Querying Complex Federated Clinical Data Using Ontological Mapping and Subsumption Reasoning

The secondary use of clinical data is an essential part of translational and comparative effectiveness research. In designing user interfaces for querying clinical data, we found challenges in managing disease classification codes, medication history, and computed indices, which were amplified by the longitudinal nature of clinical care. In order to address these challenges, we integrate domain-specific and standardized ontological/terminological systems as plug-and-play components in the system architecture for data exploration and user interface support. We present an ontology-driven query interface framework called VISAGE+ (Visual Aggregator and Explorer Plus), which is capable of: (a) mapping multiple data sources and database columns to single concepts in an ontological system, (b) querying multi-dimensional data using concept attributes, and (c) supporting queries that require transitive closure over the ontology class hierarchy (a.k.a. subsumption reasoning). Three VISAGE+ instances have been deployed for large-scale multi-center projects in Sleep Medicine, Epilepsy, and for exploring Medicare and Medicaid data. These instances demonstrate the generality and flexibility of the VISAGE+ framework in supporting large-scale data federation, novel user-interface features, and the usability of the interfaces and systems for clinical research.