Testing a necessary evil of food safety: FDA’s microbiological surveillance sampling

As part of the U.S. Food and Drug Administration’s risk-based and preventive approach to food safety, which is at the core of the FDA Food Safety Modernization Act, the agency began developing a new, more robust surveillance sampling approach in 2014. As the agency moves forward with this approach, it will continue to refine procedures based on lessons learned. The goals of the surveillance sampling are to keep contaminated products from reaching consumers and to facilitate a greater understanding of hazards.

food.lab.testingFDA will publish information regarding test results on the web, including total number of samples collected/tested, and collection date, sample type, and pathogen detected for positive samples.

The Sampling Approach

Under the new sampling approach, the FDA is collecting a statistically determined number of samples of targeted foods over a shorter period of time—12 to18 months—to ensure a statistically valid amount of data is available for decision making.  The sampling approach will help the FDA determine if there are any common factors among positive findings such as season, region, and whether the product was produced domestically or imported.  The FDA’s past approach to microbiological surveillance sampling has been to collect a relatively small number of samples for many different commodities over many years. 

The sampling design for each food represents what U.S. consumers are likely to find in the marketplace. Accordingly, the agency has considered the volume of the target food that is imported and produced domestically and the number of states/countries that produce the target food.

During the first year of this new effort, the FDA focused on sprouts, whole fresh avocados, and raw milk cheese (aged 60 days). The FDA collected more than 800 samples total and tested them for Salmonella, Listeria monocytogenes and E. coli O157:H7. For fiscal year 2016, The FDA will sample and test cucumbers and hot peppers for Salmonella and E. coli O157:H7, taking 1,600 samples of each commodity. The agency also plans to test hot peppers for Shiga toxin producing E. coli. The FDA will conduct whole genomic sequence testing on any samples that test positive. In the future, the number of samples collected of a targeted commodity may vary, depending on the question(s) the FDA intends to answer. Data from the sampling assignments already conducted will be released soon.

The FDA will evaluate the data or results generated throughout the sample collection period and use the data to inform the agency’s short and longer term decision making. By developing these data sets, the FDA seeks to identify potential vulnerabilities and ways to enhance the food safety system.

Depending on the results, the FDA may react or take certain steps, such as:

  • Decreasing sampling, if few positive samples are obtained;
  • Implementing more targeted sampling if trends are identified; for example, if positive samples come from a specific geographic region, a specific facility, or during a particular season;
  • Follow-up inspections;
  • Working with state or international regulatory partners to take corrective actions and implement preventive controls;
  • Developing new or enhanced industry guidance; and
  • Conducting outreach and information sharing to better protect consumers.


A labelling mess and a technology fix, turkey edition

My latest column for Texas A&M’s Center for Food Safety:

I have this weird affliction (among many): The more I read about a food involved in an outbreak, the more I crave it.

mr-bean-turkey(6)Mad cow disease, I want beef

Salmonella in eggs; I want an omelette

WHO cancer report? Had a steak the next day, and gave the kid a salami sandwich for lunch.

Salmonella in peanut butter? Won’t go there, never liked peanut butter.

The point is that crises or occasions are opportunities to get compelling food safety information into the public discourse.

Unfortunately, most of it sucks.

The U.S. glutton-fest known as Thanksgiving, which kicks off the six-week shopping orgy until Christmas, has appeared on calendars again.

As you do.

And simultaneously, the U.S. Food and Drug Administration has finally approved genetically engineered salmon that has been in the works for over a decade (or two, I can’t keep track).

This has sparked a call for labels on all things genetically modified (I prefer engineered, all food is genetically modified).

FDA says, there’s no legal requirement for companies to label foods as genetically modified.

turkey.headAs you do.

Because FDA’s job is to regulate based on safety, not on consumer whims.

If retailers and consumer groups want to make a fuss, go ahead.

But your arguments suck.

I’ve always been a fan of full disclosure whether it’s labeling, point-of-sale info, a web url, provide full information on how food is produced.

Most people don’t care, but some do, and they can make a lot of noise.

When we sold genetically-engineered and conventional sweet corn and potatoes at a local market in Ontario (that’s in Canada) back in 2000, people preferred the GE stuff – because it required no pesticides.

The more info the better – for those who care.

With turkeys, consumers are, according to NPR , inundated with labels: natural, fresh, no hormones, young, premium and so on.

Fresh has nothing to do with the time between slaughter and sale. Instead, it means that the turkey has not been cooled to below 26 degrees Fahrenheit. In other words, it was never frozen.

