Pulse Rate Guided Oxygen Saturation Monitoring Using a Wearable Armband Sensor.

Area of Science:

• Biomedical Engineering
• Physiological Measurement
• Wearable Technology

Background:

• Continuous clinical-grade SpO2 monitoring outside hospitals is difficult.
• Wearable photoplethysmography (PPG) devices are common for health but lack clinical accuracy for SpO2.

Purpose of the Study:

• To introduce and evaluate two novel SpO2 algorithms: PRR (pulse rate derived ratio-of-ratios) and GPDR (green-assisted peak detection ratio-of-ratios).
• To assess algorithm performance using PPG signals from an upper-arm wearable device in controlled and uncontrolled clinical settings.

Main Methods:

• Developed PRR and GPDR algorithms utilizing unique pulse rate frequency estimations.
• Isolated pulsatile (AC) components of red and infrared PPG signals.
• Evaluated algorithms with green, red, and infrared PPG data from healthy subjects (n=36) and COVID-19 patients (n=52).

Main Results:

• GPDR showed the lowest RMSE: 1.6±0.6% (exercise), 3.6±1.0% (hypoxia), and 2.2±1.3% (clinic).
• PRR had higher error but broader coverage.
• Mean errors across all datasets were 0.2±2.8% for PRR and 0.3±2.4% for GPDR.

Conclusions:

• Both PRR and GPDR algorithms demonstrate high performance with low error and good coverage in both controlled and uncontrolled environments.
• These algorithms offer a promising solution for continuous, clinical-grade SpO2 monitoring in out-of-hospital settings.