Clinical Trial and Disease Search With Ad Hoc Interactive Ontology Alignments

Presented at: 9th Extended Semantic Web Conference (ESWC2012)

by Daniel Sonntag, Jochen Setz, Maha Ahmed-Baker, Sonja Zillner

We will explain how an Linked Open Drug Data (LODD) application based on diseases, drugs, and clinical trials can be used to improve the (ontology-based) clinical reporting process while, at the same time, improving the patient follow-up treatment process. Specific requirements of the radiology domain let us aggregate RDF results from several LODD sources such as DrugBank, Diseasome, DailyMed, and LinkedCT. The idea is to use state-of-the-art string matching algorithms which allow for a ranked list of candidates and confidences of the approximation of the distance between two diseases at query time. Context information must be provided by the clinician who decides on the “related”-mappings of patient context and links he wants to follow in order to retrieve disease and medication information. The industrial prototype is implemented as a facetted browsing tool which allows for all these aspects of interactive ontology alignment.

Keywords: LODD for radiology, ad-hoc ontology matching, facetted browsing, healthcare, interactivity

Resource URI on the dog food server:

Explore this resource elsewhere: