What if one day, out of the blue, you find yourself sick—really sick—and no one knows what’s wrong. This is a podcast about a sleeper illness and what one team of data scientists led by Elaine Nsoesie is doing to reduce its reach.
Sam Williamson: “It felt as if I were on some kind of hallucinogenic drug. I felt really, really hot. Really cold again. The room started spinning. I got tunnel vision. I was about to black out.”
Ginette Methot: “I’m Ginette Methot-Seare, and you are listening to Data Crunch, a Vault Analytics production. Today we’re going to talk about something that could affect you or someone you love if it hasn’t already.”
Shawn Milne: “It still is a pretty vivid memory for me just because it was such a, such a terrible thing.”
Ginette: “This is Shawn Milne.”
Shawn: “Both of us just booked for the bathroom because we were both throwing up.”
Ginette: “He’s describing a sickness that both he and a friend suffered from.”
Shawn: “On the way home, we had to keep pulling the car over, and we were just both throwing up on the side of the road. It was absolutely terrible. We were just both up all night just throwing up. Just so beat.”
Ginette: “While Shawn’s experience lasted about 48 hours, Samuel Williamson, the person you heard speak at the beginning of our podcast, had one that lasted for about a month.”
Sam: “I did go to a doctor for it after a while. They convinced me to go to a doctor. He in fact told me that my stomach was just tired, which I thought was a very strange diagnosis. So he suggested that I don’t eat anything for a week. I think I lost about ten to twelve pounds in the first week, and so I went a week without eating anything, and came back a week later, and he asked me if the symptoms had gone away, and I told him ‘no, they were about the same,’ and he said, ‘okay, well you can’t eat anything else for another week.’ I went about three days and then pigged out.”
Ginette: “While everyone’s body reacts differently to this type of sickness, stomach pain was one symptom that everyone we interviewed described.”
Amy Smart: “I remember at one point, lying on my couch in excruciating pain, and thinking, ‘this is like having a baby, only with a baby, I know it’s going to end.’”
Ginette: “Amy had two little girls when she got sick, and she became so ill and weak that she couldn’t take care of them. Fortunately, her mom lived nearby and could take her girls during the day, and her husband was able to stay home from work to take care of her.”
Amy: “I couldn’t, I couldn’t eat. I wanted to because my body was so depleted, but I couldn’t drink. I couldn’t keep anything down. We went to the ER because I was so weak, and they put me on IVs and gave me morphine for the pain.”
Ginette: “But for Amy Smart, the person speaking here, things got a lot worse.”
Amy: “All that was coming out both ends was blood. And I remember feeling like, ‘this is what it feels like to die.'”
Ginette: “Amy described to me that it literally felt like life was leaving her body.”
Amy: “I didn’t know when it would end, when I would feel better again. If it would take days or weeks or ever. I remember thinking, ‘I’m so glad it’s me and not one of my little kids’ because I don’t know how they would have survived it.'”
Ginette: “Now put yourself in her shoes for a second: you’re sick and only getting worse. When you go to the doctor, the doctor isn’t sure what’s wrong.”
Amy: “They first thought it was stomach flu, then maybe Giardia, then maybe salmonella, and then they cultured it and found I had E. coli.”
Aside: “E. coli contamination. Possible E. coli contamination. E. coli contamination.”
Amy: “By then, once it was diagnosed as E. coli, it was a relief because then they knew how to treat it, and they put me on Cipro. By then the Center for Disease Control gets involved and is interviewing and trying to match the strain.”
Ginette: “Now as an interesting side note, this is where Amy’s case deviates from the typical victim of foodborne illness. When they tested her E. coli, they found no matching strains within a thousand miles. Ultimately she was labeled a CDC mystery, and to corroborate that, years later when her parents were visiting a family friend in DC, who was the Secretary of Health and Human Services, they end up touring the CDC war room . . .”
Amy: “where it’s all these maps tracking disease, basically. My mom brought up that ‘oh, our daughter had E. coli’ because that was one of the things that was on these maps.”
Ginette: “It turns out that the guy in there lists off her case number and the strain of E. coli that she suffered from and then says, ‘she’s a CDC mystery; we’re still trying to figure her out.’”
Ginette: “So while nobody knows how she really got it, Amy has her own theory:”
Amy: “It did match a strain where there was an outbreak in the Great Lakes area, in Minnesota, Wisconsin, North Dakota, and my husband had just come home from a business trip that was in Wisconsin, and somehow I got it from my husband I guess. Somehow he brought it home from Wisconsin. That’s the only explanation.”
Ginette: “While Amy’s case isn’t typical, her experience is similar to the experience of other E. coli victims, but E. coli is only one cause for food poisoning. There are many other sources. In fact, a lot of us get sick from foodborne illness whether we know it or not.”
Bill Marler: “You’re talking about 48 million Americans every year due to foodborne illness, 125,000 hospitalized, and 3,000 deaths. If you look at it from that point of view, it’s a significant problem.”
Ginette: “The CDC estimates that one in six Americans get sick from domestically acquired foodborne illness every year. Now to give you a clearer image of how many people that is, imagine the entire population of New York City, multiply five and a half times, and that’s how many people get sick annually from domestically acquired foodborne illness in the United States.”