The U.S. Department of Agriculture does not define young for turkeys, but it requires that turkeys that lived more than a year be labeled as yearling or mature.

USDA says natural means no artificial ingredients have been added to the turkey meat, and the meat is only minimally processed.

Free-Range are raised in the standard, crowded houses but have access to the outdoors.

Premium means nothing.

No Hormones Added means nothing: By USDA law, turkeys (and other poultry) are not allowed to be given growth hormones.

And so it goes.

A possible fix is using smart phones and QR codes, so those who care can find out everything – and I mean everything, including if the seed was derived from radiation mutagenesis, a primal form of genetic engineering – if they want.

Meanwhile, we have enough food safety idiots practicing the things that actually make people sick.

During a cooking segment on the Today show this month, Matt Lauer handled an uncooked turkey, wiped his hands with a towel, then grabbed a piece of the cooked turkey that was sitting nearby and gobbled it down.

The tweets said, “Enjoy Salmonella for the next 24 hours, idiot,” and “We were screaming at the television set. Did you not hear us?” Lauer apologetically explained all of this on the next day’s show.

Other holiday tips:

Do not wash turkey.

Do not place a whole turkey over your head.

Do not pass babies with leaky diapers around the table.

In 2005, one American recalled how, when dessert arrived, the family started passing around the newborn baby. As recounted on the Internet site, fark.com, “Apparently, the baby had a pretty full diaper, and it was kinda leaking. He was passed to my uncle, and then passed to someone else. What my uncle didn’t notice was that a little something rubbed off of the baby as he was passed. He looks down on his tie and sees what he believes is some pumpkin pie filling, so he scrapes it off, and takes a bite. He spent the rest of the night in the back yard throwing up.”

We’ll be having turkey and duck with friends on the weekend. It’ll be safe.

Dr. Douglas Powell is a former professor of food safety who shops, cooks and ferments from his home in Brisbane, Australia.


Top-10 turkey questions the Butterball hotline has ever been asked

Butterball’s Talk-Line has helped confused cooks with Thanksgiving turkey prep since its inception 35 years ago — and while the service has successfully churned out thousands of responses to common questions, which the company so

botterball.hotline1.So I’m looking at a turkey from 1969 sitting here in my father’s freezer … any tips on the best way to cook a 30-year-old bird?

The Talk-Line suggested the man throw out the old turkey and purchase a new one. Then, the Talk-Line suggested to cook the turkey in the open roasting pan method.

  1. How do I roast my turkey so it gets golden brown tan lines — in the shape of a turkey bikini?

The experts helped to create a “bikini look” by using aluminum foil in certain places on the turkey.

  1. How to carve a turkey when all of its bones have been broken?
  2. I carved my turkey with a chainsaw … is the chain grease going to adversely affect my turkey?
  3. Why does my turkey have no breast meat?

A disappointed woman called wondering why her turkey had no breast meat. After a conversation with a Talk-Line operator, it became apparent that the woman’s turkey was lying on the table upside down.

  1. It’s my first Thanksgiving and I have a tiny apartment-sized oven … how much will my turkey expand when cooking?
  2. How do I get my turkey to stop sudsing? Is a soapy turkey recoverable?

A first-time Thanksgiving chef called after she had washed her turkey with dish soap. You don’t have to clean your turkey, simply pat the extra juices dry with paper towels before stuffing or roasting the turkey.

  1. For the sake of delicious smells, can I cook my turkey over the course of four days?

The Talk-Line doesn’t recommend slow-cooking your turkey over the course of multiple days. You are able to use a slow cooker if needed, but experts would recommend 6-8 hours in the slow cooker. If cooking in the oven, it should only take a few hours to cook.

  1. How do I baste a pre-basted turkey?

Some folks love to baste the turkey while it’s cooking. If you’re one of them, the Talk-Line suggests basting only a few times during the cooking process so you don’t continuously let out the heat of the oven.

  1. My turkey thawed on my lap … can I eat it?

A gentleman won a turkey at the casino, and brought it home on the bus where it had thawed. The safest way to thaw your turkey is in the refrigerator — it takes one day for every four pounds of turkey.

Lifehacker covers the science of Thanksgiving

Lots of folks like to say that food safety in the home is simple. It isn’t. There are a lot of variables and messages have historically been distilled down to a sanitized sound bite. Saying that managing food safety risks is simple isn’t good communication; isn’t true; and, does a disservice to the nerds who want to know more. The nerds that are increasingly populating the Internet as they ask bigger, deeper questions.

