Researcher delves into natural language processing to improve patient care
September 20, 2017 / by Nanci Hellmich
I want to help the patients get the treatment they need, get it when they need it, and get it faster.
Özlem Uzuner, a new associate professor in engineering, speaks three languages—English, French, and Turkish—but it’s her fourth language that could help save lives.
Uzuner is an expert in natural language processing, a field of computer science that involves turning human language into coded form. She is working on algorithms that translate physicians’ narratives about their patients into data that can be analyzed to find better treatments.
“The algorithm’s aim is to get the computer to understand human language the way we do,” says Uzuner who joined Mason Engineering’s Department of Information Sciences and Technology (IST) this semester. “Once we achieve a certain level of performance, we can give the algorithm all of our narrative text data, and it codifies everything for us.”
Uzuner and other researchers, including her collaborators at the Massachusetts Institute for Technology (MIT), are developing natural language processing methods to find insights into the unexplained deterioration of patients’ health, adverse prescription drug reactions, and problems with combinations of medications.
They are also extending their methods to help figure out which patients are best suited for specific clinical trials to test new therapeutic treatments.
Uzuner’s other recent work involved using computer methods to remove private information from patients’ records—such as names, ages, addresses, and insurance details—so the remaining health information can be used by medical researchers who study different diseases and treatments. "Those who could not study the data before because of privacy concerns now can, which means science moves forward faster,” she says.
This is a field where computer science meets medicine, and it’s the perfect career path for Cyprus-born Uzuner, who comes from a family of medical doctors, including her mother, brother, and five cousins.
She’s contributing to health care but in a different way than her family members. “I want to help patients get the treatment they need, get it when they need it, and get it faster,” Uzuner says. “Health is the most important thing. If you are not healthy, you don’t have anything.”
Her decision to pursue this area of research “was a little bit of inclination and a little bit of luck,” she says. She loves learning languages. When she went to MIT, she studied computer science and was looking for ways to combine her passion for languages and science.
She decided to bring her research to Mason because of the university’s emphasis on data analytics as well as the possibility of collaborating with faculty members in different departments who have expertise in health informatics. She also values Mason’s proximity to the National Institutes of Health and the National Library of Medicine.
Andre Manitius, the acting chair of IST, says, "Dr. Uzuner is a very well-known expert in natural language processing, data analytics, and in their applications to the narrative texts of medical records. She has pioneered this area, and she currently is a tremendous leader in the research community.”
Her specialty will make her work attractive not just to the Volgenau School of Engineering but also to health science groups on campus and potentially to Inova, he says.
Next semester, Uzuner will teach a master’s level course in natural language processing, and she hopes to recruit graduate students who are as excited about this research as she is.
“You need two students who really love what you do and they tell their friends,” she says, “and then you don’t need to recruit ever again.”
Dr. Uzuner is a very well-known expert in natural language processing, data analytics, and in their applications to the narrative texts of medical records.