Nurse Stress Dataset (Hosseini 2022)
Citation: Hosseini, S.M., Gottumukkala, R., Katragadda, S., et al. (2022). Scientific Data, 9, 330.
DOI: 10.1038/s41597-022-01361-y
Access: Dryad — free download
Tier: 2
Overview
Real-world stress detection dataset from 15 nurses working in a hospital during COVID-19. Naturalistic recordings with end-of-shift stress self-reports.
Signals
- Empatica E4 wristband
- EDA (electrodermal activity)
- HR (heart rate)
- TEMP (skin temperature)
- ACC (accelerometer — not used in original model)
Labels
- Stress events (detected algorithmically + validated by end-of-shift surveys)
- Survey responses on stress contributors
Relevance
Only real-world healthcare stress dataset available. Natural environment noise profile matches deployment conditions. Shows that EDA features (mean, min, max) were better predictors than HR/HRV in this context.
Linked To
- Physiology: asd-autonomic-dysregulation