What this page covers: the per-source license terms that apply to each piece of this repository (code, model weights, source clinical data, annotation releases, synthetic samples), and the secondary citations a publication building on this work should include alongside the main paper. The root README has the high-level summary; this page is the authoritative breakdown.
The source code in this repository is released under the MIT License.
external/plm_ca/ is third-party code, vendored for convenience. It is a copy of JoakimEdin/explainable-medical-coding at the commit linked in the overview, included here only to make pipeline setup effortless and to pin the exact upstream version this work was developed against. It carries its own MIT license (external/plm_ca/LICENSE, (c) 2023 Joakim Edin). The handful of new scripts added on top of the vendored tree (listed in external/plm_ca/README.md under "Additions in this fork") are MIT under this repository's root LICENSE; any other modifications are minor adaptations of upstream code. Any work that uses or builds on code under external/plm_ca/ must cite Edin et al. (2024); see Citations.
The NER, AC, entity-only ICD-10 coding, and full-text ICD-10 coding model weights distributed via data_download.py were trained on data subject to non-commercial licenses (MIMIC-III, MIMIC-IV, MIMIC-IV-Note, and the i2b2/n2c2 challenge corpora for AC). Use of these weights for commercial purposes is forbidden. They are also not released for clinical decision support or billing automation. This restriction is inherited from the underlying training data and applies regardless of the MIT license that covers the code that produced them. The same constraint applies to the BioLM RoBERTa-PM encoder weights used as initialisation (Lewis et al. 2020; full BibTeX in Citations). If you need a commercially-usable model, retrain on data licensed for that use.
MIMIC-III, MIMIC-IV, and MIMIC-IV-Note are distributed by PhysioNet under the PhysioNet Credentialed Health Data License and require a PhysioNet account, completion of a free human-subjects research training course (CITI "Data or Specimens Only Research"), and a signed Data Use Agreement. The planned MIMIC-IV-Ext-EntityCoding annotation release will follow the same access model. This repository does not redistribute any MIMIC data; users must obtain it from PhysioNet directly. See data/README.md and external/plm_ca/data/raw/MDace/README.md for the expected layout once credentialing is complete.
i2b2/n2c2 challenge data (used to train the AC model) requires registration with the DBMI portal and a separate signed DUA; see ac/README.md.
The AC model is additionally trained on the public clinical-assertion label CSVs from bvanaken/clinical-assertion-data (van Aken et al. 2021). The label files are downloadable from the public GitHub repo, but they index MIMIC-III note text and only become useful when joined to NOTEEVENTS.csv from MIMIC-III v1.4, which remains under PhysioNet credentialed access. See ac/README.md for the staging layout. If you use these labels, please cite van Aken et al. (2021); BibTeX is in Citations.
The 400-note annotation release accompanying the paper is intended for PhysioNet distribution under the same credentialed-access terms as the underlying MIMIC-IV-Note text. Until that project is published, it is not inspectable from this repository. If you use the dataset after publication, please cite the paper and include the PhysioNet release in your data availability statement.
The MDACE annotation tree under external/plm_ca/data/raw/MDace/ was vendored by the upstream PLM-CA repository, not by this project. MDACE was created by Cheng et al. and is released under CC BY 4.0 at 3mcloud/MDACE. The note text it annotates still comes from MIMIC-III and is subject to the PhysioNet terms above. If you use these annotations, please cite Cheng et al. (2023); BibTeX is in Citations.
data/sample_data/sample_notes.csv contains GPT-4o-generated synthetic clinical notes authored for this project. They contain no real patient data and are released under the same MIT license as the code.
The main paper BibTeX lives in the root README. The entries below are the secondary citations a publication building on this work should include alongside it.
For the annotation dataset (exact citation pending PhysioNet publication):
@article{unpublished,
author = {Douglas, James C and Gan, Yidong and Hachey, Ben and Kummerfeld, Jonathan},
title = {{MIMIC-IV-Ext-EntityCoding: Clinical Named Entity Recognition and Assertion Classification Dataset for ICD Coding}},
journal = {{PhysioNet}},
year = {2026},
month = may,
note = {Version 1.0.0; PhysioNet publication pending}
}Any work using code under external/plm_ca/ must additionally cite Edin et al. (2024):
@inproceedings{edin-etal-2024-unsupervised,
title = "An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records",
author = "Edin, Joakim and
Maistro, Maria and
Maal{\o}e, Lars and
Borgholt, Lasse and
Havtorn, Jakob Drachmann and
Ruotsalo, Tuukka",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.280/",
doi = "10.18653/v1/2024.emnlp-main.280",
pages = "4869--4890"
}If you use the AC model or its training data, please also cite the bvanaken assertion labels:
@inproceedings{van-aken-etal-2021-assertion,
title = "Assertion Detection in Clinical Notes: Medical Language Models to the Rescue?",
author = "van Aken, Betty and
Trajanovska, Ivana and
Siu, Amy and
Mayrdorfer, Manuel and
Budde, Klemens and
Loeser, Alexander",
editor = "Shivade, Chaitanya and
Gangadharaiah, Rashmi and
Gella, Spandana and
Konam, Sandeep and
Yuan, Shaoqing and
Zhang, Yi and
Bhatia, Parminder and
Wallace, Byron",
booktitle = "Proceedings of the Second Workshop on Natural Language Processing for Medical Conversations",
month = jun,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.nlpmc-1.5/",
doi = "10.18653/v1/2021.nlpmc-1.5",
pages = "35--40"
}If you use any of the trained model checkpoints distributed with this project, please also cite the BioLM RoBERTa-PM encoder used as initialisation:
@inproceedings{lewis-etal-2020-pretrained,
title = "Pretrained Language Models for Biomedical and Clinical Tasks: Understanding and Extending the State-of-the-Art",
author = "Lewis, Patrick and
Ott, Myle and
Du, Jingfei and
Stoyanov, Veselin",
editor = "Rumshisky, Anna and
Roberts, Kirk and
Bethard, Steven and
Naumann, Tristan",
booktitle = "Proceedings of the 3rd Clinical Natural Language Processing Workshop",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.clinicalnlp-1.17/",
doi = "10.18653/v1/2020.clinicalnlp-1.17",
pages = "146--157"
}If you use the MDACE annotations:
@inproceedings{cheng-etal-2023-mdace,
title = "{MDACE}: {MIMIC} Documents Annotated with Code Evidence",
author = "Cheng, Hua and Jafari, Rana and Russell, April and Klopfer, Russell and Lu, Edmond and Striner, Benjamin and Gormley, Matthew",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.416",
doi = "10.18653/v1/2023.acl-long.416",
pages = "7534--7550"
}