The European Food Safety Authority (EFSA) and EU Commission’s Inter-European Union Reference Laboratories (EURL) Working Group on Next Genome Sequencing (NGS) organized the second Science Meets Policy conference this week. More than 100 people from 20 countries attended in-person while online viewers peaked at 257.

Stefano Morabito, from the Italian National Institute of Health (ISS); George Haringhuizen, at the Dutch National Institute for Public Health and the Environment (RIVM); João André Carriço, of bioMérieux; Katja Alt, from the German Federal Ministry of Food and Agriculture and Coen van der Weijden, from the Netherlands Food and Consumer Product Safety Authority (NVWA) were featured on a panel discussion.

Panelists highlighted the need for flexibility, especially as there are 27 countries in Europe at different stages in use of Whole Genome Sequencing and as outbreaks can involve nations outside the EU. They discussed the types of data needed, data quality and comparability, resource issues, how to build trust and the role of different parties including regulators and industry.

EFSA and ECDC’s One Health WGS system has been operating since July 2022 and while things are going well, not all member states are contributing equally.

Data sharing dilemmas

Bernhard Url, EFSA’s executive director, said genomic data sharing is now at a turning point.

“We believe that from a technological and methodological point of view we are ready to use WGS data more widely and with more impact. Many of the problems have been solved and the technical infrastructure has been built. There is no doubt, at least within the community, that data sharing adds value, because it leads to faster outbreak detection and better tracing. It increases the likelihood of connecting sporadic cases to clusters and to detect outbreaks and there is measurable economic impact,” he said.

However, despite the knowledge that sharing data helps, there are still some obstacles that prevent wider use, said Url.

“There are technological gaps as not all member states or organizations use WGS on a routine basis. There is also a worry that people and countries say we don’t have a robust legal basis to share data. There is a fear that people lose control of data, they produce the data, share it but don’t know what happens afterwards. There is a concern that if this technology would be used widely, many more clusters would be detected, which is good from a public health point of view, but it also would increase the workload of national authorities to follow-up and deal with these clusters.”

Url said it would be “unwise” to wait for legislators to define the rules of the game.

“The WGS community must do our part to create the conditions for success. We think there is a lot we can do to move data sharing forward in the current legislative framework. We still have to work on creating a mutual understanding about benefits and limitations of this technology. We have to agree on common guidelines, processes and procedures, otherwise we wouldn’t know how to compare different outcomes,” he said.  

“We want to act as openly as possible but as confidential as needed, there is a fine line that we have to find. EFSA has invested resources in creating a technological infrastructure for enabling WGS data sharing, mainly to tackle foodborne threats. We will continue to do our part to move genomic data sharing forward.”

United States perspective

Eric Stevens, from the U.S. Food and Drug Administration, said the GenomeTrakr network is the result of 12 years of work. At the end of 2021, there were 600,000 genomes in the public database, today it is more than 1,000,000 sequences.

“After more than a decade of experience, it is not the sequencing that is the challenge when transitioning to this data, it is how you are going to analyze it, train staff, gain the skills and enable the entire system to utilize it effectively,” he said.  

“Metadata helps to tell a complete picture, without it you have a DNA sequence, which can only tell you some stuff. Contextual data gives that data life, it tells you where those bacteria came from, how they were living and when we are starting to think about the interventions we can make, we need that information to understand the complete picture.

“For us, the best use is making it open data available to anyone because somebody maybe interested in Salmonella, somebody else in E. coli and sometimes they overlap with interventions you can make for preventive controls and reducing contamination.”

Stevens said once data is in the database, a variety of things can be looked at.

“When you start thinking of the global food chain you can think where do we need more data from and start doing some projects to tackle those problems to better understand how food becomes contaminated in the first place. You wouldn’t know any of this unless you had the data that can help point the way,” he said.

“GenomeTrakr is responsible for almost 100,000 food and environmental isolates to tell a more complete picture of linking clinical isolates back to their sources, so we can not only respond to foodborne outbreaks but then try to prevent them. When you start looking at where your sources of food and environmental isolates linked to human illnesses come from, you can start doing source attribution and more preventive targeting. If we can get to a point where we can upload data in real time we can start to make those connections as early as possible to get a contaminated product out of the market.”

It can also help in moving from responding to outbreaks to trying to prevent contamination occurring.

“In a facility for example, you are not going to do WGS to identify a pathogen, you can do a quick culture method to see presence or absence. But if you have a facility that is concerned about whether they have a resident pathogen you would 100 percent like this information from WGS. You could expand that further to farms and potential water sources,” said Stevens.  

“When you start doing projects in different parts of the world you start understanding that everybody has issues that maybe aren’t issues for you. We’ve done a lot of work in Latin America and the big problem in getting started in sequencing is availability of reagents. We hear it costs five to seven times more than what we pay. When we talk about this being utilized by the world we have to start focusing on those issues that are going to make the most impact.”

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

The event, organized by UNIDO, the Department of Agriculture, Water and the Environment of Australia and the Standards and Trade Development Facility (STDF), also raised concerns about data ownership, privacy and quality, as well as trust between stakeholders.

Donald A. Prater, associate commissioner for imported food safety at the U.S. Food and Drug Administration, said the agency was talking about data and information every day.

“Increasingly, we are using predictive analytic tools, artificial intelligence and machine learning to drive our risk-based resource allocation so in those oversight activities such as inspections and sampling, we want to use data and information to guide us. We have constrained resources, we can’t be in every place all the time so we are looking to do regulatory activities that are the most impactful for public health. We want to go places where there is increased risk,” he said.

