Syndromic Surveillance in Bioterrorist Attacks

he article by Nordin et al. (1) in this issue of Emerging Infectious Diseases describes the use of syndromic surveillance to detect inhalational anthrax resulting from a hypothetical covert release of Bacillus anthracis spores at a major shopping mall. This study is an important evaluation of syndromic surveillance's utility in detecting an inhalational anthrax epidemic against a background of real patient presentations. Based on historical clinical data from a large health maintenance organization (HMO), the authors evaluated the sensitivity of a syndromic surveillance system to detect an incident by season of the year, day of the week when the release occurred, and attack rate in mall patrons. Although numbers of persons exposed and becoming ill, as modeled in the study, are not specified, the effect can be inferred from the specified methods. On the basis of information from the mall's Web site (2) and the methods stated in the article, the number of cases associated with a 15% attack rate in mall visitors (115,000 daily average) and workers (12,000) would be ≈19,000 (if no additional exposures occurred after day of release). Of these patients, 59% would be from the metropolitan area in which the mall was located, an additional 6% would reside within a 150-mile radius of the metropolitan area, and the remainder would be from more distant points, including international visitors. Syndromic surveillance, with the HMO patient database, would detect 50% of such incidents by day 5, with only 20% detected by day 4. Lesser attack rates would notably lower the probability of detection. Even more problematic, the syndromic surveillance systems, as modeled, would fail to detect the outbreak in 13% of releases in summer and 47% of releases in winter. Performance would improve markedly with higher attack rates. After detection of an aberrant signal, the occurrence of a syndrome must be investigated to determine the cause, and exposure history of patients must be determined to discover the source. These investigations could result in additional delay before a targeted response could be mounted to prevent more illnesses. Such delays are problematic because the effectiveness of postexposure prophylaxis for inhalational anthrax is related to speed of implementation (3). The authors point out that an astute clinician might diagnose inhalational anthrax in a patient before syn-dromic surveillance detected that an outbreak of some type was occurring. If, as the 15% attack rate scenario suggests, >100 patients had onset of illness on …