Advances in PPG Wearables: From Novel Digital Biomarkers to Racial Bias Correction

Smart watches, rings, or accessories utilize optical sensors and accelerometers combined with powerful machine learning create exciting algorithms allowing parameters such as steps, calories, heart rate and its derivatives, blood oxygen, and sleep pattern to be predicted. The Quantified Self industry blossoms and wearable gadgets are now everywhere. We believe that wearable technology is still in its infancy and there are great potentials for the technology to disrupt existing clinical practices and aspire to diagnose or predict the onset of diseases that were once believed to be unpredictable. This presentation discusses pathways and activities needed to build wearable products beyond their current consumer-grade paving the way for wearable as a medical-grade device. Two specific examples will also be presented. The first example relates to how we use our wearables to discover a unique novel parameter that correlates to acute stress and how it is being utilized to improve the service of mental health counselling. The second example focuses on characterizing and correcting skin tone biased, an inherent problem associated with optical sensor, in pulse oximetry which has recently been reported for its racial and ethnic discrepancy.