Value of providing data
Regulators sit on a mountain of data but industry also has quite a bit and there are other sources, said Prater.

“Data quality is a big issue. One challenge in sharing data and information is ensuring confidentially and creating that environment where the fear of punitive action is minimized, that will help us to get better together. We’re looking at data-sharing platforms, leveraging reliable third-party audits and partnerships with international food safety regulators. How we get data and share it is a challenge. We’re looking at techniques like aggregation, de-identification and anonymization to provide levels of confidentiality but still monitor trends and public health outcomes,” he said. 

Julie Pierce, director of openness, data and digital at the Food Standards Agency, said a lot has been learned through years of experience in the UK.

“Now, we have a different mindset. It is much more about having the right information available at the right time to make a decision. It is about predicting the future, we don’t want to be looking backwards as to what has happened. We learned the importance of data governance, data owners, those who need to use or access the data and understanding those roles. We need to improve accuracy of the data, there is a lot out there, some of it is good and some is not. Where it isn’t good we can improve it. Timeliness of data and getting hold of it as close to real-time as possible is important. We need to get better at overcoming some of the trust issues,” she said. 

“In discussions with businesses we are finding that having standardization, consistency and knowing what is required is valuable. That is one of the roles we can play to make it easier for businesses to share data with ourselves and others. We have to try to demonstrate where the value is downstream if someone invests this piece of data. We need to show the value that can be pushed back upstream. So, I provide this data and get some value back either to help ensure my next production run is safer, or to improve yields or get a better price. Data needs to be an asset rather than it just being an overhead with the regulator demanding data for a certificate.”

Trust in third parties
Friedrich Sövegjarto, of the Austrian Agency for Health and Food Safety (AGES), asked as there is a high level of food safety in Europe, is data from private sources needed?

“On the other hand, we have reduced resources and new challenges like sustainability and the problem of food fraud. The EU system means inclusion of self-control data and our risk assessors ask about the possibility of getting this private data to do a better risk assessment. They only have data from official controls. We just started these discussions but we face some challenges. The main one is confidentiality and also trust in data,” he said.

“In Austria, the food safety authority gets data from the food business and there are no private assurance systems in between. Whenever we have an outbreak or some food crisis, the food business is responsible but the second most responsible is the authority. If data is taken from private systems, how can we make sure that we don’t lose our independence?”

Dubai’s digital work
Bobby Krishna, from Dubai’s food safety department in the UAE, spoke about how the region had been using digitalization in recent years.

“Most regulators use data from foodborne illness notifications, surveillance and inspections but these data sets are lagging because you get the information after a problem has occurred. If you want to be preventative you need data that tells you something is going to go wrong,” he said.

“Compliance costs can be bearable if it allows companies to export their food elsewhere or show evidence to get more business but that value is not there for digitalization. Everything on a record is a data source, provided it is in a digital format and not trapped on paper. It is anything that can be converted into a useable decision-making or predictive tool.”

Nima Bahramalian, industrial development expert at UNIDO, said it is key to ask what drives the decision of a food company to record, self-report and share data.

“One theory is the perceived benefits and costs of participating in schemes or adopting a new practice. Whether it will bring additional commercial benefits is not always visible. In developing contexts, the return on investment is not immediate in many cases. Would sharing data result in reducing controls and cost to get assurances and certification? Second, is the perceived ease of using the technology. Digitalization can facilitate access to training materials and certification but it is important that technical assistance creates an environment that makes use of technologies easier.”

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

Juno Thomas, from the National Institute for Communicable Diseases (NICD), said there were a lot of unknowns about food and waterborne disease in the country.

“The data we have really represents the tip of the iceberg. The true burden of disease is the great unknown. We try to improve on sources of data and completeness but it is a mammoth task and it is going to be a labor of love for many of us for years to come,” she said during a presentation at the South African Food Safety Summit, which was sponsored by Marler Clark, the publisher of Food Safety News.

“Until there is a coordinated approach to looking at the data more holistically, we will continue to have these narrow perspectives and not be able to get a true understanding of what the food safety issues are in South Africa and where we need to focus research on foodborne disease.”

Lack of a complete picture
Thomas said one of the problems was very little data was available.

“The data sources we do have vary in completeness for particular pathogens. The best data, the most complete epidemiological, laboratory and molecular data, is for Salmonella Typhi and Salmonella Paratyphi A, B and C and Listeriosis. For non-typhiodal Salmonella, Shigella, Campylobacter and E. coli we rely on data from lab-based surveillance through the GERMS-SA platform at the NICD. This relies on voluntary submission of isolates from public and private sector labs. We really have no way of knowing how many cases of these organisms we are missing,” she said.

“The use of PCR and cultural-independent diagnostic tests means there won’t be isolates for those cases and they won’t be counted by us. There are no sources of aggregated national data for enteric viruses, for toxin producing bacteria, such as Staphylococcus aureus, Bacillus cereus and Clostridium perfringens, and enteric parasites. We know that just relying on the notifications that come through the system underrepresents even the recognized outbreaks that are there.”

Data and isolates from private sector labs are no longer always shared with the NICD after the 2017 to 2018 listeriosis outbreak linked to polony made by Tiger Brands and Thomas said this “unfortunate collateral damage” meant everyone was working in siloes.

