User Tools

Site Tools


8a.006.UL - Visual Analytics of Real-time Hospital Emergency Department Operations: A Human-Centric Approach

Project - Summary

Nurses and medical practitioners have a lot of stressful events in their daily life due to patient crisis, workload etc. This can lead to burnout [3] which is a well-known problem in nurses lives that results in a lot job abandons every year. Moreover, detecting high-stress moments in daily life can reduce the risk of threatening effects of stress. However, identification of accurate stress level is challenging due to variability of signals from person to person. Moreover, the performance of stress detection using existing wearables is still not excellent. In this project, we do a comprehensive analysis of the different types of stress signals, their statistical features, and their utility in stress prediction and investigate our prediction on nurses in a hospital. We also detected stressful events and send the results to the participants to validate the results. We observe that multi-modal sensing, combined with lag characteristics of bio-metric signals, can substantially improve stress detection performance. The accuracy of predicting stress increased to 98% using Skin temperature signal using 10-fold cross-validation in within subject and 84% for between subject stress detection tasks.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Raju Gottumukkala PI (337) 482-0632 UL Lafayette
Christoph Walter Borst Co-PI (337) 482-1023 UL Lafayette
Prabhakar V Vemavarapu PhD Student Not Available Not Available UL Lafayette
Syedmajid Hosseini PhD Student Not Available Not Available UL Lafayette
Tomi Tiekko Project Mentor Not Available Tieto
Mike Lucito Project Mentor Not Available Not Available Schumacher Clinical Partners
Funded by:
Schumacher Clinical
Partners & Tieto

Project - Novelty of Approach

  • Digital twin tracks various metrics at multiple granularity levels (department, facility, city, and total)
  • Digital twin in not limited to basic descriptive analysis like patient inflow but advanced analytics like patient forecasting, change in patient volume, and tracking
  • Integrates all the sensor data into a virtual 3D environment
  • This pilot is cost effective with minimal changes to the infrastructure

Project - Deliverables

1 Machine learning techniques for stress detection from wearable GSR, heart rate and skin temperature
2 Machine learning techniques to detect activity from accelerometer and heart rate sensor
3 Visualization of stress data in 3D environment
4 Software for automated detection of stress and activities using wearable, and real-time streaming processing system
5 Testing the prototype on actual hospital setting on nurses

Project - Benefits to IAB

The proposed approaches can be used to detect stress in real-time from various physiological signals collected from a wearable device. This information can be used to identify employees that are under stress and various situations that increase the stress of the employees. This can be used to improve employee retention by improving the quality of like in the workplace.

Project - Documents

FilenameFilesizeLast modified
8a.006.ul_ip_disclosure.docx22.6 KiB2020/09/08 14:25
8a.006.ul_final_report.docx2.2 MiB2020/08/17 16:11
8a.006.ul_mid-year_report.docx238.1 KiB2020/01/21 12:27
8a.006.ul_executive_summary.docx54.1 KiB2019/10/28 12:20

Life Form Feedback

Year 8 Project Poster Session Feedback (Fall 2019)
Real name Great Progress On Course Needs Change Off Course Abstain
Kimmo Valtonen (kimmo.valtonen)     
Sumit Shah (sumit.shah)     
projects/year8/8a.006.ul.txt · Last modified: 2021/06/02 15:29 by sally.johnson