A Cloud Based Epidemiology Network to Investigate Geographical Dynamics of Respiratory Disease

Background: Real-time data collection of respiratory disease is important for understanding the spatiotemporal dynamics of disease transmission in the US. Healthcare professionals use tools such as FluView to help identify local pathogen circulation; however, these tools are limited to syndromic surveillance, and track a limited set of pathogens. Understanding respiratory disease dynamics requires 1) a large, pathogen rich data set 2) geographically dispersed data sources, and 3) fine temporal resolution. Here we utilize FilmArray® Trend, a research epidemiology system containing exported data from FilmArray® Respiratory Panel (RP) tests, to investigate geographic patterns of 20 common pathogens.

Methods: Over 6,000,000 individual pathogen assays from 19 clinical sites were exported to the Trend database from 2013 to present. Trend data was smoothed and normalized to produce the time series of pathogen incidence. A cross-correlation analysis was performed to compare sites to one another and determine offset of pathogen incidence. The results were plotted on a map of the US with visual indicators of correlation strength and directional movement as defined by cross correlation lag values.

Results: The respiratory pathogens detected by the FilmArray RP test show a diverse set of spatial and temporal behaviors Most striking was the spread of the virus Coronavirus OC43, and Respiratory Syncytial Virus (RSV), with RSV traveling from east coast sites to west coast sites across the US over 20 days. In contrast Parainfluenza virus 3 (PIV3) shows a small cross correlation lag across all of the Trend sites during the regular summer season, indicating near simultaneous onset of detection nationwide. A localized cluster of PIV3 in the winter of 2016 was observed in the midwest and west, identifying the significance of localized regional trends.

Conclusion: FilmArray Trend shows great promise in deciphering spatiotemporal dynamics of these common respiratory pathogens. This system can identify localized outbreaks and directional movement of pathogens over time. Future work with finer geographic distribution of contributing sites will aide in making conclusions regarding spatial dynamics of all 20 RP pathogens. Other pathogen transmission models may also be explored using this data set.