|
To combat this problem, Ben Y. Reis, Ph.D., first author of the paper who performed this research while he was in the Children's Hospital Informatics Program, and colleagues developed a system that uses multiple ''filters'' to analyze visit rates over a continuously shifting seven-day period. To address the fact that outbreaks can affect hospital visit rates in different ways, the researchers developed four different filters to look for different patterns: a fixed increase in visits over seven days; a week-long, steady increase; a week-long, exponential increase; and a simple one-day increase, typical of current systems.
After testing this multiple-filter approach on 10 years of hospital visit data spiked with simulated bioterror attacks, the researchers found that each of the weeklong filters worked better than the one-day approach. Researchers also found the weeklong filters detected disease outbreaks even better when working simultaneously. The researchers suggest that the detection capabilities of a multiple moving filter system could help enable a timely and effective response in the event of a deadly bioterrorist attack.
Kenneth D. Mandl, research director in the Division of Emergency Medicine, at Children's Hospital Boston and faculty member at the Children's Hospital Informatics Program, is senior author of this paper. Marcello Pagano, Ph.D., is second author. Reis is now at the Markle Foundation in New York, N.Y.
|
