News Release

Forecasting influenza outbreaks

Peer-Reviewed Publication

Proceedings of the National Academy of Sciences

Real-time Influenza Forecast for New York City on Jan 27, 2018.

image: This is a real-time influenza forecast for New York City on Jan 27, 2018. view more 

Credit: Image courtesy of Columbia Prediction of Infectious Disease, Columbia University, <a target="_blank"href="http://cpid.iri.columbia.edu/">cpid.iri.columbia.edu</a>.

Researchers report a method for forecasting the spread of influenza in the United States. Infectious disease spread is closely associated with human mobility, but incorporating detailed real-time data on human movement into disease forecasting methods remains challenging. Sen Pei and colleagues developed a model for forecasting influenza spread based on commuter data from the US census survey. For each state, the model divides the total population into subpopulations based on where people live and commute to work, and calculates changes in the numbers of infected and susceptible individuals in each subpopulation over time. The model also estimates the effect of noncommuting travel between states. Using the model, the authors performed state-level retrospective forecasts of the 2008-2009 influenza season through the 2012-2013 influenza season in 35 US states. The model predicted the week of disease onset in a state up to 6 weeks in advance of the observed onset. The model also predicted disease onset week, peak week, and peak intensity more accurately than forecasts in which each state is modeled in isolation. According to the authors, the model might help predict future outbreaks of influenza and other respiratory viruses in real time.

###

Article #17-08856: "Forecasting the spatial transmission of influenza in the United States," by Sen Pei, Sasikiran Kandula, Wan Yang, and Jeffrey Shaman.

MEDIA CONTACT: Jeffrey Shaman, Columbia University, New York, NY; tel: 212-305-3590; e-mail: <jls106@cumc.columbia.edu>; Sen Pei, Columbia University, New York, NY; tel: 347-502-9026; e-mail: <peisen1988@gmail.com>


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.