In 1993, 623 people in the western U.S. fell ill with a little-known bacteria called E. coli O157:H7. Ultimately, four children would die from their infections; many others suffered long-term medical complications. The bug was later traced to undercooked hamburger served at Jack in the Box restaurants. This outbreak thrust foodborne illness onto the national stage as a real and present threat, sparking a sea change in the way Americans and the government treat this issue. To commemorate the 20th anniversary of the 1993 Jack in the Box outbreak, Food Safety News has produced a series of retrospective stories chronicling the outbreak itself and how food safety in America has changed since that time.
In some ways, it was good that the 1993 Jack in the Box E. coli O157:H7 outbreak happened in Washington state. Tragic as it was, it could have been much, much worse.
When children started falling seriously ill, the cause was a mystery, but state health officials were uniquely positioned to crack the case.
To start, the state was pretty good at foodborne illness surveillance. In the 1990s, Washington reported more foodborne illness outbreaks than any other state except for New York–not because it had more outbreaks, but because health officials there were good at detecting them.
State epidemiologists had also been collaborating with the University of Washington and the local Children’s Hospital on an E. coli O157:H7 study. Long before the pathogen made national headlines, they knew it was turning up in the state and they knew it was a risk in undercooked hamburgers.
Armed with this knowledge, Washington became the first state to make E. coli O157:H7 a reportable disease, which meant doctors had to report their cases to the state health department. (By 1993, only a fifth of states had followed suit).
“We had a strong history of foodborne disease surveillance. That really helped us out,” says John Kobayashi, the head epidemiologist for the Washington State Department of Health at the time. “When you look at 50 to 60 clusters and outbreaks per year, you get really good at it.”
The experience came in handy, because Kobayashi’s team had to rely solely on old-fashioned epidemiology to figure out what was making people so sick.
This was before PulseNet, the national network of public health laboratories that use pulsed-field gel electrophoresis (PFGE) to connect illnesses that have the same pattern, or DNA “fingerprint.”
PFGE technology existed in 1993, but it was too time consuming and expensive to be useful during outbreaks. Having a network like PulseNet would also have been next to impossible without the Internet or access to email.
Even with relatively good surveillance in Washington, it took 39 days to determine that there was a serious outbreak afoot. And this discovery, too, happened due to some luck. Phil Tarr, a gastroenterologist at Children’s hospital, who happened to be the nation’s leading pediatric expert on E. coli O157:H7, noticed an uptick in children coming in with bloody diarrhea and immediately alerted Kobayashi at the state health department.
In mid-January 1993, the state’s team of epidemiologists (which was bolstered by Centers for Disease Control and Prevention fellows serving at the time), sprang into action, alerting emergency rooms to look for symptoms of infection and interviewing patients about where they had eaten recently. They pinpointed undercooked hamburgers served at Jack in the Box as the likely source of the outbreak in less than a week.
Their rapid sleuthing prevented 250, 000 potentially contaminated hamburgers from being consumed. It’s estimated that this prevented another 800 E. coli cases and an unknown number of deaths.
The power of PFGE
After the national attention on the outbreak subsided, it took several months to determine the subtype, or fingerprint, of the bacteria that had wreaked havoc in the Pacific Northwest.
To get the PFGE patterns from clinical and beef isolates, epidemiologists in Washington had to send them to the federal Centers for Disease Control and Prevention (CDC) in Atlanta. When they got the results back over the summer, all the samples matched.
“It became clear that if we’d had the method, we would have been able to detect these more quickly,” says Kobayashi. “We may not have even had a Jack in the Box outbreak, because there were a small number of cases in December that matched the outbreak strain.”
Dr. Bala Swaminathan at CDC, who had recently been tapped to lead the agency’s new food safety and diarrheal disease branch, had the same revelation.
“Why can’t we use this powerful technology [PFGE] to detect outbreaks when they’re happening?” asked Swaminathan. “Not too many people were on board with my idea that this could be done. People didn’t think this could be done in a decentralized fashion.”
The Internet, improvements in technology and software and federal funding eventually came together to make a network like PulseNet possible.
President Clinton’s Food Safety Initiative, launched in the fall in 1997, would provide much of the funding needed to build the capacity for the program, which initially focused on E. coli O157:H7.
Under the initiative, CDC launched the Epidemiology and Laboratory Capacity for Infectious Diseases (ELC) program, which doles out funding to improve foodborne illness outbreak detection at the state and local levels.
“ELC was a great source of funding for PulseNet,” says Swaminathan, who helped build the program from the ground up. “Money started to go out to the states for the equipment they needed.”
In those days, the equipment a state health lab would need to conduct PFGE and participate in PulseNet ran between $50,000 and $60,0000, he recalls.
“The Jack in the Box outbreak really played a major role in building PulseNet,” says Swaminathan. “No one was questioning the need for something like this. We didn’t have to sell the program. It sold itself.”
Building a national system
In 1996, when PulseNet first became operational, participating labs uploaded about 350 patterns. By 1999, the program was uploading more than 10,000 and it had been awarded the prestigious Innovations in American Government Award.
By 2001, the network covered all 50 states. Within another year, public health labs were uploading more than 25,000 patterns and the program was recognized as one of the top 15 programs to have ever been awarded the innovations award.
PulseNet was effective from the get-go, in part because there was a very strict, standardized protocol for submitting PFGE patterns from early on. At the time, variation in processes made it difficult to compare PFGE patterns between different labs or across states. With standardization, comparing and matching patterns became possible.
PulseNet is also more than a network; it’s a community. There’s a lot of online communication and a sense of purpose among the public health experts who contribute to the system.
“It’s a pretty tight knit group,” says Dave Boxrud, the Molecular Epidemiology Supervisor at Minnesota Department of Health, who has been at the health department since 1992. “It’s been an extremely effective system because of the level of collaboration.”
For the past several years, between 50,000 and 75,000 patterns have been uploaded to PulseNet each year. Using this detection system, health officials are monitoring between 5 and 15 new clusters each week. Today, labs across the country can perform PFGE in 24 hours.
At any given point, according to Peter Gerner-Smidt, who now heads PulseNet at CDC, epidemiologists are dealing with somewhere in the neighborhood of 28 outbreaks or disease clusters of various sizes.
“The beauty of PulseNet is that a large number of people can access a large amount of information in a way that’s useful,” says Hugh Maguire, Program Manager at the Colorado Department of Public Health and Environment, whose team worked on the deadly Listeria cantaloupe outbreak in 2011.
But building the infrastructure that makes PulseNet possible took many years, and now many in public health worry that budget constraints could weaken the system going forward.
Challenges
The biggest challenge facing PulseNet, says Maguire, is “first and foremost: funding.”
“I know it’s become almost eye roll-inducing to talk about funding, but it’s the reality,” he adds. “If you want to have well trained people doing this testing, you have to have the funding.”
On top of the general pressure to cut government while the U.S. economy struggles to recover, it’s hard for states to plan when the federal budget continues to be made up of a patchwork of continuing resolutions, instead of yearlong budgets. In Colorado, for example, the state lab relies heavily on CDC funding, with less than 10 percent of the its budget comes from general state funds.
The ELC program, which still supports PulseNet and other disease-tracking activities at state health agencies, peaked at about $78 million in grant funds in 2002. In 2010, the grant money was down to $52 million and the future funding levels are uncertain, as is CDC’s overall budget. The uncertainty comes as state and local governments are also making public health cuts.
“One thing people need to realize is that because [PulseNet] is a decentralized system, it needs the resources to type these isolates in a timely fashion,” says Swaminathan. “PulseNet is only as strong as its weakest link.”
