11/14/2017 Claudia Lutz, Carl R. Woese Institute for Genomic Biology
Written by Claudia Lutz, Carl R. Woese Institute for Genomic Biology
Embedded in our society is a cultural memory of the old-time family doctor, a medical practitioner who knows of your family, your history, and your daily life, and uses that knowledge to provide the most optimal care. One Illinois faculty member and his research team have been working to move closer to that goal by exploiting a piece of familiar technology—the smartphone that can now be found in the average American’s pocket.
Dr. Bruce Schatz, head of Medical Information Science and an affiliate professor of Computer Science, and his coauthors, including Qian Cheng (PhD CS '17), previously developed software for Android phones that uses the phone’s native motion sensor to predict a lung patient’s disease state. That prediction was based on the patient’s movements during an exam at a hospital. In a study published in Telemedicine and e-Health (DOI: 10.1089/tmj.2017.0008), the official journal of the American Telemedicine Association, they described their latest step forward—a demonstration that the new version of their software can be used to monitor a patient’s status while they perform everyday tasks outside of the hospital.“The question I started working on ten years ago was how you could capture everything going on in a person’s life, their environment, in enough detail to be clinically relevant, to help predict things,” said Schatz, who is also a member of the Carl R Woese Institute for Genomic Biology. “By watching people moving with their phones . . . you might be able to tell what's special about the ones that do poorly.”
Traditionally, patients with respiratory conditions, including chronic pulmonary obstructive disease (COPD), can be assessed via several measures. Patients may be asked to breath into a spirometer, a device that measures the volume of air that a patient can breathe out in a certain time. A measure of health that relates to everyday functioning is the six minute walk test, in which patients are evaluated based on walking pattern as they walk down a corridor.
Schatz’ team had already developed and tested software that can predict the result of either of these tests with high accuracy. But to quickly identify when a patient’s condition may be deteriorating, or to collect the volume of health data that a precision medicine effort would require, medical practitioners need an easier, more scalable way to monitor patients.
“All of the efforts with fancy sensors that you stick onto people have failed because people don’t stick them on properly, or only use them sporadically, or they only last for a certain period of time,” said Schatz. “If you asked, what sensors does nearly everyone have that you can measure things with, you end up with phones.”
To fully take advantage of the ubiquity of cell phones, Schatz’s team needed to adapt their previously developed software to effectively monitor relevant data about gait during everyday life. The original version of their software used the motion sensors and accelerometers in Android phones to track body motion, pauses during which an individual might be catching his or her breath, overall speed, and other features.
Many of these features are complicated if the individual is doing tasks around the home or running an errand. In a clinical setting, any change from a healthy person’s gait might be related to lung function; in other settings, it might instead indicate setting down an object, interacting with another person, or pausing while considering what to do. Researchers improved their software’s ability to focus only on movements that occurred during intervals when walking was the only activity.
Schatz hopes that by continuing to expand the functionality of this software, his group can make it possible to collect much more robust data sets to further efforts in predictive or precision medicine. As methods for collecting, analyzing and interpreting genomic data advance, methods for evaluating an individual’s environment, behavior and physical condition must do likewise, in order to make accurate predictions about their future health. Schatz is currently collaborating with researchers at the University of Illinois at Chicago on one branch of the National Institutes of Health Precision Medicine Cohort Program.
“[Health] depends on both genes and environment . . . if you want to do prediction of heart and lung disease or any of those major chronic diseases, you have to do this,” he said. “You have to look at diet and exercise and stress.”
As the capabilities of his group’s software mature, Schatz is seeking new opportunities and partnerships to apply their work to these larger goals. His hope is that by building wellness technologies into everyday devices, he can help build better health into the life of every individual.