Keywordsblockwise missing patterns
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractAnalytics is the discovery, interpretation, and communication of meaningful patterns in data. Analytics is an integral component of health information systems (IS), showing promise in various areas such as disease risk modeling, clinical intelligence, pharmacovigilance, precision medicine, hospitalization process optimization, digital health, preventive care, etc. In my dissertation, I focus on analytics in two important application areas of health IS, namely digital health and preventive care. Digital health analytics focuses on enhancing individual wellbeing via continuous tracking of health indicators, while preventive care analytics is the science of extracting insights from electronic health records to assist clinical decision-making towards preventing illness or diseases. With rapid development in healthcare big data and IoT technologies, research in digital health and preventive care (DHPC) analytics is increasing in importance and complexity. Limited predictors, incomplete data, non-linear input-outcome relationships, the presence of multiple outcomes, and heterogeneity in effects are some of the key challenges in DHPC analytics. My dissertation consists of three essays that introduces a collection of novel quantitative methods to address these challenges. The first essay presents a new feature engineering method that uses network science to predict high-cost patients at the point of admission in hospitals with limited information. The second essay describes a novel method to analyze incomplete data containing block-wise missing patterns using a reduced modeling approach. The third essay leverages a wearable devices-based study and introduces three new quantitative methods to model the effects of sound level on an individual’s physiological wellbeing in the workplace. The set of predictive and explanatory modeling methods proposed in these essays not only address important modeling challenges in DHPC analytics, but also more broadly contribute to business analytics, design science, and health IS research.
Degree ProgramGraduate College
Management Information Systems