You are here

Modeling Sampling Strategies for Detecting Intentional Contamination of Imported Food Systems

Award No.: 
Principal Investigator: 
Randall Singer
PI Organization: 
University of Minnesota-Twin Cities
Significant quantities of fresh and frozen food are imported into the U.S. each year. In particular, fruits, vegetables and seafood are imported from countries around the world, and once these products arrive into the U.S., they are broadly distributed and either consumed directly or used as ingredients in further-processed food items. Without a solid understanding of the production chain of these foods, it is impossible to perform a vulnerability assessment of these food systems to intentional contamination events. Furthermore, it is impossible to develop sampling strategies that can quickly and accurately detect the contamination event in order to trace its origin and minimize its impact. Our research team believes that domestic food protection efforts need to be global in scope. Because the U.S. imports large quantities of fruits, vegetables and seafood from Latin America, we will focus this project on the importation of grapes, berries, frozen concentrated apple juice and salmon from Chile. Chile is currently the number one salmon supplier to the United States. In addition, more than $1.3 billion in fruits and fruit preparations were imported from Chile in 2007. The goals of this project are to model the distribution chains of these items, to evaluate the vulnerability of these systems to intentional contamination, and to evaluate sampling strategies throughout the food systems for the rapid detection and removal of contaminated products from the food supply chain. We will focus on intentional contamination with Bacillus anthracis spores, Clostridium botulinum toxin, Yersinia pestis, Shiga toxinproducing Escherichia coli, Francisella tularensis, and chemicals including ricin. To accomplish these goals, we will proceed in three phases. First, we will characterize each of these four food systems in terms of their production, processing, storage, and transportation within Chile and the U.S. Vulnerability profiles will be developed for each item based on accessibility to critical production stages, feasibility of large-scale contamination, chemical and physical properties of the agent / toxin and the food vehicle, and the dose of the agent or toxin that can be delivered to consumers through this food vehicle. In the second phase, we will develop detailed food system and distribution models for the two highest risk food systems, based on the crude vulnerability profiles developed in the first phase. Special emphasis will be placed on identifying contamination methods and access points in the food production and distribution chain. In the third phase, we will develop risk assessment models for each agent / toxin in each food system. Stochastic simulation models will be used to evaluate sampling strategies, incorporating information about diagnostic test accuracy, analytical sensitivity, seasonality, and the use of pooling methods.
Analytic Methods Utilized: