, the way to do it is to make the count go down, not to tell people how safe the food is.

The New Yorker ran a cartoon (see above without permission) this week that says it better than I can. A couple stands in the produce section with a basket of greens, and one of them tells the other that they have enough groceries that could make them sick right away, and it is time to go pick up some ultra-processed food that will kill them more slowly. The joke lands because everyone standing in front of the bagged salads has had the same thought. That is where thirty years of the safest-food-supply line has gotten us.

What the purchase data shows

The numbers are not anecdotes. Numerator reported that more than 6.5 million American households stopped buying salad mixes and kits during the month that ended July 26, that ten percent of households said they had cut spending on salads and fresh greens, and that more than a quarter said they had stopped buying them altogether. The firm put roughly 280 million dollars a month of produce spending at risk. NielsenIQ had dollar sales of prepackaged salads down 14 percent for the four weeks ended July 25 against the same period a year earlier. Fetch had fresh lettuce down 10 percent, salad mixes down 23 percent and salad kits down 27 percent over two weeks, with the shift going to frozen vegetables. Total Quality Logistics, which handles around 140,000 produce shipments a year, saw shipments drop about 20 percent year over year in the week of July 13 as buyers cancelled orders and some growers dumped product rather than risk it.

The price signal is the sharpest one. Lettuce prices fell 16.4 percent from June to July in the Bureau of Labor Statistics seasonally adjusted data, the largest one-month drop on record. Prices do not fall like that because supply went up. They fall like that because nobody wants it.

Restaurants felt it the same week. Sweetgreen said the outbreak cost it about six percentage points of same-store sales in July and cut its full-year outlook. Chipotle put its hit at around two points in the back half of July even though its lettuce was never implicated. Salad and Go filed for Chapter 11 and closed every location, citing the outbreak on top of its existing problems. Traffic at Chopt fell 24 percent the day after FDA named the outbreak, and Taco Bell, where the traceback began, was still running roughly 20 percent below its normal visit pattern into August.

The Grower Shipper Association said late last month that demand is starting to rebound, and it is honest about why: schools are back in session, and institutional buying is up. That is a calendar effect. The association also said growers remain worried about whether people feel comfortable buying fresh produce. They should be.

What the government said while that was happening

On August 11, with the purchase data already in hand, Acting FDA Commissioner Kyle Diamantas told CNBC that we have the safest food supply in the world, that Americans should have full confidence in it, and that they should feel confident eating fresh produce including leafy greens. HHS posted the same day that the outbreak was contained and American food was safe. CDC published its weekly surveillance count that same Tuesday: 13,895 laboratory-confirmed cases of cyclosporiasis since May 1 and more than 10,000 more under investigation.

Three weeks later, the outbreak is bigger. As of August 27, CDC and FDA put the lettuce outbreak at 11,458 illnesses in 20 states, 495 hospitalizations and 2 deaths, with illness onsets running from June 14 through August 15. Georgia, Tennessee and Texas were added that day. CDC’s national surveillance count stood at 17,180 laboratory-confirmed domestic cases and 922 hospitalizations as of August 26, against 1,180 cases for the same stretch of 2025. Michigan alone reports 14,510 cases and 356 hospitalizations. FDA moved its inspections upstream to iceberg growers in Mexico in the third week of August, five weeks after the recall.

That is one outbreak. FDA’s active investigations table carries nineteen rows this week: the lettuce, a 431-person Salmonella outbreak in 32 states tied to imported jalapeños that pulled products off shelves at seven national grocery banners, a multi-pathogen alfalfa sprout outbreak Minnesota named in August, an 84-person Salmonella Newport cluster posted August 26 with no food identified, and a string of smaller Cyclospora files with the product column blank. A shopper who reads any of this is not being irrational when she reaches for the frozen peas. She is reading the same pages the Commissioner is.

Why the slogan makes it worse

The safest-food-supply line is not new. Every administration has used it, and I have spent most of my career pointing out what it costs. CDC’s own consumer page still says 48 million Americans get sick from food each year, 128,000 are hospitalized and 3,000 die. The most recent Global Food Security Index put the United States third on food quality and safety, behind Canada and Denmark, and thirteenth overall. Third is a respectable score. It is not first, and the index grades laws on paper, not sick people in hospitals.

The deeper problem is that reassurance without results reads as spin, and consumers price spin into their decisions. When the agency says full confidence on a Tuesday and adds 1,449 people to the outbreak the following Thursday, the lesson people take is not that lettuce is safe. It is that the people telling them lettuce is safe are not the ones to listen to. Trust does not come back because someone in Washington says a sentence. It comes back when the count goes down and stays down, and when the reasons it went down are visible.

What would bring people back

I have watched this happen once. Hamburger nearly killed the fast-food business in 1993. USDA declared E. coli O157:H7 an adulterant in ground beef in 1994, the industry fought it, lost, and then built the testing, the hold-and-release programs and the supplier requirements that made the rule work. My firm’s hamburger cases went from a steady docket to almost none after 2002. Consumers came back to hamburger because it stopped making them sick, not because anyone told them it was safe. That is the model, and it took a regulator willing to set a standard and an industry willing to meet it.

For produce, the list is short and none of it is new. FDA should implement the FSMA traceability rule in 2028 without another delay and fund the field staff to enforce it, because a traceback that takes two months from first illness to first advisory is the reason a recall could not catch the lettuce before it was eaten. FDA should be inspecting foreign growers and their water, not just the packing plants, before an outbreak and not five weeks after the recall. CDC and FDA should publish one set of numbers on one schedule, so the public is not reading three counts that do not agree. And the agency should stop declaring outbreaks contained on days its own surveillance page is still climbing.

Industry has the bigger job, because it has the product. Growers and processors should test agricultural water and product for the pathogens that actually show up in leafy greens, including the parasite that has now sickened tens of thousands of people twice in this outbreak’s history, and they should publish what they find. The buyers, the restaurant chains and the grocery banners that lost sales this summer, should write those tests into their supplier specifications the way the beef buyers did after 1994, and refuse product from suppliers that will not run them. A company that finds the problem and says so in public earns more trust than any agency statement can give it.

The couple in the cartoon is the market. They are not going to be talked back into the produce aisle. They will walk back in on their own when the shelves stop sending people to the hospital, and government and industry both know exactly what that would take.

Three Federal Outbreak Pages, One Evening: CDC Is Working 40, FDA Lists 19, USDA Lists None

By Bill Marler on September 2, 2026

Posted in Case News

Every Wednesday I read three government pages in a row. Tonight they told three different stories, and the difference between them is the most useful thing on any of them.

The CDC page: forty

CDC updated its current outbreaks page today. The agency does not name the investigations, only counts them by germ: two Campylobacter, seven E. coli, seven Listeria and twenty-four Salmonella. That is forty active multistate investigations. CDC says it typically coordinates between 17 and 40 at any given time, so forty is the top of the range the agency itself calls normal. None of the forty is described. There is no product, no state list, no case count. There is a number, and a note that some investigations become public notices and most do not.

The FDA page: nineteen rows, seven with no food named

FDA’s CORE investigations table was last modified August 26 and has not had its Wednesday refresh as I write this. It carries nineteen rows marked Active. Thirteen are ongoing and six are outbreaks FDA says have ended but has not closed out. Seven of the thirteen ongoing investigations have no product named, and together they account for 486 reported illnesses.

Two of those seven deserve a closer look. Salmonella Newport, reference 1403, was posted August 26 with 84 cases, a blank product column, traceback initiated, and no inspection and no sampling. Salmonella Oranienburg, reference 1387, was posted July 8, sits at 99 cases, and after eight weeks still shows traceback only, with neither an inspection nor a sample collected. Nearly 200 people are sick across those two files and the public has not been told what to avoid.

The row that did move is the sprout file, reference 1394. FDA merged what had been two anonymous E. coli rows, O103:H25 and O26:H11, into a single incident after CDC found four people infected at the same time with three E. coli strains and Salmonella Agona. The advisory names alfalfa sprouts distributed by Everything Sprouts, LLC of Minneapolis, and the company’s recall was added August 24. That is the only row on the table where FDA has initiated all three of traceback, inspection and sampling and also has a recall in hand.

The remaining Cyclospora rows crept: reference 1398 from 17 to 22, reference 1392 from 201 to 203, reference 1385 from 27 to 28. The iceberg lettuce advisory has not been touched since August 20. The requesón Listeria outbreak, the one on this table with a death, is now marked ended, with final information posted August 26.

The USDA page: three for the year, all closed

The FSIS outbreak investigations page lists three investigations for all of 2026, and the Active column says No on every one. In May, Listeria in retail-sliced headcheese, with a public health alert and an Illinois notice. In May, E. coli O157:H7 in beef kofta served at The Kebab Shop, with a public health alert and a California notice. In June, Salmonella Enteritidis in suspected chicken at the Broome County Correctional Facility in New York. That is the whole list. Nothing is open. Marler Clark represents people sickened in the Kebab Shop outbreak, which I note here so you can weigh it.

In fairness, FSIS explains the emptiness. It posts an investigation only when the evidence points to a product it regulates, and a meat or poultry item that got sick from an FDA-regulated ingredient, as happened with the jalapeño public health alert in August, never appears on this table at all. The page describes what gets posted, not the whole food supply.

Even with that caveat, the contrast holds. Here is what the three pages say tonight, side by side.

