Although I was very young at the time, I can still vaguely recall occasional pre-dawn trips to the San Francisco Produce Terminal Market with my grandfather and father to buy produce for their mom-and-pop grocery store and a few other neighboring merchants. Deals back then, on this small scale and larger, were made with handshakes, cash and a person’s word. Paper rarely entered into the transaction. Even when transactions were written down, it was often on little pocket notebooks in undecipherable shorthand codes only meaningful to the individual, just done as reminders. Years later, as I gained broader, first-hand exposure to the full spectrum of the produce industry, it became obvious that relatively loose record keeping and personal relationships governed many levels of production, sales and other business interactions. Despite the obvious problems that these arrangements have always created, there has existed a long history of shared pride among producers, suppliers and buyers in the “handshake and a promise” subculture.
Although much has changed due to evolving business needs and regulatory mandates, one still encounters these types of arrangements today. It is, in part, a reflection of this operational subculture that barriers have persisted to rapid and broad adoption of detailed documentation, record keeping, uniformity of data collection and other components necessary for effective internal traceability within the produce business. These same elements are essential for both accurate and rapid external trace-back and trace-forward systems, ideally conducted by the industry in conjunction with regulators, in the event of an outbreak investigation.
With attention focused on issues surrounding traceability vs. trace-back in response to the challenges of trying to contain the recent Salmonella Saintpaul outbreak, both from a concern for limiting further risk to consumers and the already damaged reputation of the produce industry as a whole, there has been an outpouring of articles and information detailing the wide variability in the status quo for produce traceability and the specific Produce Traceability Initiative (PTI, see “Continuous Improvement Trends in Produce Traceability” on page 42 of this issue).[2-4]
Equally, technology, data management and traceability solutions providers have been predictably aggressive about extolling the virtues of various technology platforms to support the execution of the PTI. As these resources represent a rich source of details on what the industry has been doing, this article will focus on some obstacles that are characteristic of the fresh produce supply chain, benefits that may be derived from a uniform standard and some common concerns about cost and the logistical burden of implementation.
Expectations of Trace-back
It is important to distinguish between a) a perception within the produce industry as to the general adequacy of traceability and trace-back/trace-forward capabilities and b) U.S. Food and Drug Administration (FDA) officials’ expectations regarding the information required to execute a timely trace-back of product to supply source(s). Neither are the current traceability compliance standards uniformly functional for all purposes in food safety risk reduction nor as woefully inadequate and mismanaged as purported by spokespersons on both sides. Beyond the media hyperbole, a thoughtful analysis of some of the key issues regarding the variable language of trade, so steeped in the culture and customs of the traditional produce sector and flavored by rich regional traits, reveals some significant challenges for the regulatory agencies. It is no wonder that problems have been encountered when trace-back investigations are, in fact or by practice, constrained to sort through sometimes complex and quirky commodity distribution channels without the benefit of industry insights.
Traceability in the produce industry, from elaborate handwritten ledgers to sophisticated electronic data-capture and data-sharing among trading partners, is certainly not new. The Perishable Agricultural Commodities Act (PACA) of 1930 required the documentation and retention of records that covered produce lot identity and all transfers and transactions between shipping point and destination receiver. However, even the most ardent produce industry supporters recognize that the system has always been uneven. Inconsistent information, lack of details or inadequate lot identifiers on bills of lading often form the basis for problems in dispute resolution between trading partners (e.g., shippers, retailers) and the transportation and logistics provider.
In addition, there are numerous examples of produce moving through unconventional channels to consumers. Various long-accepted practices that divert product, unacceptable for mainstream handling and long-haul shipping to distant markets, to be pulled from cull lines or harvested fields by small operators, represent clear red flags for potential problems in food safety and traceability. Add to this the practice of putting unwashed product from Packer A in used containers from Packer B and a cash and carry sale with no records. Some of these activities, often referred to as the underbelly of the produce industry, have never been condoned or accepted as a whole and are illegal in some states. While it seems unlikely that these types of backdoor deals would result in large, multistate outbreaks, one outlet for these products has been small fresh processors of specialty mixed-component items, such as salsas. It is possible to derive plausible and significant risks to consumers and a serious obstacle to trace-back for regulators from such sources. With the increased stringencies of voluntary and emerging mandatory, commodity-specific, food safety management and audit systems, packers and handlers are shutting down outlets for this subset of produce distribution.
