Use Internet search data to accurately track state level influenza epidemics

Area of Science:

• Epidemiology
• Public Health Surveillance
• Computational Biology

Background:

• Traditional epidemic surveillance methods often have a time lag, delaying crucial insights.
• Real-time data is essential for effective epidemic control and prevention strategies.
• Internet big data offers a valuable alternative for monitoring current epidemic trends.

Purpose of the Study:

• To develop and present ARGOX (Augmented Regression with GOogle data CROSS space), a novel methodology for accurate, real-time tracking of state-level influenza epidemics in the United States.
• To integrate diverse data sources, including internet search data and traditional surveillance data, for enhanced epidemic monitoring.
• To account for spatial correlations and evolving search patterns in influenza activity estimation.

Main Methods:

• ARGOX methodology combines national, regional, and state-level internet search data with Centers for Disease Control and Prevention (CDC) influenza surveillance data.
• The model incorporates spatial correlation structures of state-level influenza activities.
• It also addresses the dynamic evolution of public internet search patterns related to influenza.

Main Results:

• ARGOX demonstrated an average of 28% error reduction compared to the best alternative methods for real-time state-level influenza estimation.
• The methodology proved robust and reliable in tracking influenza activity from 2014 to 2020.
• Achieved accurate real-time tracking of state-level influenza epidemics.

Conclusions:

• ARGOX provides a significant advancement in real-time epidemic surveillance, outperforming existing methods.
• The methodology's accuracy and reliability make it a valuable tool for public health officials.
• ARGOX has the potential for broader application in tracking county- and city-level influenza activity and other infectious diseases.