IFT's Traceability Improvement Initiative aims to advance work in the area of food product tracing through several means including hosted events where thought leaders exchange knowledge and ideas. In August 2011, the Initiative, in collaboration with GS1 US, convened a group of 50 product tracing stakeholders, as a follow-on to a successful event the month prior. Representatives conducting pilots or implementation studies in produce, seafood, dairy, and other industries discussed the objectives, challenges and learnings. Some of the learnings from on-going initiatives included the sense that better information management provides a return of investment; data often exist but may not necessarily be appropriately linked through the supply chain; and enhanced product tracing enables better accountability and quality control. Challenges identified in enabling traceability throughout the supply chain were the distribution complexity; the need for training, communication, and collaboration; improving the reliability, quality and security of data captured, stored and shared as well as the importance of standards in data and interoperability of technology. Several approaches to overcoming these challenges were discussed. The first approach incrementally improves upon the current “one up/one down” system by requiring electronic records and tracking internal as well as external critical tracking events. The benefits of this approach are its similarity to existing regulatory requirements and low cost of implementation; resulting in a higher probability of adoption. The major disadvantage to this process is the longer response time required during a trace (back or forward). The second approach is similar to a “pedigree” approach where historical information about the food travels with it through the value chain. A major advantage of this approach is the quickest response time during a trace. Some of the disadvantages of this approach are potential for misuse of data, the volume of data required to be maintained at value chain end points, and data privacy concerns. The third approach requires individual nodes within the value chain to maintain electronic records for its own data and make them available for querying during a traceback for outbreak investigation. The major advantage of this approach is the protection of confidential information and the potential for quicker access during a trace. However, the primary disadvantage of this approach is the need for greater computational power and a more complex mechanism to linking the value chain through the data. As next steps, a subgroup will work on clarifying the approach to meeting the goals of traceability, better defining critical tracking events, and articulating the strategy and return on investment from a regulatory and industry perspective. This will result in improved alignment of on-going traceability pilots and initiatives as well as a more actionable guidance document for public review.
In January 2011, President Obama signed into law the Food Safety Modernization Act (FSMA), which in addition to a number of new requirements related to food safety, carries new product traceability requirements aimed at enabling faster and more targeted product identification and recall in the event of a public health threat. The U.S. Food and Drug Administration (FDA) must oversee implementation of the FSMA product traceability provision and specifically has been required by Congress to conduct 2 traceability pilots (1 for fresh produce and 1 for packaged foods) and submit its recommendations to Congress by mid-2012.
The Institute of Food Technologists (IFT) in collaboration with GS1 US convened a group of industry leaders with the goal of defining a vision for product traceability that can be widely disseminated and adopted. The group consisted of approximately 50 leaders from the supply and demand side of the food industry, as well as a number of academics, technology providers, and regulators (federal, state, and local).
To date, the group, with support from IFT's Traceability Improvement Initiative, has held 2 meetings to align on the general scope of a voluntary industry-wide product traceability approach that could be implemented in cooperation with the FDA and that might preempt more stringent requirements that are likely to emerge if the FDA were to act without the participation and cooperation of industry. The immediate next step identified at the July 2011 Traceability Research Summit was “the development of a common vision for a simple, low cost, acceptable and implementable traceability system model”. It should be noted that for the purposes of discussion and this document, the term “system” can be used interchangeably with “approach”. System is not intended to mean a proprietary technology or specific process. There was also a desire to highlight on-going pilot initiatives within the industry in an effort to not re-invent the wheel. To meet these 2 objectives, at the August 2011 Traceability Research Summit, IFT invited several industry and association representatives to describe their involvement with pilots and real-world projects. In addition, several approaches were presented to the group in an effort to gain consensus on a common vision and goal for traceability and a framework within which to achieve it.
Food Safety Modernization Act's (FSMA) Traceability Requirements and a State Regulator's Perspective
The FSMA requires the FDA to recommend traceability and record keeping requirements with the following parameters.
Cannot require records of recipients of a food beyond the immediate subsequent recipient of such food; Full pedigree (that is, farm to fork) cannot be required
Case level tracking cannot be required
Enhanced recordkeeping requirement will apply to “high-risk” foods.
In meetings to date, the FDA has stated their interest in partnering with industry to define an appropriate approach to product traceability. They have also strongly encouraged industry to reach beyond the very narrow scope of the FSMA in defining its traceability approach. At this point, it is not clear how the FDA will go about rulemaking given the requirements provided in the FSMA; however, when considering comments from State and Federal regulators in meetings to date it is becoming clear that:
Regulators prefer a solution that allows them to target potentially affected product with precision and accuracy
Factors that could be considered for differentiating between high-risk and low-risk products include frequency and severity of outbreaks, shelf life, and value and volume of products in commerce.
Formal guidance from the FDA on these points is not expected until the completion of FSMA-required pilots and perhaps not until mid-2012 when their recommendations are due back to Congress.
Ben Miller, who manages the Rapid Response Team and leads traceback investigations at the Minnesota Dept. of Agriculture, provided some real-life examples of the issues faced by investigators. He began by describing the investigative process, beginning with the onset of illness, and noted the time that it takes before a problem may even be realized. This lag time owes to the time it takes between consumption and illness, the time for medical testing (if done at all), submission to public health laboratories, and additional testing and analysis. Epidemiologists look at the totality of illness, relying heavily on pulsed-field gel electrophoresis patterns and PulseNet, the system that collects the “fingerprints” of bacteria causing illness around the country, to determine if an outbreak is underway. It is only after this recognition, and interviews with affected consumers, that regulators begin collecting records to determine if food(s) causing illness came from a common source. Miller noted that in many cases, good records stop before the supply chain stops, which hinders the ability of regulators to rapidly evaluate supply chains and identify points of convergence of products. In other cases, he commented that the quantity of data can be overwhelming. In most instances, investigators need to manually match records from 1 supply chain point with the next, which is a laborious process, but 1 that requires skill. A system that allows data to be more readily assimilated would ease the burden on regulators to evaluate records and should allow more rapid resolution of a traceback investigation.
Recent Industry- and Association-Led Pilots and Projects
Several technology solutions providers, food industry and trade association representatives were invited to present their traceability-related pilot studies and their findings. The collaborative environment was created to learn from existing efforts and discuss not only the challenges being faced by all stakeholders, but also the merits and cautions of proposed solutions.
The following traceability studies and initiatives were presented during this research summit:
Product Traceability Initiative (PTI) by Produce Marketing Association
FDA- contracted Mock Tomato Pilot by IFT and Darden
Produce Pilot by Underwriters Laboratory and FoodLogiQ
Fresh Food Pilot by GS1 US
Manufacturing Pilot by NWFPA and OSU
Food Service Pilot by UFPC
Technology Interoperability Study using Seafood by IFT and Bumble Bee
Seafood Projects by Trace Register
Dairy Products Traceability Program by US Dairy Export Council
Industry Pilots by TraceTracker.
The food products represented by the previous list of pilots include: raspberries, carrots, strawberries, tomatoes, oranges, processed foods, beef taco, canned tuna, and dairy, among others. The breadth of stakeholders involved with these initiatives included: growers, transporters, distribution centers, retailers, food-service, processors, bulk suppliers, technology providers, standards organizations, regulators, academia, and trade associations.
One of the primary motivators for conducting a pilot was the need to improve the speed and accuracy of any recall or trace-back investigation. Stakeholders recognized the need to reduce the current response time from a few days to a few hours. Everyone also understood the importance of fostering collaborations between regulators and the industry in the event of an epidemiological investigation or recall. Conducting pilots and similar mock exercises was 1 way to develop these relationships in a low-stress environment. It also affords the opportunity for mutual understanding and respect for each other's roles and responsibilities during a crisis situation. With so many options available (from a technology standpoint), some pilots attempted to evaluate the applicability of these techniques (like bar codes and RFID) to best suit their specifics needs.
