BACKGROUND: Florida conducts ILI surveillance using multiple surveillance systems including two syndromic surveillance (SyS) systems: Electronic Surveillance System for the Early Notification of Community-based Epidemics-Florida (ESSENCE-FL) and ILINet. ESSENCE-FL syndromes and subsyndromes have been validated for conducting surveillance for influenza and influenza-like illness (ILI) in Florida. Florida’s ILINet participation is primarily made up of outpatient providers. Both SyS systems measure ILI activity around the state, though the two use different case definitions. Of the all the participating facilities, twenty-nine participate in both SyS systems. This is a correlational study of ILI data between these two SyS systems to determine the appropriateness for combining data sources.
METHODS: The 29-facility aggregate weekly percent ILI visits for ILINet data, the ESSENCE-FL ILI syndrome, and ESSENCE-FL influenza subsyndrome were compared for week 40, 2012 to week 46, 2016. All retained ILINet facilities are part of a single chain of urgent care centers. ILINet data entry is performed by an individual who accesses the chain’s medical record system and applies the ILINet case definition (application has not been reviewed for correctness). ESSENCE-FL data were queried by facility for all visits meeting the ILI syndrome definition (any mention of flu/influenza, or fever and cough, or fever and sore throat) and the influenza sub-syndrome case definition (any mention of flu/influenza with some specified negations).
RESULTS: The correlation coefficient between the ILINet and ESSENCE-FL ILI syndrome data was 0.59. The correlation coefficient between the ILINet and the ESSENCE-FL influenza sub-syndrome weekly data was 0.54. Within the ESSENCE-FL SyS, ILI syndrome and influenza subsyndrome was 0.95. Comparing the aggregate percent of visits for ILI syndrome and influenza subsyndrome accessed via ESSENCE-FL were four and two times greater, respectively, than the percent of visits due to ILI identified in the ILINet data.
CONCLUSIONS: The ESSENCE-FL SyS data were substantively different from the ILINet SyS data. This suggests that the different SyS data are measuring ILI activity differently enough that they should not be combined, or substituted for one another. Given the important differences in the current ESSENCE-FL and ILINet SyS ILI case definitions, the combination of the two SyS should not move forward until an evaluation of whether a new ESSENCE-FL ILI query can be identified that would more closely match the ILINet case definition.