Context: Early in the 2009 pandemic influenza A (H1N1) experience, children aged 5 to 17 years were determined to be disproportionately affected compared with recent influenza seasons.
Objective: To characterize the pandemic among school-aged children, to enable timely influenza outbreak identification, and to determine which school-based influenza surveillance indicator correlated most closely with a laboratory-based standard influenza indicator (standard) and, therefore, might be most useful for future school-based influenza surveillance.
Design: During the 2009–2010 school year, we monitored students using 3 different surveillance indicators: (1) all-cause absenteeism, (2) influenza-like illness (ILI)–related absenteeism, (3) and ILI-related school health office visits. Thresholds were set for each indicator to identify individual school outbreaks. Each surveillance indicator was compared with the standard, confirmed influenza cases among hospitalized patients.
Setting: Tri-County (Denver metropolitan area), Colorado.
Participants: Prekindergarten through 12th-grade students in public schools.
Main Outcome Measures: Correlation coefficients comparing each influenza surveillance indicator with the standard and graphs comparing weekly rates for each influenza surveillance indicator or weekly outbreak counts with the standard.
Results: Correlation between the surveillance indicators and the standard varied greatly. All-cause absenteeism correlated most poorly with the standard (Pearson's r = 0.33) and ILI-related health office visits correlated moderately well (r = 0.63). Influenza-like illness–related absenteeism correlated best (r = 0.92) and could be improved (r = 0.97) by shifting ILI-absenteeism data later by 1 week. Graphs of weekly rates or weekly outbreak counts also illustrated that ILI-related absenteeism correlated best with the standard.
Conclusions: For influenza surveillance among school-aged children, when feasible, we recommend using ILI-related absenteeism, which correlated best and its rate peaked more than 1 week sooner than the standard. The other 2 surveillance indicators might be useful in certain situations, such as when resources are limited.