Study of Factor Associated with Fatigue in Emergency Medicine Professionals Using Wearable Sensor and Machine Learning

EasyChair Preprint 15567

Abstract

This study investigates the factors associated with work-related fatigue among emergency medical professionals (EMPs) in government and private hospitals in Thailand. Utilizing a cross-sectional descriptive study design, the research employs questionnaires addressing social factors such as age, gender, and economic status, as well as work-related factors including experience, hours worked, and workload. A total of 50 EMPs, comprising nurses, nursing assistants, and paramedics, participated in the study. Health data were gathered through wearable sensors, capturing physiological metrics such as heart rate and electrocardiogram (ECG) data alongside questionnaire responses. To assess work-related fatigue, the Multidimensional Fatigue Inventory (MFI-10) was utilized. The analysis of correlations between various factors and levels of fatigue was conducted using Chi-square tests. Findings indicate that both personal and work-related factors significantly correlate with fatigue in emergency medical personnel, particularly smoking habits, supervisory support, and engagement in emergency work, with P-values < 0.05. Following the identification of related factors, the study further explored the relationship between social factors, sensor data, and post-work fatigue using machine learning models, specifically Gradient Boost, XGBoost, and CatBoost. Results showed that the CatBoost model provided the highest performance, achieving an accuracy of 93.33% compared to the other models. This research contributes valuable insights for health promotion initiatives aimed at improving the quality of life for EMPs, ultimately supporting the sustainability of the healthcare system.

Keyphrases: Fatigue Assessment, Work-related Fatigue, Workload, machine learning, wearable sensor

@booklet{EasyChair:15567,
  author    = {Natnicha Tohchalee and Wares Chancharoen and Wiriya Mahikul},
  title     = {Study of Factor Associated with Fatigue in Emergency Medicine Professionals Using Wearable Sensor and Machine Learning},
  howpublished = {EasyChair Preprint 15567},
  year      = {EasyChair, 2024}}