Skip to content

Team, funding and citation

Organizations behind MedDeID

MedDeID is developed through a clinical–academic collaboration at the University of Antwerp and Antwerp University Hospital (UZA). The work brings together data science and software engineering from Adrem Data Lab, clinical data infrastructure and technical expertise from the UZA Data Innovation Lab, and research support from the Research Foundation – Flanders (FWO).

Adrem Data Lab Data science, machine learning, software development and research supervision. University of Antwerp Academic home of Adrem Data Lab and the MedDeID research collaboration. UZA Data Innovation Lab Secured clinical research environment, technical setup and data-innovation expertise. Research Foundation – Flanders Supports the research through an FWO SB Fellowship, grant 1SA3226N.

The research team includes Stig Hellemans, Tom Stroobants, Elyne Scheurwegs, Pieter Meysman, Philippe Jorens and Kris Laukens, with affiliations across Adrem Data Lab, the University of Antwerp and UZA. Correspondence: stig.hellemans@uantwerpen.be.

How to cite MedDeID

Once the MedDeID article has a DOI, cite the article for the overall framework. Also cite the exact software, model or dataset release used so the work can be reproduced.

Until then, use this project-level citation and replace the bracketed fields:

Hellemans, S., Stroobants, T., Scheurwegs, E., Meysman, P., Jorens, P., & Laukens, K. (2026). MedDeID: an open framework for locally deployable de-identification of Dutch clinical text using real or synthetic training data [Software]. Version [version]. [repository or DOI].

For a reproducible citation, include:

  • the component name and version;
  • the model repository and immutable revision;
  • the dataset split and immutable revision or DOI;
  • the annotation-guideline version, when relevant.

The MedDeID collection on Hugging Face links to the current model and dataset cards. Prefer an artifact's own citation block when available.

Licence

MedDeID software repositories use AGPL-3.0-only unless stated otherwise. Each repository's LICENSE and NOTICE files are authoritative.

Models, datasets, annotation guidelines and lookup resources may use different terms. Check the licence and attribution information attached to each artifact before redistribution. External comparison systems retain their own licences.