An official website of the United States government.
Traceability has been an item of concern in the risk-management of industries such as pharmaceuticals, aircraft components and, more recently, food and food ingredients. In most cases the traceability is regarded as a distribution expense to be minimized, and as a result the amount of information carried is usually minimal unless legislation or threat of litigation causes it to be more complete. What are poorly understood in the milieu of factors affecting whether a contamination incident can be contained are the roles of information level and level of information transparency. <P>
The proposed study will seek to model a distribution system and determine what information is required to interrupt product delivery in the event of a contamination incident. This will take into account time lags that are usual in the case of manufacturers distributing contaminated product at the ingredient level as well as intentional contamination at various levels by outside agents. The utility of this study will be to provide the foundations of a predictive model for the use of intervention, assessment of detection utility and resulting action response time and the effects of base rate variability on low-probability contamination events. <P>The objectives of this study are therefore to develop an initial, simplified model of a multi-component manufactured food product system from commodity-level ingredient to packaged product. This will be use to simulate the effects of the addition of a contaminant at various points above the usual, local product-tampering level. Additionally, the probabilistic factors of test-inaccuracy and identification failure will be included in the simulations so that the effects of human errors and other factors may be estimated. The completion of this work should lead to the development of a higher degree of complexity and accuracy of the model as well as a higher precision and accuracy of results. While this type of simulation will not accurately recreate the shifting realities of the broad range food productions, it can provide direct, clear and quantitative data indicating where useful real-world trials may be conducted and shifts in policy and procedure would be most effective.
NON-TECHNICAL SUMMARY: Many of the efforts in large scale food contamination have been directed at detection of outbreaks of food poisoning in the population and then remediation after an outbreak occurs. The food industry, which is usually quite careful about quality and safety, already has coding and recall management practices in place. These have historically worked very well after problems are detected, but assume that the producer is acting in good faith, that the inspection, notification and recall systems operate as they are supposed to, and that the product itself is not counterfeit. The implementation of Hazard Analysis and Critical Control Point (HACCP) can also provide optimal points for assaying for contaminants or inspection for disrupted seals or counterfeit goods. Increased registration and security requirements for food processing plants have reduced access to the production facilities and mitigated the threat of bioterrorism. Balancing this is the broad range of ingredients from multiple sources that may be shipped without tamper indication or verification systems, as well as reliance on Certificates of Analysis for ingredient safety rather than verifiable in-house testing. Many of these factors leave the system open to attack. Previous Work and Present Outlook: Previous work in the Packaging Laboratory has focused on productivity improvements, particularly for production and packaging operations that lack the necessary engineering infrastructure and expertise to do extensive modeling studies. The results of this have been good with initial implementation by a local food manufacturer having shown good results and being able to highlight production bottlenecks both in physical capacity and scheduling/queuing methods. While this has been useful work, the tools developed for this and some of the results obtained have highlighted the possibility of using similar modeling techniques to conduct studies on the effects of a distributed contaminant through a larger food production system and distribution system. This has the potential for assessing the effects of all types of contamination episodes, as well as similarly highlighting critical intervention points that might be used in the event of a large-scale contamination episode. The far-reaching effects of this are not only in the obvious use in determining the best means for getting contaminated products out of circulation, but in estimating an optimal means for minimizing the, sometimes-substantial economic impact to the producers, processors and retailers. <P>
APPROACH: Procedure: This project will be developed in several distinct steps: <OL> <LI> The relevant components, interactions and dynamic characteristics of a simple food processing and distribution system will be estimated in order to give a central framework to the study. Data taken from the recent Peanut Corporation of America incidents with contaminated products being intentionally shipped as ingredients as well as other contamination incidents will be used as both general guidelines and benchmarking data for the resultant model system. <LI>Once a useful basic model has been constructed, Monte Carlo simulations will be run on the system with differing degrees of uncertainty in the distribution of products through the system as well as different points of origin for the contaminant. Results from the previous step will be used to develop an aggregation of results that should indicate specific areas of vulnerability both for the incursion of contaminants as well as vulnerability of end users. Similarly, points of intervention will be developed to provide the highest downstream 'shadow' of effect. The effects of non-traditional pathways (relabeling and re-importation) will be addressed to understand the further effects of unforeseen factors. These will be evaluated in the light of the Monte Carlo simulation to provide insight into both weaknesses in the model and weaknesses in the food security system. <LI> Costs and broader impacts such as market-segment sales depression will be estimated from the effects of real-world contamination episodes as well as the results of the modeling done in the study.