Related Experiment Video
Updated: Jun 30, 2026
07:31
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Importance ranking and predictive model construction of WMSDs risk factors among shift-working nurses based on
Li-Chong Lai1, Hai-Chen Wu1, Xiao-Ying Cao1
1The Second Affiliated Hospital of Guangxi Medical University, No. 166 Daxue East Road, Xixiangtang District, Nanning City, Guangxi, 530007, China.
BMC Nursing
|May 28, 2026
Summary
Work-related musculoskeletal disorders (WMSDs) affect 85% of nurses, influenced by factors like sleep quality and night shifts. Machine learning identified key determinants, enabling targeted interventions to reduce nurse injuries.
Area of Science:
- Occupational Health
- Nursing Research
- Biomedical Engineering
Background:
- Work-related musculoskeletal disorders (WMSDs) are a significant occupational hazard for nurses, impacting patient safety.
- Circadian disruption from shift work may exacerbate WMSDs.
- Identifying WMSD determinants is crucial for developing effective interventions.
Purpose of the Study:
- To identify and rank individual and environmental determinants of WMSDs in shift-working nurses.
- To develop and validate a prediction tool for WMSDs using machine learning.
- To enable targeted, proactive interventions to reduce occupational injuries.
Main Methods:
- A cross-sectional study collected data on general information, lifestyle, psychosocial factors, and shift characteristics.
- Seven machine learning algorithms (LDA, PLS, RDA, GLM, RF, SVM-Radial, SVM-Linear) were employed.
- Model performance was evaluated using AUC, accuracy, and specificity; Random Forest showed the highest predictive power.
Main Results:
- WMSD prevalence was 85.19% among 1,080 shift-working nurses.
- Key determinants included perceived control, social support, sleep quality (PSQI), chronotype, bending, and night-shift characteristics.
- The Random Forest model achieved a high predictive performance with a median AUC of 0.919.
Conclusions:
- WMSD occurrence in nurses is influenced by a combination of individual, lifestyle, occupational, and psychosocial factors.
- The Random Forest algorithm demonstrated superior predictive capability for WMSDs.
- Implementing interventions based on identified determinants and the Random Forest model is recommended.
