The number of multi-state WGS clusters of foodborne pathogens (e.g., Salmonella enterica, E. coli O157:H7) identified by PulseNet has grown in recent years, and investigations of these clusters are becoming increasingly complex. Traditional analytic epidemiologic investigation methods based on subject interviews (e.g., case-control studies) are often not sufficient to identify and confirm the vehicle of outbreaks caused by commercially distributed food items.
Consequently, product tracing has emerged as an increasingly important part of the epidemiologic process for the identification of a foodborne outbreak vehicle. The overall goal of product tracing to aid an epidemiologic investigation is to determine whether a food item consumed by multiple case-patients in a cluster has a source or distribution point in common. Product distribution patterns are determined to add specificity to exposures and therefore assess the plausibility of one or more vehicles. This page briefly summarizes the process of product tracing conducted to aid an epidemiologic investigation.
The decision about when product tracing is warranted in an epidemiologic investigation, and for what food item(s), depends on the timely collection and evaluation of detailed epidemiologic data. Depending on the retail source and distribution network, epidemiologists, environmental health specialists, and investigators from regulatory agencies may all assist in the traceback effort. With the number of agencies potentially involved, it is essential to maintain clear lines of communication in order to coordinate efforts efficiently. A lead epidemiologist investigator should be designated who can represent all cooperating epidemiology agencies. This lead investigator is the logical person to request that a regulatory agency conduct a product trace. A meeting or conference call between the public health investigators and the involved regulatory agencies should be conducted to go through the information in detail, so that questions may be answered and a firm plan agreed upon.
Exposures to trace should be prioritized based on: the likelihood that the exposure is truly the exposure of interest for a case-patient; the availability of clear, documented details on the exposure; whether or not other case-patients share specific commonalities (e.g., ate at the same restaurant, shopped at the same grocery store, reported the same brand or variety of the suspect food item); and geographic and/or temporal dispersion of case exposures.
Conducting a product trace on any given exposure is conceptually straightforward: it is determining and documenting the producer, manufacturer, supplier, and distribution pathway(s) for the food item(s) of interest. A key goal in a traceback is to determine if there is a supplier or other point in the distribution chain in common. The investigator doing the trace should generally start by talking to the manager who orders food at the facility that provided the food to the case-patient. Make sure the manager is clear on the specific meal/purchase date(s) of interest (these dates are provided by the epidemiologic investigators based on the food consumption dates provided by case-patients).
A great deal of useful information often can be provided by a food manager during a 10 minute phone conversation; this information can be sufficient to move the investigation forward even if additional investigatory efforts are not possible or feasible. However, as the manager is providing information on the suppliers of the food item of interest, the investigator should also request written (or ideally electronic) documentation of that information—typically invoices from the distributor corresponding to the shipment(s) most likely associated with sale of the suspect food item and corresponding bills-of-lading. Because the purpose of these tracebacks is to provide detailed exposure information for the epidemiologic investigation, the speed of the traceback is critical. Thus, gathering information by fax and/or e-mail in addition to telephone is likely to be more efficient than dispatching inspectors to gather physical records from each establishment.
As the data are being gathered, a flow chart or diagram should be made illustrating the distribution pathways and timeline from the exposures that were traced. Agencies conducting the traces should regularly share updates on the traces with the lead epidemiologic investigators using this data presentation format.
Product tracing data are just one of the many sources of information that should be considered in an outbreak investigation and are best interpreted through a joint analysis by the epidemiologist investigators and the regulatory agencies that conducted the traces. These entities should be in constant communication throughout the process and agree on how the product tracing data relate to the rest of the epidemiologic, laboratory, and environmental assessment information.
To begin building a functional network and system for conducting product traces as part of epidemiologic investigations, epidemiology agencies should develop a contact list for all potential regulatory agencies that could be asked to conduct an epidemiologic trace and a protocol for engaging these collaborators in such an effort. Ideally, relationship building should be conducted ahead of time and entail development of joint protocols, MOU’s, etc. (However, the absence of such protocols or MOU’s should not preclude or impede the conduct of the trace back.)
Each regulatory agency should develop its own specific response protocol for conducting a product trace in this context. This protocol should include the names and training requirements of staff responsible for conducting that agency’s part of the trace, how they should be notified of a product tracing request, procedures to gather the necessary information from the involved food facility, and a mechanism for sharing the information with the requesting epidemiology agency.