There were many real-world benefits that resulted from these pilot studies. Some were expected, like improved speed and efficiency of response; while others were ancillary, like better cost controls due to more streamlined processes and improved traceability through the transportation system (after the product leaves the shipper and before the product arrives at the receiver). Reduction in waste and improved yields from better tracking was another benefit identified in 1 pilot. Capturing data electronically resulted in significant costs savings in the long run but also improved the quality of their products and processes. A literature review conducted by the dairy industry also revealed that certain countries had better traceability systems in place relative to the United States, partly driven by more stringent regulatory requirements, and there was a concern about maintaining competitiveness internationally.
Several challenges and unresolved issues also emerged during the execution of these pilots. Lack of standards for data capture, storage and sharing resulted in a less efficient traceability system. Missing links and data across the supply chain as well as a difference in terminology resulted in some confusion during a mock traceback. There was a need to tweak the traceability requirements based on the characteristics of the product being tracked (for example, shelf life). Non-technical factors such as appropriate employee training on modified business practices or updated standard operating procedures were crucial to the success of a pilot. Another significant barrier to improving traceability was the varying requirements from trading partners on what each one believed to be key data elements and critical tracking events for their value chain. Harmonizing these efforts would result in lowering the cost for compliance with the various requirements as well as speeding up any potential recall or trace investigation. While some pilots focused their efforts on identifying appropriate data collection practices at each point in the supply chain, others looked at how these data may be shared or synchronized more effectively across trading partners while maintaining privacy and confidentiality of the data. Due to the wide variations in the capabilities and processes of existing commercial traceability technology solutions, a need was also identified for the development of an audit standard to ensure any future traceability system, at a minimum, enables whole-chain traceability. Finally, there was consensus that as long as there is interoperability, a number of technological approaches can be adopted by various sectors within the supply chain and result in an effective tracing system.
The following table summarizes some of the conclusions and challenges that emerged as themes through the discussion of the various pilots and studies (Table 1).
Table 1. Themes of product tracing pilots and studies.
Better information management provides return on investment
Complexity of value chain
Data often exist but can be better used to provide product tracing
Training, communication, and collaboration between all stakeholders
Enhanced tracing enables better accountability and quality control
Reliability, quality and security of data captured, stored and shared
Pilot learnings need to feed operational projects
Standards in data and interoperability of technology
Approaches to Improve Product Tracing
Before identifying an approach to traceability, the participants of the summit sought to define the goals of an ideal traceability system. There was general consensus within the group that the goals, at a minimum, should be based on a food safety and traceback perspective in an effort to protect public health. Regulators would need access at all points in the supply chain—however, access and needs would vary within industry participants. It is of paramount importance to identify how to positively affect public health and consumers and adopt a system or approach that maximizes public health benefits. It was noted that the motivation for the implementation of some systems that are now used for traceability may or may not have initially related to traceability, and may or may not be fit for multiple purposes. There may be a need to balance the sometimes conflicting goals of business return on investment (ROI) and the public health, noting that industry will not change their practices solely to improve traceability for the purposes of protecting public health. A strong ROI case along with regulations may be required to encourage adoption.
In order to achieve these goals, the 3 approaches depicted in Figure 1 and 2 were discussed (Approach 1 and 2 are methodologies and Approach 3 is a collaboration platform that could be used in conjunction with the other 2).
This approach is similar to the current US Bioterrorism Act of 2002 which requires the maintenance of records for 1 level up and 1 level down a supply chain. Each supply chain participant is responsible for collecting incoming shipment records from its suppliers, capturing any transformations within its own facilities, and retaining outgoing shipment records to its customers. The primary difference between this approach and the status quo is the requirement for electronic record keeping of traceability data and standards built around the processes for improved efficiency. Internal traceability (for example, keeping track of which ingredient lots went into which finished good lots) is also a key component of this approach and cannot be excluded.
