Framework demonstrates improved performance in healthcare data integration, indicating potential for enhanced clinical applications.
Key Points
The neo4j-based framework greatly improves query performance for clinical data, achieving up to 48.4x faster execution times compared to traditional databases.
It integrates MIMIC-IV clinical data with SNOMED CT medical ontology, creating a graph of over 625,000 nodes and 2 million relationships.
Three clinical applications were successfully demonstrated, including analysis of pneumonia rates among ventilated patients and quality measure monitoring.
Results revealed that 96.7% of eligible diabetic patients were without necessary statin prescriptions, highlighting critical areas for healthcare improvement.