Contaminated romaine lettuce is the culprit of an E. coli outbreak that has sickened 121 people in 25 states and killed at least one unidentified individual in California, according to the latest announcement from the Centers for Disease Control and Prevention on Wednesday. People first began experiencing symptoms on March 13, and the CDC released its initial announcement on the outbreak on April 10. This is the largest E. coli flare-up since 2006.
Yet, details regarding the source of the tainted lettuce and the cause of contamination are still fairly scant. Thus far, the CDC and Food and Drug Administration’s investigation has revealed that lettuce from Harrison Farms in Yuma, Arizona, led to a cluster of eight sick inmates at a prison in Nome, Alaska, in early April. But the CDC also warned that the Harrison Farms romaine was only responsible for those eight cases, and the cause of the other poisonings is still a mystery.
So, in the close to two months since people started getting sick, we’ve learned almost nothing about how the romaine was contaminated or which farms were responsible for a majority of the poisonings. Lettuce doesn’t even last for two months. Why has it taken the CDC and FDA this long to track down any source of the tainted lettuce?
The answer can be found in how extremely complex the system of U.S. agriculture has become. It’s true that all of the details of this outbreak are not yet public, so it’s difficult to know if there is a certain snag that is holding up this particular investigation. But this is nothing out of the ordinary, considering the circumstances. As Jaydee Hanson, a senior policy analyst at the Center for Food Safety, told Slate, “It is not unusual for it to take so long to track down the source of a microbial outbreak.”
There are two main systems to consider in tracing an outbreak to its source: the government’s processes for detection and investigation, and the produce supply chain itself. It’s difficult to make these systems work in concert because the former has largely unavoidable delays while the latter has an incredibly quick turnaround (see: lettuce doesn’t last two months).
Let’s take, for starters, the detection of an outbreak. When the earliest victims of this bout of E. coli first began feeling ill back in March, it’s unlikely that they went straight to the CDC. Instead, they probably visited doctors, who have regulatory obligations to notify the CDC if they suspect the presence of E. coli. The path from a patient getting sick to the CDC learning of the case is long and winding.
“With each of those steps there is a little time lag,” said Patrick Baur, a postdoctoral fellow at UC–Berkeley’s Department of Environmental Science, Policy, and Management, “Time lag from when you eat it to when you get sick. Time lag from when you get sick to whether you go to the doctor. Time lag from the doctor’s office to the laboratory. And then again there’s a time lag from the laboratory to when it goes to the county or state-level aggregation authority. And then there’s a time lag until it gets to the Centers for Disease Control.”
All in all, it takes two to three weeks from the time a person gets sick to the time that the CDC enters it into the data system. And even when the initial data from an outbreak becomes available, the agency still has to wait a bit longer to recognize any trends. The Epi Curve that the CDC has provided for this spate of E. coli speaks to the difficulties of spotting an outbreak in the first place:
When the CDC first became aware of that handful of cases from early March, it still had to wait another several weeks before becoming aware of the cluster of cases later in the month. It was that set that indicated there was an outbreak afoot. Once the CDC, along with the FDA, begins the investigation, it could be that a month has passed since people first came in contact with the spoiled produce.
“The challenge is that by the time that the data from any case ‘catches up’ to the CDC, the case is several weeks old and individuals have a hard time remembering what they ate yesterday, much less three weeks ago,” said Hanson.
By the time that investigators finally begin to try to trace the path and origin of a contamination, the produce supply chain has moved on. The U.S.’ produce operations are designed to move fruits and vegetables from farms to people throughout the country as quickly as possible due to the perishable nature of these goods. For example, Yuma region provides a vast majority of the lettuce that the U.S. consumes in the winter, and the supply chain has to process and transport all of these greens for the entire country within the span of three weeks.
Growing season for romaine is already over in the Yuma region, and agricultural operations there are now preparing their fields for different crops. CDC and FDA officials are left to investigate an environment that has radically changed since the romaine was present. “If they’re taking soil samples, that soil is going to have been completely moved around and regraded. There’s going to be new plants growing there. If it was animal or water source, those animals may have moved on and the water would have been flushed out probably down to the Gulf of Mexico,” said Baur. “In most cases, they’re not going to find the smoking gun.”
Produce supply chains are also designed to ensure that contaminations are discrete events, rather than ongoing ordeals. There are sanitation mechanisms along the way that regularly flush out E. coli and other toxins, and reduce the risk that a single batch of tainted lettuce will wreak havoc on an entire supply chain through cross-contamination. These safeguards are intended to minimize the spread of an outbreak—and most of the time, they work. But they also make it more difficult for investigators who are swabbing packing houses and processing plants for signs of the bacteria weeks after the lettuce was there.
The speed of the supply chain also necessitates that the produce change hands a large number of times, thus introducing complexity into the system. Tracing a product from a grower to a processor or distributor is fairly straightforward thanks to the use of barcodes, QR codes, and identification numbers (though Baur indicated that there is always a chance that information will be lost whenever produce is transferred from one party to another). But the sheer number of parties in this chain requires investigators to talk to more people to gather all the information.
Finally, once the produce reaches individual consumers and restaurants, traceability becomes particularly difficult. “The major processors know who they’re buying from, and whose oak leaf lettuce they’re mixing with their romaine,” said Hanson. “What makes it harder is when it gets into the restaurant business and they’re dumping this week’s bags with last week’s bags with the week before’s bags. Tracking that is much harder.” Once an investigator finally makes it to the restaurant—if the sick person even remembers that they ate lettuce there three weeks prior—the packaging containing the source info has likely been discarded or mixed up with other bags. It’s hard to know when they served what to who.
Improving the traceability of produce would essentially require us to slow down the supply chain, make it shorter and simpler, or invest more money into the system. Each of those options has associated trade-offs, as Bauer pointed out: It could increase the cost of produce, or limit the variety that is currently available throughout the year in every part of the country. It all depends on what we would be willing to sacrifice for better food safety.