Answering the Unknown

Baylor faculty, students and alumni apply statistical knowledge to the fight against COVID-19

As the world battles the coronavirus, several people associated with Baylor University are involved in research related to the pandemic. Much of this research is in the field of biopharmaceutical drug research through the University’s partnership with Eli Lilly and Company, a pharmaceutical company headquartered in Indianapolis.

James D. Stamey, M.B.A. ’97, Ph.D. ’00, professor of statistical science at Baylor, is one of several faculty involved in the research partnership. Due to contractual agreements with Lilly, specifics about the research are not for public knowledge. In fact, Stamey says Baylor faculty and students involved with the research are always on a need-to-know basis with the pharmaceutical company.

“The tiny little bits we learn, we can’t divulge,” Stamey says. “They give us scenarios, and we give them statistical help.”

The Baylor-Lilly partnership, which is more than a decade old, received the 2012 Statistical Partnership Among Academe, Industry and Government (SPAIG) Award. Dozens of Baylor doctoral candidates have participated in statistical analysis through the partnership.

John W. Seaman, Ph.D., professor of statistical science, is one Baylor faculty member who provides statistical research support to Lilly.

“We support a team whose purpose is to develop and implement advanced statistical methods across Lilly’s research efforts,” Seaman says. “We have advised them on a number of mathematical, statistical aspects — clinical trials and the like. We don’t analyze data; we try to help them build statistical models that can be useful.”

David Kahle, Ph.D., associate professor of statistical science, is another Baylor faculty member involved in the Lilly partnership. Kahle describes statistical models as mathematical simplifications of reality.

“They help us reason about various aspects of the world in which we live,” Kahle says. “That’s generally what we do with any of our collaborators, whether that’s in biopharmaceutical drug development or in any other aspect or domain of statistics.”

The Lilly partnership is one of several collaborative efforts between Baylor’s statistical science department and professional fields.

Amanda S. Hering, B.S. ’99, Ph.D., associate professor of statistical science, works with various industry partners in wastewater treatment research. Jane L. Harvill, Ph.D., professor of statistical science, has supported applications in the solar energy industry. Joon Jin Song, Ph.D., associate professor of statistical science, who is involved with the Lilly collaboration, previously worked in rainfall data research.

“There are application domains in statistics everywhere,” Kahle says. “Take someone like Dr. Song. One year, he’s working on rainfall data, and the next year he’s working on coronavirus-type things, generally speaking. That’s unique to our field, and it’s what drew a lot of us to statistics. We describe it as being able to play in other people’s back yards.”

The Baylor Collaborative on Hunger and Poverty (BCHP) provides meals during summer months to children who rely on school-provided meals during the academic year. BCHP expanded their efforts last spring when schools closed due to the pandemic. Kahle and statistical science students analyze data to help make BCHP’s efforts as efficient as possible.

In many ways, efficiency is the ultimate goal for statistical scientists in their collaborative efforts. This includes Baylor’s work with Lilly.

“One of the purposes for our models is to enable either more rapid enrollment of subjects into a clinical trial or to reduce the number ultimately required for that clinical trial to be useful without compromising safety but still understanding the efficacy of the drug,” Seaman says.

“In general, that’s what statistics is about — trying to model uncertainty in ways that allows us to more completely understand the scientific problem at hand. One way you can find out if all the light bulbs you’re producing at your plant are working is to run them all until they die. That’s not very efficient. Statistical methods are meant to be efficient in ways that allow us to draw conclusions without looking at an entire population, which is usually not possible.”

According to Seaman, Baylor’s partnership with Lilly has generated more than $2 million in research grants. Furthermore, the partnership leads to internships and postgraduate work.

“Our first student went to Lilly in 1996, and we’ve routinely sent students there ever since,” Seaman says. “Most people stay once they get there.”

One such alumna is Karen Price, Ph.D. ’01, a Lilly research fellow and statistical innovation center lead. Price is the company’s liaison with Baylor researchers.

“It’s common for doctoral candidates to defend their dissertations and then go work for companies before they’ve actually walked,” Kahle says. “Many have gone to Lilly; others have gone to the Food and Drug Administration and other pharmaceutical companies.”

Seaman says one recent graduate has already been given substantial responsibility in the fight against COVID-19.

“I can’t tell you her name or the company, but she works for a major pharmaceutical firm involved with developing vaccines,” Seaman says. “She defended her dissertation a year ago, and she’s recently been made responsible for all randomization efforts for their clinical trials.”

Not all Baylor statistical science alumni work under as much secrecy, but most work in fields that limit dissemination of specific information about their work.

Somer Blair, Ph.D. ’17, was a part of a four-person research team at Baylor that engaged with partners at Lilly to study Bayesian modeling, especially prior elicitation methods, simulation and associated coding. Blair was an Outstanding Graduate Student Teacher Award nominee while at Baylor.

Today, Blair is manager of data analytics for Fort Worth’s JPS Health Network, where her team analyzes medical records data registries. JPS Health is the county tax-supported hospital and, therefore, serves many vulnerable people in the community, including uninsured and homeless patients. Blair appreciates that her work leads to better understanding of how to help those in need, especially during the pandemic.

“We also use that data to analyze protocols around COVID-19 throughout our hospital — descriptive-type studies where we might summarize outcomes or characteristics across different subgroups,” Blair says. “As we’re moving more toward telehealth, we analyze certain procedures and outcomes related to that, including patient satisfaction with telehealth. Sometimes, our data analysis involves things we don’t immediately think about when discussing coronavirus research.”

Blair says Baylor prepared her professionally by teaching her how to think creatively, something that has been most beneficial during the pandemic when she and others in her field have been asked to answer questions surrounding a topic about which limited data exists.

“I used to think statistical science was ‘by the book,’ but most real-world problems require much more,” Blair says. “At Baylor, I learned how to address statistical problems in different ways from different angles, and that I could create my own ways to address questions and test these methods logically and systematically. My mentors at Baylor also emphasized the importance of communication and people skills.”

Additionally, Blair says she gained professional confidence through her Baylor studies and her work with Lilly.

“I was encouraged to think creatively and assess my methods critically, and they trusted and supported me to do so,” she says. “In every problem, confidence in your work is crucial to making headway.”