Skip to content

EvidenceOSS/NINDS-CDE-Tools

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 

Repository files navigation

NINDS-CDE-Tools -- Utilities for NINDS Common Data Elements in TBI Research

Status: Incubating License: Apache-2.0 Part of: Evidence Commons

"Multi-cohort TBI research fails at harmonization, not at modeling. Without standardized variable definitions, even large datasets produce incomparable results."

Attribute Value
Status Incubating
Maturity Design Phase
License Apache-2.0
Part of Evidence Commons
Mission Pillar Pillar 2 (UQ & Conformal) + Pillar 1 (Evaluation) -- cross-cutting

Overview

NINDS-CDE-Tools is intended to provide developer-facing utilities for validating, transforming, and mapping clinical datasets to NINDS Common Data Element (CDE) compliant schemas. The NINDS CDE program standardizes data collection across TBI clinical studies, and NINDS CDE v3.0 reclassifies biomarkers under "Neurodiagnostic Technologies: Biomarkers." Consistent CDE compliance is a prerequisite for reproducible multi-cohort TBI research, yet no open-source tooling exists to validate datasets against CDE definitions or harmonize variables across studies programmatically.

The core ontology that will power these tools already exists in the private evidenceos-research/evidenceos-ontology package: 13 YAML schema files totaling 7,780 lines. The parent codebase also contains a loader registry (991 LOC) with 247+ CDE mappings across 8 datasets. This repository is designed to contain the extracted, standalone tooling layer -- a CLI and Python library that wraps these schemas for community use. None of the planned tools exist yet; the schemas exist, the tooling does not.

Architecture (Planned)

Component Description Parent Code Exists This Repo
CDE Validation CLI Validate dataset columns, value ranges, and coding against NINDS CDE definitions Schemas: 13 YAMLs, 7,780 lines Not started
Mapping Utilities Transform non-CDE datasets to CDE-compliant format Loader registry: 991 LOC, 247+ mappings Not started
Dataset Harmonization Align variables across multi-cohort studies (IMPACT, CRASH, CENTER-TBI) to shared schema Partial (loader handles per-dataset) Not started
CDE Browser Searchable catalog of CDEs with definitions and value sets Not built Not started
FHIR Mapping Bridge CDE definitions to FHIR resources for health system interoperability Not built Not started

Current State

What exists in the parent codebase (evidenceos-research/evidenceos-ontology):

  • 13 YAML schema files (7,780 lines total): cbim_framework, clinical_entities, endophenotype_atn, imaging_cdes, implementation_science, knowledge_translation, methodology_extraction, provenance, temporal_phases, universal_schema, lsr_to_modeling_pillar, tbi_modeling_taxonomy, living_systematic_review
  • Loader registry: 991 LOC, 8 datasets, 247+ CDE mappings

What does not exist yet:

  • CDE validation CLI
  • Mapping utilities as standalone tools
  • Cross-cohort harmonization helpers
  • Searchable CDE browser
  • CDE-to-FHIR mapping layer
  • Any standalone code in this repository

Extraction Plan

  1. Extract CDE schema definitions from ontology YAML files into a standalone, documented format
  2. Build validation CLI that checks datasets against extracted schemas (column presence, value ranges, coding)
  3. Implement mapping layer for common input formats (REDCap exports, CSV, FHIR Questionnaire)
  4. Add harmonization utilities tested against known multi-cohort alignment challenges
  5. Build searchable CDE catalog with definitions, value sets, and cross-references
  6. Publish as standalone Python package (PyPI) with no dependency on the private research codebase

Ecosystem Context

graph LR
    A[NINDS CDE v3.0<br/>13 YAMLs] --> B[NINDS-CDE-Tools]
    B --> C[All Research Packages<br/>CDE compliance]
    B --> D[Lab-in-a-Box<br/>data quality gate]
    B --> E[AIDA<br/>CDE-to-FHIR mapping]
    style B fill:#2A9D8F,stroke:#1E3A8A,color:#fff
Loading

NINDS-CDE-Tools is designed to serve all research packages (CDE compliance), Lab-in-a-Box (data quality enforcement at ingestion), and AIDA Infrastructure (interoperability via CDE-to-FHIR mapping). As standards tooling for global TBI research, this is a strong Digital Public Good candidate. Canonical source: evidenceos-research/evidenceos-ontology. See Evidence Commons for related projects.

Contributing

This project is in the design phase. The underlying schemas exist in the parent codebase, but no standalone tooling has been extracted. Contributions are welcome once the initial CLI and validation layer are in place.

License

Apache-2.0 -- see LICENSE for details.

About

NINDS Common Data Element tooling — CDE extraction, mapping, validation for TBI research datasets

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors