Product Tracing in Food Systems: Legislation versus Reality
By Jennifer C. McEntire, Ph.D.
It is easy to justify procrastination when you are in Washington, D.C., since you know that Congress won’t pass legislation until after your article is nearly complete. By all accounts, food safety legislation should have passed by now, and yet other major societal issues keep taking a front seat.
Since the Institute of Food Technologists (IFT) published our reports on product tracing in November 2009, the most commonly asked question has been, “What’s next? What will the U.S. Food and Drug Administration (FDA) do?” At first, it was easy to refocus the question on the FDA (and the U.S. Department of Agriculture) since a public meeting was scheduled for December 2009 and a comment period was open until March 2010. The FDA was collecting input before deciding what to do. More recently, the question is refocused to Congress: until Congress decides what to do, the FDA may not act. However, as time has passed and S510 and HR 2749 have been dissected, there seems to be a general sense of where food safety legislation—at least in terms of product tracing requirements—might wind up.
Product tracing is the ability to trace potentially contaminated product, the consumption of which may cause an adverse health impact, through the supply chain. There is some debate whether product tracing is related to food safety. It is. When a potentially contaminated product is in the supply chain, and perhaps in the marketplace, there is the potential for consumers to become ill—a food safety issue. Rapid, accurate product tracing allows that product to be removed from distribution, protecting public health. Some in the food safety community argue that product tracing is strictly reactive, since it isn’t activated until after a contamination event has occurred. That’s not necessarily true. The ability to trace back and identify points of convergence can help identify the root cause of an issue and, hopefully, put controls in place to stop that type of issue from occurring again. A good traceback can help prevent a future event.
Outbreaks Drive Recognition of Issues
Too many outbreaks, each with individual nuances, have exemplified the inability to trace products throughout a supply chain. Many firms report that they are able to trace products. When pressed for details, it might be revealed that they may have very good internal controls or may keep great records at receiving and shipping. This doesn’t mean that their supply chain partners have these systems in place. The supply chain is only as strong as its weakest link.
Hopefully, product tracing systems and practices will continue to improve so that the Peanut Corporation of America incident will remain the prime example of product tracing gone wrong. If all firms in the longest supply chain used the maximum 24 hours to identify one step forward and one step back (as required by the Bioterrorism Act of 2002), all affected product should have been identified within days, perhaps weeks. Instead, for months after the problem was identified, product recalls continued to be announced, culminating in over 360 firms issuing recalls of nearly 4,000 products.
In contrast, the infamous tomato/pepper outbreak might be viewed as product tracing gone theoretically right (with the epidemiological investigation going wrong). Although the recordkeeping practices of the produce industry were blamed for the inability to identify the point of convergence of contaminated tomatoes, and certainly, paper records slow things down, the fact that common points along the supply chains were not found could have suggested that tomatoes were not the culprit.
COOL Is Not Product Tracing
Still, these high-profile cases gave American consumers a glimpse into the complexity of the food supply chain and left them wondering where their food actually comes from. When IFT conducted a study of product tracing in food systems, done under contract with the FDA in 2008–2009, we spoke to 58 food companies along several points in the supply chain for the produce, ingredient and processed food industries. Country of origin labeling (COOL), while intended as a marketing tool, was cited by many firms as providing consumers with information about product origins. The loopholes associated with COOL won’t be discussed here, except to say that COOL does not provide product tracing. As mentioned previously, product tracing is a requirement of the system, with visibility needed at each point in the supply chain. Simply knowing the country of origin is insufficient, and for some products, the complexity of the supply chain is such that the list of supplying countries is long. As illustrated in Figures 1 and 2, even something as American as a cheeseburger can have ingredients sourced from around the world.
