Before the 2004 G8 Summit of industrial nations in Savannah, Ga., the number of patients showing symptoms consistent with viral meningitis increased in the emergency rooms of hospitals around the area.
Dr. Susan Lance, the state epidemiologist, said the one-day increase in patients with the symptoms turned out to be consistent with case reports that had already verified the movement of viral meningitis around Georgia.
The alert was available from the ERs, though, thanks to syndromic surveillance—a tool for discovering disease before establishing a diagnosis. Syndromic surveillance could be a valuable adjunct to traditional disease surveillance, especially for detecting bioterrorism or emerging diseases.
"With bioterrorism, if we wait until we get a laboratory diagnosis, we may be pretty far down the road as far as being able to prevent anything," Dr. Lance said.
Of course, an enterovirus such as the type that causes viral meningitis would be an ineffective weapon because only a small percentage of people who contract an enterovirus develop symptoms.
But epidemiologists want to detect outbreaks of any disease in people—and in other animals—as soon as possible. And syndromic surveillance might provide an early warning system.
Surveillance of syndromes
Syndromic surveillance is the tracking of syndromes, groups of symptoms or signs of disease, rather than the tracking of specific diseases. A system for syndromic surveillance might track influenzalike illness, among other syndromes. A spike in cases of influenzalike illness in people or other animals could indicate an outbreak of a respiratory illness such as influenza itself or another disease.
Funding for syndromic surveillance systems became available more widely after the terrorist attacks of Sept. 11 and the anthrax cases of the same year. Most programs function at a local level rather than nationally, so far, and more of them track symptoms in people than signs of disease in animals.
Yet, many emerging diseases are zoonotic, according to Dr. Tracee Treadwell, epidemiology chief with the Epidemiology, Surveillance, and Response Branch of the Bioterrorism Preparedness and Response Program through the Centers for Disease Control and Prevention.
"As a veterinarian, I was trained to keep a very open mind about things and realize how a disease presents in one species may be dramatically different than how it presents in another," Dr. Treadwell said.
The CDC and state governments are developing some of the programs for human syndromic surveillance. Syndromic surveillance systems under development for animals include the National Companion Animal Surveillance Program and the Rapid Syndrome Validation Project–Animal.
The National Companion Animal Surveillance Program at Purdue University's School of Veterinary Medicine draws on the database of Banfield, The Pet Hospital, which stores pets' medical records from hundreds of hospitals across the country.
Dr. Larry Glickman, an epidemiology professor at the veterinary school, said that CDC funding for syndromic surveillance was what allowed Purdue to develop the ability to draw on the Banfield database.
And the Department of Homeland Security has contacted Purdue to ask about signs of disease in animals, he said. The program also has picked up clusters of distemperlike disease in dogs.
Dr. Glickman said the program does not provide syndromic surveillance in the truest sense of the term, though. No one monitors or analyzes the data in real time, and no follow-up process exists for clusters of unusual signs of disease. The program has primarily been the basis for other research.
"We needed this in veterinary medicine to do lots of different types of studies," Dr. Glickman said.
But the Banfield database, with information about 12 million animals over the past four years, still provides an opportunity for syndromic surveillance that is rare in the fields of animal and human health. Health systems tend to be regional but not national, and almost every private practice collects data differently.
Dr. Glickman said incorporating information from private practices into a single system for syndromic surveillance would be like herding cats. He said for such a system to succeed, it would need to be electronic and automatic.
Dr. Bradley D. De Groot, an epidemiology professor at Kansas State University, said the Rapid Syndrome Validation Project–Animal is an electronic system for enhancing communication between veterinarians and officials about unusual signs of disease in livestock.
The idea for RSVP-A began with the human Rapid Syndrome Validation Project at Sandia National Laboratories in New Mexico. Collaborators on the RSVP-A pilot for cattle include Sandia, New Mexico State University, the New Mexico Department of Agriculture, and Kansas State University—with funding from the Kansas Department of Animal Health, U.S. Department of Homeland Security, and U.S. Department of Agriculture's Animal and Plant Health Inspection Service.
