trial-metadata-extractor

To extract clinical trial metadata, provide the Skill with your Protocol and SAP documents via file paths or uploads. It parses study design, endpoints, populations, visit schedules, and analysis plans, then outputs a standardized JSON contract for downstream consumption.

The JSON output format serves as a contract-like structure capturing study design, endpoints, populations, visit schedules, and analysis plans from your Protocol and SAP. This standardized JSON can be directly consumed by downstream tools to plan ADaM datasets and generate mock TLG shells.

Yes, you can extract trial metadata from a Protocol without an SAP, or vice versa. The Skill processes either document independently to capture study design and endpoints, but providing both enables cross-document validation to flag discrepancies between Protocol and SAP elements.

Discrepancies are flagged by comparing Protocol and SAP elements for consistency during extraction. The Skill identifies mismatches between study design, endpoints, or analysis plans across both documents and surfaces them in the structured JSON output for your review.

Yes, the extracted trial metadata JSON is designed for downstream readiness and works for generating mock TLG shells. The contract-like JSON output can be fed directly into a mock TLG generator to plan ADaM datasets and create shell tables, listings, and graphs.

Limitations include reliance on the quality and completeness of your Protocol and SAP source files. The extraction preserves fidelity to source documents and flags discrepancies but does not resolve them, and the JSON output requires downstream tools to generate final mock TLGs.