Agency	Open files	Product not yet named	Oldest open unnamed cluster
CDC	40 active multistate: Salmonella 24, E. coli 7, Listeria 7, Campylobacter 2	Not disclosed	Not disclosed
FDA CORE	19 rows; 13 ongoing, 6 ended	7 of the 13 ongoing, carrying 486 illnesses	Salmonella Oranienburg, 99 sick, posted July 8, traceback only
USDA FSIS	3 for all of 2026; 0 active	0	None

What the meat table used to look like

I built a law practice on the meat table. From the 1993 Jack in the Box outbreak through the 2002 ConAgra recall, hamburger E. coli O157:H7 cases were the core of what this firm did. They are now close to zero for us. That did not happen by accident. In 1994, USDA declared E. coli O157:H7 an adulterant in raw ground beef. In 1996, the Pathogen Reduction and HACCP rule made every slaughter and processing plant write down its hazards and prove it controlled them, under an inspector who was in the building. In 2011, the agency extended the adulterant rule to six more Shiga toxin-producing serogroups. Test-and-hold followed. The industry fought most of it, then built its operations around it, and the hamburger cases went away. It is the best thing that has happened in food safety in my lifetime, and the empty FSIS page is what it looks like from the outside.

Which is why last Friday matters

On August 28, the President posted on Truth Social that he was authorizing legal documents to be drawn to give farmers and ranchers the right to process their own food, calling the four largest beef processors a monopoly. The post came two days after a radio host told him that inspections are insane and regulations are killing ranchers. Agriculture Secretary Brooke Rollins said announcements would start Monday, including waiving red tape in processing, expanding the ability to sell across state lines, and support for small processors, and on Tuesday said a further announcement on labeling and cross-state sales would come later this week. As of this evening, no rule, order or bill text has been published, and existing federal inspection requirements remain in effect.

The reaction from the people who actually run the meat business was quick. The Meat Institute warned that uninspected meat reaching consumers could damage the country’s food safety reputation, and the National Cattlemen’s Beef Association said weakening federal inspection and food safety standards is not the answer. Representative Thomas Massie, who has spent years pushing the PRIME Act to loosen custom-slaughter rules, called the post a nothing-burger because farmers can already process their own animals; what they cannot do is sell that meat without inspection.

I have no quarrel with more regional processing capacity. Ranchers hauling cattle hundreds of miles because four companies own the plants is a real problem, and Congress has been funding small processors for five years. The way to get there is more inspected plants, more state programs recognized as equivalent to federal inspection, and more inspectors, not fewer. What cannot be on the table is the inspector. The FSIS page is empty because someone is standing in every plant that sells meat across state lines. Take that away and the page fills back up, and I know exactly what it fills up with, because I have the files from 1993 to 2002.

The bottom line

CDC is working forty investigations and will not say what they are. FDA has seven open files with no food named and nearly 500 people in them, including a 99-case Salmonella cluster that has gone eight weeks without an inspection or a sample. USDA has nothing open at all. One of those three agencies has a system that works, and it is the one the President was asked last week to loosen. Whatever gets announced this week, the right question for Secretary Rollins is simple: will every animal sold as meat still be slaughtered under inspection? If the answer is yes, the plan deserves a hearing. If the answer is anything else, look at the FSIS page tonight, because it will not look like that for long.

Sources: CDC current outbreaks (updated September 2, 2026); FDA CORE investigations table (modified August 26, 2026); FSIS outbreak investigations; CNN, Agri-Pulse and Iowa Capital Dispatch on the August 28 post and the Secretary’s follow-up.

The Atlantic Says Our Summer of Contamination Could Have Been Much Worse. It Is Right. Put E. coli O157:H7, Salmonella, or Listeria on the Same Lettuce and the Arithmetic Is Thousands in the Hospital and Hundreds Dead.

By Bill Marler on September 2, 2026

Posted in Case News

Daniel Engber’s piece in The Atlantic this morning, The Truth About American Food Safety, closes with the thought that our summer of contamination could have been much worse. He is right, and the reason is not the system. It is the bug. Cyclospora cayetanensis is a parasite that sends about one confirmed patient in twenty to the hospital and almost never kills anyone. Run this summer’s case count through E. coli O157:H7, Salmonella, or Listeria monocytogenes, on the same lettuce and through the same supply chain, and you are no longer counting weeks of diarrhea. You are counting hospital beds, dialysis chairs, and funerals.

Start with the numbers as they stand. CDC’s case data page, updated September 1, reports 18,445 laboratory-confirmed cases of cyclosporiasis acquired in the United States since May 1, with 990 hospitalizations and 2 deaths, in 49 states and the District of Columbia. Thousands more are still being sorted or confirmed. CDC’s outbreak page ties 11,458 of those illnesses, in 20 states, along with 495 hospitalizations and both deaths, to iceberg lettuce from Taylor Farms de Mexico. This summer’s hospitalization rate is 5.4 percent of confirmed cases. That is what Cyclospora does. FoodNet, CDC’s active surveillance network, put the Cyclosporahospitalization rate at 6 percent for 2016 through 2018 and the death rate at 0.2 percent. For scale, FoodNet’s hospitalization rates run 27 percent for Salmonella, 40 percent for E. coli O157:H7, and 96 percent for Listeria.

Salmonella: about 5,000 hospitalizations and 75 to 130 deaths

The same FoodNet report puts Salmonella at 27 percent hospitalized and 0.4 percent dead for 2016 through 2018, and at 30 percent and 0.7 percent in 2023. Apply those rates to 18,445 confirmed cases and you get 5,000 to 5,600 hospitalizations and 75 to 130 deaths. That is not a stretch for produce. The 2023 cantaloupe outbreak, Salmonella on melons imported from Mexico, reached 407 confirmed cases, 158 hospitalizations, and 6 deaths; 44 percent of the people with information available were hospitalized. In CDC’s summary of the multistate outbreaks it solved in 2023, 35 percent of the people in Salmonella outbreaks went to the hospital.

E. coli O157:H7: roughly 7,400 hospitalizations and more than a thousand cases of kidney failure

FoodNet counted 276 domestically acquired O157 infections in its catchment area in 2023. Of those, 111, or 40 percent, were hospitalized, and three died. Forty percent of 18,445 is roughly 7,400 hospitalizations. Deaths run from about 75 at FoodNet’s average for all Shiga toxin-producing E. coli to about 200 at the 2023 O157 rate. Then there is hemolytic uremic syndrome, the kidney failure that O157 causes, mostly in children. In the 2018 Yuma romaine outbreak, 210 confirmed cases produced 96 hospitalizations, 27 cases of HUS, and 5 deaths. In the 2006 spinach outbreak, 205 confirmed cases produced 103 hospitalizations, 31 cases of HUS, and 3 deaths. Earlier O157 outbreaks ran about 6 percent HUS; Yuma ran nearly 13. Apply that range to this summer’s count and somewhere between 1,100 and 2,400 people, most of them children, are in kidney failure. I spent the first decade of my career on E. coli O157:H7 in hamburger, from Jack in the Box in 1993 to the ConAgra recall in 2002. I know what a few hundred cases look like from inside a pediatric intensive care unit. I do not want to learn what 18,000 look like.

Listeria does not need the case count

FoodNet puts Listeria at 96 to 97 percent hospitalized and 19 to 22 percent dead. In 2011, cantaloupe from one Colorado farm produced 147 confirmed cases; 143 of the 145 people with information available were hospitalized, 33 died, and one pregnant woman miscarried. I represented the families of all 33. Take one percent of this summer’s confirmed count, about 180 people, and at the 2011 ratios you have about 40 dead. Nobody has ever seen a listeriosis outbreak of 11,000 people, and I hope nobody ever does, because at FoodNet’s death rate the toll would be more than two thousand.

It could absolutely have happened

Cyclospora is a human parasite. It reaches a lettuce field one way: human waste, usually carried by water. On August 20, FDA said its inspections and sampling had moved from the Taylor Farms de Mexico processing plant to the iceberg lettuce growers in Mexico. A field that takes in untreated water does not have a parasite problem. It has a water problem, and water does not sort its pathogens. In 2018, CDC found the Yuma outbreak strain of E. coli O157:H7 in irrigation canal water in the growing region. Product from Taylor Farms has been linked to E. coli O157:H7 three times since 2015, in celery, slivered onions, and romaine, a record I laid out in July. Lettuce and spinach carried O157 across the country in 2006, twice in 2018, and again in 2019. The distribution system that put this lettuce into restaurants and stores in at least 31 states moves whatever is on the lettuce, and it moves it fast.

The lag does not care what the bug is

The first illness in this outbreak began June 14. The recall came July 17. CDC says it takes about six weeks to determine whether a sick person is part of a Cyclospora outbreak. With a pathogen that hospitalizes one patient in twenty, a five-week gap between first onset and recall costs weeks of diarrhea. With one that hospitalizes two in five and puts children on dialysis, every week of that gap is measured in kidneys. Engber reports research projects lost, a CDC genetics team hollowed out, and experienced staff gone; HHS told him no FDA investigators were lost. Barb Kowalcyk, who directs George Washington University’s Institute for Food Safety and Nutrition Security and who is a friend and one of the brightest people in this field, squares the two better than I can: the system went into this summer underfunded and years behind on changes it knew it needed, and the loss of experienced people since then tied its other hand behind its back. She told Engber that Jack in the Box changed USDA in 1993 and that this summer will drive the next shift; I would add that the 2006 spinach outbreak drove FSMA. I hope she is right, and I hope it happens before the arithmetic above stops being hypothetical. What is not in dispute is that the traceability rule that would let FDA follow a case of lettuce back to the field in days rather than weeks, finalized in 2022 and due this past January, has been pushed to July 2028 by FDA and by Congress. That delay applies to E. coli and Salmonella exactly as it applies to a parasite.

Engber offers his last line as both a solace and a warning. I will take the warning. We got the gentle bug this year. Nothing about how it reached the field, how it moved through the country, or how long it took to find it guarantees we get the gentle bug next time. Luck is not a food safety system.

Four E. coli Cases Tied to Dungeness Valley Creamery Raw Milk. No Recall. No Headlines.

By Bill Marler on September 2, 2026

Posted in Case News

Washington named the Sequim dairy on a web page on July 2. Two months later, not one news outlet has written a word. It is the dairy’s fourth outbreak since 2009. Its milk has been pulled from shelves six times before. Not this time.

