Leaf

Leaf generates dynamic SQL queries from a flexible biomedical concept system to enable cohort discovery and translational biomedical research across heterogeneous clinical databases and enterprise data warehouses (EDWs).

Key Features:

• Flexible Biomedical Concept System: Defines hierarchical items and ontologies with textual representations and associated SQL query building blocks for representing clinical concepts.
• Model-agnostic, data-driven querying: Compiles queries that can target single or multiple clinical databases with differing data models.
• Abstract Syntax Tree-based Query Compilation: Transforms query representations into abstract syntax trees that are compiled into dynamic SQL queries.
• Dynamic SQL Query Generation: Produces executable SQL for querying enterprise data warehouses (EDWs) and clinical databases.
• Support for Heterogeneous Data Sources: Enables integration of clinical data across multiple, diverse databases and clinical data warehouses.

Scientific Applications:

• Cohort Discovery: Identifies patient cohorts by querying clinical criteria defined via biomedical concepts and SQL building blocks.
• Health System Analytics: Enables analysis of clinical data across institutional datasets for system-level metrics and studies.
• Clinical Research Data Warehouses: Facilitates querying and extraction of research-ready datasets from EDWs and clinical repositories.
• Precision Medicine Biobanks: Supports selection and characterization of biospecimen-linked cohorts using clinical concept definitions.
• Large-scale Clinical Studies: Enables analysis of extensive patient cohorts across heterogeneous data sources for translational research.

Methodology:

Defines hierarchical concepts and ontologies with textual representations and SQL query building blocks; generates abstract syntax trees from query representations and compiles them into dynamic SQL executed against enterprise data warehouses (EDWs) and clinical databases.