Yelp reviews could be a new way for health inspectors to catch onto food poisoning outbreaks early on, according to a new report by the Centers for Disease Control and Prevention,
Between July 2012 and March 2013, the New York City Department of Health and Mental Hygiene worked with Columbia University and Yelp to analyze about 294,000 reviews to identify potential references to foodborne illness.
Researchers used a software program to analyze a data feed provided by Yelp for keywords such as “sick,” “food poisoning” and others. A foodborne disease epidemiologist then examined the selected results to see whether they were useful for the purposes of the study.
Of the 893 reviews that were flagged by the software program used to analyze the reviews, 56 percent, or 499 reviews, described some kind of food-related illness, such as diarrhea or vomiting.
Of those 499 reviews, 468 reviews described some kind of illness within 4 weeks of the review, or without specific dates. Just about 3 percent of illness reported in this set of reviews had also been reported to the DOHMH, leading researchers to interview 27 reviewers by telephone. As a result, three unreported restaurant-related outbreaks linked to 16 illnesses were discovered.
The DOHMH led an investigation of these establishments, and found “multiple food-handling violations” in the three restaurants, including live roaches, failure to wash vegetables, improper cold food storage and other issues. In each of the three cases, the likeliest culprit (shrimp and lobster cannelloni, macaroni and cheese spring rolls and a house salad) were identified.
According to the NYC DOHMH, about 3,000 311 calls regarding food poisoning are received each year, with about 1 percent identified as outbreak-related. The researchers note in the study that social media surveillance could be a useful way to capture information from people who are more likely to post a restaurant review than to contact 311.
“Because foodborne cases have a common exposure, a restaurant patron review-based system can identify small, point-source outbreaks that are not easily found by systems reviewing large sources of data,” the report says. Other programs in the future, for example, might more specifically look out for the presence of vermin in restaurants.
However, the NYC DOHMH noted that “substantial resources” were required to run the system, including staff members who were asked to read reviews, send e-mails, interview reviewers and perform follow-up inspections.
The agency is planning to continue working on the Yelp project, with daily instead of weekly feeds of reviews, as well as additional information from other review sites.
Other organizations have also explored the potential of analyzing social media to gain insight into public health trends, including a project at the Boston Children’s Hospital analyzing Twitter to find insomnia-sufferers, or the Google Flu Trends project which analyzes search terms to predict flu activity in over 25 countries.
Date: May 25, 2014
Want to publish your own articles on DistilINFO Publications?
Send us an email, we will get in touch with you.