Four people got sick with Shiga toxin-producing E. coli after drinking raw milk in late May 2026. The Washington State Department of Health investigated, concluded that raw milk from Dungeness Valley Creamery, a state-licensed dairy in Sequim, Washington, was the most likely source, and on July 2 posted its findings on an outbreak page. Then nothing. No DOH news release that I can find. No news story anywhere, not even in the Peninsula Daily News or the Sequim Gazette, which have covered every recall, alert and all-clear at this dairy for more than a decade. I missed it too. It sat there for two months.

What Washington DOH found

Four people were infected with the same highly related strain of STEC between May 28 and June 4, 2026. Two reported drinking Dungeness Valley Creamery raw milk shortly before they got sick: an Oregon resident who bought it on May 25 and a Snohomish County resident who bought it on May 28. A King County resident drank raw milk but declined to name the brand. A Texas resident tested positive while visiting Washington and could not be interviewed. DOH calls the creamery’s raw milk “the most likely source of exposure.” No one was hospitalized. No one died. There was no recall.

Why there was no recall

The Washington State Department of Agriculture samples every licensed raw milk dairy once a month. Its routine Dungeness Valley Creamery sample, collected June 16, tested negative for STEC, and on that basis DOH says no recall was recommended.

Look at the dates. The last known purchase was May 28. The sample came nineteen days later. Raw milk lasts days, not weeks; the milk those people drank was gone long before the sampler arrived. A negative result on June 16 says nothing about milk bottled in late May.

DOH knows this. In 2018, when two people got sick after drinking this dairy’s milk and every sample came back clean, state health officials told the Peninsula Daily News that the batch that infected them was unavailable for testing. That is the raw milk problem in one sentence: the investigation takes longer than the milk lasts. It happened again in May 2025, when three people were infected with STEC after drinking Cowlitz Hill Guernseys raw milk, one was hospitalized, WSDA’s May 5 sample was negative, and there was no recall.

How Washington announced past outbreaks, and this one

In December 2009, the state put out a news release naming the dairy. In September 2018, DOH issued a press release; the owner told the Sequim Gazette he learned of it from a retailer. In April 2021, DOH issued a release, and KIRO 7, KING 5, MyNorthwest and the Skagit County Health Department carried it within days. In July 2026, DOH posted a page that says it is sharing the information “for consumer awareness.” Awareness requires that someone be told.

Dungeness Valley Creamery outbreaks and recalls, 2009 to 2026

Dungeness Valley Creamery has held a WSDA raw milk license since 2006. The public record:

2009. At least three people were infected after drinking the creamery’s raw whole milk; E. coli O121 was confirmed and O157:H7 suspected. WSDA’s December 2 release said no E. coli was found in the current batch, but its investigation at the dairy turned up the same bacteria that caused one of the illnesses. The Peninsula Daily News has since described that outbreak as under previous ownership: the farm was then run by Jeff and Debbie Brown, whose daughter and son-in-law, Sarah and Ryan McCarthey, worked with them then and own it today.

2013. In February, WSDA warned consumers off the creamery’s raw whole milk, skim milk and cream with best-by dates of March 2 or later after a February 19 cream sample tested positive for STEC. The creamery stopped making cream. No illnesses.

2014. The owners warned customers off milk and cream with a July 2 expiration after a preliminary WSDA result on cream; follow-up tests were negative and sales resumed after three days. No illnesses.

2018. On September 7, DOH announced that a child under 5 from Island County and a person in their 70s from Clallam County had E. coli infections after drinking the creamery’s raw milk; later described as a 2-year-old and a 72-year-old woman, both recovered. The owners said DOH was speculating prematurely. WSDA tested 21 samples, fifteen from stores and six from the dairy; all were negative. Production never stopped. No recall.

2019, twice. In April the creamery recalled whole milk, skim milk and cream with best-by dates of April 6 through 20 after STEC turned up in whole milk dated April 6, one of two positives from late March and early April. More than 1,000 gallons were dumped and WSDA cleared them April 22. In June a sample dated June 5 was positive, the third of the year, and the creamery recalled product with best-by dates of June 5 through 14. Ryan McCarthey told the Peninsula Daily News the dairy had been testing through an independent lab since April. The state cleared them in three days. No illnesses.

2020. In October, after WSDA found STEC in raw cream dated September 29, the creamery advised customers to stop using milk and cream with best-by dates of September 29 and 30. Seven follow-up samples were negative. No illnesses.

2021. In April, five people who drank the creamery’s raw milk, bought in Clallam, Skagit, Kitsap and Clark counties, tested positive for Campylobacter. The creamery recalled all raw milk with best-by dates of April 13 or earlier and stopped distribution. The bacteria were traced to a single cow, which was put down. The owners told the Peninsula Daily News it was their first Campylobacter and added Campylobacter to their testing. Food Safety News called it the dairy’s fifth recall and third outbreak in twelve years.

2026. Four STEC infections. Most likely source. No recall.

Since 2009, Dungeness Valley Creamery has been linked to four outbreaks and fourteen reported illnesses, and its raw milk has been pulled from sale six times. State samples were positive for STEC in 2013, three times in 2019, and in 2020. In three of the four outbreaks, 2009, 2018 and 2026, the state never found the bacteria in the milk. A license and a monthly test do not make raw milk safe. They tell you, after the fact, that something got through, and sometimes not even that.

What should change

First, when DOH names a dairy as the most likely source of an outbreak, it should say so with a news release, as it did in 2009, 2018 and 2021. A page nobody sees is not a warning.

Second, WSDA should publish each licensed raw milk dairy’s sampling and outbreak history in one place a shopper can find. The label on every bottle warns that raw milk may contain harmful bacteria. It does not say this dairy has had four outbreaks and six pulls. Buyers can decide what to do with that. They cannot decide what they are never told.

Third, DOH’s advice bears repeating: children under 5, pregnant people, older adults and anyone with a weakened immune system should not drink raw milk. Mine is simpler, and it has not changed. Nobody should. If you are going to anyway, know who you are buying from.

Four people got sick in May. The state named the dairy, tested a bottle three weeks later, found nothing, posted a page, and moved on. No reporter noticed. Neither did I, until now.

Frequently asked questions

Is Dungeness Valley Creamery raw milk linked to an E. coli outbreak?

Yes. Washington DOH says four people were infected with the same strain of Shiga toxin-producing E. coli between May 28 and June 4, 2026, and that Dungeness Valley Creamery raw milk is the most likely source. No one was hospitalized or died.

Was Dungeness Valley Creamery raw milk recalled in 2026?

No. A routine WSDA sample collected June 16, 2026, nineteen days after the last known purchase, tested negative, and DOH says no recall was recommended.

How many outbreaks and recalls has Dungeness Valley Creamery had?

Four outbreaks with fourteen reported illnesses: E. coli in 2009, 2018 and 2026, and Campylobacter in 2021. Its raw milk has been pulled from sale six times: 2013, 2014, twice in 2019, 2020 and 2021.

How many raw milk outbreaks has Washington had?

DOH counts at least eleven from 2005 through 2024: seven Shiga toxin-producing E. coli, three Campylobacter and one Salmonella. Cowlitz Hill Guernseys in 2025 and Dungeness Valley Creamery in 2026 make thirteen, not counting the ten people sickened by Twin Sisters Creamery raw milk cheese in 2025.

Harvard Read the FRESH Act Line by Line. Its Preemption Clause Reaches Salmonella, Restaurant Inspections, and the Right to Sue.

By Bill Marler on September 1, 2026

Posted in Case News

The Harvard Law School Food Law and Policy Clinic released a report on August 25, 2026 taking the first comprehensive look at the preemption clause inside the draft FRESH Act. Margaret Lynn and Emily Broad Leib conclude that a provision superseding any and all State requirements or prohibitions relating to the use, labeling, sale, or marketing of food would displace state additive bans, allergen and menu disclosures, heavy metal limits, PFAS rules for packaging, retail and restaurant inspection, cottage food laws, and state manufacturing licensure, in most of those areas with no federal rule required to take their place. They are right. The reach is worse than the headline suggests, and the record of who asked for this language is a matter of public filing.

The clause preempts whether or not a federal rule exists to replace it.

Congress has preempted state food law before, and it has done so with a scalpel. The Nutrition Labeling and Education Act of 1990 lists the specific categories of state labeling requirements it displaces, and a state law falls only if it is of the type covered by a federal requirement named in the statute. That structure exists for a reason. It means preemption arrives with a federal rule attached, and nothing drops into a hole.

The FRESH Act carries no such limit. As the Harvard report reads it, the draft supersedes state requirements across the board without conditioning that on the existence of any corresponding federal standard. Where a federal rule exists, the state rule yields to it. Where no federal rule exists, the state rule yields to nothing at all. That is not uniformity. That is subtraction.

The phrase added or naturally occurring substance in food reaches pathogens, not just dyes.

Nearly all of the public argument about this bill has been about food chemicals: dyes, PFAS, the GRAS loophole. That is where the fight started, and it is where the coalition backing the bill would prefer the fight to stay. CSPI and EWG both came out against the draft on those grounds within days, and both were right to. But the draft does not stop at intentionally added ingredients. It also sweeps in any substance that may be expected to become a component of food, and any added or naturally occurring substance in food.

Read that as a lawyer reads it. The Harvard report points to United States v. Anderson Seafoods, which litigated whether mercury in swordfish was added or naturally occurring, and to Continental Seafoods v. Schweiker, which litigated the same question about Salmonella in shrimp. Courts have been construing that phrase for more than forty years, and they have construed it to reach microbiological contamination. Which means a state limit on Listeria in ready-to-eat food, a state testing requirement, a state action level for a pathogen in a product sold inside its own borders, all of it sits in the preempted zone as drafted. Nobody at the April 29 hearing described the bill that way. That is what the words say.

The FDA does not inspect restaurants, and the three thousand agencies that do would be preempted.

By the FDA’s own account, more than 3,000 state, local, and tribal agencies hold primary responsibility for regulating food service and retail food in this country. The agency publishes a model Food Code, but the Food Code has no force anywhere until a state adopts it, and the FDA has no direct authority over a grocery store or a diner. That division is not an accident of history. It is the architecture.

