You log into your Yelp account after dining at a new restaurant. But instead of scribbling down your thoughts about the flavours of the beef carpaccio or pasta alle vongole, you record when and how the dishes made you sick.
Your griping, it turns out, can play an important public health role: researchers in Boston are now using Yelp to monitor food poisoning outbreaks.
Looking at posts from about 5,824 food-service businesses across the US, researchers in the Boston Children's Hospital's Informatics Program extracted reviews that described food poisoning to find out which ingredients made people sick and whether they mapped on to data from the Centers for Disease Control and Prevention Foodborne Outbreak Online Database (FOOD) of health department reports.
Their findings? "Ten percent of all reports on Yelp are food-borne poisoning related, which is worrying but useful," said one of the study's authors, John Brownstein who also co-founded the epidemic tracking tool HealthMap. What's more, when the researchers compared which foods were implicated on Yelp to the CDC data — they found that they mostly matched.
The same foods made people sick in the Yelp and CDC surveillance
In a new study in the journal Preventive Medicine, the researchers reported that 32 percent of Yelp users complained about poultry, while 33 percent of reports in the CDC system had to do with poultry. Sixteen percent of Yelp users had trouble with seafood, while 12 percent of the CDC reports did. Eggs were implicated in 23 percent of the Yelp reviews; the same proportion in CDC's surveillance system. Overall, at both Yelp and CDC, beef, dairy, grains-beans, poultry, and vine-stalk were the most troublesome food categories.
At a time when about 48 million Americans have food poisoning every year, this data could be very helpful in mitigating some of that burden, says Brownstein. "We can use these data for disease prevention and investigation. If someone is saying they had a bad meal at a restaurant, that could have a direct public health response because a public health person could identify the violation, and the food that made someone sick, and investigate."
Brownstein's group is already developing a database that will aggregate Tweets with the Yelp information and feed it into an interface that public health agencies across the US can use. It'd support the social-media tracking efforts major US cities — such as New York and Chicago — have been experimenting with already. In New York recently, for example, the department of health managed to uncover three outbreaks using that Yelp data that would have otherwise been missed.
What if people use Yelp to sabotage each other?
In the hyper-competitive restaurant world, it seems possible that people could start using Yelp to log false reports about food poisoning and catch the attention of health department inspectors.
Brownstein said this is indeed possible. "An inspector from public health may show up to look into what's happening but they're not going to shut the restaurant down on a false report." Plus, he added, "We're going to have to be very careful about monitoring the system. Beyond false reporting, we want to make sure we're providing public health with something useful."