Quick Emergency Response Casts Light on Science Of Detection, Prevention and Control
On Monday evening, April 27, a Harvard dental student with a high fever and other flulike symptoms visited the urgent care clinic at University Health Services. Primary care physician Martha Katz, aware of Mexico’s spreading H1N1 outbreak, collected a deep nasal swab sample and started the student on oseltamivir (Tamiflu).
By evening, a coordinated emergency management team, including key people at the Dental, Medical, and Public Health schools, temporarily closed the Dental School. As an added precaution, they also suspended classes and clinical activities for medical and dental students. The 10 sick students were told to stay home from work and school for seven days after falling ill.
Putting Illness on the Map
The dental student flu cluster was first detected the old-fashioned way: an alert physician examining a patient in the context of an emerging flu strain identified a suspicious infection. Yet teams of HMS researchers are also developing and deploying surveillance systems to find new ways to detect and analyze the spread of infectious diseases of all types. For example, the “HealthMap” system is an automated webcrawler that cruises hourly through 20,000 local, national and international sites looking for keywords (see box for all links in this story). Every few hours, the real-time data are aggregated by disease and displayed on a map. There is now a special page for swine flu and a Twitter feed to provide rapid updates.
On April 1, this system first detected and disseminated a local news story in La Jornada of a “mysterious” influenza-like illness in the town of La Gloria, Mexico. There, 60 percent of the 3,000 inhabitants had been infected since early March, say HealthMap creators John Brownstein, HMS assistant professor of pediatrics at Children’s Hospital Boston, and Clark Freifeld, an MIT student. Their report was published online May 7 in The New England Journal of Medicine.
“To know what might turn into a major problem is what needs a lot of work,” said Brownstein, an epidemiologist.
Two other systems developed by HMS researchers track flu and other outbreaks locally. AEGIS, a real-time automated system, integrates data from emergency departments around Massachusetts and analyzes patterns of disease in space and time, said founder Kenneth Mandl, HMS associate professor of pediatrics at Children’s and a CDC biosurveillance adviser.
Another system monitors electronic medical records in physicians’ offices every day by age, sex and ZIP code for the state health department and the CDC, said creator Richard Platt, chair of the HMS Department of Ambulatory Care and Prevention. These offices are mostly in the Harvard Vanguard and Atrius practices around Boston and in the Western Berkshires. “We’re trying to understand which data sources are the best at identifying clusters soonest,” Platt said. The group has also used electronic health records to study surveillance methods in Northern California, Texas, Minneapolis and Denver.
“This is a bit of a miracle,” said Michael Farzan, HMS associate professor of microbiology and molecular genetics at the New England Primate Research Center. “With HIV, the time from isolating the virus to determining its sequence was years. With SARS, it was on the order of months. For swine flu, it’s been weeks.”
Farzan studies how influenza and related viruses enter cells and ways to block their entry. He is particularly interested in resistance to Tamiflu, which blocks transmission of the virus from cell to cell but does not cure influenza, which explains why the drug is effective only within the first 48 hours after symptoms occur.
H1N1 remains sensitive to Tamiflu, but the specter of widespread drug resistance worries many people. Based on a recent mathematical model published online April 30 in PLoS Medicine, HSPH epidemiologist Marc Lipsitch and his colleagues estimated that treating the earliest cases first with a secondary drug, the spray zanamivir (Relenza), could greatly delay the onset of resistance and ensure that the world’s stockpiles of Tamiflu would be more effective against a pandemic.
In Boston, health care protocols changed daily in response to the data generated around the world, said emergency physician Paul Biddinger, an HMS assistant professor of surgery at Massachusetts General Hospital and an HSPH assistant professor in Health Policy and Management. “We all knew a new flu strain was coming,” he said at a May 6 lunchtime “teach-in” for students at HSPH on the science and public health of type A influenza H1N1. “But we didn’t know how well it was transmitting, to which populations, and exactly how or if it could cause serious health consequences.”
It is too early to understand the true threat and reach of H1N1. The death rate could be as low as the 1918 pandemic—which was fatal only in about 2.5 to 5 percent of cases and killed tens of millions because it infected so many—or as high as the 60 percent fatality rate of the recent H5N1 avian flu outbreaks in Asia. But one thing is certain: the knowledge about the virus and effective responses is starting to evolve almost as quickly as the virus itself.
Web Resources on Flu Research at HMS and HSPH