How digital technology can improve mental health treatment

Like other faculty members of the Link Lab at the University of Virginia School of Engineering, Laura Barnes works at the intersection of data science and physical systems. In this case, though, the systems she is interested in are not bridges or buildings but human beings. Barnes, an associate professor of engineering systems and environment and director of the Sensing Systems for Health Lab, is exploring devices that can produce a continuous, real-time record of an individual’s health, setting the stage for more effective personalized care.

“Traditional methods for assessing and treating illness often require patients to travel to appointments with a trained clinician. This presents both cost and scalability issues,” Barnes said. “Furthermore, these in-person assessments often fail to capture the real-life experience of individuals suffering from memory limitation and recall bias.”

Smart devices such as smartphones and wearables offer a way to put these in-person assessments in context. Working in the background, they could produce data streams that would give clinicians a broader view of their patients’ health — and even alert them to changes that might require immediate intervention, saving lives. For patients without access to a clinician, these electronic health interventions could make much needed treatment a reality.

In essence, the enabling technology for the next great advance in medical technology may already be on peoples’ wrists or in their pockets. “Using advanced data science and machine learning techniques, the data generated by our smart devices could reveal a great deal about the health of individuals,” Barnes said. “They might give you an idea, for instance, about their physical activities, their moods and even their social networks.”

Laura Barnes with Sensing Systems for Health Lab students in group photo

In the Sensing Systems for Health Lab, Laura Barnes (far left) and her students are designing intelligent systems for understanding the dynamics and personalization of health and well-being. Fusing computational methods with technology such as smartphones and wearables, her group studies how macro- and micro-level human behaviors relate to health. Their technologies and methods are applied to chronic diseases such as anxiety, depression, cancer, infectious disease and traumatic brain injury.

Fusing computational methods with technology such as smartphones and wearables, her group studies how macro- and micro-level human behaviors relate to health. Their technologies and methods have been applied to chronic diseases such as anxiety, depression, cancer, infectious disease and traumatic brain injury.

In one such project, researchers in her lab are collaborating with UVA psychology professor Bethany Teachman’s group to explore the challenges involved in using technology to monitor and treat mental illness. Awards from the Hobby Postdoctoral and Predoctoral Fellowships in Computational Science and the National Institutes of Mental Health support the research.

In the case of mental illnesses, smartphones and wearables hold more than the promise of more attentive management and effective treatment. They can extend care to more people. Experts believe the number of people with diagnosable mental illnesses significantly outstrips the number of practicing psychologists available to treat them. As a result, technology-enabled interventions could emerge as a first line of treatment for a subset of conditions.

As a test case, Barnes and Teachman are developing systems to monitor and automatically deliver appropriate eHealth interventions to people with social anxiety, a condition that is particularly prevalent among college students. This will require a series of breakthroughs. The primary one is identifying digital biomarkers that accurately reflect a person’s mental state so the appropriate intervention can be delivered at the appropriate moment.

For instance, the accelerometer on a smartwatch might indicate that an individual’s hands are shaking while their GPS data reveals that he or she is at home. Taken together, this combination might be a biomarker, indicating that the individual is feeling anxious and has retreated from others. While Barnes’ group draws on computational methods to determine the most relevant fusion of data from existing smart devices, the researchers are open to collaborating with others to develop new sensing systems. For example, more sensitive digital biomarkers capable of distinguishing between degrees of anxiety could help enable timely intervention.

“If we could determine when people were feeling anxious, our system could automatically deliver an intervention — a deep breathing or a visualization exercise — that might reduce their anxiety,” Barnes said. “This is particularly important because, if untreated, anxiety tends to escalate.”

Digital biomarkers would also help clinicians understand whether their interventions were effective and whether an in-person consultation might be necessary.

Ultimately, Barnes hopes to go beyond digital biomarkers for anxiety to characterizing the social situations that, on an individual basis, generate anxiety. If the system could identify the triggers, it could act proactively by delivering an intervention that might eliminate or minimize the anxiety a person feels.

“Ideally, we would have a menu of interventions that the system could produce, customized for the context and for the individual,” she said. “Over the long term, this stream of digital interventions could significantly improve a person’s quality of life.”