Coronavirus impact
Overall, 355 outbreaks were reported to NICD from 2018 to 2021 but only 146 were investigated. The majority occurred in the household followed by educational settings like schools and universities. 

“2019 was a bumper year for foodborne outbreak reporting and in the first few months of 2020, preceding the hard lockdown, there were a number of outbreaks. Since the lockdown, and continuing into 2021, there have been relatively few outbreaks and months at a time were not a single outbreak was reported,” said Thomas.

“We aren’t sure whether this is due to changed health-seeking behavior during the COVID-19 pandemic or whether the non-pharmaceutical measures we implemented as part of the controls, which includes hand hygiene, restaurants not open and events being closed, are factors that resulted in a real decline in foodborne disease.”

Concerning trends

There were a number of common themes in these outbreaks, said Thomas.

“Firstly, the most commonly identified pathogen is always non-typhiodal Salmonella and is typically associated with meat, poultry and eggs. We haven’t yet seen an outbreak where the Salmonella in patients can be directly linked to another type of food such as fruit or vegetables,” she said.

“We’ve seen an association of Salmonella outbreaks with informal and ritual slaughter of food animals, in particular goats. We’ve seen an emergence of Salmonella Newport associated with goats used in ritual slaughter but we see a range of Salmonella serovars from informal slaughter of goats and cows.”

Eating meat not fit for human consumption is another growing problem.

“This speaks to food insecurity and issues around animal health. We’ve seen an increasing number of outbreaks associated with eggs, in some cases we’ve had informal caterers that buy cracked eggs from producers at a reduced price and then use them in mass catering. We’ve also seen outbreaks in hospitals were non-irradiated eggs are used and patients are allowed to choose whether they would like their eggs soft or not, which is a major food safety issue in a healthcare facility environment,” said Thomas.

“Poor food safety practice in the home is an ever present problem. There is much to be done in terms of health education and improving basic food safety practice throughout the country. We’ve seen an increase in outbreaks associated with events like funerals and weddings. Informal caterers or members of a family that do mass catering and often there are major problems with hand hygiene, food storage and reheating. School feeding schemes are also of great concern as these types of outbreaks are second only to household outbreaks, again there are problems with basic food safety practice.”

The virtual event also had sessions on big data, cybersecurity and ransomware and food defense as well as panel discussions on food fraud and product recalls in South Africa.

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

Frank Yiannas was speaking at a Health Talks webinar on digitalization, food safety and trade with other panelists from Ghana, India and Ireland.

“The world around us is changing rapidly. Part of this rapid change is that data and information have become digitized and can be shared at the speed of thought. And new and emerging technologies are increasingly taking big and real-time data and putting it to good use,” he said.  

“For example, advances in artificial intelligence, the Internet of Things, sensor technologies, and blockchain are improving business processes. And the food system is reshaping itself, using these technologies, to meet the expanding global supply chain and the changing needs of consumers.

“I believe the new digital technologies offer the potential to help us predict and prevent food safety problems and better detect and respond to problems when they do occur. Leveraging the power of data is going to allow us to solve some of our food safety challenges that I never thought we could solve.”

Traceability developments

People often think that food traceability is a reactive tool but this isn’t true, said Yiannas.

“In the event of a foodborne outbreak, better traceability leads to better safety by enabling rapid traceback to the source of a contaminated food, the ability to speed up recalls, and better fueling the root cause analyses to understand how the incident happened and prevent reoccurrences again in the future,” he said.

“A digital, traceable food system will be a safer food system. But we, as food safety professionals and regulators have to be very aware, we can’t create a digital divide. If we do this right, it will allow small and medium enterprises to compete better with large institutions.”

The FDA is set to publish a final rule on food traceability in November 2022 and held a no or low-cost technology traceability challenge in 2021 that received 90 submissions and had 12 winners from the United States, Canada, and New Zealand.

Yiannas also spoke about a data analysis tool called 21 Forward.

“Unleashing the power of data is an overarching goal in FDA’s work to modernize food safety. We are now using this tool in the infant formula crisis that we are seeing in the United States with shortages. Analyzing high volumes of data is enabled by the scalability of this platform. This in turn has helped guide discussions with industry on how to increase production of various types of infant formulas,” he said.

“Better food safety begins and ends with better data. We have a lot of food safety data, it was often kept on paper but now we have these new tools that can bridge the gap between data and converting that into information.”

Predictions and seafood pilot

FDA is working with the private sector to create public-private data trusts to share data better, said Yiannas.

“In the U.S., we’ve seen repeated outbreaks with fresh leafy greens and there’s an organization called Western Growers that is doing just that. Getting the private sector to share all the data they have on fresh leafy greens, anonymously, is a great example of work that will strengthen predictive capabilities and inform risk-management decisions,” he said.

“It is clear that the FDA and food producers should also be looking at ways to tap robust, high quality data sources to strengthen our predictive analytics. We are continuing to explore the use of artificial intelligence, specifically machine learning, in a pilot designed to strengthen our ability to predict which shipments of imported seafood pose the greatest risk of violation. We are in the third phase of the pilot.

“Initial findings suggest that machine learning could greatly increase the likelihood of identifying a shipment containing potentially contaminated products. Doubling or tripling the ability to predict which shipments potentially are violative through the screening process is expected to result in much more effective utilization of resources to examine, sample, and test products at the port of entry.”

Irish perspective

Orla Moore, from the Food Safety Authority of Ireland, said digitalization has increased since the COVID-19 pandemic.

