Healthcare Information Networks: Discovery and Evaluation

Large amounts of healthcare information have been made available over the Internet by various sources. To better assess both the quality and credibility of online healthcare information, it is essential to understand the properties of such information and its sources. This paper introduces a first attempt to study healthcare information distributed on the Web. In this project, we first implement a focused crawler to collect information from a variety of different healthcare-related web sites. We explore and analyze the crawled network, where patterns that were not previously visible can be observed. Our results provide insight into how healthcare information sources are related, as well as the sort of sites that are most important based on network topology. We further introduce two case-studies on user-based evaluation of quality of search results provided by commercial search engines. We show that trustworthiness is not employed in ranking search results, so that suspicious sites may sneak into top results, potentially confusing users.