One of the stumbling blocks for RSVP-A has been finding a way to fit the reporting of syndromes into the routine activities of veterinarians, Dr. De Groot said, because the system is not automatic.
"So we're trying to develop within this surveillance system things that provide value to their routine activities," he said. "And that's been a challenging task."
The project started with 17 veterinarians in Kansas in 2003 and added 11 veterinarians in New Mexico in 2005. Most of them submit information via a Web site. Dr. De Groot said the veterinarians report signs of disease in cattle for six syndromes that could indicate bioterrorism, emerging disease, or foreign animal disease.
"We want to find something early in the course of its development," he said.
To promote participation in the project, RSVP-A collaborators have tried sharing more data from syndromic surveillance with veterinarians—who wanted to see what signs of disease other veterinarians were reporting. But adding case details to the Web site has not prompted veterinarians to visit several times per week, as the collaborators had hoped.
The collaborators also are trying to improve methods for submitting information. The propensity for veterinarians to record cases varies, Dr. De Groot said, and the six syndromes can overlap.
The vision is to refine RSVP-A for cattle and expand syndromic surveillance to other production animals. Currently, the collaborators are working on surveillance of swine fatalities.
Connections among species, surveillance systems
Syndromic surveillance remains much more widespread for people than for domestic animals.
BioSense is the CDC's new, national system for human syndromic surveillance. The BioSense strategy is to collect data about symptoms and indications of disease from hospitals, laboratories, pharmaceutical companies, and other sources.
Dr. Treadwell said the Bioterrorism Preparedness and Response Program also has conducted syndromic surveillance for some time at events, such as sports competitions. The CDC collaborates with the USDA and National Animal Health Laboratory Network to watch for signs of potentially zoonotic disease.
Dr. Treadwell said further evaluation will be necessary to determine the efficacy of systems for syndromic surveillance. With animals, syndromic surveillance is even more difficult because the signs of a disease can manifest differently among species. Nevertheless, she said, current systems for syndromic surveillance sometimes detect disease earlier than traditional surveillance.
"We've not had bioterrorism, thank goodness, but they've picked up outbreaks," she said.
Syndromic surveillance also has the potential for false alarms, but Dr. Treadwell said health departments conduct only initial investigations on the basis of a spike in signs or symptoms of disease. Before launching a full inquiry, investigators check other surveillance systems or simply call the local hospital.
She said the government also has been pushing to create connections among separate systems for syndromic surveillance.
At the Georgia Division of Public Health, Dr. Lance's epidemiology branch continues to have access to several systems of syndromic surveillance—though they're mostly local and specific to humans.
Dr. Lance can consult BioSense for information from area military clinics and the University of Pittsburgh for over-the-counter drug sales, and she receives reports from ambulance crews as well as from ERs. Soon, she also hopes to have the state's school districts reporting absenteeism to their county health departments.
All of these data can indicate disease, either through symptoms or signs of disease or secondary indicators. Of course, no system is perfect. For example, Dr. Lance's epidemiology branch accepts text files from the ERs. A program identifies symptoms and groups them under a syndrome, yet the program might not detect diarrhea as a symptom if someone in the ER misspells the word.
But the program has another component that collects diagnoses at the time of discharge from the ER or admission to the hospital. Dr. Lance said the diagnoses and syndromic surveillance, together, have identified apparent notifiable diseases that no one reported in a timely manner to the district health department.
When she was working at the G8 Summit two years ago, Dr. Lance had hoped syndromic surveillance would be a way to detect bioterrorism events early. Now she believes syndromic surveillance acts as a supplement to traditional disease surveillance, and it should be able to help delineate the magnitude and direction of an outbreak. Syndromic surveillance could identify the beginning of influenza season a little bit sooner, for example, and it might be a real aid during a pandemic of influenza.
But health officials still rely on astute clinicians and laboratorians to detect most outbreaks of disease in people—and animals.