“We’ve seen pop up kitchens, dark kitchens and new businesses, some registered and some unregistered, emerge during COVID and Brexit. We’ve seen changes in how people order food and get food delivered,” she said.

Applications of digital technologies include remote auditing, e-learning, horizon scanning, early warning systems, licensing and registration and e-certification for health certificates.

“Artificial intelligence makes for quick, easy and in some cases cheap, use of data and information. It reduces bias out of decision making quite often. A few examples are risk assessment, quality control, predictive modelling and hygiene monitoring of equipment,” said Moore.

“The key benefit with blockchain is the transparency, it is secure in that the files can’t be edited, so from an audit point of view it is a solid system. One of the main difficulties is that everyone along the supply chain has to adopt and use it.”

Moore also spoke about seeking out emerging threats and risks through horizon scanning.

“We want to be aware of the next thing that is going to happen. It boils down to data mining tools. We have an emerging risk screening group that meets quarterly, and there could be sub-groups if there are emerging threats. This is to acquire data, monitor social media platform and publications,” she said.

“The challenges of digitalisation are the financial cost or the perceived cost to businesses, a lack of information, training and resources, fragmentation in governance or regulatory frameworks, confidentiality and data protection, access, connectivity, data quality and ownership of the data. There is quite a difference in the level of adoption of digitalization across organizations, industries and countries. Whatever system is used by businesses they must be able to provide information to authorities on demand. Digitalization has already happened, everyone is trying to get on board, so they are not left behind.”

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

The organization was launched in 2011 when three U.S. agencies — the Centers for Disease Control and Prevention, the Food and Drug Administration, and the USDA’s Food Safety and Inspection Service (FSIS)—created it to improve certain efforts related to food safety. 

Specifically, its purpose is to improve coordination of federal food safety analytic efforts and address cross-cutting priorities for food safety data collection, analysis and use.

“Since its inception, IFSAC’s focus has been foodborne illness source attribution: identifying which foods are the most important sources of selected major foodborne illnesses. As part of this effort, IFSAC now produces annual estimates for four priority pathogens: Salmonella, Escherichia coli O157, Listeria monocytogenes, and Campylobacter,” according to the group’s 2022–2023 Interim Strategic Plan.

Federal agencies and food safety experts use IFSAC’s analyses to help form strategic planning and risk-based decisions; estimate benefits of interventions; and evaluate the impact of interventions, such as new or revised regulations, policies, and performance standards. By bringing together data from a variety of sources, broadly exploring an array of methods and disciplines, and developing sound analytical methods, IFSAC scientists can improve estimates of the sources of foodborne illness.

The activities of IFSAC have been substantially affected by the COVID-19 global pandemic, according to a statement accompanying the group’s 2022-23 interim plan. During 2020 and 2021, many staff members from CDC, FDA, and FSIS who lead and participate in IFSAC projects and their oversight were deployed to COVID-19 response efforts, had to focus on pandemic-related activities, or covered agency program activities while other staff was deployed. 

IFSAC continues to publish annual estimates of foodborne illness source attribution, but resource limitations have delayed some projects.

As a result of the resource constraints, IFSAC officials developed an interim strategic plan, which describes the group’s accomplishments during 2017-2021 and identifies key activities for 2022-2023. Near the end of the interim period, the group intends to share information about its direction, goals, and approaches to to future work.

“During the years 2022–2023, we will continue to publish annual reports on foodborne illness source attribution for priority pathogens. We will continue to improve methods for estimating foodborne illness source attribution using outbreak and sporadic — non-outbreak-associated — disease data, pursuing external collaborations as needed to maximize capabilities and access to data sources,” according to the group’s interim plan.

Specifically, IFSAC leaders intend to give priority to the following activities:

• Analyzing trends in foodborne disease outbreak-associated illnesses over the past 20 years and submitting a manuscript to a peer-reviewed journal describing its methods and results.

• Continuing to develop and refine machine-learning approaches to predict the food sources of human illnesses with unknown sources by using whole genome sequencing (WGS) to compare Salmonella isolates of known sources with those from ill persons whose source is unknown.

• Adapting the WGS-based methods developed for Salmonella to attribute sporadic Campylobacter illnesses to food sources.

• Assessing the frequency of multi-year outbreaks and their impact on source attribution analyses and deciding whether to improve the methods for using them in outbreak-based source attribution models.

• Collaborating with FoodNet staff to estimate population attributable fractions for key food sources of sporadic Salmonella Enteritidis and Campylobacter illnesses by developing case-control studies using FoodNet case exposure ascertainment data and FoodNet Population Survey data.

• Continue to develop a method to incorporate into attribution estimates those outbreaks associated with complex foods (i.e., multi-ingredient foods) for which the contaminated ingredient is unknown.

“IFSAC’s primary focus continues to be improving estimates of the food sources of illness caused by major pathogens. Whereas IFSAC has always focused on estimating the sources of all — not just outbreak- associated — illnesses, the methods have thus far only used data from outbreaks. In recent years, we developed methods to use data from sporadic illnesses to make these estimates,” according to the interim plan.

“In this interim period, we will continue to evaluate our approach to attributing Campylobacter illnesses to specific food categories. Our recent reports have highlighted the challenges associated with attributing Campylobacter illnesses to foods based on outbreak data, due to the outsized influence of outbreaks from foods not widely consumed but with high risk of illness, such as unpasteurized milk and chicken livers.”

