Beyond influenza-like-illness: data-driven syndromic surveillance for fine-scale monitoring of seasonal respiratory viruses.

Effective seasonal respiratory virus surveillance in the United States is critical to minimize healthcare burden and maintain economic productivity. Syndromic surveillance aims to monitor disease trends based on symptom occurrence but suffers from outbreak identification issues since case definitions are not disease-specific. We leverage a high-volume medical claims database, representing over 40% of the U.S. population annually from 2016-2023, to develop data-driven syndromic profiles for both influenza and RSV with a regression modeling approach. We then apply these syndromic profiles to estimate total symptomatic disease dynamics and burden by geography and demography. Symptom-inferred disease estimates show strong agreement with existing surveillance systems for both spatiotemporal trends and disease burden. Across all seasons examined, influenza prevalence was largely spatially homogeneous at the county-level, with consistency between seasons and age groups. Data-driven syndromic surveillance can provide fine-scale, representative disease tracking that strengthens forecasting, informs resource allocation, and enhances public health preparedness.