In the views of many, justification for the PTI has been building for a long time. Key reasons for establishing a whole supply-chain tracking system include the following:
• Current track and trace capabilities, following the movement of specific shipments through distribution to the point of sale or consumption, are generally poor
• Multiple standards and product identifiers, even for the same lot as it moves from a shipping point to multiple receivers, can become significant barriers to restricting the scope of an FDA outbreak investigation
• Traceability gaps are created as unit loads are broken and reformed with commingled product
• Much of current traceability information gathering and sharing relies heavily on cross-referencing, which may greatly slow a trace-back effort
• Recent lessons have shown that problems of traceability in one sector create economic vulnerabilities for the entire supply chain and ripple effects in all secondary industries
Origins of Current Traceability Initiatives
Even if one were to start the clock at the relatively recent elevated recognition and concern for the role of fresh produce in foodborne illness due to pathogens, in the early to mid 1990s, many sectors of the produce industry had established complete and robust traceability initiatives well before the U.S. Bioterrorism Act of 2002 required tracing “one step forward and one step back” throughout the supply chain. Though most farms are exempt from the registration and mandatory record-keeping requirements, several commodity groups and individual businesses implemented rapid traceability and trace-back/trace-forward capabilities that included coding information back to a grower or even a ranch harvest-block. Examples include California cantaloupes and stone fruits. In 1997, a Food Safety Committee of the California Cantaloupe Advisory Board was added to the standing committee structure of the melon-handler USDA Marketing Order. In 2002, this board added traceability to include a carton-level trace-back code for every handler. However, it is recognized that the ease of traceability may be lost at retail level, as commingling often occurs at the point of sale.
The California Tree Fruit Agreement (CTFA), covering peaches, plums and nectarines, developed shipper-implemented trace-back capabilities in the mid-’90s to identify the source of fruit back to the orchard where it was grown, often back to the harvest crew, printed on each container. This information, in combination with the box label, can be used to quickly locate the fruit packer and, with separate codes on the box, identify each individual grower and the specific orchard where the fruit was grown. In mock trace-back exercises that the CTFA performs on fruit samples taken from retail stores throughout the U.S., it has taken less than 15 minutes to identify the packer, grower and orchard.
Naturally, in any traceability system, once the labeled carton or shipping unit is gone, the system may have great difficulty in rapidly isolating the source of an issue to a particular region, packer or farm. An additional concern, even with these and other exemplary systems, is that the language of lot identification has not been uniform, and lot integrity may not be constant during distribution. Carton re-use and re-sale is another recognized issue that may interfere with effective traceability.
A primary developmental need of positive lot identification and produce traceability arose from the need to develop internal tracking and documentation capabilities for business analysis of unit operations and adjusted payment for raw lots delivered to a packing facility, once quality, net yields and market price were calculated. The language and alphanumeric systems of these internal tracking mechanisms were often largely uninterruptible by external parties. This dilemma often involved the growers themselves, leading to payment disputes due to the nuances of commingling of lots to meet pack-out requirements or buyer specifications. More recently, effective internal lot definition and traceability have become increasingly critical as shippers and handlers react to the evolving landscape of food safety management.
Traceability is the key to executing a proactive retrieval of lots implicated in a suspected or confirmed pathogen-detection event. A quiet (notification not required) non-release or reverse-distribution of product that remains under the control of a produce shipper or handler is highly preferred, given the negative consequences of a recall notification of buyers or public health regulators. Though frequently referred to as a “market withdrawal,” this term is not strictly accurate for most produce shipping situations in the context of its FDA definition. Market withdrawal occurs when a product has a minor violation that would not be subject to FDA legal action, as would be the case in knowingly shipping adulterated food. This situation has become more common, though infrequent, as a consequence of the increased application of pathogen testing to fresh produce. This is especially problematic when lots are artificially created from a larger production or processing unit but only individual sub-lots are tested for specific buyers. Test results and their implications relative to a specific farm or production block may not be known prior to shipping of the linked lots. Arguably, product in a refrigerated transport vehicle may not be considered under the control of all parties. Newly enhanced traceability capabilities have undoubtedly played a key role in the rapid electronic notification and lot identification to prevent suspect product from reaching retail shelves and consumers.