IFSAC’s work from 2017-2021 included:

• Developing a recency-weighted statistical modeling approach to estimate the sources of foodborne illness caused by specific pathogens and published the method in a peer-reviewed journal.

• Updating IFSAC’s scheme for categorizing foods implicated in foodborne disease outbreaks, and described it in a peer-reviewed published article.

• Producing annual estimates of the sources of foodborne illness for Salmonella, Escherichia coli O157, Listeria monocytogenes, and Campylobacter for 2015–2019.

• Analyzing outbreak data on Salmonella illnesses acquired from consumption of pork products to support draft USDA performance standards for the minimum acceptable Salmonella contamination of pork cuts and ground pork products.

• Continuing to explore new methods and models for foodborne illness source attribution, including random forest and other machine learning algorithms.

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

The aim is to help develop a picture of the food system, its safety, authenticity, and risks and vulnerabilities, so issues can be better managed.

Speaking at IAFP Europe, Julie Pierce, director of openness, data and digital, said she had to persuade the FSA that putting funds and faith into data was a good idea.

“I started with a narrow question, answered it and then expanded it into other regions and commodities. We started off with the observation on the amount of aflatoxins in figs from Turkey. We noticed a seasonal variation in the number of alerts we were seeing and could determine the weather was impacting on the aflatoxin level,” she said.

“It was a relatively straightforward model that we built but it was relevant to those grappling with the real life issue. It proved to be relatively straightforward to extend the model beyond figs to other commodities like Bolivian Brazil nuts.”

Prediction to manage resources
Pierce said the cost doesn’t have to be high and data systems can be fast.

“Our experience is the cloud services are not hugely expensive. The technology that manages the transport of data around the system is not where the huge cost lies. The cost and time comes from poor data quality and trying to understand the data you are looking at, do I need to invest any time and effort in validating or cleansing collected data? We spend more time at that upfront part of the process than we do on the downstream development of tools.”

Using machine learning and artificial intelligence, the agency has developed tools to identify risky imported food and feed products coming into the country and scan incidents being reported globally.

Pierce also spoke about a signal prioritization and a risk likelihood dashboard.

“What we are doing now is trying to see further afield and into the future. From our point of view, having that time to predict and plan for and take action to mitigate any risk is so valuable. We have a system that is collecting data from regulators like ourselves but also other sources like trusted websites, translating it into English and combining it all together into an assessment of the risk we are seeing and presenting that to the user. We update that on a daily basis and speed of response has proven to be really important,” she said.

“The risk likelihood dashboard helps our import team, port health and local authorities trying to determine where the risk is and which commodities should be sampled, when and from where. This has been in operation for a year or so and we are continuing to develop it building on the experience of the data science team and our users as to what they now wish to see. Once you give them some insight it is inevitable they will want more.”

Role in helping local authorities and during pandemic
One proof of concept is helping local authorities prioritize their inspections.

“We used artificial intelligence to build a tool that learns from the past to be able to predict which food establishments at the time of their registration would be likely to be most risky. They don’t necessarily have to have started operating but we can predict with a high degree of accuracy what their likely Food Hygiene Rating Scheme (FHRS) score is going to be. That helps local authorities understand the risk and prioritize the inspections they undertake,” said Pierce.

Some of the approaches and tools were repurposed when the COVID-19 pandemic hit.

“The sorts of questions we were posed just over a year ago now where things like what might happen to the resilience to trade? What was happening in real time for our consumers? We’d been listening to social media for a while for potential norovirus outbreaks but we found we could use the same techniques to listen to consumers voicing concerns around whether or not the virus was conveyed through food packaging and if they should wash their food in bleach,” she said.

“Through the pandemic, as hospitality was closed down, we saw many businesses where still trying to operate but changed their operating model and some new ones started up and came online. It’s often very fast and easy to start selling on a digital platform. But where is that physical shop window to display your FHRS sticker? Are you even visible to the regulators? Yes, you are. You have a digital footprint and we can scan the platforms to see who is operating a food business, we can then check to see whether you are displaying your FHRS online and it is the right rating.”

Example projects
Collaborators include the Food Industry Intelligence Network, Food Standards Scotland, British Retail Consortium, British Meat Processors Association, Red Tractor, and the Food and Drink Federation.

Pierce said the FSA is mindful of the sheer volume of data available across the food system.

“We are starting to work with organizations like FIIN and drawing up a data sharing agreement to see whether or not we can share sampling data the industry are taking and share our data back with the industry. With Red Tractor we’ve been exploring whether the digital remote audit approaches work and if application might be taken forward in the longer term. We’ve used BRC accreditation ratings as a potential input to our ability to segment and understand what is going on in the industry at large beyond individual businesses but trying to group together different parts of the sector.”

The agency has done a number of research projects on IoT, sensors and blockchain.

“The blockchain pilot main learning was the technology was easy but governance was hard. We took that learning into a project on data trusts with the University of Lincoln and the Internet of Food Things building on concepts developed by the Office for Artificial Intelligence and Open Data Institute in the UK,” she said.

“While there is a lot we can do with those smaller datasets, really AI is dependent on access to large comprehensive data and information collected from multiple sources. We see the value when we can link different datasets together. What can we do to ensure the right data is being shared in the right way at the right time?”

The FSA is looking at data trusts for sharing data with different parties in the food system and referred to the work as the Open Ecosystem Federation.

