Syndromic surveillance systems have been developed for early detection of bioterrorist attacks, but few validation studies exist for these systems and their efficacy has been questioned. To assess the capabilities of a syndromic surveillance system based on community clinics in conjunction with the WSARE algorithm in identifying early signals of a localized unusual influenza outbreak. This retrospective study used data on a documented influenza B outbreak in an elementary school in central Israel. The WSARE algorithm for anomalous pattern detection was applied to individual records of daily patient visits to clinics of one of the four health management organizations in the country. Two successive significant anomalies were detected in the HMO's data set that could signal the influenza outbreak. If data were available for analysis in real time, the first anomaly could be detected on day 3 of the outbreak, 1 day after the school principal reported the outbreak to the public health authorities. Early detection is difficult in this type of fast-developing institutionalized outbreak. However, the information derived from WSARE could help define the outbreak in terms of time, place and the population at risk.