Different parts of the food industry are subject to different laws that are often administered by different government agencies. However, the 2002 Bioterrorism Act provided a major recordkeeping requirement that most equate with product tracing. The Act requires what is commonly referred to as “one up/one back” recordkeeping. Specifically, recipients of a product are required to record from whom they received a product, and upon shipping, they must document to whom they sent product. The assumption that the requirements of the Act provide product tracing is wrong for at least three reasons. First, some are exempt. Exemptions at the ends of the supply chain, such as the farm and retail, are not ideal, but with the one up/one back system, the primary recipient of a product from a farm should know where the product came from, making a matching record at the farm level redundant. The greater issue with the exemptions is when they apply to those in the middle of the supply chain, such as brokers, who play a major role in the food production system. A lack of records results in a break in the chain. The second issue with the Act is the inadequacy of the data required to allow for product tracing. Lot numbers are critical pieces of information to distinguish product. Language in the Act reads that lot numbers shall be recorded “if available,” and this provides a substantial amount of wiggle room. In fact, in the IFT study, a major barrier to improving product tracing was that suppliers did not provide the necessary information. Lastly, had the Act been written to provide perfect product tracing, requiring all in the supply chain—at least the part in the U.S.—to keep the appropriate information, the FDA could not impose these requirements around the globe, nor could the FDA enforce them. Everyone involved in food understands the tangled web of food production and procurement, and the concept of worldwide food tracing, though necessary, seems daunting.
At several points over the past few years, a la carte product tracing bills have been proposed by various members of the U.S. Congress, and are included in Table 1. It now appears that the greatest change to product tracing requirements will result from the passage of some form of the Food Safety Modernization Act (S510, which has not yet passed the full Senate as of this writing, and HR 2749, which passed the full House in the Summer of 2009).
Tracing in Senate Bill S510
The product tracing requirements outlined in the current version of S510 give the Secretary of the Department of Health and Human Services (HHS), who oversees the FDA, a great deal of flexibility regarding the implementation of product tracing systems. It must be noted that the Senate bill has not yet passed the full Senate, with a manager’s amendment released in August 2010. Despite the overall general language on product tracing, there are two specific elements that may decrease the ability to trace food: 1) the legislation would not require product tracking to the case level and 2) additional recordkeeping requirements would be imposed for “high-risk foods.” Part of the definition of “high-risk foods” pertains to the processing steps that reduce likelihood of contamination. Clearly, lack of compliance with Good Manufacturing Practices is not accounted for, and many foods involved in recent outbreaks would not have been designated “high-risk.”
Tracing in House Bill 2749
In contrast to the Senate bill, the House bill, which has already made its way through the full House, provides much more detail on product tracing requirements. It would authorize the Secretary of HHS to “(i) identify technologies and methodologies for tracing the distribution history of a food that is, or may be, used by members of different sectors of the food industry, including technologies and methodologies to enable each person who produces, manufactures, processes, packs, transports or holds a food to (I) maintain the full pedigree of the origin and previous distribution history of the food; (II) link that history with the subsequent distribution of the food; (III) establish and maintain a system for tracing the food that is interoperable with the systems established and maintained by other such persons; and (IV) use a unique identifier for each facility owned or operated by such person for such purpose, as specified under section 1011 and (ii) to the extent practical, assess (I) the costs and benefits associated with the adoption and use of such technologies; (II) the feasibility of such technologies for different sectors of the food industry; and (III) whether such technologies are compatible with the requirements of this subsection.” As a result, “the Secretary shall issue regulations establishing a tracing system that enables the Secretary to identify each person who grows, produces, manufactures, processes, packs, transports, holds or sells such food in as short a timeframe as practicable but no longer than 2 business days….The Secretary may include in the regulations establishing a tracing system (i) the establishment and maintenance of lot numbers; (ii) a standardized format for pedigree information; and (iii) the use of a common nomenclature for food.” This bill also specifies that public meetings and pilot projects should be conducted.
Of particular concern is the authority of the HHS Secretary to perform studies on and standardize formats for pedigree. While the term pedigree is not defined in the bill, the concept seems extremely attractive: with a pedigree, you can see the entire history and path of a product at a glance. There is no need to do a laborious trace, since each handler of a product adds their name to the list as the product moves through the supply chain. In an industry where the barriers to entry are very low, resulting in numerous small operators (particularly in the world of blenders, repackers and the like), where price is a deciding factor in purchasing decisions, where rework is a fact of life and where substitutions are not uncommon, accurately handling information to provide a pedigree seems unlikely to be successful or economical in the near term. Currently, S510 does not explicitly require a full pedigree.
