Header graphic for print

Food Safety News

Breaking news for everyone's consumption

Armchair Epidemiology: How to Solve an Outbreak

Like most people, I wear a lot of different hats — mom, wife and what I am most known for — food safety advocate. I am also wearing another hat that I am hoping to put on a shelf by the end of the summer — graduate student. Before I was tragically thrust into the world of food safety, I was a master’s level statistician and had spent my career working in clinical research finding new treatments for schizophrenia, high cholesterol and other health issues. After I became involved in food safety, I realized that if I wanted to be viewed as more than a grieving mom, I needed to take my education up another notch. So, I went back to school to get my doctorate in Environmental Health — in Molecular Epidemiology to be exact.

“What’s that?” you may ask. Well, when it comes to food safety and foodborne illness, molecular epidemiology is critically important. It is the backbone of outbreak investigations and foodborne disease surveillance. Molecular epidemiology uses molecular techniques to study the impact of genetic and environmental risk factors on the causes, trends and prevention of disease. In less scientific terms, it means using DNA to find out what causes disease and what can prevent it. Basically, epidemiologists are disease detectives, and molecular epidemiologists use genetic information to do their detective work.

Over the past weeks, as we watched the E. coli outbreak in Europe unfold, there has been a lot of commentary on the efficiency and effectiveness of the European and German responses. I read criticisms of the amount of time it took to identify the outbreak, the premature announcement that cucumbers were the source, the lack of a definitive source, and so on. As I was reading, it was clear that many people do not understand the challenges that epidemiologists face in investigating foodborne illnesses. As an epidemiologist-in-training, I want to share with you a crash course in epidemiology. Welcome to Foodborne Illness Investigation 101.

Investigating foodborne illness is a really tough thing to do. It takes a lot of time, rarely follows a linear path and often is unsuccessful. It’s not surprising, since the deck is stacked against the investigator. When most people are sickened with foodborne illness, they immediately assume that the last thing they ate made them sick. If only it were that easy. Most likely it was something eaten several days before. For E. coli, it usually takes 3 to 10 days for symptoms to appear. For listeria monocytogenes, it can take up to 60 days — that’s right, two months.

Of course, most people don’t head to the doctor right away. Think about it. How sick would you have to be to go to the doctor and have a stool sample taken? Once the stool sample was taken, it would have to be sent to the lab and cultured which takes another couple of days — assuming the physician ordered the right test. Many foodborne illness tests are not automatically done — even though many physicians assume that they are. Once the culture is complete, the physician and/or laboratory are often required to report positive results to public health officials but, again, there are often delays.

(Maybe now you can see why the foodborne illness numbers are just estimates. Public health officials really only hear about the sickest people — the tip of the iceberg.) So, by the time you went to the doctor, had the lab test run, and public health officials were notified, the question is more likely what did you eat two to three weeks ago (and where). I’m sure you would be hard put to remember.

Enter the foodborne illness detective a.k.a. state or local health department epidemiologist. The bacteria have at least a two-week head start, and somehow the epidemiologist is expected to get ahead of it immediately. The investigation starts with an extensive interview with the patient about everything they ate, where they went, how many animals and sick people they had contact with, etc. The questionnaire is probably 10 pages long and takes a lot of time to complete. Meanwhile, a microbiologist will run additional tests so the genetic fingerprint of the bacteria can be uploaded into CDC’s PulseNET database to see if it matches other cases. Of course, entering the data into PulseNET is voluntary — completely voluntary. And since everyone is overloaded with work due to state budget cuts, it may take a while to get around to it even with the knowledge that timely entry into PulseNET is critical to identifying outbreaks. In fact, it’s how the 2006 spinach and 2009 peanut butter outbreaks were identified.

A couple of days later CDC announces a national outbreak related to the strain the epidemiologist uploaded. CDC hosts a conference call. All the people sickened have to be re-interviewed, because every state uses different questionnaires and the data can’t be combined to do the analysis that will help identify a source. (A while back, there was an effort to get all the states to use the same form, but only seven agreed to participate.) Re-interviewing patients will take time — especially if the state has a de-centralized surveillance system. Investigation activities must be coordinated across different divisions and branches of government so it is not uncommon for “turf” issues to arise.