At this point in the epidemiologic investigation, the most crucial aspect of an effective product trace is speed; therefore, the timeliness of investigations is more important than complete standardization. Process recommendations should not be used by any agency to develop rigid criteria or to require detailed justifications or priority schemes from another agency before it will participate in a product trace as part of the epidemiologic investigation. Rather, these recommendations should be used to increase the application, speed, and efficiency of product tracing in outbreak investigations.
Following are multiple examples of tracebacks conducted in support of epidemiologic investigations.
Example #1: Cyclosporiaisis in British Columbia (Emerging Infectious Diseases 2009;15:1286-1288)
Shah et al. (2009) provide an excellent example of tracebacks conducted during the epidemiologic investigation of a cyclosporiasis outbreak in British Columbia. The authors do an elegant and concise job of explaining the rationale, approach, execution, and interpretation of product traces conducted as part of an epidemiologic investigation, providing a “must read” for anyone interested in this topic. In brief, case interviews, population control comparisons, and product distribution information limited suspect foods to strawberries, cilantro, and basil. Interviews of grocery store owners, restaurant managers, and distributors were used to trace the produce items to suppliers (Figure 4). The trace implicated Mexican organic basil from a particular distributor as the outbreak vehicle. Once the vehicle was identified, the authorities conducted a full traceback of organic basil by using formal documentation.
Example #2: Multistate E. coli O157:H7 outbreak associated with hazelnuts (Journal of Food Protection 2012;75:320-327)
In this multi-state outbreak of E. coli O157:H7 infections, tracebacks were used by state regulatory agencies to complement traditional epidemiological cluster investigation methods to confirm hazelnuts as the outbreak vehicle. Bulk in-shell hazelnut or mixed nut (including hazelnut) consumption was documented during epidemiological interviews of the first seven cases in three different states; no other strong hypotheses emerged. In part because hazelnuts had never before been identified as a source of E. coli O157:H7, investigators agreed that demonstrating a common source of hazelnuts was paramount to the conclusion that they were indeed the vehicle. Based on case onset dates, purchase dates, and purchase locations, regulators in Minnesota, Michigan and Wisconsin traced product back through the supply chain. Six (86%) retail locations received the suspect hazelnut or mixed nut shipments from a Minnesota distributor, and one retailer (14%) received their products from a Wisconsin distributor. Both distributors received 100 percent of their bulk in-shell hazelnuts and mixed nuts from a distributor in California, and a recall of nuts from the California distributor was issued. The outbreak strain of E. coli O157:H7 was subsequently isolated from hazelnuts or mixed nuts in three states.
Example #3: Use of global trade item numbers in the investigation of a Salmonella Newport outbreak associated with blueberries in Minnesota, 2010 (Journal of Food Protection 2013;76:762-769)
In August 2010, the Minnesota Department of Agriculture and Minnesota Department of Health investigated an outbreak of six cases of Salmonella Newport infection occurring in northwestern Minnesota, which identified fresh blueberries as the cause. Initially, traditional traceback methods involving the review of invoices and bills of lading were used to attempt to identify the source of the outbreak. When these methods failed, novel traceback methods were used. Specifically, supplier-specific 12-digit Global Trade Item Numbers (GTINs) and shopper-card information were used to identify a single blueberry grower linked to cases, corroborating the results of a case-control study in which consuming fresh blueberries was statistically associated with illness (5 of 5 cases versus 8 of 19 controls, matched odds ratio [MOR] undefined, p = 0.02). Consuming fresh blueberries from retailer A was also statistically associated with illness (3 of 3 cases versus 3 of 18 controls, MOR undefined, P = 0.03). Based on initially incomplete evidence in this investigation, the invoices pointed to wholesaler A and grower A, based on first-in–first-out product rotation. However, when point-of-sale data were analyzed and linked to shopper-card information, a common GTIN was identified. This information led to an on-site record evaluation at retailer A, and the discovery of additional records at this location documented the supply chain from grower B to wholesaler C to retailer A, shifting the focus of the investigation from grower A to grower B. This investigation demonstrates the emerging concepts of Critical Tracking Events (CTEs) and Key Data Elements (KDE) related to food product tracing. The use of these shopper-cased data and the event data that were queried by investigators demonstrates the potential utility of consciously designed CTEs and KDEs at critical points in the supply chain to better facilitate product tracing.