Solving new era food safety challenges requires advancements in data sharing

— OPINION —

The battle against foodborne disease has been waged since the beginning of time.  Historical literature provides evidence of how earlier civilizations tried to protect themselves, their families, tribes, and villages from getting sick from the food they ate. The examples range from the emergence of cooking with fire to initial attempts to preserve foods with salt and spices to the invention of the first refrigeration units.  While they weren’t the type of refrigerators you and I have at home today, prior civilizations did construct wooden boxes that kept foods cold with blocks of ice.

Fast forward through time and we witnessed steady advancements that improved food safety, such as through the discovery of pasteurization by one of my heroes, Louis Pasteur, to the canning of foods, and – in more recent times – the invention of HACCP (Hazard Analysis Critical Control Point), a risk management framework that was created by NASA and Pillsbury to strengthen foodborne illnesses prevention among astronauts in the U.S. space program.

The Foodborne Disease Battlefield is Changing

However, the battle against foodborne disease is a never-ending race between detection and prevention.  The food system continues to change, there is a growing at-risk population, and new and emerging foodborne hazards are being identified.  For example, if you look at the foodborne pathogens that we battled as a society in the first part of the 20^th century, they look quite different than the foodborne hazards we’re concerned about today.  A review of the literature lists Amoebiasis, Botulism, Brucellosis, Staphylococcal food poisoning, Tapeworms, and Trichinosis, among other foodborne pathogens, to be of public health concern during the first half of the 1900s.

Fast forward to 2024 and you’ll see that many of the foodborne hazards that we’re concerned about today are not the same issues we were concerned about then.

Just think about the food safety issues you’ve seen recently in the news – which I refer to as new era food safety challenges – and you’ll notice that most of them are missing in the original list above.  What you’ll notice missing then but here now are foodborne hazards such as Listeriosis, Shiga toxin-producing E. coli (STEC), Norovirus, Cyclospora, chemical hazards such as heavy metals, food additives, Per- and Polyfluoroalkyl Substances (PFAS), and more.

Modern Times Require Modern Approaches – The Need for Data Sharing

Solving new era food safety problems is extremely challenging and it will require new ways of thinking and working.  It will require us, as a profession, to go beyond the traditional tools we’ve used, such as testing, inspections, and training.  It will increasingly require data sharing.

This reality is unavoidable and inevitable. To put this into context, the 20^th century has been referred to as the industrial revolution.  There were many advancements over that time in industrial manufacturing.  In contrast, the 21^st century is being referred to as the digital revolution.  And there is one critical currency that the digital revolution is completely dependent on – and that currency is data. 

In fact, that’s why we’ve all heard phrases like “data is the new oil.”  In today’s economy, better use of data and digitization is critical for continued business success.  And the need to keep up is an imperative. New data, in almost all areas, is being created at an accelerated rate.  It’s now estimated that 2.5 quintillion bytes of data are being generated every day.  And the generation and use of new data are fueling almost every part of the modern economy. 

Bridging the Food Safety Data and Preventive Information Divide

However, we’ve all heard it said that data does not equal information.  It’s only when data is analyzed and given context and meaning that it can be converted into information.  And it’s only when information is put into preventive food safety action, that we can call it food safety knowledge.

Historically, as a profession, our ability to turn data into information has been challenging, primarily because of two key factors.  First, food safety data was (and still is) often collected on paper.  As Bill Gates stated in his classic book, Business @ the Speed of Thought, “Data on a piece of paper is a dead-end; data in digital form is the beginning of meaningful action.”  The second is that, up until recently, it was challenging to analyze large volumes of data, even if it was in digital form, and convert it into information.  In other words, there’s been a big divide between the large volume of food safety data that exists and our ability to make sense of it.

But that huge divide is increasingly being bridged through new and emerging technologies such as artificial intelligence, especially a branch called machine learning.  Despite these new and powerful tools now available to us, “the biggest obstacle to using advanced data analysis isn’t skill base or technology; it’s plain old access to the data,” writes tech expert Edd Wilder-James in the Harvard Business Review.

Data Sharing – Our One Best Hope to Gain Dramatic Food Safety Insights

Many of our remaining food safety challenges exist because of significant gaps in knowledge, ranging from the prevalence of specific hazards, an understanding of their geographical distribution, and their relationship with other dependencies within complex agricultural ecosystems.  