Local health departments might be asked to re-interview the patients using the standardized form. They too are understaffed and overworked but may be reluctant to seek state or federal assistance with patient follow-up, citing confidentiality concerns. There might also be a sense that patients will be more receptive to additional inquiries from a local health department person with whom they have spoken with previously. These factors can contribute to delays in the investigation progress. Remember the 2009 Salmonella tomato/pepper outbreak that lingered on for nearly three months? One of the reasons it took so long to find a source was that local health departments refused to share information with their state health departments, and both refused to share information with CDC. It’s hard to do detective work if you don’t know where the victims are.

Now, if you are lucky, you are an epidemiologist in the state of Minnesota, where you’ll be able to take advantage of a centralized food safety system and Team Diarrhea (yes, it’s really called that). Team D — made up of graduate students from the University of Minnesota’s School of Public Health — swings into action at their command center and conducts standardized phone interviews. It’s super efficient and, once they get involved, Minnesota usually figures out the source. It is unbelievable that Team D is on the chopping block in Minnesota’s budget debate. Actually, every state needs a Team D.

Hopefully, by now, you are beginning to see why it can take so long to “solve” an outbreak and why so many end up being a “cold case” with only a list of potential suspects. As with regular detective work, centralized systems work in your favor and de-centralized ones increase the likelihood that you’ll end up with no clear-cut answer. The recent outbreak in Germany is a perfect example. An editorial in Nature published last week explains how Germany’s complicated, de-centralized surveillance system impacted that country’s ability to respond to the E. coli outbreak.

Sadly, the German system sounds a lot like the system we have in the United States. The Germans have two federal ministries, two federal technical institutes, and 16 state ministries involved in foodborne illness surveillance. In the United States, there are 15 federal agencies and thousands of state and local agencies involved. In both countries, the coordinating agencies — CDC and RKI — receive information indirectly and have no direct authority over the
sources of their information. At least the United States has PulseNET — for now. Already under-funded, PulseNET is in serious jeopardy as state governments furlough staff and/or choose not to replace staff who leave. Without timely data being uploaded into the database, PulseNET cannot identify outbreaks and prevent others from being sickened.

Early on, some lamented that if a definitive source was not found the German outbreak would be a lost opportunity to learn about the epidemiology of foodborne disease. I disagreed. One of the lessons learned is that sometimes centralization and standardization is a good thing — especially if you are a disease detective.


Barbara Kowalcyk is the CEO of the Center for Foodborne Illness Research and Prevention. She was voted the Huffington Post’s Ultimate Game Changer in the food category in 2010. This article was first published June 29, 2011 as “Welcome to Armchair Epidemiology: The How-To’s of Solving a Foodborne Illness Outbreak” on Huffpost Food on Huffington Post. Republished with permission.


© Food Safety News
  • Doc Mudd

    Sure, the outbreak epidemiologist always works at a disadvantage, for all of the reasons you have explained. That’s part of what makes it exciting work.
    But, environmental surveillance (the un-exciting routine grunt work) can effectively reduce outbreak incidence and speed outbreak investigation. Consider, in the German example, if bacterial cultures had been occasionally performed and reported on sprouts, cucumbers, tomatoes, all of that. The “new” e. coli bug would probably have been noticed lurking in the environment, likely would have been described somewhere already. Fecal contaminated sprouts would have been documented from time to time and would be suspected early on in the outbreak investigation (along with meat and raw milk). The “goose chase” over E. coli contaminated Spanish organic cucumbers would not have been so distracting – instead of being a surprise finding, those grubby organic cukes would have been common knowledge and either be cleaned up ahead of the outbreak or ruled out early.
    So, true, outbreak epidemiologists must enter the race behind the first wave of runners. The Germans, it seems, not only had to start from a dead standstill, but they appeared to be facing backward with their shorts down around their ankles when the starting gun was fired. Complacency.

  • Barbara,
    Thanks so much for giving us an “armchair view” of the challenging world of molecular epidemiology. It was really helpful. Best of luck in your “sleuthing” career. And many congrats on your award, “Ultimate Game Changer in the food category in 2010,” from Huffington Post.

  • doc raymond


  • Barbara,
    Thanks so much for giving us an “armchair view” of the challenging world of molecular epidemiology. It was really helpful. Best of luck in your “sleuthing” career. And many congrats on your award, “Ultimate Game Changer in the food category in 2010,” from Huffington Post.