News Release 

CDC designates UMass Amherst a flu forecasting Center of Excellence

Biostatistician awarded grant to develop innovative, collaborative models

University of Massachusetts at Amherst

IMAGE

IMAGE: Flu forecasts can inform both individuals' behavior and target the response by public health experts. view more 

Credit: UMass Amherst

A University of Massachusetts Amherst biostatistician will receive up to $3 million in funding over the next five years from the U.S. Centers for Disease Control and Prevention (CDC) to operate a UMass-based CDC Influenza Forecasting Center of Excellence, one of two in the nation.

Nicholas Reich, associate professor in the School of Public Health and Health Sciences, whose flu forecasting collaborative has produced some of the world's most accurate models in recent years, leads a team that will work closely with the CDC, identifying new methods and data sources to sharpen the accuracy and improve communication of seasonal and pandemic flu forecasts.

"We know there are a lot of groups that have done trailblazing work in this field, so it's really a great honor to be selected," Reich says. A research group from Carnegie Mellon University, led by Roni Rosenfeld, was chosen as the other CDC Influenza Forecasting Center of Excellence. Rosenfeld's group has collaborated closely with the Reich Lab at UMass Amherst as part of the FluSight Network, a multidisciplinary consortium of flu forecasting teams.

Improving the precision of infectious disease forecasting is lifesaving work. Each year since 2010, in the U.S. alone, influenza has caused between 9 million and 43 million illnesses, hospitalized between 140,000 and 960,000 and claimed the lives of between 12,000 and 79,000 people, according to the CDC.

These new predictive tools could more effectively target the public health response to a potential flu outbreak, helping to determine the timing for flu vaccine campaigns, potential school closures and travel restrictions, as well as the allocation of medical supplies and antiviral medications. They could also help hospitals make the most efficient staffing decisions.

Reich is aiming to communicate more accessible and user-friendly information to the public, perhaps via a smart phone app. "We want to convey the forecasts in ways that people can understand, as it relates to their everyday lives," Reich says. "If we can communicate the data effectively, we might change behavior."

An app could help people gauge their risk, based on their location and their individual characteristics. "Maybe they'll tell grandma not to go to the shopping mall in the next two weeks, or maybe your kid with asthma won't visit the children's museum in Holyoke, where so many kids go to play in the winter," Reich says. "Those are the things I can foresee, where you're making everyday choices with this information."

The UMass Amherst Center of Excellence includes collaborators Evan Ray, assistant professor of mathematics and statistics at Mount Holyoke College, who completed postdoctoral research at the Reich Lab; Caitlin Rivers, senior associate at the Johns Hopkins Center for Health Security; Anna Thorner, an infectious disease specialist and the leader of the biosurveillance research team at UpToDate, an online clinical decision support resource; and two industry companies that run diagnostic tests for respiratory viruses - BioFire and Quidel. The companies serve as data providers for the UMass Center of Excellence, sharing anonymized test results from across the nation.

In recent years, flu forecasters have been spreading a wide net for their models, using Google search trends, HealthTweets and other nontraditional sources of information. "There is quite a bit of data floating around - Internet data and anonymized, cloud-based clinical test records - that aren't part of the public health data record," Reich says. "These data help us understand what's happening right now and make more accurate projections."

Reich's group uses ensemble methodology, incorporating 21 models in an open platform that shares data and coding to maximize forecasting capabilities. "Pooling the strength of many models together, collaboratively with multiple teams, results in a more consistent and more accurate forecast," Reich explains.

###

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.