Loyalty data solves an outbreak of foodborne illness.
By Tom Weir
Although experience tells us that there is often contaminated food on the market, we never know for sure until clusters of people start getting sick. When that happens, the faster the response the better. The longer the source goes undetected, the more people are likely to suffer illness or even death as a result.
A few months ago, the U.S. Centers for Disease Control and Prevention enlisted the help of supermarkets in pinpointing the origin of salmonella bacteria that sickened a couple hundred people spread throughout most of the nation. It marked the first time the CDC used frequent shopper card data to find the brands responsible for an outbreak of foodborne illness, and the success of the operation suggests that it shouldn’t be the last time.
Using interviews and questionaries, the CDC was able to narrow the list of suspect items until investigators were fairly certain the bacteria came from an Italian meat product. When victims couldn’t remember the brand names they purchased, the agency turned to frequent shopper databases.
With the victims’ permission, investigators asked the grocers about some of their purchases and got enough hits to trace the bacteria to salami from Daniele, Inc. of Pascoag, R.I. and then to the imported black and red pepper used on the meat. Recalls were issued.
That’s a success story, but it raises the question of whether success would have come sooner and some victims spared if the investigation had started with the frequent shopper data. The first of the 252 cases was reported last July 4, and though the outbreak was widespread by fall, the source wasn’t determined and the recalls issued until January. The exhaustive nature of the CDC’s investigation is detailed on its website. It seems, however, that in some cases computer science might produce faster results than medical science.
If the CDC, the Food and Drug Administration, the Agriculture Department or any other agency probing an outbreak of foodborne illness started building a database of victims’ frequent shopper information right at the outset, a very good list of suspect products might be developed quickly and point investigators in a productive direction much sooner.
This would not work in all cases. The recent salmonella recall involving hydrolyzed vegetable protein, an ingredient in thousands of products, is an example.
There are definite privacy concerns, but this doesn’t come remotely close to being a Big Brother issue. First, the authorities have to get the victims’ permission to use their data. Anyone can say no. But others will say yes and the process can continue. Second, only people who got sick would be asked to share their purchase history, so the vast majority of consumers would never be affected. It’s hard to imagine a bigger waste of investigators’ time than plowing through a retailer’s entire loyalty database.
While grocers rightly guard their customers’ frequent shopper data, cooperating with a food-safety investigation—when the customer has given permission—seems like a public-spirited thing to do. It’s very much like some retailers’ existing practice of matching product-recall announcements to purchase histories and notifying frequent shoppers who bought the recalled products.
Tom Weir can be reached at firstname.lastname@example.org.