One appealing aspect of improving internal traceability has been the recognition that data capture and analysis using manual bar code reading or automated identification systems, such as radio frequency identification (RFID) labels, is capable of significantly improving efficiencies and becoming a driver for improving unit operations. Tracking lot-associated data over time from site history, seed lot, all crop inputs, seasonal weather fluctuations, maturity and other factors may lead to the identification of patterns that affect quality, profitability and safety. Currently, data-rich internal traceability is being explored most intensely in the postharvest handling phases to capture the specific timing of key transfers and movements from harvest to cooling to staging for shipment. Linking this data to similar information from shipping docks to a food service or retail distribution center may be instrumental in localizing problem routes, carriers or receiver situations that result in avoidable losses. Not all parties are enthusiastic about the prospect of being able to isolate the cause of product compromises and cold-chain deficiencies.
A Call to Action
As the potential for fresh and fresh-processed produce to cause foodborne illness has become increasingly evident, one of the industry responses has been to seek a uniform and harmonized mechanism to effectively minimize the acute and chronic burden of recalls and to accelerate the effort to get to a common distribution, processing and supply point.10 Narrowing the focus of a voluntary product recall to a single supplier and avoiding commodity-wide consumer advisories from public health agencies is seen as a compelling argument for accepting the total supply-chain management milestones.[2,4] Very briefly, these milestones include the following:
1. Obtain a GS1-issued company prefix, which allows for the unique identification of products from that company. This unique prefix, held by the brand owner, is incorporated into all global trade identification number tags (GTINs) assigned to produce shipping case or carton units that are recognizable/readable throughout the supply chain. Repackers, depending on specifics of commingling and other parameters, need to obtain their own unique GS-1 prefix.
2. Brand owners assign specific 14-digit GTIN numbers to all of their product shipping-unit configurations based on the combination of the assigned prefix and a reference number. The expectation is that the reference number assignments follow a uniform strategy and logic across the industry for better consistency in communication.
3. Brand owners provide the specific GTINs and corresponding data to buyers for information recognition at the receiver location.
4. All of those packing product should provide bar code and human-readable information specifying the GTIN and lot number on every case of produce shipped by late 2010.
Extensive details of the history of development, terminology, the GS-1/GTIN system, timeline expectations and supply-chain member endorsements of the PTI are provided in the references at the end of this article. As with any program in which significant start-up and implementation costs are a certainty, as well as divergent expectations for benefits and success, not all supply-chain participants embrace the program.
Currently, one of the concerns regarding the effectiveness of the PTI in reducing produce- or commodity-wide economic hardship and loss of confidence among consumers is the lack of standards for data detail to be captured and shared. This is typically referred to as the granularity of information behind the lot code. For example, in the event of an outbreak investigation, the trace-back may implicate a single farm, based on a lot code. Assume that during the expansion of the investigation, linked data identifies a suspect soil amendment and its supply source. A trace-forward effort from that soil amendment supplier would, ideally, lead rapidly to other farms, potentially affected lots and back up the supply chain to all receivers. With rare exceptions, this may be a significant challenge today.
In contrast to the desire to execute such a rapid identification of common inputs or proximity, a frequently expressed concern is that such rapid traceability data-sharing technologies will result in the arbitrary and capricious removal of all products even remotely or tangentially connected with an implicated lot. Due to the often highly perishable nature of produce, it is likely that insufficient time would be available to determine whether the concern about co-contamination is substantiated.
Industry Needs for Traceability Systems
Traceability systems may be designed to include varying levels and approaches to improving breadth, depth and precision. A longstanding function of traceability has been to differentiate one supplier from another by providing their buyers and prospective buyers the assurances of a detailed and well-implemented program. Exactly how much breadth, the degree of information regarding the product’s production attributes, is needed in a particular timeframe is a matter of debate. In general, the consensus is that the first flood of information should be well controlled and limited by design to surgically isolate one potential source from all others. In the heat of a crisis, such as an outbreak investigation, it is easy to overload the investigators with too much information. Delving into additional layers, such as a common seed lot or specific harvest equipment, would only come later, once the legs of the trace-back have been confirmed to a reasonable degree.
