By the Numbers: USU Data Scientists Chase Respiratory Illness Outbreaks, including COVID-19

Utah scientist Lindsay Meyers leads a team of colleagues laboring to refine a syndrome-based, but pathogen-specific, rapid disease-reporting system, installed at hospitals, to get ahead of respiratory scourges that plague individuals and communities with seasonal regularity. Her project connects diagnostic systems to a central database, where test results can be surveyed.

In the back of her mind, Meyers, director of medical data systems at Salt Lake City-based BioFire Diagnostics, LLC, was keenly aware of the threat of a novel respiratory virus popping up among the usual suspects of cold and flu-causing viruses. She didn’t know 2020 would be the year such a pandemic would envelope the globe, causing a real-life threat to herself and her community.

“I’d have been very disappointed with myself, if I hadn’t had a system in place before something like this erupted,” says Meyers, who briefly attended Utah State University, before earning a bachelor’s degree from the University of Utah in 2005.

About a year before SARS-CoV-2, the novel coronavirus that cause COVID-19, was identified in China, Meyers ran into BioFire colleague and USU alum Jay Jones (BS’09, MS’09, Mathematics and Statistics), who was hosting a visit with his former mentor, Chris Corcoran, head of USU’s Department of Management Information Systems.

“It was a chance meeting,” says Corcoran, David B. Haight Endowed Professor of Analytics. “Jay and I were discussing possible projects for students in USU’s new Master of Data Analytics (MDATA) program, Jay spotted Lindsay and immediately saw a potential fit.”

Meyers welcomed the idea.

“BioFire first implemented its real-time, automated monitoring system, BioFire Syndromic Trends, at hospitals in 2014, and it’s now installed on about 10 percent of the more than 5,000 BioFire Systems throughout the United States, as well as overseas,” she says.

Collected data is uploaded to a cloud database, which forms a huge, and growing, dataset about the prevalence, seasonality and co-infections of dozens of respiratory pathogens detected in millions of patient samples. Harnessing the data to distill reliable information, including disease surveillance, is a formidable challenge.

“Enlisting USU graduate students and faculty really bolstered our resources,” Meyers says. “Tackling our dataset is like looking at a tree laden with cherries and only being able to pick a few.”

Corcoran, with USU colleague Richard Cutler, professor in the Department of Mathematics and Statistics, selected 16 MDATA students for the project and, starting in November 2019, set to work with Meyers. An interdisciplinary program coordinated by USU’s MIS, Mathematics and Statistics, Economics and Finance and Computer Science, MDATA includes specialization options in MIS, statistics, as well as economics and finance.

“The diversity of backgrounds among our team members has strengthened this project,” says Spencer Perry, a student team member, who earned a master’s degree in MDATA, specializing in statistics, this spring and who also completed a bachelor’s degree from USU in public health in 2018. “Data analytics combines mathematics, statistics, data science, machine learning, computer science and more, plus our team members represent an even more diverse range of academic disciplines.”

Perry delved into his background in life sciences to approach the project. Viruses, he says, are deceptively simple entities; basically protein shells enclosing genetic data.

“Viruses are not even classified as life,” he says. “They’re not made of cells, they can’t maintain themselves in a stable state and they can’t make their own energy, grow on their own or survive without living hosts – including humans.”

Coronaviruses, Perry says, along with many other viruses, including influenza strains, rotaviruses and rhinoviruses, are endemic, meaning “they’re always with us.”

That miserable cold that wore out its welcome and the “stomach bug” that wrecked a family reunion were caused by viruses (or by, or in combination with, bacterial infections) that health practitioners see and battle year in and year out.

But viruses mutate and, when they do, can yield novel strains like those that caused the so-called Spanish Flu of 1918 or the situation vividly unfolding with this season’s COVID-19.

“And that’s what we’re trying to stay ahead of,” Perry says. “The challenge is in coming up with a way of standardizing the flood of data coming in from very different sites, collected by varied methods, to the BioFire system.”

Among the team’s aims is developing ways to organize the data to maximize use of every morsel of information, including identifying signals potentially related to emergent pathogen strains.

“Precision is key to enable scientists receiving this data to develop better tests to detect pathogens, to develop vaccines ahead of outbreaks and to enable development of antivirals, antibiotics and other remedies to accurately target specific pathogens,” Perry says.

Meyers says she has full confidence in the USU team’s efforts.

“USU’s graduate students have the intellectual capability and intense curiosity we need to pursue these big questions,” she says. “Because of this and because they have oversight and guidance from excellent faculty mentors, I’d had no hesitation sharing their analyses with the Centers for Disease Control.”

Corcoran says the project is exactly the kind of real-world experience he strives to cultivate for his students.

“These kinds of partnerships are crucial to prepare our students for the workforce,” he says. “Nearly every company is now a data company. Employers need employees, who can take on messy problems with no back-of-the-textbook answers.”

Perry sees data analytics as a pivotal tool for public health.

“I think, someday, we’ll look at epidemiology not as its own field, but as an area of statistics that focuses on public health,” he says. “It’s very much based on mathematics and probability.”