Cornell researchers have created a computer model that can help produce farms and food processing facilities control COVID-19 outbreaks, keep workers safe and the food chain secure.
With the Food Industry CoVid-19 Control Tool (FInd CoV Control), users input details like work environment, number, age, infection and vaccination history, physical proximity and living conditions of employees to create a simulation of how an outbreak might spread. Then they can compare the effectiveness and cost of intervention strategies, like testing, vaccination and physical distancing.
"If this next pandemic is a respiratory disease, we can relatively easily adapt the model to it," said Renata Ivanek, professor of population medicine and diagnostic sciences in the College of Veterinary Medicine, whose lab led the development of the tool. "If it's a slightly different disease, it will take a little more work, but having this whole framework already developed is a huge help."
They published their findings in "An Agent-based Model of COVID-19 in the Food Industry for Assessing Public Health and Economic Impacts of Infection Control Strategies," on April 24 in Scientific Reports.
In testing, the tool revealed patterns about the effectiveness of interventions that could be used to inform best practices and policy. For instance, they found that infrequent testing was expensive and failed to stem an outbreak.
"You either do it really well, like very detailed and very frequent, or don't at all," Ivanek said.
Frequent testing and isolation meant more work hours lost at the start of an outbreak, but it ended quicker. However, if a company doesn't test frequently enough, it still incurs the costs without stemming the outbreak.
They also found that vaccination is only effective when used preemptively, before an outbreak has begun. Physical distancing, face coverings and ventilation improvements were effective interventions, they found.
To design the model, the team focused on produce, meat and dairy processing facilities, and fruit and vegetable growers. They populated the tool with information from research studies, national databases on workforce characteristics, and data about vaccination and infection rates from the U.S. Centers for Disease Control and Prevention (CDC).
They inputted details of real-life COVID-19 outbreaks at five food operations to validate the model's performance and found that the tool's predictions matched the real outcomes with a high degree of probability.
The model can generate thousands of comparable simulations of an outbreak, creating "synthetic" data that can help food operations make evidence-backed decisions. "It's really dangerous to make decisions based on anecdotal reports from single operations," Ivanek said.
During the pandemic, the team made a simpler version of FInd CoV Control available for food operations to use. The version detailed in the paper is more sophisticated and reflects changes in the data over time.
"The literature kept growing as the outbreak was unfolding," Ivanek said. "That was actually one of the big hurdles, because we would have the model in some decent state, and then the new information would come out and we'd have to expand it."
The newest version requires some know-how to run it, but "industry could take it and make it commercial," Ivanek said. It could also be modified to support other essential industries.
The paper's lead author is former postdoctoral associate Christopher Henry '07. Co-authors include former postdoctoral associates Ece Bulut and Sarah Murphy, Ph.D. '20; Martin Wiedmann, Ph.D. '97, professor of food science in the College of Agriculture and Life Sciences (CALS); Samuel Alcaine, associate professor of food science (CALS); Aaron Adalja, assistant professor of food and beverage management in the Cornell Peter and Stephanie Nolan School of Hotel Administration; Claire Zoellner, Ph.D. '17, of iFoodDecisionSciences in Seattle; and Diane Wetherington of iDecisionSciences in Seattle.
The Industry Advisory Council, composed of executive-level leaders from major U.S. produce farms and food processing companies and an official from the CDC, provided guidance to the modeling work and feedback on the tool's structure.
This study was funded by the Agriculture and Food Research Initiative from the U.S. Department of Agriculture's National Institute of Food and Agriculture. Additional funding came from the Cornell Institute for Digital Agriculture.