Depth of traceability is related to the logic of requiring one step back, one step forward at each point in the supply chain; this point is illustrated by the ability of the FDA, for example, to laboriously walk back, one leg at a time, along a marketing channel for peppers that led to specific farms in Mexico that were reported to have supplied the original samples with a genetic match to the case isolates of S. Saintpaul.11 Additional genetically matched strains were obtained from environmental samples and pepper samples at a regionally related farm. In some long-existing traceability systems as described above, the carton-associated lot codes would have allowed that step to occur in one giant leap, right back to the farm block. The availability of individual consumer units in domestic refrigerators, though not generally the case, was instrumental in facilitating the rapid trace-back of implicated bagged spinach to a processor and a narrow group of farm sources in 2006.10 One objective of the PTI would be to bring that degree of depth identification transfer of implicated product to all sectors. A significant but manageable barrier to a one-step concept are those commodities, such as mature green tomatoes, that may be commingled at various points from initial packing to re-pack operations.
Lastly, the precision of the traceability system will be largely dependent on how the industry collectively—or its members individually—wishes to define a functional lot and the scope of the losses if a lot is withdrawn or recalled. The balance between segregation and practicality in handling products that are not precision-manufactured but subject to the influences of season, soil type, variety and other factors remains to be worked out in full.
Other concerns about the consequences of implementing the PTI revolve around the milestone expectation that inbound and outbound GTIN readings be performed at the case level. Taken at face value, concerns for serious compromises to cold-chain management and a potential for elevated food safety concerns were included in the initial reaction by those not endorsing the PTI. In reality, the time and labor constraints that make this an impractical expectation are well recognized, and pallet-level coding and other practical solutions are being developed.
As with other aspects of food safety management in the fresh produce sector, the scope of investments and ongoing costs to meet the expectations embodied in the PTI have raised further concerns among small and limited-resource growers and handlers. Simple systems for implementation among smaller suppliers that allow at least a reasonable level of voluntary participation and compliance in traceability were outlined previously.12 Claims that the PTI is readily adopted by small entities and uses current internal traceability systems would appear to overestimate the degree of adoption and resources available in the small volume but large number of small farm operations across the U.S. Reaching the first milestone—obtaining a GS-1 prefix—has an entry-level cost of over $800 and may exceed $50,000 depending on gross sales, employees and products (e.g., stock-keeping units or SKUs) marketed.13 An annual fee, currently around $200, is also required.
Detailing the direct and indirect costs of obtaining a GS-1 prefix, maintaining the GTIN, and preparing and practicing for a recall event are beyond the scope of this article but well described in several information resources compiled by the various PTI partners.3 Here again, the PTI committees are aware of the concerns and are working on solutions within the context of the need for uniform traceability at all levels of production and handling. At the same time, some of those expected to embrace the PTI are equally concerned that double standards for implementation are looming. It has long been claimed that many retailers have circumvented their own food safety-management requirements of suppliers when product availability is limited or for their “Locally Grown” sourcing programs. Many retailers have, in recent years, increased their efforts to extend much of the audit compliance standards to smaller suppliers, but the perception of “waivers” and an uneven playing field remains. For those interested in delving into the debates within the produce industry, a rich source of information and opinion on the concerns and controversy surrounding the PTI milestones, costs and “waivers” for small-scale growers may be found at the on-line publication, Perishable Pundit.
A Work in Progress
The responsibility for the safety of fresh produce does not reside with one sector of the supply chain, but extends from seed to shelf and beyond to the responsible and sensible behavior of consumers. Rapidly developing traceability initiatives that include putting source-tracking codes in the hands of consumers at point of sale or making them accessible by entering human-readable codes found on item labels into a website are remarkable developments. The degree of transparency and intimate data-sharing is already transforming major players and mid-size early adopters.
In various grower and shipper meetings, industry insiders lament the painful process of conversion to a more uniform and pervasive system of external traceability. It is painful in that it alters not just long-standing practices but truly a culture and way of life. However, the greatest predictions of gloom relate to the impact on the bottom line. An often-heard phrase, “I only see my costs, I don’t see my benefits,” is reflected in adversarial attitudes toward implementing microbial food safety documentation, record keeping and the audit process. Though lagging behind by years, an equally common observation by growers and packers has been that maturing into a comprehensive food safety program has resulted in substantial gains in quality and has reduced both losses and adjustments at destination. Pinpointing returns on investments in traceability and new systems to enhance regulatory trace-back is a challenge, but it is reasonable to anticipate tangible returns in efficiency from better internal tracking. The greatest identified benefits are not individual but will result from addressing and removing the current inefficiencies and barriers to outbreak investigations. Eliminating the chance of a nationwide consumer advisory against a commodity or region is strong incentive for the full supply chain to move to a new paradigm and culture.