Forecasting the Future
No one knows what the final legislation on product tracing will look like. However, there has been speculation that some key elements of the House bill might win out.
It appears likely that the Secretary of HHS would be required to conduct some pilots, and the only argument heard from industry on this aspect is that more pilots should be conducted. It will be very difficult to capture the nuances and specific practices of each industry if only a few pilots are conducted. When IFT was contracted by the FDA to perform a mock traceback of the tomato supply chain, industry and trade associations argued that the simplicity of the tomato supply chain was not representative of the complexity of other produce supply chains. If the diagrams that the FDA shows illustrating the tomato supply chain are simple, I fear what complexity looks like! A pilot study of a complex processed food will surely face some challenges—primarily in the acquisition of so much data.
Industry, with support from trade associations, has begun making enhancements to the way products are traced in anticipation of pending legislation, and in response to the deficiencies illustrated by numerous outbreaks. Table 2 shows the vision for and objective of product tracing in various sectors of the food industry, the way products are currently identified and the way they might be identified moving forward. It appears that many industries are adopting the GS1 system of global standards. An overview of the types of standards that GS1 supports (e.g., standards for expressing data and transmitting data) can be found at www.gs1us.org for those in the U.S.
For tracebacks and trace-forwards to become more rapid, some parts of the food industry will need to change the way they record information, and in some cases, the way they view product. IFT found that some firms assigned identifiers to pallets, or in some cases even entire truckloads, making it very difficult to distinguish how the components of the pallet or truckload traveled through the supply chain. Additionally, the “first in, first out” inventory rotation system, relied upon to a great extent by some segments of the food industry, is wrought with exceptions, necessitating accurate recordkeeping and not simply deduction of when a product was “likely” in inventory.
IFT recommended to the FDA that product be distinguished at the level of the lot and be traced at the case level. Recognizing that the rate product moves throughout the supply chain–particularly in and out of distribution centers and warehouses–is incredibly high, we hope that technological innovations will enable foods to be traced at the case level in an economical and efficient way.
However, industry will not be alone in needing to make a change. While most recognize that providing records in an electronic format will increase the ability of regulatory agencies to analyze data orders of magnitude faster than today, IFT was surprised to learn that during the course of our study, the FDA’s Center for Food Safety and Applied Nutrition was resistant to collecting these types of records. It is discouraging to hear FDA representatives complain about receiving information on paper, when IFT also heard industry report that they hit the “print” button when federal or state authorities asked for information, because hard copies were desired. Clearly, some segments of the food industry have information in an electronic format, and mechanisms for industry to share and government to receive this information must be developed. The FDA has taken positive steps with the Reportable Food Registry and PREDICT tool for import screening, which are both electronic systems.
Some argue that the technology to manage food system data for the purposes of tracking products is not available. The technologies commercially available may not easily integrate into an existing system and may not readily accommodate paper batch logs, but the third-party systems explored by IFT have phenomenal capabilities. Additionally, many companies already use electronic systems that can be tweaked to enable the capture of data needed to trace products (such as Enterprise Resource Planning, Warehouse Management and accounting systems).
A main reason cited for the lack of implementation of third-party product tracing systems is that individual firms believe they are already in compliance with the Bioterrorism Act and do not view these systems as providing anything the company cannot already do. However, they understandably fail to see the “bigger picture” of product tracing as a system-wide issue that transcends any single firm. With the exception perhaps of vertically integrated supply chains, it is rare to find a company that can truthfully say they are able to trace their product forward and backward. Some systems that “connect the dots” will surely be superior to the pen-and-paper system used today.
Industry Implications: Cost
It is impossible to predict the cost to improve product tracing when the current practices are so incredibly varied and when the “standard” has not been set. Needless to say, most firms will likely need to make some change to their operations, and with change usually comes cost. It is important to note that more rapid and accurate product tracing can save both lives and industries. Saving lives is clearly a primary concern. Being able to identify the cause of illness and remove that suspect product from the marketplace has a public health impact that can be quantified. Some preliminary attempts at assessing how a change in the timing of a traceback could have mitigated the public health burden have been made, but more work should be done in this area.