Bill: “The vast majority of people who get sick with foodborne illness, norovirus is the main culprit, but it tends, fortunately, not to cause severe illness.”
Ginette: “This is Bill Marler. He’s the top foodborne-illness attorney in the nation.”
Bill: “Over the last twenty plus years, I’ve been involved in every major foodborne illness outbreak that’s occurred in the US. I’ve represented tens of thousands of people over that period of time. The bugs that you have to pay a lot of attention to, Listeria, E. coli O157, other Shiga toxin-producing E. coli—those are the bugs that cause very severe long-term complications and death.”
Ginette: “I asked him to describe the worst case he had ever seen, and it was hard for him to pinpoint a worst—he’s seen a lot of bad cases—but he described several cases that surprised me.”
Bill: “If you exclude death cases, and unfortunately I’ve been in rooms with families who’ve removed their children from life support—and certainly those are something you just never forget whether you are a lawyer or a human being, but I’ve represented people who have become really brain damage, who no longer can care for themselves, suffer stroke, people who become paraplegic, people who’ve had their large intestines removed, who’ve had kidney transplants. I was just e-mailing back and forth with a client of mine who developed a campylo from a campylobacter illness, develop Guillan-Barré syndrome, and she was paralyzed for four years, completely paralyzed.”
Ginette: “He said through a lot of hard work and the help of a computer-aided exoskeleton on her legs, she is now able to walk a little bit.
“Considering how many people food poisoning affects according to CDC estimates, it’s likely you’ve thought of someone you know personally who has been affected by it. In fact, as we were preparing for this show, I suddenly realized I knew someone who died from foodborne poisoning while he was traveling abroad for work. It was a shock to his friends and to his young family, and I saw the ripple affects. It could happen to any of us if we’re not careful and sometimes even if we are. So is there a way to prevent it or reduce the number of people it touches? The University of Washington’s Data Science for Social Good summer program is harnessing big data to work on this issue.”
Elaine Nsoesie: “The underlying idea is that if we can identify unsafe products faster, then the recall process can occur a lot sooner.”
Ginette: “This is Doctor Elaine Nsoesie. She’s a professor of global health at the Institute for Health Metrics and Evaluation at the University of Washington.”
Elaine: “We wanted to go back a little bit and look at data on unsafe food products and whether there’s a way for us to predict whether there’s going to be and FDA recall based on what people were talking about.”
Ginette: “Elaine and her team focused on Amazon. They scraped thousands of product reviews on the site to see if those reviews could predict FDA product recalls. But it was tricky. For one thing, the FDA tracks food through Universal Product Codes, or UPCs—those little bars you find on the back of your yogurt—while Amazon tracks products through product IDs, so they had to match the UPCs from the FDA recall reports to Amazon Product IDs. Another thing they had to account for was that sometimes people are sarcastic and give a product five stars only to rip it apart in the review.”
Ginette: “But once they successfully matched and collected their data, they”
Elaine: “realized that the proportion of recall products was actually a lot smaller compared to products that were not recalled, so then we had this unbalanced data set.”
Ginette: “So there’s more work to do to get this project to a point where it can actually predict FDA recalls, but”
Elaine: “if we can successfully develop a process where we can actually predict this—whether a product is going to be recalled or not—then the next step would be, is to work with local departments of health and build a tool that they can use to track these reviews, and hopefully it can help them in early detection of unsafe food products, so that they can do product recalls a lot faster.”
Ginette: “It turns out this project stems from another interesting foodborne illness project.”
Elaine: “I was a postdoc at the Computational Epidemiology Group at Boston Children’s Hospital, and there we started a project looking at how we can do surveillance of foodborne illness using data from Twitter, and we also did an initial evaluation using Yelp data.”
Ginette: “They ended up basing their project on a tool called Foodborne Chicago that the Chicago Department of Health developed. Foodborne Chicago uses Twitter to track people talking about food poisoning and asked them to submit more information.”
Elaine: “The idea was to work with local public health departments to develop a dashboard that they can use to track reports of foodborne illness on Twitter and then they can use that in decision-making regarding restaurant inspections and outbreak surveillance and things like that. So this project that we completed this summer is basically an extension of that.”
Ginette: “This quest to reduce foodborne illness continues with no definitive completion date, but Elaine and her team, along with others, will be working hard to make it happen. So, the next time you’re at the store, think twice about your mung-bean sprout purchase.”
Ginette: “If you want to see what foods Bill Marler doesn’t eat, go to vaultanalytics.com/datacrunch to see our show notes and links to interesting articles.
“We depend on your feedback to make a good podcast, so go to vaultanalytics.com/datacrunch, and submit your feedback to us, and if you have a data science project that you’d like us to feature, please contact us.
“Vaultanalytics.com has free resources to help you become a better data scientist. Check it out.”
A big thank you to Elaine Nsoesie, Bill Marler, Amy Smart, Sam Williamson, and Shawn Milne.
Links
Six foods Bill Marler never eats
Team page for Elaine Nsoesie’s project
Elaine Nsoesie’s postdoc project