全球智库信息集成服务系统

Introduction: School absenteeism data might serve as an early indicator of disease outbreaks. However, before resources are committed to prospective surveillance, absenteeism data should be evaluated.

Objectives: This study evaluated the usefulness of school absenteeism data for early outbreak detection.

Methods: Data obtained from the New York City Department of Education on 1.2 million students (1,160 schools) for the 2001--02 academic year consisted of the number of students registered and absent by grade, school, and day. Reason for absence is not routinely collected. Citywide trends were examined separately for elementary and secondary students. Linear regression models predicted the expected percentage absent after controlling for day of week and pre- or post-holidays. Geographic clustering was assessed by the spatial scan statistic.

Results: Average daily absenteeism was higher among secondary students (13.7%) than elementary students (7.6%). No sustained increase in absenteeism was associated with the peak of the 2001--02 influenza A season (this period overlapped with winter break). A 2-week increase in absenteeism in March among elementary school children corresponded with peak influenza B season. Spatial analysis detected 790 clusters of absenteeism at p<0.01 (where only two clusters would have been expected by chance alone). Two of these clusters occurred during a previously reported gastrointestinal outbreak at one school.

Conclusions: A multiday, citywide increase in absenteeism among elementary students coincided with peak influenza B activity, but school absenteeism data were not useful for detecting the influenza A season. Although the system was able to detect one known localized gastrointestinal outbreak, this cluster did not stand out among other major clusters. Information on reason for absence and improved analytic methods might make absenteeism data more useful for early outbreak detection.