Preempt state requirements relating to the sale of food and you have preempted state food codes, state permitting, and state inspection, with no federal inspector standing behind them and no federal appropriation to hire one. The bill does not give the FDA that authority, does not fund it, and could not compel the states to keep doing the work for free, because the federal government cannot commandeer a state into administering a federal program. I have spent thirty years reading health department inspection reports in outbreak cases. Those documents exist because a local sanitarian walked into a kitchen under state law. Take away the state law and the sanitarian has no reason to be standing there.

Any and all State requirements is the phrase that has ended tort claims before.

Here is the part that lands closest to what I do. The Federal Food, Drug, and Cosmetic Act creates no private right of action. A person injured by contaminated or misbranded food recovers, if at all, under state law: negligence, strict products liability, breach of the implied warranty of merchantability, state consumer protection statutes. Every one of those is a state requirement.

The Supreme Court held in Riegel v. Medtronic that New York common law duties count as requirements within the meaning of a federal preemption clause and preempted them. It reached a comparable result under ERISA in Pilot Life v. Dedeaux. Whether the FRESH Act clause would be read that way is not certain, and I will not pretend it is. But the phrase any and all State requirements or prohibitions is broader than the language that did the work in Riegel, and a defense lawyer would be committing malpractice not to raise it. If it works, the family of a child with hemolytic uremic syndrome is left with a federal statute that contains no remedy and a state remedy that Congress erased. That is not a technical consequence. That is the consequence.

The coalition that wanted this published its specification a year before the bill appeared.

Representative Kat Cammack of Florida, who sits on both Energy and Commerce and Agriculture, released the discussion draft on April 22, 2026 and presented it at the Energy and Commerce Health Subcommittee hearing on April 29, where it was one of twenty-eight food bills on the agenda. Her stated case is a patchwork argument: differing state requirements create real costs for manufacturers, for retailers, and ultimately for families.

I am not interested in guessing at a member’s motives, and I would not print a guess if I had one. I am interested in the paper trail, and the paper trail here is not hard to follow.

In October 2025, six months before the draft appeared, a group called Americans for Ingredient Transparency launched. Its website states the ask in a single sentence: the Federal Food, Drug, and Cosmetic Act should be amended to establish the FDA as the sole entity setting the floor and the ceiling for regulations on the marketing and sale of foods and beverages, including safety assessments, ingredient approvals, registrations, reporting requirements, and labeling requirements. Read the two nouns again. Floor and ceiling. A floor is a minimum that states may exceed, and that is what federal food law has been since 1906. A ceiling is a maximum, and it means a state may not do more to protect its own residents than Washington has gotten around to doing. The group asked for both. The FRESH Act delivers both. There is no mystery about the drafting influence, because the customer published the specification.

The group’s own partner page names the backers: the Coca-Cola Company, PepsiCo, Nestlé USA, Kraft Heinz, General Mills, Conagra, Tyson Foods, Hormel, McCormick, Mondelez, Ferrero, Keurig Dr Pepper, Ocean Spray, Sargento, Sysco and Ken’s Foods, alongside the Consumer Brands Association, FMI, the National Restaurant Association, the National Retail Federation, the Meat Institute, the American Bakers Association, American Beverage, the Corn Refiners Association, the American Farm Bureau Federation and a dozen more. Its two senior advisors are Julie Gunlock, who directs the Independent Women’s Network and wrote a book arguing that alarm about food chemicals is itself the problem, and Andy Koenig, a founding partner of Kwinn Consulting who served as a special assistant to President Trump in the Office of Legislative Affairs and was previously vice president of policy at the Freedom Partners Chamber of Commerce. The footer of the site reads, paid for by Americans for Ingredient Transparency. Whatever else that is, it is not a consumer group.

One member company disclosed lobbying on state additive bans three months before the coalition existed.

Conagra Brands filed a lobbying report for the second quarter of 2025, signed by Megan Garcia on July 21, 2025, listing under its food issue area: issues related to FDA’s Generally Recognized as Safe regulations, and issues related to state food additive bans and labeling laws. Two in-house lobbyists are named on the page. That filing predates the launch of Americans for Ingredient Transparency by three months and the FRESH Act draft by nine.

I have made the point before that the useful move is to stop arguing about intent and go read the disclosures. When the Guardian did exactly that on the Food Traceability Rule, the filings named the organizations that had asked Congress to push compliance to 2028, in their own words, on their own forms. The same method works here, and it will keep working, because the law requires these people to write down what they are lobbying for.

This is the same play, run twice, eighteen months apart.

In 2024 and 2025, industry asked Congress to delay the traceability rule and then to defund its enforcement, and Congress obliged through the appropriations process. In 2026, industry is asking Congress to erase the authority of every state to regulate food inside its own borders. The delay argument was that one rule was too hard to comply with. The preemption argument is that fifty standards are too many to comply with. Both arguments end in the same place, which is fewer obligations and less traceable accountability, and both were made by many of the same trade associations.

I will grant industry its strongest point, because it is a real one. A different warning label statute in every state, each with its own list and its own threshold, is a bad way to run a national food supply, and a manufacturer cannot print a different box for every state line. Uniformity has value. The Association of Food and Drug Officials made the right observation at the April hearing, which is that preemption should be applied carefully and deliberately, so that it strengthens rather than strains the federal and state partnership. Careful and deliberate describes the NLEA. It does not describe this draft.

A federal floor is worth building. A federal ceiling is not.

The states did not wander into food chemical regulation out of boredom. They moved because the FDA did not. California banned ephedra before the FDA acted. New York City and California banned trans fats years before the federal ban. California acted on four additives in 2023, West Virginia acted on dyes, and more than thirty states introduced legislation of that kind in 2025 alone. The FDA’s own mandatory GRAS proposal, published this month, still does not require premarket approval and would not bind anyone until roughly 2029. I represented 339 of the people sickened by the Daily Harvest crumbles, and the ingredient in that product was never reviewed by anybody before it was sold. State legislatures noticed. That is what a laboratory of democracy looks like when the laboratory upstairs has the lights off.

The FRESH Act is still a discussion draft. It has no bill number, it has not been marked up, and there is time to fix the clause without abandoning the parts of the bill worth keeping. Take the preemption provision out or narrow it the way the 1990 Congress narrowed its own, and the GRAS registry and the post-market review program can be argued on their merits. Leave it in and Congress will have voted to take away nearly every tool the states have, replace almost none of them, and close the courthouse door on the way out.

Read the report. It is footnoted to the statutes and the cases, and it is the most careful thing written about this bill so far.

The Cyclospora Season Closed Yesterday With 17,180 Confirmed Cases, Two Deaths and a Fifty-State Count of 32,857. The Calendar Stopped. The Parasite Did Not, and Three States Have Already Stopped Publishing.

By Bill Marler on September 1, 2026

Posted in Case News

The cyclosporiasis season that CDC defines, May 1 through August 31, closed yesterday, August 31, 2026. It closes with 17,180 laboratory-confirmed cases of cyclosporiasis acquired in the United States, 922 hospitalizations and two deaths across forty-eight states and the District of Columbia, against 1,180 for the entire 2025 season. It closes with 11,458 of those illnesses in twenty states tied by CDC and FDA to recalled Taylor Farms de Mexico iceberg lettuce, with at least 495 hospitalizations and both deaths inside that outbreak. And it closes with a fifty-state count, built from what each state health department has been willing to publish, of 32,857 as of this morning, September 1. Cyclosporiasis is the intestinal illness caused by Cyclospora cayetanensis, a parasite that reaches produce when human waste reaches a field or the water that runs through it. The calendar closed on Monday. Nothing about the parasite did.

The federal record closes the season at roughly fifteen times last year, and it is still counting July.

CDC posted the last weekly update of the season on August 25, with data through August 24: 17,180 laboratory-confirmed domestic cases, 922 hospitalizations, two deaths, cases in forty-eight states plus the District of Columbia, and at least 11,844 more that are not laboratory confirmed or still need work to decide whether they were acquired here. Another 1,805 confirmed cases are travel-associated and sit on top of that. The median onset date is July 7 and the range runs from May 1 to August 19. On June 18 the same surveillance pageshowed 145 domestic cases in seventeen states. That is more than a hundredfold in ten weeks.

The outbreak pages moved last on Thursday, August 27, as I wrote that night: 11,458 illnesses from twenty states, illness onsets from June 14 through August 15, at least 495 hospitalizations and the two Michigan deaths, with Georgia, Tennessee and Texas added that week. CDC says genotyping shows the parasites from sick people in this outbreak are genetically related, which ties the patients to each other and not yet to the lettuce, and FDA says its inspections and sampling at iceberg lettuce growers and the processing facility in Mexico are ongoing and its status is still Ongoing. Neither agency has said what the end of the season means for this investigation, and neither page mentions it.

Most people with cyclosporiasis never get a stool test, because the parasite is missed by the routine ova and parasite exam and has to be asked for by name. The standard estimate of that gap comes from Scallan and colleagues in Emerging Infectious Diseases in 2011, which puts the underdiagnosis multiplier for Cyclospora at 83.1. Applied to the 17,180 confirmed cases, that puts the real number near 1,428,000.

Seven state rows moved in the last week of the season, and one of them is a correction I owe the reader.

Michigan reported 14,510 cases and 356 hospitalizations as of August 27, up 233 cases and 21 hospitalizations from the week before, its second-smallest week of the outbreak. Indiana stands at 2,090 as of August 26, up 21, with Allen County at 508. Kansas is at 829 as of August 26, up 22, with 36 hospitalizations; the state counted 687 cases in July and 66 in August. West Virginia is at 315 as of August 28, up two, with 21 hospitalized. Washington is at 76 as of August 28, up five, and it still publishes the breakdown nobody else does: of 25 cases with no international travel, twelve traveled to an outbreak state, three of them are counted in the Taco Bell outbreak, and nine never left Washington. New Hampshire stood at 94 as of August 25, up 25 from the 69 my chart had carried since August 4, and the hospital at the center of its cluster declared that cluster over the next day.

