Toward Unobtrusive Patient Handling Activity Recognition for Injury Reduction Among At-Risk Caregivers.

Area of Science:

• Biomedical Engineering
• Occupational Health
• Wearable Technology

Background:

• Nurses face high overexertion injury risk from patient handling activities with awkward postures.
• Current workplace activity recognition relies on inaccurate and non-scalable human observation.
• Accurate recognition of patient handling is crucial for mitigating caregiver injury risk.

Purpose of the Study:

• To develop an automated system for recognizing patient handling activities.
• To reduce the risk of overexertion injuries among healthcare providers.
• To introduce a novel spatio-temporal warping (STW) pattern recognition framework for activity classification.

Main Methods:

• Utilized a smart wearable device, Smart Insole 2.0, with integrated sensors for unobtrusive data capture.
• Developed a novel spatio-temporal warping (STW) pattern recognition framework.
• Calculated a warped spatio-temporal distance to quantify similarity for activity recognition based on plantar pressure.

Main Results:

• Achieved an overall classification accuracy of 91.7% for eight common nursing activities.
• Classified qualitative profiles of activities with 98.3% accuracy.
• Classified load levels of activities with 92.5% accuracy.

Conclusions:

• The proposed Smart Insole 2.0 and STW framework effectively recognize patient handling activities.
• This technology offers a promising solution for reducing healthcare provider injuries.
• Automated activity recognition can improve occupational safety in nursing environments.