And when it comes to understanding these knowledge gaps to better inform prevention, a fundamental statistical principle still holds true.  Larger volumes of good quality data generally equal greater knowledge due to enhanced statistical power.

Thus, in a large and decentralized food system, no single entity, whether it be a grower, processor, retailer, or regulatory agency, can be as smart as all of us working together when we’re sharing data and what we think we know.

That’s why data sharing is so critical to advancing some of our greatest, new era food safety challenges.  

Data Sharing Use Cases Abound

While I sincerely believe data sharing can enhance almost every aspect of food safety, I want to share just a couple of use cases to illustrate this point.

Safety of Fresh Leafy Greens

Fresh leafy greens play a critical part in an overall healthy diet.  And most of the time, fresh leafy greens are grown and produced safely.  In fact, it’s reported that Americans consume over 100 million servings of fresh leafy greens safely on a daily basis.  

Having stated that, many outbreaks of foodborne illness related to produce have been linked to fresh leafy greens.  Public health officials report that between 2009 through 2018, 40 STEC outbreaks were linked to fresh leafy greens in the United States and Canada.  

Moreover, some of these outbreaks have been high-profile events that resulted in broad consumer advisories. Health officials advised Americans to avoid eating fresh leafy greens, regardless of source, resulting in significant industry losses and damage to consumer trust.  

Concurrently, during this timeframe, many growers have implemented more aggressive preventive measures, such as treatment of irrigation water, and ensuring greater buffer areas between crops and grazing cattle on adjacent lands, as well as increased sanitation efforts on harvest equipment.  They’ve also tested more samples of pre-harvest product and water than ever before.  Although these measures are believed to have strengthened the safety of fresh leafy greens, there remain knowledge gaps in terms of all of the potential sources of STEC, and routes of contamination.   

As a result, regulatory scrutiny continues to become more stringent related to the growth and processing of fresh leafy greens.  This enhanced scrutiny has resulted in FDA issuing a Produce Safety Rule, including fresh leafy greens on the list of foods that require additional record keeping as part of FDA’s new Food Traceability Rule, and the recent issuance of a final pre-harvest agricultural water standard.  In addition, it has resulted in repeated, annual, routine surveillance assignments in which regulators test fresh leafy greens to determine if they are contaminated with pathogenic microorganisms.

The FDA has over several years conducted microbiological surveys of fresh leafy greens already in commerce.  Unfortunately, these regulatory surveys are extremely disruptive to commerce. Growers will place on hold any products taken by FDA until results come back to avoid a potential scenario in which they learn of a positive test after the product has already entered commerce.  Over the years, the number of positive findings detected by the FDA has been rare, and because of the length of time it takes to get back FDA results, these assignments have resulted in tens of millions of dollars’ worth of food waste due to fresh leafy greens spoiling while waiting for FDA test results.

As an example, during the 2022 harvest season, the FDA coordinated a sampling assignment at 14 farms in the Salinas Valley, CA, agricultural region.  In this assignment, a mere 62 samples were taken.  Of the 62 samples of leafy greens analyzed, the FDA reported that they detected Salmonella Enteritidis in one sample of romaine lettuce.  Again, this is not intended to be critical of FDA.  They’re simply trying to fulfil their public health mission.

However, and in contrast, imagine if we were able to take a more data-informed approach, through data sharing between the private and public sector, to achieve the same or an even better public health outcome.  

As an example, in 2022, Western Growers, a trade association, established the GreenLink™ Data Sharing platform to allow their members to share their test results in an anonymous manner in an attempt to see broader trends of pathogen detection across growers and the broader growing region.  I believe in this work and last year joined GreenLink™ as an advisor.

The hypothesis is that fresh produce growers, collectively, generate hundreds of thousands of food safety data points each growing season, ranging from pre- and post-harvest product test results, analysis of water samples, and numerous other qualitive observations.  However, the massive volumes of food safety data being generated and collected are often not digitized, exist in siloed systems and, therefore, hinder the sector’s ability to turn data into collective information and to see larger patterns and trends.  Such an ability, to turn big data into information, could serve as an early detection system and a potentially better understanding of patterns or trends, such as whether or not seasonality plays a role in contamination events.  Moreover, the transparency of the large body of data being generated by growers could potentially lessen regulatory scrutiny based on the industry’s self-monitoring practices, as well as enhance consumer confidence and lead to broader industry best practices.  

