Stop the Stresses

UT Tyler Team Creates Algorithms for Wearable Devices to Decrease Physician Burnout

Publication Date: 04/20/2020

Many physicians, particularly those working in emergency rooms, work in fluid, high-stress environments. When they aren’t feeling well in terms of occupational stress or burnout, who is caring for them? One College of Engineering research team hopes to improve the mental health of physicians with an algorithm — a set of coded digital instructions — developed for wearable devices to detect stress.

Graduating senior Joshua Stapp of Tyler and assistant professor Dr. Premananda Indic collaborated with a team at the University of Massachusetts Medical School on this project. The researchers focused on detecting stress in emergency medicine (EM) physicians using the Empatica E4, which is a wearable and commercially available biosensor.

Burnout is considered a long-term consequence of high-level stress in the workplace. Among physicians, burnout is associated with increased medical errors, depression, substance abuse, mental health issues, early retirement and an increased rate of suicide, according to the American Journal of Medicine. The current estimate of burnout among United States physicians is estimated to be around 50 percent.

“Stress has several adverse effects on health, and in particular, physicians who are in emergency rooms can have burnout effect, which can significantly affect not only their health, but also the decision-making process that can impact patient care,” Indic said. “By reducing stress, the negative effects of burnout in physicians and other high-stressed professionals can also be mitigated.”

A Team Effort: The UT Tyler and UMass Experiment

The team’s preliminary study included eight EM physicians from the UMass Memorial Medical Center in Worcester, Massachusetts. Each subject in the experiment wore the Empatica E4 on their wrist during work shifts and would self-report times that they felt stressed. Each device, with an appearance like an Apple iWatch, collected data including heart rate and electrodermal activity, which is the measurement of changes in electrical properties of the skin.

Indic said the team determined the analysis of the data and the development of machine learning algorithms that could distinguish between stress-related and normal sensor data in order to detect stress in real time.

Even with the small sample size, the resulting algorithms exhibited a stress detection accuracy of about 70 percent when comparing the data prior to a reported stress event to data during periods of no stress, according to Indic. This means that the developed algorithms were able to detect stress prior to the subject realizing they were stressed, he said.

“Further study into this trend can give rise to interventions that can help physicians reduce stress and/or receive notifications of incoming stress,” Indic said.

Making an Impact

Assisting Indic on the project since last March, Stapp processed the collected biometric data and developed the stress detection algorithms used. He enjoyed providing solutions to a real-world issue while working alongside the professor.

“Seeing the algorithms I developed being used to help physicians is very fulfilling for me, and I’m proud to be part of such important research started here at UT Tyler with Dr. Indic,” said Stapp, who will graduate this May with a Bachelor of Science in electrical engineering and a minor in computer science.

Results of the first stress detection algorithms were completed last summer and submitted in a manuscript to the 53rd Hawaii International Conference on System Sciences (HICSS). Refinements to the algorithms and data analyses were explored, and Indic and his research team presented their research in January at the conference.

Looking Ahead

So, what’s next for the group?

Collaborators at UMass are exploring avenues to run another similar study with a larger number of subjects to further refine the stress detection algorithms. By identifying physiological markers with the Empatica E4, Indic hopes the results can then help identify appropriate intervention strategies to reduce stress. Stress is also one of the predictors for mental health problems and substance abuse, he noted.

“East Texas has high rates of mental health and substance abuse problems compared to other regions in Texas,” Indic said. “We plan to continue working with the UMass group, develop this technology and possibly have it available to help East Texas healthcare workers as well as individuals recovering from substance abuse. That’s our hope.”

After graduation, Stapp will apply the skills he learned at UT Tyler to help develop training devices in the medical field. He plans to pursue master’s and doctorate degrees. In his spare time, Stapp enjoys working on personal electronic and programming projects and playing video games.

Serving UT Tyler since 2016, Indic works with both undergraduate and graduate students in his research center, which is dedicated to developing wearable biosensor systems for the prediction of life-threatening events in the mental and physical health arenas.

Indic holds degrees in technology and electrical and electronics engineering and completed postdoctoral training in medicine. He also is a senior member of the Institute of Electrical and Electronics Engineers (IEEE). Aside from wearable technology and big data analytics, his other research interests include signal processing, machine learning, biomathematics and nonlinear dynamics.

#coe