Although societal benefits can be shown, from the perspective of an individual firm, the impact on the bottom line needs to be quantified. Perhaps one day product tracing will be so accurate that the tomato industry would have been readily exonerated in the 2008 Salmonella outbreak. The financial loss to the affected firms could have been mitigated. Another oft-reported tangible benefit from improvements in tracing comes from tighter controls on inventory and being able to link product quality with supplier information. Again, the extent of these benefits is firm-specific and depends on the difference between what is occurring today and what the firm might do in the future.
Product Tracing Challenges
Cost aside, there are other challenges that various members of the food supply chain will face and need to overcome to improve product tracing. Some are firm-specific and relate to changes in the flow of product that may result from the need to add identifiers to products. Many food processing environments are not amenable to things like labels: dust and water are two commonly encountered obstacles that complicate the ability to attach a label to a box. Within the food manufacturing environment, processes have been developed to be highly efficient, and the addition of “another step” specifically for traceability may be a challenge in some operations. Warehouse and distribution operations also deal with high throughput, and the procedures for “picking” products may require some adjustment in order to capture the information needed for product tracing.
Even if the recommendations made by IFT are fully implemented, there will still be some products that are more difficult to trace than others because of their production processes. It is impractical to expect a wheat farmer to mill his own flour or for tomatoes to ripen at exactly the same time. The extent to which facilitating product tracing may compromise efficiency needs to be discussed.
Bulk Product Tracing
The difficulties associated with tracing bulk product have been voiced in many forums. Tracing bulk ingredients such as wheat, sugar, oil, etc. is not impossible, but it’s certainly complicated. Some product-storage silos need to be emptied and cleaned regularly. These “break points” enable one to determine the finite (even if high) number of contributors to a given silo. For other products, complete emptying is done rarely, if ever. However, this does not excuse the firm handling the product from keeping appropriate records. In this instance, although the records may show that the pool of constituents is in the hundreds, thousands or more, engineers who understand the way the specific product flows through the silo can provide estimates of the amount of constituent that might reasonably be expected to remain in the system. Obviously, the more time has elapsed since the addition of a specific input, the less of that input would be expected to be in the silo. Dilution is not always a good solution, but having a sense of the relative amount of inputs in a silo or other bulk container could prove very useful when tracking products.
Tomatoes have served as the poster child for complex product tracing, somewhat unfairly. Conceptually, commingling and repacking are no different than producing multi-ingredient food products. It’s just that in the case of commingling, the ingredients happen to be the same. Therefore, tracking commingled product should not be viewed as an exceptional circumstance. Like other multi-ingredient products, the information needed to track incoming ingredients (e.g., multiple lots of tomatoes), such as lot numbers or other identifiers, needs to be recorded and linked to the outgoing product (e.g., a box of tomatoes consisting of multiple sources). The same can be said for rework, which, according to many firms IFT spoke to, is already tracked as an ingredient. Clearly, possible sources and routes of a problem increase in the case of rework, but this is not an excuse to not maintain adequate records.
Recommended Next Steps
Whether or not the House or Senate versions of a food safety bill pass, change is coming. For some, the change could be small. For others, the changes might put them out of business.
For now, IFT hopes firms will consider the recommendations it made to the FDA with respect to the concepts of Critical Tracking Events and Key Data Elements. We believe that identifying the right points at which to capture information (the Critical Tracking Events) and generating or capturing the data needed to trace products (the Key Data Elements—such as lot number, production date, shipper, etc.) will provide a solid foundation for whatever Congress may mandate and the FDA (or the Food Safety and Inspection Service) might implement. n
Jennifer Cleveland McEntire, Ph.D., is Senior Staff Scientist and Director of Science and Technology Projects at the Institute of Food Technologists (IFT) in Washington, D.C.
McEntire, J., et al. 2010. Product tracing in food systems: An IFT report submitted to the FDA, Volume 1: Technical aspects and recommendations. Comp Rev Food Sci Food Safety 9(1):92-158.
Mejia C., et al. 2010. Product tracing in food systems: An IFT report submitted to the FDA, Volume 2: Cost considerations and implications. Comp Rev Food Sci Food Safety 9(1):159-175.