Heat Stroke Prevention in Hot Specific Occupational Environment Enhanced by Supervised Machine Learning with...

Area of Science:

• Occupational Health
• Environmental Monitoring
• Biomedical Engineering

Background:

• Wet-bulb globe temperature (WBGT) is a common heat stress index but has limitations in accuracy within varied microclimates.
• Individual factors like ventilation, clothing, and body size significantly influence heat strain in hot occupational environments.
• Existing heat stroke prevention methods may lack personalization, especially in demanding work conditions.

Purpose of the Study:

• To develop and validate a novel personalized vital sign index for improved heat stroke prevention.
• To integrate personalized heat strain temperature (pHST) with other vital signs for individual-level heat stress assessment.
• To enhance the accuracy of heat stroke prevention in hot occupational settings using a machine learning approach.

Main Methods:

• A wearable device was developed integrating a pHST meter, heart rate monitor, and accelerometer.
• A personalized vital sign index was formulated, adjusting WBGT based on individual physiological data.
• Supervised machine learning algorithms were employed, trained on the proposed personalized vital index.

Main Results:

• The developed system achieved an overall heat stroke prevention accuracy of 85.2% in a summer occupational experiment.
• The system demonstrated a high true positive rate of 96.3% and a true negative rate of 83.7%.
• The personalized vital sign index effectively adjusted heat stress assessment to individual levels.

Conclusions:

• The proposed personalized vital sign index, incorporating pHST and machine learning, offers a more accurate approach to heat stroke prevention.
• Wearable technology integrated with personalized physiological monitoring can significantly enhance safety in hot working environments.
• This personalized approach addresses the limitations of traditional WBGT measurements in diverse occupational settings.