$2.16 million NIH grant will support the renowned Reich Lab
University of Massachusetts Amherst biostatistician Nicholas Reich, who rose to national prominence for his COVID-19 pandemic forecasting leadership, has received a five-year, $2.16 million grant renewal from the National Institutes of Health (NIH) to continue advancing methods for real-time forecasting during infectious disease outbreaks.
A professor of biostatistics and epidemiology in the School of Public Health and Health Sciences, Reich and his team were known even before the pandemic for producing some of the world’s most accurate models for predicting influenza trends. He leads a flu forecasting collaborative from his Reich Lab that is funded and designated by the Centers for Disease Control and Prevention (CDC) as one of two national Influenza Forecasting Centers of Excellence.
In addition to his influenza and COVID-19 work, Reich also has conducted years of research into how dengue fever, a severe viral infection transmitted by mosquito bites, is spread in Thailand.
“The need to support outbreak response through better data collection and modeling has never been clearer,” Reich says. “This project, funded by the NIH, will enable us to continue our research into the ‘basic science’ of outbreak modeling. We aim to identify key data sources that can improve real-time modeling and forecasting efforts.”
When SARS-CoV-2 began its global spread in early 2020, Reich used his expertise to develop a forecasting collaborative among more than 80 infectious disease modeling teams from around the world. That turned into the COVID-19 US Forecast Hub, which maintains an up-to-date forecasting record of U.S. pandemic cases, hospitalizations and deaths, in coordination with the CDC. Every week, these forecasts are published on the CDC website.
Forecasting, based on complex mathematical and statistical models, is a critical tool for infectious disease response and planning, capable of saving lives and quelling outbreaks by informing public health guidelines and providing information to the general population.
“The main idea behind developing a forecasting hub is that we can reduce our dependency on any one particular model and develop better understanding of where the pandemic might be heading based on the results from dozens of models. There is a ‘wisdom of the crowds’ aspect to this project,” Reich says.
Reich’s ensemble forecast representing multiple models generated the most consistent predictions of COVID-19 deaths between May 2020 and April 2021, according to preliminary research not yet peer-reviewed.
As the pandemic waves ebbed and flowed, Reich worked round-the-clock. In high demand, he juggled interviews with science journalists from around the world, explaining the complexities of modeling and the importance, as well as the limits, of the forecasts.
“At the same time as we’ve been running an operational forecasting system, we’ve also been doing research about what aspects of the pandemic are predictable. We’ve learned a lot in this regard. For example, many models make pretty accurate forecasts for about a month into the future of how many deaths from COVID-19 we might see. But it has proven much harder to predict future COVID-19 cases. Typically, only one- to two-week forecasts achieve good accuracy for cases.
“This whole experience of working closely with so many groups and in particular the CDC has been exciting from a collaborative perspective, but also exhausting. I wish I had a crystal ball that could tell me how much longer we’re going to have to keep the Hub running, but unfortunately, that’s not the kind of crystal ball we can make. We’re going to keep doing it until these forecasts really aren’t needed any more.”