“We are working with HMRC and the Cabinet Office to establish a standard set of tools and protocols that will allow us to share the minimum amount of data to allow each of us to perform our different roles at the border with only the data we need to make the decisions we need to so we are not pulling in all the data about the product as it moves across the border,” said Pierce.

“It’s using chicken imports and we are hoping to prove this approach will work for all participants. Our partners at the University of Lincoln are trialing this approach with the fresh produce industry through a program called Trusted Bytes.”

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

While the concept of overall equipment effectiveness (OEE) is straightforward, the rollout can cause huge disruptions if not done strategically. 

Below, Roger Woehl, Chief Technical Officer for SafetyChain Software, outlines five key areas of focus when successfully rolling out OEE: 

Define goals
The first part of implementing any OEE improvement plan is for a business to set clear goals of what it wants to achieve and any potential consequences it may have on other areas of production. 

“OEE is not meant to improve one single point of production. It must consider machine availability, output performance and the product quality to maximize throughput,” says Woehl. 

If the goal is to increase the number of units produced per shift, then consideration must be made to what that does to the quality and safety of end products. 

“A company may find that turning up the speed of production actually has far greater negative impacts on the overall business by jeopardizing quality or safety standards,” he says. “They may meet safety compliance standards but increase product rejections due to failed quality specifications. By analyzing all the data together, they can make a calculated decision on whether or not their goal makes economic sense.” 

Automate your data source
The biggest opportunity to improve OEE is to automate data by capturing real-time events with sensors that feed data into a centralized analysis platform. 

“Manual data is only as good as the person recording it and what they are physically able to capture,” says Woehl. “What tends to happen is that micro-events aren’t caught, like if a machine is temporarily down for two minutes.”

While not recording production gaps like that is common on paper, they can accumulate into a larger output problem and may signal a bigger issue in production that needs to be addressed. 

“Real-time data captures these small micro-events and allows operators to trace back to where the root of the problem is,” explains Woehl. “This allows for a fast and efficient improvement of OEE.”

Rollout line by line
Before implementing new operating systems, Woehl recommends a pilot program be formed with a small group of people to trial and adjust new processes before implementing them broadscale. 

“The main advantage of the pilot process is to work out the kinks and to figure out what works best for the business,” he says. “One of the biggest mistakes a company can make is attempting to improve everything all at once. This can cause huge disruptions and headaches.”

Instead, he advises new processes are rolled out one line at a time, and for practices to be made broadscale as soon as a company has reached a level of control for handling all the continuous adjustments for improvement.

Essential to the success of this process is bringing people onto the pilot program team who will buy into implementing change and will not badmouth the process as they work through issues, he says. 

“When implementing change into a business and introducing new technologies, it is really important for everyone involved to be on board with making it work. This starts with your pilot team supporting the goals and a commitment to help carry that support throughout the business,” he says. 

Develop a continuous improvement program
“The idea of continuous improvement is an important theme in the food industry. You can’t always be perfect, but you can work to continuously improve,” says Woehl. 

According to him, improvement is most quickly seen when a company establishes an attainable threshold and tackles it in small, incremental steps. 

“Let’s say a company wants to reduce its downtime from a shift that is currently 1 hour and 45 minutes. The first goal would be to shave off 15 minutes. To achieve that, we would look at all the different areas contributing to downtime and pick one to improve,” he says. “In this case, the area of focus was reducing the changeover time of product packaging materials. We can then zero in on that one process and look for ways to improve it.” 

Once the first identified area has become more efficient, Woehl says to pick another area to shave an additional 15 minutes off. 

“Setting realistic goals that can be accomplished in small bite-sized chunks will allow a company to improve OEE much quicker than if it tries to address multiple areas all at the same time,” he says. 

Cost analysis
Once a company starts seeing the monetary benefits of improving OEE, it must continue to monitor production to maintain performance. 

“OEE is like tuning up a race car. A one-time tune-up will get it running well, but only for a certain amount of time. If you want it to operate at its best continuously, then it is going to require regular adjustments,” says Woehl. “If a company invests the time and tools necessary to improve OEE and then discontinues monitoring production once they reach their target, OEE will eventually decline.” 

To justify investment into improving and maintaining OEE, Woehl says to look at the return seen within a specific time window, such as three months, and then consider the cost of lost production if it were allowed to decline. 

“When looking at pure OEE – which considers machine availability, performance and product quality – it can be used as a huge competitive advantage. Having accurate, real-time data that takes into account the entire business picture allows a company to find what type of production model is right for them and opportunities to improve output efficiency,” says Woehl.  

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

The Food Standards Agency (FSA) commissioned the study by RAND Europe in February 2018. In late 2017, the agency started work on an approach to strategic surveillance for food safety that is data-driven, proactive and flexible. RAND Europe is a not-for-profit research organization that works on public policy.

The present approach to food surveillance includes random product sampling by local authorities. Instead of traditional sampling, the FSA wants to use modern analytic techniques and diverse data sources to create a new way of identifying risks to public health from food.

Researchers created an approach to strategic surveillance, to selecting and prioritizing topics and improving communication and coordination.

An FSA spokeswoman said the agency is still developing the approach to surveillance.

“We commissioned this research to assist with the design and implementation of a strategic surveillance system for the U.K. food system and are currently working on implementing the recommendations made by RAND Europe,” she said.

