Grant: Project Grant

Pulse oximetry is a vital monitoring tool clinicians use to track patient status. By measuring the absorption of two colors of light through the patient's finger, pulse oximeters calculate arterial blood oxygenation which is a metric of cardiopulmonary function. Pulse oximeters are widely used throughout the hospital system and on nearly all in critically ill patients, including those with heart failure, COVID, and those undergoing surgery. However, studies have recently shown that some pulse oximeters overestimate blood oxygen saturation in dark-skin patients 3x more than in light-skin patients. This clinical bias can leave at-risk patients in danger of being overlooked because their blood saturation "seems" safe. To correct for this bias, we have developed an optical wearable that quantifies skin tone using novel sensors which implement a dozen colors of light. We have also created a skin-tone invariant arterial oxygenation algorithm that takes melanin into account. In this work, we will simultaneously evaluate and optimize the device and algorithm through two pilot studies, one on one study on striped pig models and one on healthy volunteers within the Toronto General Hospital. The animal model will allow us to optimize our device and algorithms and the healthy volunteers will allow for validation in humans during controlled desaturation tests. The combination of pilot trials will allow us to compare our findings in controlled settings and real-world scenarios, thus shedding light on the origin of the bias and providing us a unique dataset to test solutions. The outcome of this work will demonstrate the equitable design of pulse oximeters to the medical device industry, hospital systems, and consumer health market - enabling more equitable healthcare for the entire population of Canada.