The advantages of this approach are its familiarity due to its alignment with current mandates resulting in a higher probability of participation by the industry. Because many companies are already in compliance the cost for adoption would be relatively low, it would be relatively easier to build consensus around this approach. If executed correctly, it preserves confidential supply chain trading relationships. An electronic record keeping requirement would significantly speed up the process of investigating potential contamination events in the supply chain. There is also some data redundancy with this approach (compared to Approach 3) which could be useful for validation and verification purposes.
The primary disadvantage of this approach is that during a trace-back investigation, regulators would have to go to each node in the supply chain, (consecutively) to identify convergence and the potential source and impact of a contamination. This linear process would increase the workload of government regulators as well as cause a delay tracing products through the supply chain.
This method is similar to what many consider a pedigree approach to traceability where each downstream recipient of a food product receives a record of the history of that product's movement through the supply chain. Historically, pedigree has been used to verify the authenticity of an item and address counterfeiting concerns.
The primary advantage of this method is that it gives regulators access to the largest amount of information in the smallest amount of time. It also has some value added benefits that go beyond traceability—like verifying the point of origin. An important issue related to protecting public health is informing and educating the consumer about potential recalls. Depending on how the consumer unit was labeled, there is also the potential for consumers to more readily identify whether or not they possessed a potentially unsafe product.
However, there are several potential disadvantages with this approach. There is a risk for poor or erroneous data with no easy way to validate the pedigree information provided to a trading partner. It also results in an exponentially larger data set as the product moves towards the end-users in the supply chain. Due to the large number of transactions and the large size of each transaction, there would be an added strain on any technology infrastructure used to facilitate these transactions (for example, EPCIS). A bigger concern was the loss of confidentiality when revealing upstream trading partner information to downstream customers. Due to the increased use of technology to manage data, there is also an increased potential for hacking. Technology use may also result in a large initial capital investment while still increasing the time (compared to Approach 1) required for capturing and storing all the data.
This approach outlines the potential use of cloud computing as a technology collaboration platform to enable whole-chain traceability. Practically, this approach could be used in conjunction with Approach 1 or 2 since the latter are more closely related with business operating practices which may be enhanced through the use of technology. The system would not be a central-data repository (to minimize scaling issues) but more of a hub or registry with both people-to-people and machine-to-machine communication facilitated through standards via data request and authorization for information. A registry would show what data are available and where, along with pertinent information based on authorization levels.
The primary advantage of this approach is the responsibility of each supply chain participant to maintain records for their own data. This decreases the risk to accidental release of intellectual property or confidential and proprietary data. If implemented correctly, it has the potential to provide better information to regulators in shorter amounts of time when compared to the other approaches.
Some of the disadvantages to this approach are the greater need for computing resources and an increase in implementation complexity. This approach could also be confused with Approach 1. There may also be confusion resulting from a lack of understanding of technology like cloud computing which resulting in a lower rate of adoption. There was hesitation on how a universal querying system could be created to fully utilize the benefits of this approach (similar to how we use Google to search for web pages today). Finally, it would require more effort from the regulators to conduct a traceback investigation (when compared with Approach 2).
Data Capture and Sharing Considerations
Attendees of the summit were asked to discuss the potential data capture and sharing requirements that would need to be considered when implementing the end-state system. In attempting to do so, most engaged in a discussion of defining and describing critical tracking events (CTE) and key data elements (KDE) for each CTE.
At a minimum, KDE includes “Who, What, Where, When and Why?” for each supply chain CTE. This could include the origin of product facility, deliverer and/or receiver, the capture timestamp of date and time and the item number which could include the production date, lot number and quantity. A critical tracking event was defined as the transfer of or the change in the state of a product (be it location, ownership, or composition).
