MEDCIS: Multi-Modality Epilepsy Data Capture and Integration System.

Area of Science:

• Neurology
• Biomedical Informatics
• Data Science

Background:

• Sudden Unexpected Death in Epilepsy (SUDEP) is the primary cause of epilepsy-related mortality.
• SUDEP is most prevalent in patients experiencing frequent, intractable seizures.
• Studying SUDEP necessitates large, high-risk patient cohorts, typically found in Epilepsy Monitoring Units (EMUs).

Purpose of the Study:

• To develop an informatics infrastructure for integrating data from multiple EMUs to facilitate SUDEP research.
• To address challenges in data integration and access across distributed epilepsy databases.

Main Methods:

• Developed the Multi-Modality Epilepsy Data Capture and Integration System (MEDCIS).
• Utilized Natural Language Processing (NLP) for retrospective clinical text, an ontology-driven interface for prospective data capture, and cohort search/signal visualization tools.
• Employed the Epilepsy and Seizure Ontology (EpSO) for streamlined interfaces, enhanced usability, and federated query execution across distributed databases.
• Integrated 936 patient datasets from two EMUs (UH CMC and NMH), processing over 77GB of multi-modal signal data.

Main Results:

• MEDCIS successfully federated patient data from disparate databases using EpSO.
• The system processed and rendered over 77GB of multi-modal signal data via a web interface.
• The EpSO-driven VISual AGgregator and Explorer (VISAGE) interface accurately answered approximately 74% of 40 clinical questions.

Conclusions:

• MEDCIS provides an integrated environment for multi-EMU data capture, integration, and analysis, crucial for SUDEP research.
• The use of EpSO enhances data harmonization and query capabilities across distributed epilepsy datasets.
• The system demonstrates significant potential for advancing SUDEP research by enabling analysis of larger, integrated patient cohorts.