GWU, Office of CEHP – Watergate Office Building, 1st floor conf. room
2600 Virginia Ave. NW, Washington, D.C. 20052
SNOMED CT is the largest clinical ontology in the world, used in over 40 countries to support clinical documentation and analytics. Quality assurance of large ontological systems such as SNOMED CT is an indispensable part of their life cycle, but remains challenging. We focus on the detection of missing hierarchical relations in SNOMED CT, leveraging both lexical and structural features. We identify non-lattice subgraphs in SNOMED CT as indicative of potential errors and leverage lexical patterns to propose remediation for these errors. Preliminary results indicate that these automated methods for quality assurance can facilitate the work of human editors.