BACKGROUND: Flu Near You (FNY) is an online participatory syndromic surveillance system in which volunteers report whether they’ve experienced any influenza-like-illness (ILI) symptoms each week. The reports are analyzed and mapped by zip code to generate local and national views of FNY ILI that provide the public, public health officials and researchers with up to date information on ILI in the United States. FNY ILI has been compared to sentinel provider (SP) ILI surveillance data at the national level.
METHODS: California FNY ILI data was compared to SP ILI data and laboratory data collected by the California Department of Public Health. We calculated the Pearson correlation and root-mean-square deviation (RMSD) between FNY ILI and each data type for the 2012-2013 through 2014-2015 influenza seasons. To determine the minimum number of weekly participants needed to maintain accurate influenza activity estimates, mean ILI correlations between FNY and SP systems were computed from 1000 bootstrap samples of 10 to 300 weekly participants in increments of 10.
RESULTS: FNY received 217,041 reports from residents of 57 of California’s 58 counties during the 2012-2013 (65,097 reports) through 2014-2015 (83,441 reports) influenza seasons. FNY ILI and SP ILI had a correlation of 0.75, 0.83, 0.87 and a RMSD of 1.57, 0.78, and 0.68 during the 2012-2013, 2013-2014, and 2014-2015 seasons, respectively. FNY ILI correlated with laboratory influenza positivity by 0.53, 0.79 and 0.90 during the same three seasons, respectively. FNY ILI activity peaked five weeks before SP ILI, one week after SP ILI, and one week before SP ILI during the three influenza seasons, respectively. During the 2014-2015 influenza season, FNY ILI had a weak or negative correlation to other respiratory viruses (parainfluenza, adenovirus, human metapneumovirus, rhinovirus, enterovirus, and coronavirus) except respiratory syncytial virus (0.80). At least 110 weekly FNY participants were needed to reach a mean ILI correlation of 0.60. During the 2014-2015 season, San Francisco and Los Angeles were the only counties that had ≥110 weekly participants. Four of the five major geographic regions of California had ≥110 participants reporting each week.
CONCLUSIONS: The correlation between FNY ILI to SP ILI and laboratory influenza positivity increased from the 2012-2013 to the 2014-2015 season possibly due to increasing numbers of participants. FNY ILI may supplement SP ILI data for geographic regions not meeting CDC’s SP goals. More information is needed about California FNY participants and how to integrate FNY data into other influenza surveillance systems.