Issue 24 - Is School Absenteeism Data a Valuable Component in Early Event Detection?

____________________________________________________________________________________

Is School Absenteeism Data a Valuable Component in Early Event Detection?
____________________________________________________________________________________

Surveillance of pre-diagnostic data sources such as school absenteeism has been explored as a possible way to “enhance the timeliness of epidemic detection”. (1)

On the surface, there are a number of elements that make school absenteeism data attractive. An increase in absences might be one of the earliest indicators of potential disease outbreak; the data stream exists in all school districts and is often centralized; it can be timely if directed to the public health authority daily; and it can be a large surveillance population with wide geographic coverage.(2) Indeed, increased school absenteeism raised the awareness of Milwaukee Department of Health officials during their Cryptosporidium outbreak. (2)

However, school absenteeism data also has its drawbacks, and two recent reports describe some of these issues.

In the May/June 2004 issue of Homeland First Response magazine, Scott Vos, MPH, emergency coordinator of Johnson County, Kansas (part of Kansas City metropolitan area) Health Department, describes their different forms of syndromic surveillance, one of which is school absenteeism.

In the Kansas system, reports are faxed to the health department weekly, including the number of children absent and the reason for absence. According to Mr Vos, the system has been unreliable. Parents are sometimes reluctant to provide illness information. In addition, some schools allow parents to call into voice mail systems to report absence rather than talking to staff, which makes it more difficult to obtain the reason for absence. “Each school collects the information on illness in its own unique manner. Therefore, it is nearly impossible to make assumptions based on the information”. (5)

Johnson County is in the process of revising their system, so that each school will report “a total absenteeism of 15% or a daily increase in absenteeism of 5% of the student population.” When a report is received, additional follow up will be initiated if necessary. (5)

In a presentation at the October 2003 National Syndromic Surveillance Conference, Melanie Besculides describes the New York City Department of Health and Mental Hygiene’s (DHMC) experience with school absenteeism data. Attendance records for 1.2 million students from the entire NYC area during 2001-2002 were analyzed. Data was examined by age group, children (5-12) and teenagers (13-18). Using linear regression models, they predicted the expected number of absences, controlling for day of week, holidays and pre/post holidays. A Spatial Scan statistic with a 30-day baseline was used to identify geographic clustering of absenteeism. (2)(3)

The NYC project successfully identified a school having a known gastrointestinal outbreak; however this outbreak did not stand out among the 790 “significant clusters” identified during the school year. The data was not even helpful in identifying influenza season. The evaluation of data took substantial staff time, and of course data was not available during the summer or holidays. (2) (3)

For these reasons, NYC has elected not to pursue school absenteeism surveillance. “We’re not sure the school absenteeism will be worth adding to the system”, Dr Ed Carubis, CIO and associate commissioner of DHMC recently was quoted in Government Technology, “The problem is you know someone is absent but you don’t know why, so that really reduces the value of the information”. (4)

Whether the “why” could ever been accurately identified is also questionable. Children are absent from school for all sorts of reasons that on the surface might appear to be of interest to health officials. Is a parent going to tell a school district official that their child is home so they can baby sit other children? Or on vacation with the family? A report of illness is the likely excuse, but just as often as not the child isn’t really ill. In addition, what about situations where the child is truly ill, but the reason given is not an accurate diagnosis. For example a child might be reported ill with “the flu”, which often means a gastrointestinal illness to the public, or a child with “an asthma attack” might really be suffering from a respiratory virus.

In conclusion, whether due to issues of data collection or data reliability, school absenteeism data has significant hurdles to overcome. Whether it will turn out to be a valuable part of early event detection remains to be seen.

(1) http://www.syndromic.org/syndromicconference/2002/posterpdf/buckeridge_poster.pdf

(2) http://www.syndromic.org/pdf/poster-besculides01.pdf

(3) Besculides M, et al Evaluation of school absenteeism data for early outbreak detection, New York City. [Poster Abstract] Annual National Syndromic Surveillance Conference, New York, New York October 23-24 2003

(4) http://www.govtech.net/magazine/story.php?id=89565

(5)Voss, Scott. Picture of Health. The Emerging Science of Syndromic Surveillance. Homeland First Response. May/June 2004. Accessed at http://www.jems.com/homelandfirstresponse/pdf/5_04_feature.pdf

Back to the top