Four Use Cases for Research Networking: A Medical School, University, Federal Agency, and Physician Network

2015-08-18T15:55:32Z (GMT) by Griffin Weber
Introduction: Profiles Research Networking Software (RNS) is an open source semantic web application built using the VIVO ontology (http://profiles.catalyst.harvard.edu). We originally created Profiles RNS in 2008 to break down silos in biomedical research by helping investigators at a medical school find new collaborators across their institution. Since then, dozens of organizations have adopted Profiles RNS, and many have found novel use cases for research networking tools that we did not anticipate when we first started. This presentation compares four implementations of Profiles RNS, which illustrate the broad range of ways that institutions benefit from research networking and the unique policies, data sources, and software customizations required for those websites. <br><br>Harvard Catalyst Profiles (HC Profiles): Our first implementation of Profiles RNS was for medical school faculty in Harvard’s Clinical and Translational Science Award (CTSA) program, called Harvard Catalyst. In addition to information obtained from internal administrative systems, the primary data source for HC Profiles is PubMed. The software automatically adds publications to faculty profiles using an author name "disambiguation engine". Data mining algorithms extract MeSH terms and other information from the publications to generate interactive visualizations that illustrate a person's research interests and collaborators. HC Profiles is open to the public, and APIs enable anyone to repurpose the data for other websites. <br><br>Harvard Faculty Finder (HFF): The Harvard Provost Office requested their own university-wide version of Profiles RNS that would include all Harvard faculty. Because few faculty outside of biomedicine publish in PubMed, we purchased commercial publication data and added many new data sources (books, courses, patents, projects, etc.). In order to gain approval from each Harvard school for HFF, we had to remove most of the faculty “profiling” functionality from the website. The remaining tool is primarily used as a cross-school search engine, which guides users to local school or department websites that present additional information about the faculty. <br><br>FDA Profiles: The US Food &amp; Drug Administration uses Profiles RNS internally to assemble committees to review medical devices. Custom data feeds import sensitive information about prior device reviews and committee members. As a result, FDA Profiles must be protected behind a local firewall and access is restricted to a small number of users. <br><br>Undiagnosed Disease Network (UDN Profiles): The UDN is an NIH-funded collaboration between seven U.S. clinical sites to help patients with rare or hard-to-diagnose diseases. These patients often suffer for years without a diagnosis and are unable to find a physician who can treat them. The UDN performs a complete clinical evaluation of its patients, including genomic analyses, metabolic studies, and evaluation of environmental exposures. They then seek to match the patients with the best possible physicians within their network. UDN Profiles is a single website containing physicians from multiple UDN hospitals across the country, which enables users to search for individuals with expertise in rare or complex conditions.