Framework for Modernizing Legacy Clinical Reporting Systems Using AI
A novel strategy has been developed to modernize legacy clinical reporting systems for AI integration in pharmacoinformatics, without altering the original programming. This innovative method was applied to a SAS reporting library comprising 558 components. It incorporates a metadata layer, including a bridge map and a typed Intermediate Representation (IR), along with an orchestrator for component management. This enables the generation of structured data compatible with large language models. Moreover, the approach allows for the gradual replacement of certain outdated elements with improved core routines, thereby easing the challenges posed by complex systems in drug development and pharmacovigilance.
Key facts
- Framework is non-destructive and does not alter legacy source code.
- Uses a metadata layer with bridge map, IR, and orchestrator.
- Validated on a 558-component SAS reporting library.
- Enables AI-driven pharmacoinformatics readiness.
- Supports optional incremental consolidation.
- Addresses bottlenecks in drug development and pharmacovigilance.
- Legacy systems produce opaque output with no machine-readable intermediate layer.
- Avoids full rewrites or incremental refactoring that preserves structural barriers.
Entities
—