The recommendations are:

• Selecting and prioritizing topics should be undertaken using a data-driven, repeatable and informed decision making process;
• FSA should undertake forecasting and horizon scanning activities to identify unknown unknowns;
• The surveillance program should continue to identify and engage stakeholders;
• There is a need to define the process for acting on and communicating assessments;
• FSA should increase its visibility amongst stakeholders to assist data collection and future action;
• The approach should continue leveraging existing data and link together data sources to exploit insights;
• FSA should compile a data catalogue based on metadata; and
• Skills, knowledge and experience should be captured (i.e. in the form of a skills and methods catalogue) to assist the FSA to structure problem solving and implement the approach.

RAND Europe analysis took a top down approach by using existing FSA models and a bottom-up look by clustering themes from interview data to produce a proposed operating model for strategic surveillance. The research team interviewed senior members of the FSA to determine system requirements.

“The overall approach was designed to be flexible, evolutionary and adaptable, instead of a rigid process; it provides a guiding structure and foundation upon which the FSA can start implementing strategic surveillance and evolve the approach as the FSA grows in maturity with respect to surveillance,” according to the report.

The seven week study designed an overall approach to strategic surveillance with five steps: plan and direct; collect and collate; analyze and produce; report and disseminate; and evaluate and review.

The initial operating model designed by the FSA had the four main steps of scan, spot, narrow and evaluate, as well as a fifth step of share/publish that leads to action such as. starting an investigation, issuing a product recall or notifying other government departments or agencies outside the surveillance system.

To ensure the approach met the FSA’s needs, the study team tested it against two cases previously looked at by the FSA under its surveillance program; olive oil adulteration and contamination of shellfish with Vibrio bacteria. Testing showed the approach was appropriate for the two cases since there were no steps undertaken in them that did not appear in the overall approach.

The FSA is aiming to have the recommendations completed by early next year.

Meanwhile, the National Audit Office is looking at effectiveness of current regulatory arrangements to ensure that food is safe to eat and is what it says it is.

The report is due this summer and builds on work from October 2013 on food safety and authenticity in the processed meat supply chain.

NAO is looking at whether the FSA has evidence that the regulatory system is effective in achieving outcomes and driving performance improvements and if the Department for Environment, Food and Rural Affairs (Defra) and FSA have established a coherent and coordinated regulatory regime that can adapt to current and emerging challenges.

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

While estimations vary to the extent of which IOT will be relied upon in the future, one thing is certain – the food industry is not about to get left behind.

Here are four ways food companies are utilizing IOT technology:

Food safety efficiency:

Perhaps the biggest payout for investing into IOT is the ability to closely monitor food safety data points, which in turn helps reduce the risk of a food illness outbreak. Topping the list is the use of real-time temperature tracking sensors that have the ability to monitor a product’s temperature from the time it leaves a farm to the moment it leaves the grocery store shelf.

This type of technology is beneficial as shippers, receivers, loaders and carriers all work to become compliant with the Food Safety Modernization Act’s Final Rule on temperature control and tracking requirements. It also carries over into assisting the tracking of Hazard Analysis and Critical Control Points, and the ability to provide food companies with automated HACCP checklists for more coherent data collection and reporting that paper-based logs are not able to monitor as efficiently.

Less waste:

According to the Food and Agriculture Organization of the United Nations, one-third of human food production is lost or wasted globally each year. While some of this is food waste on the consuming end, food waste due to compromise during the transportation and distribution portion of the supply chain contributes to this whopping 1.3 billion tons per year waste estimation. However, with a centralized cloud data base tracking real-time control points for food safety, companies have the ability to integrate it into their business management decisions. For example, if a load of broccoli from Mexico experienced a temperature compromise during shipment, a company can use that knowledge to decrease the shipment’s shelf life so a retailer can act accordingly to get it on a consumer’s plate before it spoils.

Coherent records:

To stay in compliance with FSMA, food companies exceeding $1 million in revenue each year is mandated to keep at least two years of food safety records on file.

It doesn’t stop there, with shippers and carriers required to have on hand 12 months of transportation and training records.

While paper-based records are still industry standard, companies have a lot to gain by moving towards a cloud-based portal with sensors automatically streaming into the online data base to not only cut down on labor, but also human error.

Interconnected data analysis:

While the three previous points all revolve around specific information points, IOT technology has opened the door to endless data analysis by companies.

Consumer trends, logistic data and anything else that can be measure automatically can all be combined together to make more calculated and grounded business decisions.

A decade ago, the food sourcing sister to C.H. Robinson, a third-party global logistics giant currently operating in 36 countries, was operating on spreadsheets and watching its growth rapidly outpace the capabilities of its technology.

“I was thinking, ‘This is not sustainable because we were growing and growing,’ ” said Arnold, director of food safety at Robinson Fresh. “So we looked into a lot of different systems to help manage all of our data and partnered with a company to tailor a program to not only fit our needs, but our growers’ needs and customers’ needs as well.”

Its data tracking program, which utilizes the Internet of Things, otherwise known as IoT technology, has revolutionized the way Robinson Fresh tracks essential food safety and business management metrics through efficiently connecting the complex web of channels within the company.

Want to know the results of the surprise audit completed this morning or test results from water samples taken by a grower in Brazil? Once upon a time, it could have taken days to fully answer questions like these, but now all Arnold has to do is log onto the online cloud storage base and it is all there.

Like all new and flash things, the latest and greatest technology is very attractive — sexy even. But without purpose, it is essentially worthless. Will Daniels, president of Fresh Integrity Group Inc., said for a food company to be a fit for IoT, it must first reflect and identify what is needed to bring value to the investment.

