Predicting Transitions in Oxygen Saturation Using Phone Sensors.

Area of Science:

• Cardiopulmonary Health
• Mobile Health Technology
• Gait Analysis

Background:

• Mobile devices are transforming health monitoring, but their capacity for medical-quality vital sign measurement is not fully established.
• Oxygen saturation is a key health indicator, and previous research has validated phone sensors for gait pattern analysis.

Purpose of the Study:

• To investigate if smartphone sensors can continuously monitor oxygen saturation and gait parameters in cardiopulmonary patients.
• To determine if gait data can predict oxygen saturation levels and clinical stability transitions.

Main Methods:

• Twenty cardiopulmonary patients underwent 6-minute walk tests while wearing pulse oximeters and carrying smartphones with MoveSense software.
• The software continuously recorded oxygen saturation and motion data, enabling computation of spatiotemporal gait parameters.
• A gait model was trained to predict oxygen saturation categories and transitions based on walking motion.

Main Results:

• Oxygen saturation data clustered into three categories, correlating with Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages 1 and 2, and a 'Transition' category indicating clinical instability.
• The gait model achieved 100% accuracy in predicting the oxygen saturation status categories from walking motion for all 20 subjects.

Conclusions:

• Continuous oxygen saturation monitoring via smartphones can predict cardiopulmonary status, including unstable transitions.
• Gait models utilizing smartphone sensors can accurately predict these saturation-based clinical categories from walking.
• This research supports the development of smartphone-based medical devices for passive monitoring and prediction of clinical stability.