Since its launch, GreenLink™ now has hundreds of growers participating in their food safety data-sharing collaborative and hundreds of thousands food safety data points. 

Imagine if the private and public sectors could reach an understanding of how this wonderful private sector initiative could be used as a win-win to better protect consumers.

Heavy Metals in Food

As another ideal use case, consider the concerns that exist about the potential for heavy metals, such as arsenic, to be in a variety of foods.  

As background, arsenic is an element that may occur naturally in the environment and soils where produce is grown, as well as from manmade sources. It can exist in both inorganic and organic forms.  In general, inorganic arsenic is considered more toxic.  In the literature, adverse pregnancy outcomes and neurodevelopmental toxicity are listed as adverse health effects that could occur due to exposure to inorganic arsenic.  Studies have also concluded that consumption of inorganic arsenic has been associated with cancer, skin lesions, cardiovascular disease, and diabetes in humans.

In an attempt to reduce exposure to arsenic, specifically from apple juice, in 2023 the FDA developed an Action Level for Inorganic Arsenic in Apple Juice in the form of guidance for the industry.  The FDA’s guidance provided information to manufacturers on the establishment of an action level of 10 parts per billion (ppb) or 10 micrograms/kilogram (µg/kg) for inorganic arsenic in apple juice.  The Action Level is considered achievable by industry with the use of current good manufacturing practices. 

Of interest was the data set the FDA used to establish the feasibility of industry being able to meet the goal.  Remember, trace levels of arsenic can occur naturally in soil and the environment.  When you look at the total number of samples the FDA used to conduct their feasibility assessment, as listed in their guidance document, the numbers of samples in total are quite small.

All in all, they include results from 94 samples collected in a 2011 retail survey, 160 samples taken from 2013 to 2022 as part of their Toxic Element Program, and 47 apple juices samples tested for arsenic between 2013 to 2020 as part of the Total Diet Study, in which FDA regularly tests foods for a variety of analytes.

Overall, the results of a mere total of 301 samples of apple juice were used to establish federal guidance on an action level.  From a scientific and statistical perspective, that’s not a lot of data points to determine, with statistical power, a more accurate and precise range of levels that might exist in juices, differences that are likely to occur based on the growing regions of apples, and what is potentially achievable and under what timelines.

Again, while this is in no way, shape, or form a criticism of the FDA, as we know they have a vast scope of responsibility, and – in certain instances – are resource constrained. This is, however, a call to action and realization that with better data sharing, private to public, regulators such as FDA can have a lot more data and insights when establishing policy.  And with greater data-driven insights, they can create smarter regulatory action levels or targets based on the power of the large volumes of data that the industry often creates in ways governments cannot.  

Calling Both Government and Industry to Action

The case for better data sharing, whether private-to-private or private-to-public, is clear.  As a profession, we’ve always said that food safety knowledge cannot be a competitive issue.  If you believe this statement to be true, like I do, then food safety data sharing should be viewed similarly.  

How we get there – better data sharing – might not be easy, but it’s a critical imperative and it CAN be done. 

Therefore, as a food safety community, we should all call on industry trade associations, industry collaboratives, regulators, and private-public partnerships to look at some of their most pressing food safety challenges and questions – and consider the role that data sharing could play in resolving them.  Let’s look at how at how other industries have figured out how to share data, evaluate what hurdles they overcame to be able to do so (whether it’s through anonymizing data sources to gain insights at a more aggregate view or the establishment of regulatory policies that incentivize data sharing), and begin by conducting more proof-of-concepts within our profession.  

Letting concerns or obstacles prevent us from getting started is no longer an option, when the lack of such efforts means the implementation of activities that are not sufficiently data-driven informed and, ultimately, means consumers may be put at risk of illness or more serious harm.

In closing, never before in human history has our ability to turn large volumes of data into preventive and predictive food safety information been more achievable than it is today.  Let’s do that through collaborative data sharing.  Remember, our bosses – the American consumer – deserve no less.  

(To sign up for a free subscription to Food Safety News,click here)