adam

Query ADaM dataset structures and generate dummy ADaM test data in R.

13|6|Updated Aug 9, 2026
One-click install
npx skills add https://github.com/kaipingyang/CDISC_training --skill adam-kaipingyang
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: adam
Source: https://github.com/kaipingyang/CDISC_training/tree/main/.claude/skills/adam
Command: npx skills add https://github.com/kaipingyang/CDISC_training --skill adam-kaipingyang

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and scripts (resource) components.

What problem does it solve? Clinical data programmers often need quick, authoritative answers about CDISC ADaM dataset structures—variables, keys, derivations—and realistic test data, without digging through lengthy implementation guides or writing boilerplate data-generation code. ## Core Features & Use Cases - Dataset Structure Lookup: Answers questions about 30 ADaM datasets (ADSL, ADAE, ADLB, ADTTE, ADPC, etc.) using per-dataset reference files covering variable lists, origins, keys, and derivation methods based on ADaM IG v1.3 and pharmaverseadam. - Dummy Data Generation: Runs an R script to generate .sas7bdat and .qs dummy datasets for any ADaM dataset, with configurable subject counts and optional reference to real data distributions. - Concept Explanations: Explains ADaM concepts such as BDS, OCCDS, PARAMCD, and analysis flags in Chinese. - Use Case: A statistical programmer asks "ADLB 怎么衍生 BASE?" and receives the derivation method from the ADLB reference, then requests "生成 30 个受试者的 ADSL dummy 数据" to get ready-to-use test files. ## Quick Start Ask the assistant to list the variables of the ADSL dataset or generate dummy data for ADAE with 30 subjects.

Frequently Asked Questions about adam

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate dummy ADaM datasets in R?

Run the skill's R script with the dataset name, for example Rscript generate_dummy.R ADSL 30, to produce dummy data for 30 subjects. It outputs both .sas7bdat and .qs files to a configurable output directory, defaulting to /tmp/adam_dummy/.

What ADaM datasets are covered by this skill?

The skill covers 30 ADaM datasets indexed in references/index.md, including ADSL, ADAE, ADCM, ADLB, ADLBHY, ADEG, ADEX, ADMH, ADPC, ADTTE and others. Each dataset has a dedicated reference file with variables, keys, and derivations.

Which ADaM standard version does the variable documentation follow?

The references are based on CDISC ADaM IG v1.3, supplemented with actual structures from the pharmaverseadam R package using CDISCPILOT01 example data. Some datasets also reference the OCCDS IG v1.1 and therapeutic-area supplements.

Can dummy data generation match real data distributions?

Yes, if the ADAM_REF_DIR environment variable points to a directory containing real sas7bdat files, the script infers realistic distributions from them. Otherwise it generates values purely based on variable types defined in the reference files.

What is the difference between BDS and OCCDS structures in ADaM?

BDS (Basic Data Structure) holds one record per subject per parameter per visit, used for findings like ADLB and ADEG. OCCDS (Occurrence Data Structure) holds one record per event occurrence, used for datasets like ADAE and ADCM. The skill's index explains both with dataset examples.