SAN DIEGO _ More than 1.1 million people had already passed through the gates of the San Diego County Fair this summer by the time an E. coli outbreak forced the closure of all animal exhibits and rides.
News that a 2-year-old boy had died after picking up the particularly nasty infection, which was also contracted by three other children with animal contact at the fair, stirred alarm within the community. Many had already roamed the midway, stuffed themselves full of fair food and passed through the venue's cavernous livestock barns en route to pig races, pony rides and the petting zoo.
Hundreds of emails and other documents obtained through Public Records Act requests show that, while the public health team was able to move quickly, more frequent county case reviews, a more modern medical records system and more prompt and accurate responses from families with infected children might have gotten the investigation started days earlier.
The decision to shut down the animal exhibits on June 29, records show, came after four days of a behind-the-scenes scramble by the county's public health department. With one death already on the books, they decided to notify the public even though testing had not yet confirmed that all four of those first four cases had E. coli infections.
It quickly became clear those initial instincts were accurate. In the following weeks, an additional seven people, plus two more whose infections weren't confirmed, came forward, including another young boy who nearly died after suffering severe complications that attacked his kidneys.
Records show that zeroing in on the fairgrounds was no simple task. Epidemiologists had to eliminate a broad range of possible locations, from restaurants to a busy daycare center, before they were able to zero in on the fairgrounds.
And there was plenty of other work that had to be done simultaneously. County records show that the department investigated 435 disease cases in June alone. Of those, there were 43 cases of shiga toxin-producing E. coli reported that month, forcing disease detectives to sift out the 11 eventually confirmed to be part of the outbreak.
Determining whether there are connections between cases requires interviews with each subject or their legal guardians. Depending on the type of pathogen involved, it's a process that often relies on frail human memory to recall the finest possible details of possible exposure routes from foreign travel and foods consumed to places visited and close contact with others.
The investigative process doesn't get started until the health department is notified, usually after a test result administered in a doctor's office or hospital comes back positive.
Subjects often aren't interviewed until weeks after they got infected because many infections have incubation periods measured in days or weeks and it usually takes time before individuals decide to seek medical attention and additional time for medical providers to make a diagnosis.
Often, those charged with reading these particularly fragile tea leaves learn to trust their instincts, and that was certainly the case with the fairgrounds outbreak.
Emails show that the county's epidemiology team first began to suspect that it might have an outbreak on its hands on June 24, the day that 2-year-old Jedidiah King Cabezuela was admitted to the intensive care unit at Rady Children's Hospital with severe kidney problems.
While discussing his condition, an epidemiologist noted that the boy had visited the fair before he got sick. And, she said, the county had been notified just before Cabezuela's death of another boy, this one 9 years old, who tested positive for the type of toxin produced by the E. coli strain causing so much difficulty for Cabezuela.
Though the 2-year-old and 9-year-old hadn't eaten the same foods at the fair, and the older boy's parents said he didn't visit animal areas, the fact that both visited the same location was enough for the epidemiologist to suggest that the department "should at least keep an eye on" the 9-year-old, even though he never got sick enough to need hospitalization.
By Tuesday, June 25, the public health department received the news that Cabezuela died overnight, and that information pushed the team to begin a relentless search for similar shiga toxin-producing E. coli infections.
They quickly found a report of a 13-year-old girl who had a positive toxin test after visiting the fair on June 8. Her parents had told interviewers that she had contact with animals and had eaten fair food afterward without first washing her hands.
Another girl, this one age 11, had also had a positive test but her parents had not returned repeated calls for an interview. It would not become clear that she, too, had visited the fair and had contact with animals until her parents were finally reached on June 28, the same day that the county announced it would close all public animal exhibits.
Then there was the 9-year-old boy mentioned in that prescient June 24 email. His parents initially said that he had no animal contact at the fair, but in subsequent interviews those parents remembered that, yes, their son did visit the livestock barn when the family visited on June 13.
With three, then four cases all reporting food consumption and animal contact at the fairgrounds, food inspectors descended on five different food vendors who sold items that the kids reported eating and found no traces of E. coli contamination.