There was agreement within the group that CTE/KDE should only be shared when there is a formal request for information (RFI; ideally an electronic RFI). This should be done in a standardized format with a predetermined, agreed upon list of information that could be requested. Some attendees currently use the Data Distribution and Interchange Protocol (DDIP) which have fields preselected that are approved to be shared. DDIP fields include CTEs and KDEs as well as other identifying information. Depending on the end-state, this information could be uploaded to a secure site upon FDAs request with an electronic confirmation of receipt. Regulators could also provide a downloadable easy-to-access excel file designed based on guidance from the industry. The group felt that RFIs should only be made during a limited set of circumstances, like an on-going public health emergency, an epidemiological or trace-back investigation, recalls, or site inspections and audits. RFIs could also come from trading partners who found a potential contamination during routine quality assurance / quality control tests and would like to investigate its cause. Data sharing with trading partners could be done on a contractual basis with specific defined criteria for when RFIs can be made and what data would be shared.
Finally, a discussion about the timeline and implementation of the end state vision was pursued. The first step identified was a clearer definition and description of CTEs and KDEs. The development and adoption of standards was required. The need for development of a collaboration platform that enabled sharing while maintaining data privacy concerns was recognized. Once these basic foundational elements have been developed, pilot implementations would need to be conducted for further refinement, followed by training (of industry stakeholders and regulators), eventual implementation and finally, ongoing verification and audits.
Evaluation of Approaches
Over the course of 2 meetings, the IFT Traceability Group has been working to align on the basic scope and approach of a traceability solution that will meet FSMA requirements and is also achievable both operationally and economically by industry. A number of key “scoping” parameters were debated by the group; the details of each are outlined subsequently.
One up, one back versus full pedigree
In a one up, one back traceability scenario, each supply chain participant is required to maintain records of ingredients and finished product movement 1 step forward and 1 step back in the supply chain. In a full-pedigree scenario, each supply chain participant is required to maintain records of ingredients and finished product movement for themselves and each supply partner ranging all the way back to the raw source (for example, the farm). Essentially the trace back information flows with the product from 1 supply partner to the next.
Although the group acknowledged the positional benefits of full pedigree to provide complete trace back information more quickly and without the intervention of a third party (for example, the FDA or another regulator) to compile information from multiple sources, this approach was eliminated due to:
Increasing risk of missing or inaccurate trace back information as data is passed from 1 supply partner to the next;
Need to communicate track and trace information for all transactions rather than limiting reporting to products that may be contributing to a potential public health threat at a specific point in time;
Risk of confidential supply relationships potentially being disclosed to downstream supply partners; and
Increasing burden placed on supply participants at the end of the supply chain to manage and report full trace back information for all products that they trade.
Ultimately, the group agreed that the one up, one back scenario, perhaps implemented using Approach 3 (resulting in a hybrid approach) was the preferred approach since it is most similar to the Bioterrorism Act requirements that companies comply with under current law. The challenge will be for the industry to align on standards for track and trace data that is being captured and to implement systems to automate the data capture, storage, and reporting of track and trace information to regulators.
Case level versus pallet level
Although the FSMA has limited the scope of traceability, there is a growing opinion among some members of industry, regulators, and academics who participated in the Summit discussions that pallet level tracing may not be sufficient, based upon the following factors:
The “true” logistics unit is a case, not a pallet
Traceability needs to support product flows throughout the full supply chain, including as pallets are broken down by distributors and retailers into individual case level shipments
Supplier operating procedures cannot guarantee that pallets are limited to only 1 batch/lot (for example, prebuilt displays, repacks, and so on).
The increasing number of retailer requests (including some due to the Produce Traceability Initiative and other aligned initiatives) for case level labeling for the purposes of traceability also support a case level approach. While the debate over case level versus pallet level is far from over, an indication from the majority of stakeholders is that we may be heading towards a case level traceability solution.
