Tracking and predicting U.S. influenza activity with a real-time surveillance network

Area of Science:

• Epidemiology
• Public Health Surveillance
• Biostatistics

Background:

• Influenza causes millions of illnesses and thousands of deaths annually in the U.S.
• Current Centers for Disease Control and Prevention (CDC) surveillance data have a 1-2 week delay.
• Timely influenza data are crucial for effective public health response and modeling.

Purpose of the Study:

• To evaluate the validity of real-time data from a commercial influenza diagnostic network for estimating and predicting U.S. influenza activity.
• To assess the potential of the Influenza Test System (ITS) for improving influenza surveillance.

Main Methods:

• Utilized de-identified, real-time influenza test results from Quidel Corporation's national Influenza Test System (ITS) network.
• Developed linear logistic models to estimate CDC influenza metrics (positive test proportion, ILI-related physician visits).
• Integrated estimated ILI proportions into transmission models for improved national and regional influenza trend predictions.

Main Results:

• Linear logistic models accurately estimated CDC influenza metrics using ITS data on national and regional scales.
• ITS data integration improved the accuracy of influenza transmission models for predicting trends.
• The system demonstrated effectiveness for both the 2015-2016 and 2016-2017 influenza seasons.

Conclusions:

• The Influenza Test System (ITS) provides valid and timely data for U.S. influenza surveillance.
• ITS data can significantly enhance real-time "nowcasts" and short-term forecasts of influenza activity.
• Leveraging commercial diagnostic data improves public health preparedness and response capabilities for influenza.