Food poisoning ruled out, officials concluded that the E. coli exposure was most likely down to animals and, with the cooperation of the fair board, shut down all public access to animal exhibits and rides on June 29. Testing never did pinpoint the exact source of contamination, though the fair's petting zoo and pony rides were ruled out.
Once public health nurses do their phone interviews and build up as clear a picture as they can of the circumstances surrounding each individual case, epidemiologists can begin looking for patterns, keeping an eye out for clusters of patients in specific geographic areas or with other commonalities such as foreign travel or consumption of tainted food.
It is in the analysis phase that documents show a bit of a disconnect.
Public health nurses interviewed the 13-year-old girl who ended up being one of the fair outbreak cases on June 19, confirming on that day that she visited animal exhibits and ate fair food without washing her hands during a visit on June 8, according to a report sent to the California Department of Public Health.
However, the 13-year-old was clearly not on the radar of the county epidemiologist who reached out by email on June 24 to Dr. Eric McDonald, medical director of the county's epidemiology and immunization services branch. That communication referenced only Cabezuela, whose parents had been interviewed on June 23, and the 9-year-old boy whose parents eventually confirmed that he visited the fair's livestock barn.
McDonald said the 13-year-old's case wasn't mentioned because the epidemiology department hadn't yet held its weekly meeting to cross-reference cases based on factors such as location. Though her interview was completed on June 19, the next analysis meeting was not scheduled until the following Wednesday, June 26.
"We had not yet connected those dots because, at that point, only the investigator who did the interview knew those facts," McDonald said. "We did not yet have that pooled information."
He added that, though the information gleaned from each interview is entered into a computer database, the system does not have a function that automatically surfaces such geographic connections.
"I'm not aware of that kind of feature in any program," McDonald said.
Real-time analysis of incoming infectious disease cases is continuously performed, he said, but the main factor used to spot an outbreak is the incidence of higher case numbers than would be considered normal for a given time of year.
Jeff Bender, a professor in the division of environmental health science at the University of Minnesota with extensive experience in epidemiology and outbreak investigation, said that McDonald's assertion is correct. The systems commonly used by the nation's public health departments generally aren't very good at detecting common factors such as geography automatically even though such data-driven real-time analysis is becoming more common every day in modern medicine.
"Assessing exposure in real time, that's an area in public health where the technology just has not yet caught up," Bender said. "What we need are the same kinds of artificial intelligence tools to support us that are available in other areas of health care."
There were also some aspects that were simply outside of the local public health department's control.
Had the parents of the 11-year-old whose case was reported to the county on June 18 responded to requests for an interview, McDonald said, it would have been more likely that someone's ears would have perked up. It's a simple fact of doing this kind of work, he said, that some will not see much urgency in returning a public health call, especially in a case like this one where no hospitalization was required.
"As far as their parents are concerned, their kid is fine, and they're maybe wondering why the health department is calling them," McDonald said.
Misinformation, as occurred with the parents of the 9-year-old boy who turned out to have visited the livestock barn even though his parents initially said he hadn't, is also a common issue that can make it take longer for a clear pattern to develop. Those in public health know that human memory can be fallible. It's difficult to remember everything you ate and everywhere you went days or weeks after you did so, and that's why, McDonald said, public health nurses often re-interview their subjects.
"Sometimes it's best to be the third person to interview the subject, because then they have their story straight," McDonald said.
At the end of the day, the county's team of disease detectives was able to identify four connected cases and make the recommendation to curtail all public animal contact at the fair within four days of recognizing that two individual cases had both visited the fairgrounds.
While it may be the case that the investigation could have started sooner, McDonald noted that an earlier investigation would not have prevented infections. County records show that all 11 of the people who were confirmed to have fair-linked cases, and two others with suspected cases, visited the fair no later than June 22.
Bender, whose work on fair-related infectious disease outbreaks has helped create a manual of best practices for fairgrounds operators, said that pace looks fast enough.
"To me, that's a reasonable amount of time as you try to piece together the pieces of a complex puzzle like this," Bender said.
Both the U.S. Centers for Disease Control and Prevention and the California Department of Public Health, which oversee public health operations in the state and nationwide, declined to comment on the relative speed of San Diego's E. coli investigation.
