Caring for Caregivers: Monitoring Mental Well-being in Healthcare Workers and Educators Through AI

Healthcare workers and teachers are invaluable to our society, a fact that became especially clear during the COVID-19 pandemic. However, these essential professions are facing severe understaffing and overwhelming workloads, leading to significant stress and burnout. In fact, between one-third and half of the employees in these sectors are already showing signs of burnout. Unfortunately, current methods to monitor mental well-being are used infrequently and often rely on cumbersome questionnaires, providing only occasional snapshots of employees’ mental well-being.
The ‘Caring for Caregivers’ project addresses this issue by exploring an innovative system called Hummin’. With Hummin’, users can report at any time feelings of gratitude or vulnerability through a clickable button or app, helping them become more aware of their emotional state. We tested Hummin’ over 8-16 weeks at a school and a care facility, gathering data through the Hummin’ system, surveys, interviews, and a focus group. Our findings show that Hummin’ can boost self-awareness and psychological safety, there is also an indication that it might have a positive influence on burnout and stress. In its current design, Hummin’ does not encourage long-term engagement in tracking mental well-being.
We also used the real-time data collected through Hummin’, along with demographic information, to predict burnout and stress using AI, with mixed results. Although these predictive models are not yet ready for practical use, they are promising and offer valuable insights for future research on burnout and stress monitoring.
This project contributes towards continuous and close to real-time mental well-being monitoring, which could help detect and prevent burnout earlier. Additionally, it contributes to the design community by assessing the effectiveness of a tangible interface for self-reporting emotional states and to the AI community by exploring how such data could be used to predict burnout and stress.