Wednesday, 5/23/2007
5:30 PM - 6:30 PM
Level: Case Study

Clinical research and decision support systems require that patient medical records be stored as values for variables so that pertinent data can be readily accessed and understood by computers. However, conventional database systems present significant challenges to managing the frequent additions and alterations to data models necessitated by rapid changes in medical knowledge and the multiple perspectives brought to bear on medical information. For this and other reasons, most electronic medical record systems (EMRs) capture a large amount of medical data as narrative reports, which can be easily altered and extended as needed, but cannot be used directly in research or computerized decision support. When narrative medical records are converted to values for variables, inadequate database extensibility leads to the production of an ever-expanding number of data silos designed to store specific kinds of medical data, or, occasionally, the same data modeled in different ways. The costs associated with the production, management, and integration of these disparate, often proprietary, data silos create substantial obstacles to conducting clinical research and implementing sophisticated decision support systems. This presentation reports on the experience of Cleveland Clinic in using semantic technologies to overcome these obstacles. Specifically, the presentation provides overviews of the following topics:

• Development and features of SemanticDB, a semantic application for managing extensible semi-structured content (RDF and XML)
• Characteristics and use of a flexible ontology for capturing patient medical data
• Integration of foundational and domain-specific ontologies to support reasoning and query capabilities
• Dynamic generation of user interfaces for entering, viewing, and managing patient medical data
• Implementation of data validation and quality control
• Management of workflows and other process information
• Plans for future research and development.

Christopher Pierce is a research scientist and software engineer employed by Cleveland Clinic to manage cardiovascular and thoracic surgery research projects and apply semantic technologies to associated data and knowledge management problems. He holds a Ph.D. from the University of Washington, has many years of database, Web application, and semantic technology development experience, and has made over 30 presentations at professional conferences and other public venues.