Social Media Driven Public Health Informatics: Applications in Regulatory Science

Health information technology (health IT) is an emerging interdisciplinary research field that brings innovative and unique opportunities as well as challenges for information systems (IS) researchers. Public Health Informatics (PHI) is an important subdomain of health IT but yet extensively studied by IS researchers. The limited IS literature on PHI surveillance, especially on regulatory science opens new gates for IS researchers to test existing and develop new theories, to design and configure innovative methods and models, and to provide practical health managerial insights. The emerging information technology innovations provide novel insights and opportunities to improve public health surveillance. Specifically, social media analytics and intelligence, broadly accepted and applied in the current IS domains, has motivated a new branch of research in regulatory science that may bear prolific fruits both in theoretical and pragmatic perspectives. This dissertation seeks to address the problem and fills the research gaps by proposing a systematic research framework for regulatory science surveillance using social media-driven approaches. Five essays in the design science paradigm are developed. Essay I aims to provide a basic understanding of social media user-generated content on regulated products by text mining and social network analysis techniques. User-generated content can further motivate research on the user level. Essay II takes this perspective and models user features based on text. Essay III focuses on another important feature of social media – network structures, and develops a computational algorithm for large-scale social network analysis on modeling social influence. After proposing innovative IS solutions to real regulatory public health problems, essay IV endeavors to validate the use of social media datasets by combining survey and social media data for data triangulation. Finally, essay V, motivated by the keyword-based social media data collection processes, strives to automatically and accurately detect sensitive regulated product street names, utilizing rich text information in the ever-changing linguistic environment on social media. Beyond the practical insights of health decision support and solutions provided in each essay, this dissertation offers a systematic social media-driven regulatory science informatics research paradigm to guide future PHI and other analytical IS research