Google creates AI to detect foodborne illnesses

A team of Google researchers and public health experts are working on an artificial intelligence model that detects the spread of foodborne illnesses on the city level.

Researchers from Google created the model — dubbed FINDER — with machine learning, a type of AI that learns over time, rather than having to be programmed like typical software. Using anonymous and aggregated web search data from Google users who save their location history on their smartphones, the model pinpoints restaurants that are potential sources of foodborne illnesses.

The model identifies people who completed a Google search for terms indicative of food poisoning, such as "stomach cramps," to help discern which restaurants these users have recently visited.

The researchers subsequently tested the model in two cities — Chicago and Las Vegas — with help from the Harvard T.H. Chan School of Public Health in Boston, the Veterans Affairs Boston Healthcare System, and public health agencies in Illinois and Nevada. The model provided health departments in both cities with a list of restaurants the model identified as potentially unsafe. The departments then dispatched health inspectors to the locations.

In study results published in npj Digital Medicine, the researchers said restaurants identified by FINDER were more than twice as likely to be deemed unsafe during a restaurant inspection than restaurants that were identified using traditional methods, which rely on consumer complaints.

The study authors suggested cities use FINDER as a supplement to their existing procedures, since the model can help public health agencies prioritize which restaurants to inspect prior to an outbreak.

"Foodborne illnesses are common, costly and land thousands of Americans in emergency rooms every year," Ashish Jha, MD, corresponding author on the paper and professor of global health at the Harvard Chan School, said in a news release. "This new technique … can help restaurants and local health departments find problems more quickly, before they become bigger public health problems."

More articles on clinical leadership & infection control:
42 hospitals make Leapfrog's 'straight-A' list for safety
Louisiana health system detects sepsis onset with Epic EHR tool
AI tools to detect infection fall short when analyzing data across health systems

© Copyright ASC COMMUNICATIONS 2018. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 


IC Database-3

Top 40 Articles from the Past 6 Months