H1N1 Modeling Shows Short-Term School Closures May Worsen Flu Pandemics

PITTSBURGH, PA., January 19, 2010 – Closing schools for less than two weeks during a flu pandemic may increase infection rates and prolong an epidemic, according to results from computational modeling using TeraGrid resources at the Pittsburgh Supercomputing Center (PSC). The findings, reported in the Journal of Public Health Management and Practice, derive from a series of computer simulations based on several data sources, including U.S. census data, and indicate that schools may need to be closed for at least eight weeks in order to significantly decrease the spread of infection.

PSC scientist Shawn Brown collaborated on the study through the National Institutes of Health’s Models of Infectious Disease Agent Study (MIDAS) program, which supports research to simulate disease spread and evaluate intervention strategies. Brown is a co-principal investigator of the Pittsburgh MIDAS Center of Excellence, led by Donald Burke, M.D., of the University of Pittsburgh Graduate School of Public Health.

The modeling was carried out on PSC’s system called Pople, for Nobel Laureate John Pople of Carnegie Mellon, an SGI Altix system with 768 processing cores and 1.5 terabytes of shared memory. Pople is one of the leading shared-memory systems available to U. S. scientists and engineers through the TeraGrid, the National Science Foundation’s program of comprehensive cyberinfrastructure.

“Our models can be used as a virtual laboratory to undertake studies difficult to accomplish on real populations,” said Brown. “We can build a population and infect them with the flu, and then look at mitigation strategies, such as vaccinations or school closure, and see how that affects the disease spread.”

The value of school closures has been debated as a possible strategy to stem or slow the current H1N1 influenza pandemic. Indeed, hundreds of schools across the country have been closed at different periods during 2009 for fear the virus would spread more quickly if they stayed open.

“Although closing schools may seem like a reasonable way to slow the spread of flu, we found that it was not effective unless sustained for at least eight weeks after implementation,” said study lead author, Bruce Lee, MD, MBA, assistant professor of medicine, epidemiology and biomedical informatics, University of Pittsburgh. Closing schools quickly at the start of an outbreak was much less important than keeping them closed continually throughout the epidemic, he added.

According to study authors, short-duration school closures can increase transmission rates by returning susceptible students back to school in the middle of an epidemic when they are most vulnerable to infection.

The study also found that identifying sick students individually and keeping them from attending school had minimal impact on an epidemic. In addition, there were no significant differences between individual school closures and system-wide closures in mitigating an epidemic.

The study was based on an agent-based simulation model of Allegheny County, PA, which represented the county’s population, school systems, workplaces, households and communities. Simulations were based on the movement of residents each weekday from their households to designated workplaces or schools, and included 1.2 million people, 200,000 of whom were school-aged children. The study also included more than 500,000 households and nearly 300 schools.

Co-authors of the study along with Brown and Lee are Philip Cooley, MS, William Wheaton, MA, and Diane Wagener, PhD, RTI International; Ronald Voorhees, MD, MPH, Allegheny County Health Department; and Maggie Potter, JD, Samuel Stebbins, MD, MPH, John Grefenstette, PhD, Shanta Zimmer, MD, Richard Zimmerman, MD, MPH, Tina-Marie Assi, MPH, Rachel Bailey, MPH, and Donald S. Burke, MD, Graduate School of Public Health, University of Pittsburgh.

About TeraGrid:http://www.teragid.org