Friend of barfblog, and Food Safety Talk podcast co-host extraordinaire, Don Schaffner provides a microbiological catch-phrase that gets used on almost every episode of our show to combat the food-safety-is-simple mantra; when asked about whether something is safe, Don often answers with, ‘it depends’ and ‘it’s complicated’. And then engages around the uncertainties.IMG_4138

Beth Skwarecki of Life Hacker’s Vitals blog called last week to talk about Thanksgiving dinner, turkey preparation and food safety and provided the platform to get into the  ‘it depends’ and ‘it’s complicated’ discussion. Right down to time/temperature combinations equivalent to 165F when it comes to Salmonella destruction.

Here are some excerpts.

How Do You Tell When the Turkey Is Done?

With a thermometer, of course. The color of the meat or juices tells you nothing about doneness, as this guide explains: juices may run pink or clear depending on how stressed the animal was at the time of slaughter (which changes the pH of the meat). The color of the bone depends on the age of the bird at slaughter. And pink meat can depend on roasting conditions or, again, the age of the bird. It’s possible to have pink juices, meat, or bones even when the bird is cooked, or clear juices even when it’s not done yet.

So you’ve got your thermometer. What temperature are you targeting? Old advice was to cook the turkey to 180 degrees Fahrenheit, but that was a recommendation based partly on what texture people liked in their meat, Chapman says. The guidelines were later revised to recommend a minimum safe temperature, regardless of what the meat tastes like, and that temperature is 165. You can cook it hotter, if you like, but that won’t make it any safer.

There’s a way to bend this rule, though. The magic 165 is the temperature that kills Salmonella and friends instantly, but you can also kill the same bacteria by holding the meat at a lower temperature, for a longer time. For example, you can cook your turkey to just 150 degrees, as long as you ensure that it stays at 150 (or higher) for five minutes, something you can verify with a high-tech thermometer like an iGrill. This high-tech thermometer stays in your turkey while it cooks, and sends data to your smartphone. Compare its readings to these time-temperature charts for poultry to make sure your turkey is safe.

The whole piece can be found here.

Not the headline Batz would have used: Which U.S. foods are most likely to get you sick

Friend of the barfblog.com Michael Batz, says there is a difference between “which foods are most likely to get you sick?” and “which foods cause the greatest burden in the U.S.?”

So he provided commentary on a story he helped create.

The story ran in Fortune, and claimed the U.S. economy will take a $15.5 billion dollar food safety hit through lost income, lost revenue, healthcare-related costs and some intangibles, like “pain and suffering,” according to the U.S. Department of Agriculture.

Research shows the most common foodborne pathogen is norovirus, which is the leading cause of gastroenteritis. The most deadly is listeria; though it is rare, with just 123 confirmed cases in 2013, 24 led to death. The most expensive is salmonella, which is quite common — 7,277 cases in 2013, but with a small fraction (127) resulting in death.

How worried should Americans be about the safety of the food supply? Which foods are most likely to get Americans sick? A deep dive into the data offers a look at where the risks lie.

To identify the most dangerous foods, the Emerging Pathogen Institute, a research institute at the University of Florida, compiled a greatest hits of dangerous pathogen/food pairings. Using a methodology that combined the likelihood of contracting an illness and the illness’s severity to calculate total disease burden, the group identified a Top Ten list of combinations.

Rank Food and Pathogen Cost Illnesses Hospitalizations Deaths
1 Poultry (campylobacter) $1,257m 608,231 6,091 55
2 Pork (toxoplasma) $1,219M 35,537 1,815 134
3 Deli meats (listeria) $1,086M 651 595 104
4 Poultry (salmonella) $712M 221,045 4,159 81
5 Dairy products (listeria) $724M 434 397 70
6 Complex Foods (salmonella) $630M 195,655 3,682 72
6 Complex foods (norovirus) $914M 2,494,222 6,696 68
8 Produce (salmonella) $548M 170,264 3,204 63
8 Beef (toxoplasma) $689M 20,086 1,026 76
10 Eggs (salmonella) $370M 115,003 2,164 42

 The top pairings don’t necessarily map to the major outbreaks of disease; only four of the top ten outbreaks in 2014 were from pairings on the list. J. Glenn Morris, director of the Emerging Pathogen Institute, told Fortune that, outbreaks are “a small fraction” of the total number of foodborne illnesses. In most cases, foodborne illnesses are limited to a small group of people, which makes it difficult for authorities to track.

cdc.fbi.illnessBatz explains the which-foods-cause-the-greatest-burden-in-the-U.S. is about everyone. It is the summary row in a table with 300 million lines, each row representing a different person. Each row could be considered something like individual risk.

(You think you’re a special snowflake? Nope, you’re just a row in the giant spreadsheet of life.)

Our individual risks differ so greatly. Unless you’re pregnant, you don’t need to worry about transmitting Listeria monocytogenes or Toxoplasma gondii to your fetus. Your risks go up when you’re immunocompromised, particularly for some bugs. The young and old face increased risks, though again, every disease is different. People of different ages and genders have different food consumption patterns, too.

And even then, that one row represents risks faced at every meal over the course of a year. Some foods have much higher risks per serving, yet we don’t eat them that often. We consume other foods with lower risks per serving in very high quantities. When you’re telling someone which foods are riskiest, which do you mean? It’s tricky.

Five or so years ago, my colleagues and I published the results of trying to just get at that summary row. Or to be more specific, we tried to say something about which pathogens, which foods, and which pathogen-food pairs cause the greatest public health impact. This kind of information is, I believe, important for getting a handle on the landscape of foodborne disease, to help guide our efforts to reduce the burden.

fbi.batz.pie.chart.nov.15The report had a punchy “top ten” type title and got some attention (for which I’m thankful). But the attention has always come with a price, and that price is that when work like this is written up, it’s almost always presented to readers as some version of which foods to avoid.

I get it, I really do. It’s natural to frame things to readers this way, to take research and make it personally relevant to them. But it kind of butchers the work, and can do as much to misinform as to educate.

So kudos to Tamar Haspal of Fortune for mostly getting it right in an article that presents risks at the broad, national level. Boo to whoever wrote the headline, which conflates population and individual risk and asks “which foods are most likely to make you sick?”

Translating research is always tricky, and I’m never quoted quite to my satisfaction (I can’t get none). In the article, I say we value mortality at $8.7 million per life lost because “there’s also a social welfare value to a life.” Well, that’s not quite right, and I’m doubtful I put it quite that way, and my economist friends are likely groaning at the phrasing, but it’s fine. It’s mostly right, anyway. You have to learn to turn the other cheek.

But other things I just can’t let go. Like the fact that, for the record, I hate pie charts.

4 dead, 834 sick from Salmonella in cucumbers from Mexico

Who knew cucumbers could be so vile?

cucumber.spain,MEPThe U.S. Centers for Disease Control, multiple states, and the U.S. Food and Drug Administration (FDA) are investigating a multistate outbreak of Salmonella Poona infections.

Epidemiologic, laboratory, and traceback investigations identified cucumbers imported from Mexico and distributed by Andrew & Williamson Fresh Produce as a likely source of the infections in this outbreak.

Two recalls of cucumbers that may be contaminated with Salmonella were announced as a result of this investigation: Andrew & Williamson Fresh Produce and Custom Produce Sales.

838 people infected with the outbreak strains of Salmonella Poona have been reported from 38 states, an increase of 71 cases since the last update on October 14.

165 ill people have been hospitalized, and four deaths have been reported from Arizona (1), California (1), Oklahoma (1), and Texas (1).

Thanksgiving goofiness

My parents come from Ontario (that’s in Canada) every year to visit for Thanksgiving (or American Thanksgiving as it’s known to them). My mom likes to participate in the Black Friday shopping craziness; my dad likes to watch football. It’s just fun to have them around.

A couple of years ago my friend Matt Shipman and I put together some Thanksgiving meal videos – sorta our goofy take on food safety for the holidays. The content (unlike my hairline) is timeless.

And here are some food safety infosheets for the holidays.

Holiday meal food safety

Bathing birds is a food safety mess

Avoid foodborne illness during the holidays


From the Salmonella-in-low-moisture-foods files: contaminated flax leads to recall

Your organic steel cut oats & chia with flax and rye flakes might have Salmonella in it, if it’s from Homestat Farms.

Homestat Farm of Dublin, OH is recalling some of its 42-ounce packages only of “Organic Steel Cut Oats & Chia with Flax And Rye Flakes” because the flax seed ingredient has the potential to be contaminated with Salmonella.

The recalled “Organic Steel Cut Oats & Chia with Flax And Rye Flakes” was distributed on a limited basis in Sams Club retail stores located in AK, CA, FL, GA, IL, IN, KS, MD, MI, MN, MS, NC, NH, OH, PA & TX.

