If a bioterror attack should occur, early detection would be crucial for saving lives. But how could local public health officials know whether one patient's symptoms were part of a larger trend?
"Historically, disease surveillance has worked by identifying and diagnosing individuals with conditions of interest and reporting them to public health departments, but diagnosis takes a long time.Since September 11, especially, people have been looking for ways to get an early idea of when outbreaks are occurring," said Michael Stoto of the RAND Corporation, a speaker at the annual meeting of the American Association for the Advancement of Science (AAAS).
A controversial solution would be to implement a large-scale surveillance system based on information from health care providers.
Surveillance would be effective and economical according to Ken Kleinman of Harvard Medical School. In an article published in the biweekly American Journal of Epidemiology, Kleinman and his colleagues describe a statistical model used in an experimental system, funded by the U.S. Centers for Disease Control, which collects data from several large physician practices, HMOs and a national nurse call center.
"Our results indicate that there is great value to using surveillance and statistical modeling to detect early signs of outbreaks such as anthrax," Kleinman said. "The data are routinely collected during the course of business and need only statistical analysis to provide advance warning of an attack."
The system protects patients' privacy by identifying them only in terms of the number of people sick in each zip code, a
Contact: Monica Amarelo
American Association for the Advancement of Science