Comparison of the Tarrant County School Health Surveillance System to the United States Influenza-like Illness Surveillance Network (ILINet) for monitoring influenza trends

BACKGROUND: Influenza-like illness (ILI) absenteeism data have emerged as a promising source for monitoring influenza. The Tarrant County School Health Surveillance System (TCSHSS) is a recently developed system that utilizes school ILI absenteeism data, but the performance of this system has not been systematically compared to a standard syndromic surveillance system. Therefore, we aimed to compare the performance of the TSCHSS to ILI Network (ILINet). METHODS: We utilized data from the TSCHSS and ILINet collected during the 2008 and 2009 influenza seasons. The TSCHSS collected ILI absenteeism data voluntarily reported by school nurses/administrators at 202 participating elementary, middle, and high schools in Tarrant County (Texas). ILINet collected data from 3 physicians in Tarrant County who voluntarily reported ILI prevalence among patients. Data from the TSCHSS and ILINet were analyzed using changepoint regression to detect time points in each system that corresponded to significant changes in ILI prevalence. RESULTS: Change points detected in the TCSHSS and ILINet data were similar for the 2008 influenza season (TCSHSS=2/10/2009; ILINet=Week 5 [02/01/2009-02/07/2009]). In 2009, the changepoint detected in the TCSHSS data preceded the change point detected in ILINet data (TCHSS=09/29/2009; ILINet=Week 41(10/11/2009-10/17/2009). CONCLUSION: Our results suggest that the TCHSS is comparable to or potentially better than ILINet for monitoring local influenza trends, but our results are limited to 2 years of data from each surveillance stream and thus conclusive inferences are premature. Nonetheless, our analysis illustrates the utility of changepoint regression, typically used in analyzing cancer surveillance data, for analyzing influenza surveillance data.