The product comes in a 42 ounce carton UPC 8 35882 00620 4 marked with Best Buy: 03/16/17-1, 03/16/18-1, 03/19/17-1, 03/19/17-2, 03/23/17-1, 03/23/17-2, 03/24/17-1, 03/26/17-1, 03/26/17-2, 04/08/17-2. The best buy date can be found embossed on the bottom flap of box next to the bar code.

A supplier informed Homestat Farm of this after a contamination was found in flax seed provided to another customer.

That product sounds like something that Drunk Uncle might say. It probably goes well with some swiss chard on a Roku or an Amazon Prime pumpkin spice.

Foodborne outbreaks in Canada 2008-2014

Background: Enteric outbreak investigation in Canada is performed at the local, provincial/territorial (P/T) and federal levels. Historically, routine surveillance of outbreaks did not occur in all jurisdictions and so the Public Health Agency of Canada, in partnership with P/T public health authorities, developed a secure, web-based Outbreak Summaries (OS) Reporting System to address this gap.

canada.colbert.oct.14Objective: This analysis summarizes the foodborne outbreak investigations reported to the OS Reporting System between 2008 and 2014.

Methods: Finalised reports of investigations between 2008 and 2014 for all participating jurisdictions in Canada were extracted and descriptive analysis was carried out for foodborne outbreaks on etiological agent, severity of illness, outbreak duration, exposure setting and outbreak source.

Results: There were 115 reported foodborne outbreaks included in the analysis. This represents 11.2% of all outbreaks reported in the enteric module of the OS Reporting System between 2008 and 2014. Salmonella was the most commonly reported cause of foodborne outbreak (40.9%) and Enteritidis was the most common serotype reported. Foodborne outbreaks accounted for 3,301 illnesses, 225 hospitalizations and 30 deaths. Overall, 38.3% of foodborne outbreaks were reported to have occurred in a community and 32.2% were associated with a food service establishment. Most foodborne outbreak investigations (63.5%) reported a specific food associated with the outbreak, most frequently meat.

Conclusion: The OS Reporting System supports information sharing and collaboration among Canadian public health partners and offers an opportunity to obtain a national picture of foodborne outbreaks. This analysis has demonstrated the utility of the OS Reporting System data as an important and useful source of information to describe foodborne outbreak investigations in Canada.

Funding for this program was provided by the Public Health Agency of Canada.

An overview of foodborne outbreaks in Canada reported through Outbreak Summaries: 2008-2014

CCDR: Volume 41-11, November 5, 2015: Foodborne Illness

Bélanger P, Tanguay F, Hamel M, Phypers M



Pistachios following the way of the almond industry: developing validated pasteurization interventions

Nuts and other low moisture foods can be a source of Salmonella. That’s not new. But many folks in the low-moisture foods industries are now conducting risk assessments and validating interventions (like pasteurization) to keep the Salmonella out of the hands of bar patrons everywhere.

The almond industry led the way about a decade ago. The peanut industry, in the wake of two outbreaks followed. According to Growing Produce, the pistachio industry is working with friend of barfblog (and known to her close friends as at the almond queen) Linda Harris and Michigan State’s Bradley Marks on some validation work.pistachios

“At this point in our world, Salmonella is a hazard that is reasonably likely to possibly occur in low-moisture products like nuts. It’s happened before, so we have to assume it’s possible it could happen again,” Marks says. “That being the case, the proposed rules of FSMA require that the processor have a validated process that they can document that they’ve shown achieves the food safety objective.”

Marks and his colleagues are currently working on lab-scale research to evaluate the effects of the pasteurization process and product conditions on the resistance of Salmonella to heating. “We’re doing some mathematical modeling so we can understand the rate of Salmonella activation as a function of temperature, time, and conditions of the product or the process itself,” Marks says.

The second part of the project will involve similar work on a pilot scale. “We have a system where we can inoculate pistachios with Salmonella and subject them to a process like a flatbed roaster,” Marks says. “So we are looking at validating that our prediction of the inactivation of Salmonella is correct, and that a non-pathogenic surrogate (Enterococcus faecium) also is reliable as a means to validate the process.”

Marks is working closely with Linda Harris, co-principal investigator for the Western Center for Food Safety at the University of California, Davis, to develop guidelines and on-site workshop training for pistachio processors. The guidelines and training will focus on what needs to be measured and documented to meet the FSMA requirements.