The correction is Florida. My chart has carried Florida at 407 through August 8 for two weeks, while the report it links to said 450 through August 15. The number was in my own source and I did not read it. Florida moves to 450, a plus of 43, and its calendar-year count stood at 486 through August 22. North Carolina holds at 940 but its as-of date is corrected to August 25, which is when the state posted it. The net of all seven rows is 351, and the floor moves from 32,506 to 32,857.

Three states have already stopped counting out loud, and the season gave them permission.

Virginia now says at the top of its page that the season runs May 1 to September 1, that its cases have been decreasing, and that starting September 1 it will resume normal case reporting. Its last weekly table is dated August 24, at 234, and the Monday table that would have carried August 31 did not appear. New York said it would update every Wednesday through the end of August. The page still reads last updated August 19, at 939 cases from May 1 through August 17, and the August 26 update never posted. New York City, on its own page, counts 643 cases since May 1 through August 24. Indiana updated its dashboard every weekday afternoon in July. It now says the data is updated every Wednesday. Ohio has published no statewide figure since August 13, when it stood at 5,452, and that is nineteen days of silence from the second-largest state in this outbreak.

None of that is a scandal. Every one of those departments spent the summer doing work it was never staffed for. But a season end is a permission slip, and the pattern is what I said it would be when the count crossed thirty thousand: the number you get depends on who is still willing to print one. The forty-one counted rows in the table below have as-of dates ranging from July 11 to August 28. Every one is a floor, and the older ones are further below the truth than they were a week ago.

August 31 is a line on a web page. The parasite did not read it.

CDC says on its own surveillance page that the season is May 1 through August 31 and, in the next sentence, that clusters of cyclosporiasis have been detected outside that range in some years. The last onset date in the outbreak is August 15. The latest domestic onset on the national page is August 19. CDC says it takes about six weeks from the first symptom for a case to reach Atlanta, which means the September updates will still be filling in July and August. The season definition is what makes 17,180 comparable to last year’s 1,180, because both were counted in the same window. It is not a statement that anybody stopped getting sick.

What to watch this week, as I set out over the weekend: CDC’s Case Data page has moved on Tuesdays, and today is one; the outbreak pages have moved on Thursdays, which puts the next one on September 3; Michigan reports Thursday; West Virginia posts Friday; Missouri publishes on Sundays and its August 30 figure had not reached anyone I can find by this morning. The thing to watch is not which of those numbers moves. It is which of the pages stops.

Two of every three confirmed cases are tied to the lettuce. Nobody has named a product for the rest.

Set the two federal series side by side. CDC has attributed 11,458 illnesses to the recalled Taylor Farms de Mexico lettuce, as of August 27. It has confirmed 17,180 domestically acquired cases, as of August 24. The pull dates are three days apart, so I will say roughly two in three rather than print a percentage. Against the 29,024 cases CDC has either confirmed or is still working through, the outbreak is about four in ten. Against the 32,857 the states themselves publish it is about a third. The remainder is 5,722 confirmed illnesses with no attributed source, and 427 of the country’s 922 hospitalizations sit in that pile. The attribution is not by laboratory. No product sample has ever tested positive, and CDC’s genotyping ties the parasites from sick people to each other, not to a head of lettuce. It rests on interviews and on FDA’s traceback to a single supplier.

No state publishes its own split, and CDC’s first state-level table, published August 27, shows how far the two series sit apart: 4,846 Michigan illnesses attributed against the 14,510 the state reports, 2,679 in Ohio against 5,452, twenty in Texas against 315, two in North Carolina against 940. For the rest, the record is a list of things health departments have said out loud. North Carolina says its cases do not appear tied to the multistate outbreak and that parsley and cilantro were reported more often than expected. Alabama named two clusters at Mexican-style restaurants that it says plainly are not Taco Bell. New York has 939 cases and no named source at all. Hawaii is investigating a case with no travel in a state that never received the lettuce. FDA’s investigation table carries at least six other Cyclospora clusters whose sources are not confirmed. Not one of them has produced a second recall. The Taylor Farms lettuce is the only product recalled for Cyclospora in this country this year.

Taylor Farms de Mexico recalled all iceberg lettuce sourced from central Mexico on July 17, 2026. FDA puts the foodservice distribution at June 29 through July 16 to thirty-one states, and the retail product at Walmart stores in fifteen states under the Marketside label, with best-by dates through August 3. CDC publishes the epidemic curve for this outbreak on its timeline page, and I rebuilt it from CDC’s own data table on August 27 with the recall date marked. It is built on the 9,256 people, of 11,458, who have an onset date on file. Of those, 8,959, which is 96.8 percent, got sick before the recall. The week of July 6 through July 12 alone holds 5,287 people. The single worst day, July 10, holds 1,001.

Illnesses in the multistate Cyclospora outbreak by the date symptoms began. Source: CDC outbreak epidemic curve data table, downloaded August 27, 2026, built on the 9,256 people of 11,458 for whom an onset date is available. Rust bars are onsets before the July 17 recall; blue bars are on or after it. The bars sum to 9,256.

Are people still getting sick since the recall? By onset date, barely. The 297 people whose illness began on or after July 17 are 3.2 percent of the curve, and because the incubation period runs up to two weeks, an onset in the last days of July can still be lettuce eaten before the recall, and product dated through August 3 could lawfully have been in a kitchen until then. The fifteen August onsets are the ones that need explaining, and the last of them is August 15. CDC’s national page shows domestic onsets as late as August 19 and does not say whether any of them belong to this outbreak. What is rising is the count, not the illness. Every number that moved this week was a July illness reaching a database. A falling curve and a rising count are not a contradiction. They are what a six-week reporting lag looks like from the outside.

Here is where all fifty states stand this morning. The Lettuce column is whether FDA has confirmed that recalled Taylor Farms de Mexico lettuce reached that state. The last column is whether anybody has tied that state’s illnesses to Taylor Farms at all; it now reads Linked for twenty states, because CDC added Georgia, Tennessee and Texas on August 27.

State	2026 cases	As of	All of 2025	Lettuce	Taylor Farms
Michigan	14,510	Aug 27	50	Yes	Linked
Ohio	5,452	Aug 13	76	Yes	Linked
Missouri	2,152	Aug 23	none reported	Yes	Linked
Indiana	2,090	Aug 26	27	Yes	Linked
Illinois	1,437	Aug 18	301	Yes	Linked
North Carolina	940	Aug 25	300	Yes	Linked
New York	939	Aug 17	694	Yes	Unknown
Kansas	829	Aug 26	29	Yes	Linked
Kentucky	620	Jul 29	45	Yes	Linked
Oklahoma	544	Aug 18	41	Yes	Linked
Florida	450	Aug 15	210	Yes	Unknown
Iowa	327	Aug 20	64	Yes	Linked
Texas	315	Aug 11	512	Yes	Linked
West Virginia	315	Aug 28	8	Yes	Linked
Arkansas	287	Aug 24	13	Yes	Linked
Virginia	234	Aug 24	115	Yes	Unknown
Nebraska	218	Aug 6	51	Yes	Linked
Wisconsin	216	Aug 5	65	Yes	Unknown
Colorado	150	Jul 17	205	No	Unknown
Massachusetts	145	Aug 5	74	Yes	Linked
New Hampshire	94	Aug 25	7	Yes	Linked
Alabama	85	Aug 3	19	Yes	Unknown
Pennsylvania	82	Jul 29	not notifiable	Yes	Linked
Washington	76	Aug 28	not available	May have	Travel link
Maryland	69	Jul 17	123	Yes	No link found
New Jersey	46	Jul 11	180	Yes	Unknown
California	41	Jul 14	not available	May have	Unknown
Minnesota	41	Jul 17	none reported	No	Unknown
Connecticut	35	Jul 17	41	Yes	Unknown
Maine	27	Aug 7	4	Yes	Linked
Oregon	23	Jul 24	not available	No	Travel link
Arizona	19	Jul 17	49	No	Unknown
Georgia	11	Jul 17	116	Yes	Linked
Tennessee	11	Jul 17	38	Yes	Linked
Mississippi	7	Jul 22	not notifiable	Yes	Unknown
Hawaii	6	Aug 26	not available	No	No link found
Alaska	5	Jul 16	6	No	Unknown
Rhode Island	4	Jul 17	6	May have	No link found
South Dakota	3	Jul 16	3	May have	Unknown
Louisiana	1	Jul 17	99	Yes	Unknown
Utah	1	Jul 17	37	No	No link found
Delaware	no count published	—	1	May have	Unknown
Idaho	no count published	—	not notifiable	No	Unknown
Montana	no count published	—	8	No	Unknown
Nevada	no count published	—	not notifiable	No	Unknown
New Mexico	no count published	—	9	No	No link found
North Dakota	no count published	—	4	May have	Unknown
South Carolina	no count published	—	45	Yes	Unknown
Vermont	no count published	—	3	May have	Unknown
Wyoming	no count published	—	0	No	Unknown
All fifty states	32,857			31	20

Counts as published by each state health department, compiled from the linked sources on September 1, 2026. Forty-one states publish a 2026 count and those forty-one sum to 32,857, which is 351 above the 32,506 published August 26; the seven moves are Michigan, Indiana, Kansas, West Virginia, Washington, New Hampshire and the Florida correction. Nine states publish no count at all. Florida is shown on its May 1 series; its calendar-year count is higher. Every figure is a floor.

The questions I am getting this week.

How many of the cases are linked to Taylor Farms? CDC has tied 11,458 illnesses to the recalled lettuce, against 17,180 laboratory-confirmed domestic cases, so roughly two in three confirmed cases and about a third of the 32,857 the states publish. The remaining 5,722 confirmed illnesses have no attributed source. The link rests on interviews and traceback; no product sample has tested positive.

