Crosswalk is being written and revised in public. Content, predictions, and behavior will change without notice, and parts of it are certainly still wrong. If you find something that is, the repository takes issues and I would rather hear it than not.
The destination predictions here are teaching aids, not a validated mapping. The domain of the standard concept decides which table a row belongs in, and this tool guesses at that from a small built-in list. Confirm every routing decision in Athena against the OMOP CDM specification before you rely on it.
This tool and the surrounding curriculum are provided for education and training. Nothing here is medical, clinical, legal, regulatory, or professional advice, and nothing here creates a professional relationship of any kind. Decisions about your own pipeline, your institution's data, or a regulatory submission should be made with your own qualified advisors.
This software is provided as is, without warranty of any kind, express or implied, including warranties of merchantability, fitness for a particular purpose, accuracy, and non-infringement. The author makes no representation that the mappings, predictions, code references, or concept identifiers shown are correct, current, or complete. OMOP vocabulary releases change concept identifiers, standard status, and mapping targets, and content here may lag those changes.
To the fullest extent permitted by law, the author and Boyce Data Science accept no liability for any loss or damage arising from use of this tool or reliance on its output, including without limitation errors in a data pipeline, defects in a research data set, incorrect study results, or regulatory consequences.
This tool runs entirely in your browser. It makes no network requests after the page loads, and it transmits, stores, and logs nothing. Even so, do not paste identifiable health information into this page. Use synthetic or de-identified data. If you need to inspect real records, use the download button and run the file offline in whatever environment your governance requires.
You remain solely responsible for compliance with HIPAA, GDPR, your data use agreements, your institutional policies, and your IRB or ethics approvals. Nothing here relieves you of that, and nothing here should be read as an assurance that any particular use is compliant.
This is an independent educational project. It is not affiliated with, endorsed by, or produced on behalf of HL7 International, the OHDSI community, or any employer or institution the author is associated with. FHIR is a registered trademark of HL7 International. Any views expressed are the author's own.
Content is licensed CC BY 4.0 and the tool code is additionally available under the MIT License. Both are provided without warranty.
Crosswalk · Tool 03
Paste a FHIR bundle or a single resource. This walks through what it contains, predicts where each piece belongs in the OMOP CDM, and flags the places where information is about to be lost, assumed, or inverted. It reads R4, R5, and R6 ballot syntax.
Runs entirely in your browser. Nothing is uploaded or stored. Use synthetic data here; for real records, download the tool and run it offline. Instructions are below the workspace.
Each one loads a bundle and asks you one question. Answer it before you press Read it.
This page never transmits anything. The parser is inline, there are no network calls, and there is no analytics. Even so, a browser tab is not a controlled environment, and pasting identifiable records into a page served from someone else's domain is not something you want to have to explain later.
For real data, take the page with you:
It behaves identically offline, which is the whole point of a single file with no dependencies. The same applies to every tool in this series.