Read the sidebar "Key Ingredients to Successful Quality-based Traceability"
Trevor Suslow, Ph.D., is an Extension Specialist in Postharvest Quality and Safety of Horticultural Crops at the University of California–Davis. Dr. Suslow obtained a Ph.D. in Plant Pathology in 1980 from the University of California–Berkeley. He has been employed at UC–Davis since 1995, originally in the Dept. of Vegetable Crops, now part of the consolidated Dept. of Plant Sciences. Dr. Suslow has a developed an internationally recognized research and extension outreach program in microbial food safety of both vegetables and fruit, incorporating a broad spectrum of applied projects addressing key issues that span preharvest and postharvest phases. His research program has included research and extension education dealing with hazard analysis and risk assessment of diverse horticultural crops and cropping systems, including transportation and distribution. Dr. Suslow’s applied research expertise involves detection and recovery of E. coli O157:H7, non-O157 Enterohemorrhagic E. coli, Salmonella and Shigella using both conventional and advanced rapid detection methods. His past and current research has been supported by USDA’s Cooperative State Research, Education, and Extension Service, various California commodity research boards, the Western Institute for Food Safety and Security, and the Center for Produce Safety. Dr. Suslow has published over 120 articles in research journals, as peer-reviewed extension publications, and in industry trade publications on diverse topics of relevance to agricultural sciences and practical agricultural production and pest management.
3. http://www.producetraceability.org/ pdfs/CPMA_PMA_Traceability_Guide.pdf
5. http://www.fda.gov/ola/2008/traceability 073008.html
8. http://www.ers.usda.gov/publications/ aer828/aer828g.pdf
9. Lynch, M. F., R. V. Tauxe, and C. W. Hedberg. 2009. The growing burden of foodborne outbreaks due to contaminated fresh produce: risks and opportunities. Epidemiol Infect 37: 307-315.
10. http://www.ers.usda.gov/AmberWaves/ June07/PDF/Spinach.pdf
11. http://www.fda.gov/oc/opacom/hottopics/ salmonellatraceback.pdf
12. Fonasah, E.G. 2006. Traceability: Formulation and Implementation of an Economic Efficient System in the Fruit and Vegetable Industry. Choices. Amer Agric Economics Assoc 21: 243-248
14. http://www.perishablepundit.com/ index.php?date=02/11/09&pundit=4
Key Ingredients to Successful Quality-based Traceability
Integrate Quality and Traceability Data
A traceability event does not stop at one up, one down. Further due diligence includes reviewing documentation associated with incoming inspections, certificates of analysis, supplier non-conformances, supplier audits and regulatory compliance data. An integrated quality-based traceability system marries lot- and batch-level traceability information with all required quality documentation.
Involve Suppliers with an On-line System
Quality and traceability automation should extend beyond the four walls of the packer/processor. Having tier 1 and 2 suppliers, along with a single system to capture quality and traceability data, provides immediate traceability up or down stream benefiting all companies within that supply chain. On-line access to current specifications, test procedures, non-conformances and audits as well as transaction-based electronic certificates of analysis (e-COA), provide the control needed for a complete system.
Automate COA Validation
Supplier’s performance against and compliance to a packer’s/
processor’s specifications is currently a manual and time-consuming practice with no value added in most companies. With integrated quality and traceability data in an on-line system, suppliers can provide shipment and e-COA data electronically where e-COA test data is immediately validated prior to shipment.
Eliminate Manual Entry of Data
Today most quality and traceability data exist electronically in various systems. An integrated quality and traceability system can collect data directly from suppliers and multiple internal systems electronically without the need for manual data entry or duplicate effort, leading to more complete, accurate and real-time data needed for effective traceability and root-cause analysis.
Involve Multiple Tiers of Suppliers
An on-line quality system extending beyond the plant’s walls allows multiple tiers of suppliers to be connected securely to extend traceability beyond one up, one down. This level of visibility for key ingredients is becoming more and more important in today’s global marketplace.