Are any of the Taylor Farms cases outside Taco Bell, and if so, where? Yes. Since August 5, CDC’s definition has included anyone who reported exposure to the recalled lettuce served or purchased at various locations, not only Taco Bell, and the outbreak count went from 1,947 to 6,358 in the twelve days after that change. CDC has never published the split. The named places on the public record are few. New Hampshirehas tied 43 of its 94 cases to a recalled product, and 39 of those to Taylor Farms lettuce served at the Cheshire Medical Center cafeteria salad bar in Keene between early June and July 20. The hospital counts 34 cases on its own laboratory, 21 of them employees, and declared its cluster over on August 26 after two weeks without a new case. It is the only location in New Hampshire tied to an outbreak, and the only named non-Taco Bell establishment in the country. Maine tied ten of its 27 cases directly to the recalled lettuce without naming where it was served, in a state that is not on the distribution list at all. Missouri’s federal description includes people who handled the lettuce at work, not just people who ate it. Washington’s three outbreak cases all ate at Taco Bell. The other customers on the recall record, Subway, Sysco, US Foods, Jack in the Box and Walmart’s Marketside bags, have had no illness publicly tied to them by any agency.

Are the other cases linked to any other product? No product has been named. North Carolina points to parsley and cilantro, Alabama to two restaurant clusters that are not Taco Bell, and FDA lists at least six other clusters with sources not confirmed. The Taylor Farms lettuce is the only product recalled for Cyclospora in the United States this year.

When was the recall, and what did it cover? July 17, 2026. Taylor Farms de Mexico recalled all iceberg lettuce sourced from central Mexico. Foodservice product went to thirty-one states between June 29 and July 16, and Marketside-brand bags reached Walmart stores in fifteen states, with best-by dates through August 3.

Are people still getting sick since the recall? By onset date, almost none. Of the 9,256 outbreak cases with an onset date, 8,959 got sick before July 17. Fifteen have an August onset and the last is August 15. The count keeps rising because it takes about six weeks for a case to reach CDC, not because the lettuce is still out there.

Is the outbreak over now that the season is? The exposure appears to have ended in mid-July. The season is a counting window, not a finding. CDC and FDA both still call the investigation ongoing, and the counting will run into the fall.

Do I still need to worry about lettuce? The recalled product is long past its dates and FDA says it is confident it is off the market; Taco Bell stopped using that supplier on July 17. No other lettuce has been recalled in this outbreak. Wash produce, cook what you can, and know that washing does not reliably remove this parasite.

I got sick in August and never got tested. Am I too late to count? No. Call your doctor and ask specifically for Cyclospora testing; a routine stool panel does not look for it, symptoms relapse for a month or more, and it is treatable. A case confirmed in September is still a 2026 case, whatever the season calendar says.

A season is a bookkeeping convention. It tells CDC when to start a comparison and when to stop one, and it lets a state health department decide the extra work is finished. It does not tell a parasite anything. Two people are dead, 922 have been hospitalized, and the state health departments that were willing to print a number this summer printed 32,857 of them. Some of them stopped printing this week. I will keep adding them up for as long as any of them keeps publishing, and I will say so when they stop.

FDA’s Own ByHeart Botulism File: 48 Babies, 17 Strains, a Death Complaint, No Root Cause, and Six Company Names Blacked Out

By Bill Marler on August 31, 2026

Posted in Case News

The Food and Drug Administration’s outbreak response network, CORE, closed the ByHeart infant botulism investigation with an Executive Incident Summary dated May 1, 2026, filed as CARA #1350. It counts 48 infants in 17 states, 28 confirmed and 20 probable, with illness onsets from December 24, 2023 to November 29, 2025, everyone hospitalized and treated with BabyBIG. Its conclusion is a sentence the public has not seen stated this plainly: “Even though there are several hypotheses, investigational findings could not identify the source or root cause of contamination of the powdered infant formula.” I received it under FOIA request 2026-4033. It is heavily redacted. What survived is worth reading. What was removed is worth asking about.

The organism was found from the dairy dryer to the sealed can, and FDA still cannot say where it started.

FDA collected 26 samples. Two were positive for Clostridium botulinum type A: a base-mix retain from Blendhouse Allerton in Iowa, and a retain of powdered milk from Dairy Farmers of America in Fallon, Nevada. Five samples run at New York’s Wadsworth Center produced two more positives, an unopened can from Arizona, lot 251481P2, and a second DFA milk-powder retain. The California Department of Public Health found it in an opened can, lot 251131P2. ByHeart’s own laboratory reported 12 positives across finished product, base powder and organic whole milk powder, including toxin type B in finished lot 252161P2 and both A and B in the milk powder. The 35 opened consumer cans CDC tested were all negative.

Sequencing sorted 37 isolates, 16 from infants and 21 from product, into 17 distinct strains. Strain 11 ties a DFA milk-powder retain, ByHeart’s milk powder, ByHeart finished formula, the unopened Arizona can and a sick infant in New Jersey at a mean of 7.5 SNPs, the cluster FDA first described publicly in February. Strain 2 ties a second DFA retain to finished lot 251261P2, the lot seven infants reported being fed. Strain 8 ties a California infant to ByHeart product and base powder; strain 6 ties three infants to the opened California can. In the report’s words, “the distribution and comingling of ingredients likely contributed to the widespread nature of the contamination.” A full preventive-controls inspection at the DFA dryer “did not find any objectional findings.” Of 12 field assignments, one produced a Form 483, at Allerton on January 22, 2026: no qualified individual documenting corrective actions, food-contact surfaces not shown to be cleaned and sanitized, and process-specification deviations not evaluated for public health significance. Portland got none.

Three matches the report does not explain. Strain 11 also matches a 2022 isolate from broccoli and cheddar soup. The Washington infant’s isolate, strain 4, matches a 2011 isolate from Aichi, Japan at one SNP and nothing in the product. Strain 1, in finished formula and FDA’s base mix, is an “inconclusive match” to a 2015 Ohio potato salad cluster.

FDA had been inside the Allerton plant nine months before the outbreak and wrote it up.

The Incident Briefing lays out the inspection history. Allerton was inspected in February 2025 and classified Voluntary Action Indicated, with a three-item 483: an ingredient, its name redacted, “not held under conditions to prevent adulteration”; failure to eliminate harborage “when issues with rodents arose during the year 2024-2025”; and floor conditions not adequately monitored “when there were findings of confirmed Cronobacter sakazakii.” A 2022 inspection at the same plant cited insanitary conditions and inadequate written preventive controls and was followed by a regulatory meeting. The Reading, Pennsylvania plant drew a Warning Letter in August 2023, and an Official Action Indicated inspection that December, and has not run since September 30, 2023. Portland, inspected in March 2025, drew no action. I have written about this history before. The briefing adds the number that should have moved things: ByHeart held roughly 1 percent of powdered formula sales, yet of 36 BabyBIG-treated infants with any powdered formula exposure between August 1 and November 10, 2025, 15 had been fed ByHeart. FDA’s analysts called that “highly unusual” and “a significant epidemiological signal.”

The log records a death complaint the final count does not mention.

The Incident Objectives entry for November 28, 2025 reads in full: “CORE received a new consumer complaint – a second death reported FL case – case reported consuming ByHeart formula and passed away August 31st. No medical records / death certificate yet.” Nothing in the released file returns to it. The summary reports that “none died,” and Florida is not among the 17 states. It may have been resolved as no exposure or no botulism. The record as released shows the complaint arriving and never shows it closing. It also calls this a “second” death report, and the first appears nowhere. Families who fed this formula are entitled to know how both were run down.

Six companies received the contaminated milk powder lots, and every name is blacked out.

On January 28 and February 3, 2026, CORE sent six traceability assignments to firms that had received Organic West Milk lots of whole milk powder that tested positive for C. botulinum, asking for all distribution to infant formula manufacturers over the prior two years. The result: “None of the firms identified any infant formula manufacturers as customers that were supplied whole milk powder.” All six company names are withheld under Exemption 4 as confidential commercial information. Nara Organics had been selling formula through Target and its own site since July 2025, inside the window FDA asked about. Eight weeks after the last assignment closed, Nara-fed infants in California, Pennsylvania and Washington began falling ill, and on August 27 FDA reported that a Nara infant’s isolate matches a ByHeart clinical sample, a ByHeart base-powder sample and four unopened ByHeart cans, and that the shared supplier and dryer “indicate that there is a common source of contamination between the two outbreaks.” I wrote in June that an incomplete customer list left Nara off during the ByHeart trace-forward. Whether any of the six sold to Nara or its contract manufacturer is the question this file could answer and does not.

Also withheld under Exemption 4: every base-mix and milk-powder lot number; the number of lots traced at each step back to the 22 dairies; the region those dairies sit in; and three of the retailers that carried the product.

Every objective the response team set is withheld as deliberative.

The Incident Objectives form for Operational Period 7, February 17 to March 27, 2026, has its objectives box redacted in full under Exemption 5. The Incident Briefing’s “Current and Planned Objectives” page is redacted in full the same way. A separate block, withheld under Exemptions 6 and 7(D), personal privacy and confidential source, appears twice directly after the description of the plants. Exemption 5 is discretionary. On an outbreak that hospitalized 48 babies and is now genetically joined to a second, the objectives the response team set for itself are the part of this file the public has the strongest claim to read.

What the public should take from this file.

Five things. The Health Hazard Evaluation signed December 5, 2025 rated the hazard “life-threatening” and “likely to occur,” and no root cause was ever found. The contamination was in the incoming milk powder, from a dryer that serves more than one formula brand, and the file will not say where else those lots went. CDC expanded the case definition on December 5, 2025 back to March 23, 2022, the first day of production, and five days later added ten confirmed infants sickened between December 2023 and July 2025, none earlier. ByHeart told FDA one of its positive lots carried a November 15, 2024 production date, yet the recall reached only lots still within expiry: the 2023 and 2024 babies were fed formula that was never recalled. FDA had 70 consumer complaints by November 20, 2025, and had to send warning letters to four retailers on December 12 because recalled formula was still on shelves. And FDA’s Nara page, updated August 27, says its root cause investigation is ongoing, and it cannot yet tell whether the milk was contaminated before or during drying. The next day ByHeart announced FDA has “no further questions” on its root cause submission, a submission the public has not yet seen.