Other scope related parameters addressed
Several other scope related parameters were addressed by the group with general alignment being achieved as follows:
Standards (for example, GS1) should be leveraged when defining track and trace data requirements as well as data capture (for example, barcodes) and data sharing requirements
Track and trace data should be reported only to regulators and should not be made visible to parties that could use it to gain competitive insight (for example, confidential supply sources)
Track and trace data should be reported on an as needed basis, when requested by regulators
Track and trace data requests should be for a specific time period and limited to products that have been implicated in a potential public health threat
The roadmap for traceability implementation should consider what is practical and achievable by industry participants who may need to invest in equipment, systems and/or process changes
Supply chain participants will need to ensure that proper internal traceability systems and processes are in place to track inbound ingredients, conversion of ingredients to finished good by batch/lot and sale or transfer of finished goods to internal or external trading partners.
With the industry aligned conceptually on a high level approach and scope for traceability, the group proposed that the next step was to develop a more detailed working document that can be used to formalize the definition and application of key data elements and critical tracking events. Volunteers were solicited to propose details to the larger group around the following issues: overall approach, including examples; CTE/KDE definition and application; and the return on investment for public health and businesses. The group felt that a third face-to-face meeting was needed to review the deliberations of the small working group.
IFT greatly appreciates the financial support provided to the Traceability Improvement Initiative by silver level partners BASF Nutrition & Health and Underwriters Laboratories, and bronze level partner the Natl. Fisheries Inst.'s Fisheries Scholarship Fund. IFT also expresses gratitude to GS1 US for sponsoring the conference facilities for this Summit.
Finally, IFT appreciates the efforts of Dr. Bruce Welt (Univ. of Florida), Gay Whitney (GS1 US), Steve Mavity (Bumble Bee Foods), and Mary Wilson (GS1 US), who facilitated the breakout sessions, and Patricia Buccheri (GS1 US) and Rebecca Humora (GS1 US) for capturing the essence of the conversations.
In addition, the active participation of all attendees was critical to the success of this event, and IFT thanks everyone who took the time to attend:
David Acheson, Leavitt Partners LLC; Douglas Bailey, USDA-AMS; Chris Balestrini, GMA; Bob Bauer, Association of Food Industries; Tejas Bhatt, IFT; Elizabeth Board, GS1; Tom Bowman, Underwriters Laboratories Inc; Jonathan Brania, UL; Pat Buchheri, GS1 US; Greg Buckley, PepsiCo Inc; Bob Bunsey, SPEDE Technologies; Jeffrey Chester, Food Service of America; Robert Decker, Deloitte; Clay Detlefsen, IDFA; Sue Fangmann, McDonald's Corp.; Faye Feldstein, Deloitte Consulting; Angela Fernandez, GS1 US; William Fisher, IFT; Andy Furner, Trace Register; Mark Galletta, Nestle-USA; Patty Harvey, ConAgra Foods; Michael Hausman, Center of Excellence; Craig Henry, Deloitte & Touche; Ana Hooper, Darden; Rebecca Humora, GS1 US; Amanda Josey, BASF Corp.; Andrew Kennedy, FoodLogiQ LLC; Hank Lambert, Underwriters Laboratories Inc.; Qingyue Ling, Oregon State Univ.; Brenda Lloyd, UFPC, LLC; Paul Lothian, Tyson Foods, Inc.; Steve Mavity, Bumble Bee Foods; Jennifer McEntire, IFT; Sherri McGarry, FDA, CORE, Office of Foods; Joan McGlockton, Natl. Restaurant Association; Benjamin Miller, Minnesota Dept. of Agriculture; Melanie Neumann, PricewaterhouseCoopers; Vikki Nicholson, US Dairy Export Council; Jan Morritz Olsson, TraceTracker AS; Gale Prince, Your Food Safety Coach; Kevin Richardson, Heartland Solutions Group Inc; Michael Roberson, Publix Super Markets Inc; Michele Samarya-Timm, Health Educator, Somerset County, NJ; Caroline Smith DeWaal, Center for Science in the Public Interest; Hilary Thesmar, Food Marketing Inst.; Ed Treacy, Produce Marketing Association; Bruce Welt, Univ. of Florida; Gay Whitney, GS1 US; Mary Wilson, GS1 US; and Amanda Zychowski, IFT.