Estimating Physical/Mental Health Condition Using Heart Rate Data from a Wearable Device

Area of Science:

• Biomedical Engineering
• Health Informatics
• Wearable Technology

Background:

• Accurate estimation of physical and mental health is crucial for employee well-being and self-awareness.
• Heart rate variability (HRV) is commonly used for health estimation but requires high-quality, long-term data, posing measurement challenges.
• Existing methods for health monitoring face technical difficulties in data acquisition and quality maintenance.

Purpose of the Study:

• To propose and evaluate a novel method for estimating physical and mental health conditions solely from heart rate (HR) data.
• To overcome the limitations associated with traditional HRV measurements by utilizing readily available HR data from wearable devices.
• To provide a more accessible and practical approach for continuous health monitoring.

Main Methods:

• Collected over 25 million data points of HR and steps from 97 participants over 3 months using Fitbit Inspire HR.
• Integrated daily physical condition questionnaires to supplement wearable data.
• Developed a prediction model focusing on the inactive period of HR, utilizing support vector machine (SVM) on preprocessed data.

Main Results:

• The proposed method achieved a balanced accuracy of 0.582 in predicting health conditions.
• This accuracy surpasses the state-of-the-art HRV-based method, which reported an accuracy of 0.565.
• Demonstrated the feasibility of using simple HR data for health condition estimation.

Conclusions:

• Estimating physical and mental health using only HR data is a viable alternative to HRV-based methods.
• The proposed method offers a practical solution for employee mental health care and objective self-understanding.
• Future research can further refine HR-based health monitoring systems.