Everything you always wanted to know about E. coli outbreak

People stand inside a closed Chipotle restaurant on Monday, Dec. 7, 2015, in the Cleveland Circle neighborhood of Boston.
Steven Senne | AP
People stand inside a closed Chipotle restaurant on Monday, Dec. 7, 2015, in the Cleveland Circle neighborhood of Boston.

Outbreaks of food borne illness like those connected to Chipotle restaurants across the nation are more common than you might think. There were more than 800 outbreaks reported to the Centers for Disease Control and Prevention in 2013, the most recent year that complete data are available.

Those outbreaks include 13,360 people who were taken ill across the nation, 1,062 who were hospitalized and 16 deaths.

Tabulating those outbreaks through the year is no easy task. Think about it: Thousands of people get sick every day and most are not connected to particular incidents of food borne illnesses. It's outbreaks such as those at Chipotle that get headlines, but it's only after the signal has been identified amid the noise of regular flu and sickness.

So how does the CDC find particular outbreaks of food borne diseases and determine which are connected to one another? It turns out it's a long process involving your doctor, state health officials and statistical tests that go back centuries.

How an illness becomes an outbreak

When a person starts feeling ill, he or she will go to a doctor to seek treatment. Doctors can usually identify patients who have food borne illnesses rather than just a run-of-the-mill cold or flu. If the doctor suspects a patient has a food borne illness, they'll take a stool sample and send it to a clinical laboratory for testing.

The lab identifies the particular pathogen that's making you sick. The lab notifies the doctor, who in turn tells you and offers whatever treatment is appropriate. The lab also tells a state health authority if the illness is food borne, so the authority can keep track of certain diseases.

Doctors and lab technicians are required to alert health authorities if a patient is diagnosed with any of a number of diseases, including many food borne illnesses. The reports made to state health departments help to collect data and determine disease trends.

The state health departments use techniques such as serotyping, which identifies more particular types of a given bacteria, like Salmonella. They also do DNA "fingerprinting," which isolates variable elements in a string to be matched to other DNA found in a cluster.

They input their findings into a system called PulseNet.

Enter the CDC

The federal agency coordinates PulseNet, the "national subtyping network for food borne disease surveillance." So it's basically a giant database with information on all these diseases being diagnosed and identified all over the country. But people get sick all the time — even with things like E. coli. In fact, there have been more than 5,400 cases of E. coli reported in the U.S. in 2015, 300 fewer than in 2014.

Many of those cases weren't connected to any specific outbreak. Statistical tests show when a particular disease is appearing more often than health officials would expect.

The software allows investigators at the CDC to identify clusters of illnesses from pathogens with the same DNA makeup when spread over a wide area at the same time. That's how authorities identified a patient in North Dakota, who had eaten at the same Chipotle in Kansas that had sickened another person in the most recent outbreak.

CDC investigators use PulseNet as well as surveillance reports and information received from local laboratory workers to identify more cases of the illness in an outbreak, to track when people get sick and when the number of new incidents declines. Investigators use a graphic form called an epi curve to visualize the spread of an outbreak.

Running the data to identify clusters of a pathogen cropping up in different places is the easy part. Investigators also have to identify the source of the illness, which is all information that has to be collected manually. To do so, the CDC interviews patients to find out what foods they'd eaten and where recently. The amount of time they go back is dependent on the particular disease's incubation period, or how long it can remain in the human system between ingestion and showing symptoms.

The incubation period for E. coli, for example, is usually three to four days.

Once common food sources and eating establishments are identified, investigators develop hypotheses for the source of the illness. They test those hypotheses with both analytical studies and food testing. If sick people report having eaten a particular food more often than well people, to could be associated with the outbreak. Statistical testing enabled the CDC to determine the strength of a correlation for a particular food or eatery.

There could be a lot of reasons that it would be difficult to identify a particular trigger, like in the Chipotle outbreaks. The public often focuses on meat as the culprit, as happened early in the Chipotle cases. But a look at recent E. coli outbreaks in the U.S. show that vegetables like clover sprouts and spinach are often the culprits. These products can pick up the pathogens when the plants are fertilized in the field and not adequately cleaned prior to being served to consumers.

Steve Ells, co-CEO of the company, told the Associated Press that he doesn't think the company will ever know exactly which ingredient sickened customers in the larger outbreak. Still, he believes it was fresh foods like tomatoes or cilantro that had the bacteria.

One problem for the CDC in identifying fresh produce as the source of an outbreak is that by the time investigators find the food, the infected produce could be spoiled and no longer available for testing. Between physicians, state health authorities and CDC lab testing, the whole process can take weeks.

That means it's likely that once you read about people taken ill in the news, they were infected a while ago. The illnesses that were reported the week of Dec. 21 started between Nov. 18 and 26, for example.

An outbreak is considered to be over when there are no more new cases of the illness than health officials would expect in a normal population.