What the public should still see from this file.

FDA should release the objectives pages and the six company names; a supplier’s interest in keeping its customer list private does not outweigh the interest of parents in knowing what that supplier’s milk went into. FDA should say how the Florida death report, and the one before it, were resolved. ByHeart should release the submission it is citing. Fifty-two babies across two brands were hospitalized on the same supply chain, and the t-shirts go to Congress in September because the rule that would require testing this ingredient for this organism still does not exist.

Full Report:

https://www.marlerblog.com/files/2026/08/FOIA-2026-4033_Redacted.pdf

The Alfalfa Seed Behind Fifty-Five Illnesses Went to at Least Two Sprouters. FDA Asked the Seed Supplier to Recall It, and the Supplier Has Not Responded. Nobody Will Name the Supplier or the Second Grower.

By Bill Marler on August 31, 2026

Posted in Case News

As of August 31, 2026, 55 people in 15 states are sick with Salmonella Agona or one of three strains of Shiga toxin-producing E. coli after eating alfalfa sprouts, four of them hospitalized, none dead, with illness onsets running May 31 through August 8. Until today the only company in this outbreak was Everything Sprouts, LLC of Minneapolis, which recalled its Calco and Everything Sprouts alfalfa on August 22. This afternoon FDA updated its advisory to say those sprouts are “one of the sources of illnesses” and “do not explain all the illnesses,” that at least one other grower sprouted the same lot of alfalfa seed, and that the seed supplier has not responded to FDA’s request to recall it.

FDA has moved this outbreak from one sprouter to one seed lot.

Eleven days ago I wrote that four organisms in one bag of sprouts is a seed signal, not a kitchen signal, because a leaking roof gives you whatever is in roof water and does not plausibly deliver Salmonella and three separate STEC serogroups into one product in five weeks. FDA’s August 31 update says its traceback “has identified at least one additional sprout grower who received and used the same lot of alfalfa sprout seeds under investigation in this outbreak from a common seed supplier,” and that the agency “is continuing to investigate contamination of the implicated alfalfa sprout seed.” That is the agency’s phrasing, not mine.

It also explains the map. Minnesota and Wisconsin hold 38 of the 55 cases, and Everything Sprouts shipped only to those two states. The other 17 sick people live in 13 states, ten of them with a single case, from Florida to Washington to New Hampshire. A second grower with a different distribution footprint is the simplest account of those 17, and FDA has now said the Everything Sprouts product cannot account for them. Whether the second grower sits in one of those 13 states is not something FDA has said.

The seed supplier told its customers to stop using the lot. FDA asked for something else, and the supplier has not answered.

According to FDA, the supplier “has instructed its customers not to use, grow, sell, transfer, or redistribute the implicated seed lot, and to return or dispose of any remaining seed.” That is a stock recovery letter. It reaches seed sitting in a sprouter’s warehouse. It does not reach the sprouts already grown from that seed and already sold.

FDA asked for both. The advisory says FDA “recommended that the common seed supplier voluntarily recall the implicated lot of alfalfa seeds and contact their downstream customers to initiate recalls of sprouts grown from the implicated seed lot.” Then FDA wrote the sentence that prompted this post: “At this time, the common seed supplier has not responded to FDA’s recall recommendation, has not sent a recall notice to downstream customers, and has not instructed the customers who received the implicated alfalfa sprout seed lot to initiate recalls.”

Three negatives. No response, no notice, no downstream recall. I will hold the line I held on August 21 about Everything Sprouts: “has not responded” is what FDA wrote, and it is not the same word as refused. Refusal is a finding under section 423 that only FDA can make. Silence is a fact, and it is the fact on the record tonight.

What a recall produces that a customer letter does not is the point I made when Everything Sprouts was in this exact posture ten days ago: a lot number, a distribution list, notification down the chain, an entry in FDA’s enforcement report, a press release. A seed supplier’s customer list is the list of every sprouter that grew this lot. Its customers’ customer lists are the list of every store and restaurant that sold those sprouts. Neither list exists in public today, and the second one cannot be assembled until the first one is.

The second grower is not named, and its sprouts have not been recalled.

FDA named Everything Sprouts on August 21, two days after its investigators walked into the plant and before the company had recalled anything. It has not named the seed supplier, and it has not named the second grower. The advisory’s consumer advice reads, in full, “Do not eat, sell, or serve recalled sprouts.” Nothing grown by the second grower has been recalled. A person in South Carolina or Washington who bought alfalfa sprouts this week has been told nothing she can act on, because the only brand names on the federal pagesbelong to a company that never shipped to her state.

I understand why an investigation holds a name while traceback is being confirmed. That reason ran out on August 31, when FDA itself published that the grower “received and used” the lot. The grower is identified. The seed is implicated. The public is the only party still in the dark.

FDA has been able to order this recall since 2011. Seed is food.

Section 423 of the Food, Drug, and Cosmetic Act, added by the Food Safety Modernization Act in 2011, lets FDA order a recall when there is a reasonable probability that a food is adulterated and that exposure to it will cause serious adverse health consequences or death. FDA’s own recall guidance lists alfalfa sprouts contaminated with Salmonella as its example of that risk. The agency has needed to issue a mandatory recall order once, in April 2018, against a kratom seller.

Seed for sprouting is not “covered produce” under the Produce Safety Rule, and a supplier might be tempted to read that as an exemption. FDA closed that door in 2022 when it issued its seed guidance: seed used for sprouting is food under the Act. Section 423 applies to food. The supplier also has its own duty under the Reportable Food Registry to report, within 24 hours, that it holds a food with a reasonable probability of causing serious harm, and that duty does not wait for a recall.

The statute’s steps are short. A written opportunity to recall voluntarily. If the firm does not, an order to cease distribution, with an informal hearing within two days. Then a recall order, which only the Commissioner can sign. Kyle Diamantas has been acting Commissioner since May. FDA’s seed guidance says contaminated seed has been identified as the likely source of most sprout outbreaks since the 1990s, and it describes a 2016 outbreak in which the entire seed lot was recalled. The investigators who found this common lot did their job. What happens next is not their decision.

The radish mix expansion now has a stated reason, and it shipped four days past the alfalfa recall.

Everything Sprouts expanded its recall on August 28 to Robust Radish Sprout Mix in 5-ounce cups, lots 223, 226, 230 and 233. The company’s notice gave no reason. FDA’s update does: “the potential for cross-contamination” from the recalled alfalfa. Lot 233 appears in no earlier notice. The radish mix was distributed from August 13 through August 26, which is four days after the company recalled its alfalfa on August 22. Food Poison Journal carries the lot numbers for all four products.

Here is where the outbreak stands on the last day of August. The seed is implicated, in FDA’s own words. A second grower used it and is unnamed. Its sprouts are unrecalled. The supplier has been asked for a recall by the agency that can order one, and has not answered. Every one of those sentences can change tomorrow with a recall notice, a name, or a signature. Fifty-five people, four of them in the hospital, are waiting on one of the three.

FDA Testing at Happy Hens Came Back Positive for Salmonella. The Farm Announced It on Instagram. The Agencies That Have the Results Have Said Nothing.

By Bill Marler on August 31, 2026

Posted in Case News

Happy Hens, the Ramona poultry farm San Diego County named as the source of the eggs in the SalmonellaEnteritidis outbreak that has sickened twenty-three people and hospitalized seven, announced on its own Instagram account on Sunday, August 31, 2026, that FDA testing has returned two samples positive for SE, that all product from the affected sites is being held from sale, that four thousand eggs will be tested over the next two weeks, and that if more positives are confirmed, approximately ten thousand hens may need to be depopulated. As I write this, not one of the agencies that has those results — not FDA, not San Diego County, not the California Department of Food and Agriculture, not the California Department of Public Health — has published a word about them.

The regulated party announced the regulator’s test results.

The announcement came twice on Sunday: a video from the farm, and a written post on the @trulyhappyhens account headlined “FDA TESTING RESULTS ARE IN.” Screenshots of both, captured August 31, are below. The written post says the farm pulled its own samples before the outbreak was public and those were negative; that FDA has since conducted additional testing and two samples tested positive for SE; that product from the affected sites is held from sale pending more testing; and that a positive follow-up could mean depopulating roughly ten thousand hens. FDA’s CORE investigations table, last updated August 26, has no entry for this outbreak. The County’s newsroom still shows only its July 28 bulletin and its August 17 update. The only place the public can learn that federal testing at this farm came back positive is the farm’s own social media feed, with the farm’s own spin attached.

Figure 1: The Happy Hens Instagram video, August 31, 2026 — two FDA samples positive for SE, product held from sale, 4,000 eggs to be tested, roughly 10,000 hens possibly depopulated.

The numbers in the farm’s announcement are the federal egg rule running exactly as written.

Happy Hens did not invent the four-thousand-egg figure. Under FDA’s shell egg rule, 21 C.F.R. Part 118, a covered producer whose environment tests positive for Salmonella Enteritidis must begin egg testing — four samplings of one thousand eggs each — and must divert eggs away from the table market to pasteurization until the flock tests clean. “Affected sites,” “affected barns,” a hold on sale, and a 4,000-egg testing plan is that machinery, described almost verbatim. The plain reading is that FDA’s two positives were environmental samples, taken house by house, and that the farm now knows which houses. Depopulation is the farm’s own word, not the rule’s — the rule requires diversion — but ten thousand hens out of a flock its owner has put at roughly fourteen thousand tells you the positives are not confined to one corner of the operation. And FDA’s own guidance on the rule says the agency considers eggs containing SE adulterated and expects most positive egg tests to end in a recall.

