Obstructive Sleep Apnea (OSA) is a common sleep-related respiratory disease, characterized by repetitive upper airway collapse leading to episodic asphyxia and low oxygen levels. Globally there are more than 400 million middle aged adults who have moderate to severe OSA.
OSA is usually diagnosed by an overnight sleep study. While this study collects rich, complex data, very little of this is able to currently be used in clinical decision making. This complex data is essentially reduced into a single number to assess OSA severity. One of the problems posed by reducing this data to a single number is that it does not accurately predict long term health complications such as cardiovascular disease (CVD).
Dr. Najib Ayas believes that physiologic features embedded in the polysomnogram can help predict incident CVD and other long-term health complications. Extracting this data using a signal processing technique and then applied to machine learning algorithms could help to determine which features or combination of features are predictive of CVD and other long-term outcomes.
Back to Innovators’ Challenge