Twitter for health - Building a social media search engine to better understand and curate laypersons’ personal experiences

Abstract

Healthcare professionals, trainees, and laypersons increasingly use social media over the Internet. As a result, the value of such platforms as a vital source of health information is widely acknowledged. These technologies bring a new dimension to health care by offering a communication medium for patients and professionals to interact, share, and survey information as well as support each other emotionally during an illness. Such active online discussions may also help in realizing the collective goal of improving healthcare outcomes and policies. However, in spite of the advantages of using social media as a vital communication medium for those seeking health information and for those studying social trends based on patient blog postings, this new medium of digital communication has its limitations too. Namely, the current inability to access and curate relevant information in the ever-increasing gamut of messages. In this chapter, we are seeking to understand and curate laypersons’ personal experiences on Twitter. To do so, we propose some solutions to improve search, summarization, and visualization capabilities for Twitter (or social media in general), in both real time and retrospectively. In essence, we provide a basic recipe for building a search engine for social media and then make it increasingly more intelligent through smarter processing and personalization of search queries, tweet messages, and search results. In addition, we address the summarization aspect by visualizing topical clusters in tweets and further classifying the retrieval results into topical categories that serve professionals in their work. Finally, we discuss information curation by automating the classification of the information sources as well as combining, comparing, and correlating tweets with other sources of health information. In discussing all these important features of social media search engines, we present systems, which we ourselves have developed that help to identify useful information in social media.