Figure 2: The written post — “No recall has been issued” offered as reassurance, and the family “still here to tell the tale” after eating eggs from the affected barns.

This is the same account that spent two weeks calling the link impossible.

Happy Hens denied the connection on Instagram when the County named it on August 17. A co-owner told Fox 5 the family had shrugged off the investigation expecting “all negative reports,” and said the farm had just pulled its own swabs. She told another station she was confident enough in the eggs to drink them raw. A video followed saying the farm followed Salmonella prevention protocols. Sunday’s post now says the farm’s samples were pulled before anything was publicly known — which is not the sequence the owner described on camera. Both accounts cannot describe the same testing, and the difference matters, because “our tests were negative” is the entire public defense. It is also less than it sounds. SE is shed intermittently and contaminates the egg from inside the hen, so a clean swab in August says nothing about the eggs laid in June that made people sick. That is precisely why the rule treats a single positive the way it does.

The farm’s “facts” list does not match the County’s own numbers.

The post says eighteen of twenty-one reported cases are “directly associated with a kitchen in Ramona.” The County’s current accounting, reported August 21, is twenty-three cases split two ways: eighteen tied to the raw-egg mayonnaise at Ramona Family Naturals Market, and five in people sickened by Happy Hens eggs directly, with no connection to the deli. The farm kept the stale total and dropped exactly the five cases that implicate the farm rather than the kitchen. “No recall has been issued” is likewise offered as reassurance. The accurate sentence is that eggs from barns where FDA has now found Salmonella kept moving into commerce in three counties until this hold.

A wink emoji, and seven people in the hospitalization count.

The post closes with the family still eating eggs from the “affected” barns “and we’re still here to tell the tale,” followed by a wink. Seven people in this outbreak were not here to tell it from home — they told it from hospital beds, one of them still admitted in mid-August. My firm filed suit on August 18 for one of them, a thirty-six-year-old Ramona mother of three who ate a deli sandwich she did not know contained raw eggs.

The post’s last line is that integrity means showing up when things are hard. On the disclosure record, the farm has a point — just not the one it intended. In the December 2025 Vega Farms egg outbreak, CDPH issued its own press release, sampled the farm, matched a sample to the patient strain by whole genome sequencing, and the eggs were recalled. In the August 2025 Country Eggs outbreak, the health agencies announced the link themselves and a nationwide recall followed. In this outbreak, the pattern I wrote about two weeks ago has not changed: two County releases, no serotype details, no customer list, no sampling results — and now federal test results reaching the public through the defendant’s Instagram. What should happen this week is not complicated. FDA or the County should publish what was sampled, which houses were positive, and whether the isolates are being sequenced against the outbreak strain, and the County should release the list of stores that received these eggs. If the egg testing comes back positive, FDA’s own guidance calls those eggs adulterated and expects a recall. The farm says it will be transparent every step of the way. The government should not be less transparent than the defendant.

CNBC Asks How Taylor Farms Became Integral to America’s Food Supply. Seventeen Years of Outbreaks Are Part of the Answer.

By Bill Marler on August 31, 2026

Posted in Case News

CNBC’s Brandon Gomez published a piece on Friday on how Taylor Farms became integral to America’s food supply — roughly $7.3 billion in sales last year, 40 percent of the salad kits sold in American grocery stores, and a customer list that runs from Walmart, Kroger, Whole Foods and Target to McDonald’s, Taco Bell and Chipotle. The scale is real, and CNBC’s central point is the right one: when a supplier this size has a problem, the problem has the supplier’s reach. What the piece compresses into a paragraph is the record behind that point. Since 2009, eight multistate outbreaks have traced to Taylor Farms product, and in three of the largest, no recall ever reached the public.

The scale CNBC describes is national infrastructure.

Taylor Farms was founded in 1995 and grew by acquisition into one of the world’s largest fresh-cut produce companies, handling much of the chain itself — washing, chopping, packaging, labeling, refrigerating and shipping what hundreds of farms grow. Chains buy from Taylor Farms because it can deliver consistent product at a volume almost no one else can. That is precisely what makes the company something more than a vendor. A supplier moving 40 percent of the country’s grocery salad kits is infrastructure, and this summer showed what happens when infrastructure fails: Walmart recalling bagged salads, Taco Bell pulling lettuce at restaurants, and a parasite counted in every corner of the country.

Eight outbreaks in seventeen years is also part of how the company became integral.

Credit to Gomez for doing what too few business stories do — going to the record. His piece cites the list I compiled of every verifiable Taylor Farms outbreak and whether each one led to a recall, and notes the five complaints Marler Clark has filed so far in this outbreak, in Ohio, Michigan and Kentucky. The chronology in that post: Salmonella Typhimurium in shredded lettuce in 2009; Cyclospora in 2013, with 631 people sick across 25 states and a traceforward running through the same Guanajuato growing region at issue today; E. coli O157:H7 in celery in Costco chicken salad in 2015; E. coli O121 in romaine served at Panera and Chipotle in late 2021; a Cyclospora cluster at a single Colorado restaurant, Taco del Gnar, in 2023; E. coli O157:H7 on slivered onions served on McDonald’s Quarter Pounders in October 2024; E. coli O157:H7 in romaine in November 2024; and Cyclospora in iceberg lettuce in 2026. Alongside the outbreaks sits a recall roster — Salmonella, Listeria, E. coli, undeclared allergens — stretching across more than a dozen Taylor Farms corporate entities.

The gap between the outbreak list and the recall list is the part CNBC did not have room for.

Three of the largest outbreaks on that list — 2013, the 2021 romaine cluster, and November 2024 — produced no recall at all. Two more produced direct notice to foodservice customers and nothing the public could read. The two outbreaks that killed someone are also the two where the public was never given a recall notice. And 2026 added a third category: a recall that arrived on time, in public, at enormous scale — FDA’s own enforcement file puts it at 236,192 cases across products well beyond iceberg lettuce — and still could not reach food that had already been eaten, because Cyclospora’s incubation period outruns the shelf life of a salad.

The $200 million food safety claim has to sit next to the outcome.

Taylor Farms told CNBC it spends more than $200 million a year on food safety, its biggest area of funding by several orders of magnitude. Taken at face value, that is under three cents of every sales dollar — but the honest measure of a food safety program is never the spend. It is the outcome. As of FDA’s August 27 update, the agency counts 11,458 cases across 20 states with 495 hospitalizations, and on August 28 it classified the recall Class I. CDC’s August 25 update counts 17,180 confirmed cases, 11,844 more unconfirmed, 922 hospitalizations, and illness in 48 states plus the District of Columbia. My own state-by-state tally puts the floor above 32,500. CDC’s surveillance season for cyclosporiasis runs May 1 through August 31 — it closes today, and this outbreak filled it.

CNBC also notes, correctly, that FDA has never reported a positive product sample — symptoms take weeks to appear and lettuce is long gone by the time anyone looks. That is not a point in the company’s favor. It is the reason you cannot test your way out of Cyclospora. Prevention at the source — clean water, sanitation, the water testing Taylor Farms promised at this same plant in 2013 — and traceability fast enough to matter are the only tools that work.

Consolidation makes transparency and traceability non-negotiable.

None of this is an argument that big is bad. It is an argument that big carries obligations proportionate to reach. When one company’s product touches nearly every major grocer and restaurant chain in the country, its food safety failures are national events, and the rules should treat them that way: full FSMA 204 traceability implemented without further delay, a public recall notice every time contaminated product ships — not quiet direct notice to foodservice accounts — and the names of companies named while people can still act on them. Taylor Farms has now suspended iceberg sourcing from central Mexico and commissioned an outside review of the facility there. Both steps are welcome. Both are also admissions that the company’s scale demanded more than the system ever required of it.

CNBC asked how Taylor Farms became integral to America’s food supply. The answer includes the acquisitions, the refrigerated trucks, and forty percent of the salad aisle. It also includes seventeen years of outbreaks — and a recall record that too often left the public reading nothing at all.

Marler Blog

Marler Blog

Shoppers Are Walking Past the Produce Aisle. Telling Them We Have the Safest Food Supply in the World Will Not Bring Them Back.

Three Federal Outbreak Pages, One Evening: CDC Is Working 40, FDA Lists 19, USDA Lists None

The Atlantic Says Our Summer of Contamination Could Have Been Much Worse. It Is Right. Put E. coli O157:H7, Salmonella, or Listeria on the Same Lettuce and the Arithmetic Is Thousands in the Hospital and Hundreds Dead.

Salmonella: about 5,000 hospitalizations and 75 to 130 deaths

E. coli O157:H7: roughly 7,400 hospitalizations and more than a thousand cases of kidney failure

Listeria does not need the case count

It could absolutely have happened

The lag does not care what the bug is

Four E. coli Cases Tied to Dungeness Valley Creamery Raw Milk. No Recall. No Headlines.

Harvard Read the FRESH Act Line by Line. Its Preemption Clause Reaches Salmonella, Restaurant Inspections, and the Right to Sue.

The Cyclospora Season Closed Yesterday With 17,180 Confirmed Cases, Two Deaths and a Fifty-State Count of 32,857. The Calendar Stopped. The Parasite Did Not, and Three States Have Already Stopped Publishing.

FDA’s Own ByHeart Botulism File: 48 Babies, 17 Strains, a Death Complaint, No Root Cause, and Six Company Names Blacked Out

The Alfalfa Seed Behind Fifty-Five Illnesses Went to at Least Two Sprouters. FDA Asked the Seed Supplier to Recall It, and the Supplier Has Not Responded. Nobody Will Name the Supplier or the Second Grower.

FDA Testing at Happy Hens Came Back Positive for Salmonella. The Farm Announced It on Instagram. The Agencies That Have the Results Have Said Nothing.

CNBC Asks How Taylor Farms Became Integral to America’s Food Supply. Seventeen Years of Outbreaks Are Part of the Answer.

Blogs

Disease

Insight & Litigation