Extracting clinically-actionable information from wearable physiological monitors

Abstract

In this thesis I examine several ways of extracting information from wearable monitors so as to help make clinical decisions. Wearable physiological sensors are developing rapidly, and pose a possible part of the solution to the demands of an aging population and rising health care costs. It is important that the data produced by such sensors be processed into information that is clinically relevant and will have an impact on the practice of medicine. I collected data in an ambulatory setting from several wearable physiological sensors, including electrocardiogram, arterial blood pressure, pulse plethysmograph, respiration and acceleration. Using this data set, I demonstrated a few approaches - including signal processing, and algorithms based on the application of physiological models - to extract clinically relevant information. These approaches are potentially of interest to both device makers interested in developing wearable monitors, and to clinicians who will be using such monitors in the future.