CHALLENGES OF FOOD SAFETY
Recent high-profile contamination events have elevated the need to adopt a data-driven approach to assure the safety of the country’s food system. One of the most widely reported contamination events was the recent closure of more than 40 restaurants belonging to Chipotle, a US-based fast-food restaurant chain, in Washington and Oregon in October 2015 due to an Escherichia coli contamination. The Centers for Disease Control and Prevention (CDC) reported that 45 people were sickened by the E. coli O26 outbreak strain and, of those, 43 reported eating at Chipotle. Sixteen people were hospitalized although no deaths were reported. In February 2016, the CDC concluded their investigation, unable to find the source of the E. coli contaminations.
The CDC has also linked the Chipotle outbreak in the US Pacific Northwest with other reported E. coli cases in California, Ohio, New York, and Minnesota. And only a few months before, Chipotle had been linked to two other cases of foodborne contamination and resulting illness—a noro- virus outbreak in California in August and cases of Salmonella in Minnesota that have been traced to tomatoes from out-of-state farms. Then, in December 2015, 80 individuals were sickened after eating at a Chipotle restaurant in Massachusetts. And before the chain of outbreaks, Chipotle had taken the step of removing pork from its restaurant menus when one of the company’s suppliers failed to follow animal welfare standards.
Despite recent efforts and the passage of the Food Safety Modernization Act (FSMA) in 2012, foodborne infections continue to be an important public health problem in the United States. Federal data released by the Foodborne Diseases Active Surveillance Network (FoodNet) in 2015 showed little improvement in terms of foodborne illnesses when compared with data collected between 2006 and 2008, and between 2011 and 2013. The data indicated that illness due to Campylobacter—usually caused by consuming undercooked poultry—has risen by 13%. In addition, illnesses from two strains of Salmonella, javiana and infantis, typically found in undercooked eggs, milk, and meat, have more than doubled. And Listeria, the likely culprit in this year’s massive Blue Bell Creameries outbreak in the United States, was responsible for the most deaths of any strain last year. Of the 118 people who were diagnosed with listeriosis, 18 of them died.
The problems experienced by Chipotle and other food purveyors are emblematic of the challenges faced by today’s food industry. Our food supply chains are dynamic and complex—with an array of governmental agencies at different jurisdictional levels charged with regulating and supervising the safety of millions of food products produced by thousands of companies across the globe. Assuring safe food depends critically on our ability to collect, interpret, and disseminate electronic and other information across organizational and jurisdictional boundaries. The lack of visibility due to interoperability across the stakeholders of the food chain makes it difficult to quickly determine when a contamination has taken place. And once a contamination has been confirmed, the lack of visibility makes it difficult to trace contaminated food products back to the farm or country where they were produced—and also forward to locations where similar products may be waiting to be sold.
In response to these challenges, the food industry has looked to data science and “big data” for insight and a way forward. Significant efforts are being made to marshal big data tools to the cause of food safety. The definition of big data remains in flux depending on industry and application, but it typically involves the digital generation of data, often passively produced and automatically collected and stored, but also actively generated through events that serve as a trigger for marshaling response to an emerging contamination event. The premise of data science and big data for improved food safety is that, when fusing multiple types and formats of data including new and nontraditional sources, new analytics will make it possible to enhance our visibility of the food system to better monitor and respond in (near) real time to contamination threats as they occur.
