What problem does it solve? Clinical programmers must manually write complex, CDISC-conformant R code to derive ADaM BDS findings datasets (ADVS, ADLB) from SDTM domains, including baseline flagging, change-from-baseline calculations, and visit mapping — a repetitive, error-prone process requiring deep knowledge of admiral function idioms. ## Core Features & Use Cases - BDS Derivation Workflow: Generates step-by-step R code for parameter assignment, date derivation (ADT, ADY), visit mapping (AVISIT, AVISITN), baseline flagging (ABLFL), and change from baseline (CHG, PCHG) using idiomatic admiral functions. - ADLB-Specific Logic: Derives normal range variables (ANRLO, ANRHI, ANRIND, BNRIND) and supports CTCAE toxicity grade carry-through for laboratory datasets. - QC-Ready Output: Produces code with # REVIEW: annotations at protocol-specific decision points, uniqueness assertions, and required-variable checks for regulatory traceability. - Use Case: Given SDTM VS data, a completed ADSL, and an ADaM spec, generate a complete ADVS derivation program with baseline flags, change from baseline, and analysis flags ready for human QC review. ## Quick Start Ask the agent to derive an ADVS dataset from the SDTM VS domain using the study's ADaM BDS specification and the completed ADSL dataset.