“The company needs to be hungry for the data — hungry for the results of the data and a desire to use that data,” Daniels said. “In my opinion, it is somewhat of a challenging task to connect this Internet of Things to be meaningful, to make sense.”

According to Daniels, companies should implement a short-term IoT plan by going after low-hanging fruit and identify easy connection opportunities that require minimal work to activate. Then, go after the long-term plan with a bigger picture and more detailed oriented mindset.

“For example, at a freshcut processing facility, there is a lot of data being collected around wash line performance, such as temperature and chemical levels,” Daniels said.

“It is really good data when everything is working right, but the sensors collecting it can cloud over and become less reliable. So if you don’t have a maintenance crew who understands the importance of keeping those sensors maintained, you lose the value.”

This is a common problem Daniels has seen when working with food companies. Sometimes it happens when maintenance contracts change hands, sometimes when there’s turnover within a company. To avoid inconsistencies in IoT collection and interpretation, he recommends food companies build a system of process management that can transfer from one owner to the next.

While integrating IOT into a food company is a complex and resource consuming process, the payouts can be significant.

For Robinson Fresh, the ability to pinpoint and fix a problem immediately though data tracking while maintaining a connected global supply chain has made IoT indispensable.

“Do you research on a company and find one that fits your operation,” Arnold recommends to food companies exploring IoT. “This is definitely where we are going as an industry, and if you are not doing it yet, then I suggest you take the steps to start.”

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

For the amount of information gathered, the man power to make it all happen should be staggering, yet none of this data grunt work is done by a single person. How is this possible?

Through the so-called Internet of Things.

To put it simply, Internet of Things, aka, IoT, is the connection of one computing device to the internet, which may interconnect to another computing device and exchange information. For the average Joe consumer, this means turning a car on with the tap of a cell phone or receiving an alert when their washing machine rocks out of balance while they’re out of the house.

For the food industry, IoT means an efficient way to track food safety data points with things like real-time temperature tracking throughout the entire harvest to grocery checkout process, storing information in a central location for analysis.

Elliott Grant, founder of Harvest Mark — a leading traceability company in the food industry — says IoT is made up of three things:

1. Unique identity
2. Sensors
3. Ability to communicate

“The connection of the internet has gotten so low cost that you can put sensors in things and places that were previously not possible,” Grant says. “The real magic is when the sensors start to communicate clearly and the data from each sensor can be integrated across the supply chain.”

The caveats, he says, are two-fold when it comes to IOT. For one, food companies have the ability to closely collect and track essential food safety measurements to take Hazard Analysis and Critical Control Points efficiently, which can also be used to monitor quality by integrating data from multiple devices in the food chain.

“Imagine an inspection that happens at a retail store with a handheld device measuring freshness. You then correlate the readings from that device all the way back to the supply chain with the time temperatures taken from the trucks, with the measurement that might have been taken on the farm and perhaps a measurement taken during the growing season,” Grant says, adding that the integration of data points could lead to predicting any issues.

“I think that’s the specific reason for IoT. It is not necessarily responding to an event, but predicting and anticipating events in the future.”

The biggest caveat, he says, is the ability to get a return on investment for the technology. According to Grant, this is where food companies stand to gain the most, however, where they must crack down the hardest to see if IoT is right for them. His recommendation is for food companies to walk through the system manually to collect measurements to understand what can be gained, and whether that improvement would be beneficial on a large scale automatic system.

“Having more data doesn’t necessarily give you more insight. So my recommendation is to identify what are the key drivers of risk and opportunity that could potentially be improved with more measurements,” Grant says.

“Before going out and building an infrastructure, go out and take a thousand measurements by hand to see whether or not you get any benefit from that (data). Forget the cost of doing it by hand. But does the data actually give you anything? Only then, when you have some insight from a small sample, then you can go ahead and say ‘let’s spend some money and do this thing at scale.’ ”

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

• campylobacter was present on 50 percent of chicken samples, down from 71 percent in the three months from December 2014-February 2015.
• 9.3 percent of chickens tested positive for the highest level of contamination in this quarter, which is down from 21.8 percent for the three months from December 2014-February 2015.

The agency says the results are very encouraging, and one of the reasons the survey results are lower for this quarter is the action recently taken by retailers and their suppliers to remove neck skin from the bird before it goes on sale. Neck skin is the most heavily contaminated part of the chicken, so its removal is a positive step for reducing the risks to consumers. However, as the survey design has been based on the testing of neck skins, its removal means that detailed comparisons with previous results are not possible. For this reason, the most recent results have been presented as an overall figure for the amount of campylobacter on chicken sampled across the UK, in contrast with previous results which provided a breakdown of figures by retailer. It should also be noted that this survey has now stopped, and a new survey will begin in the summer with a different method for testing campylobacter levels on chicken. The results from this new survey, which will rank the results obtained for each of the retailers, will come out in January 2017. “FSS is committed to on-going research to improve our understanding of the most important causes of campylobacter in humans in the Scottish population, and I welcome the improvement in these latest results,” said Elspeth MacDonald, FSS’s chief deputy executive.  “Improving the health of consumers in Scotland is a key priority for FSS and we look forward to on-going collaboration with the FSA and industry, to continue moving in a positive direction.” FSS continues to advise consumers in Scotland that chicken is safe to eat as long as good kitchen practice is followed to help avoid cross-contamination, and chicken is cooked thoroughly. No data was available for March 2015.   (To sign up for a free subscription to Food Safety News, click here.)