Organizations: Sorbonne Université, Université Sorbonne Paris Nord, Inserm, Limics, F-75006 Paris, France · Université Paris-Saclay, UVSQ, Assistance Publique-Hôpitaux de Paris, Raymond Poincaré University Hospital, Infectious Disease Department, Garches, France · Yonsei University College of Medicine, Gangnam Severance Hospital, Department of Laboratory Medicine, Seoul, South Korea · Assistance Publique-Hôpitaux de Paris, Department of medical information, Paris, France · Health Data Hub, 75015, Paris, France · Université Paris-Dauphine, PSL, CNRS, CEREMADE, 75016, Paris, France · Université Paris-Saclay, CNRS, LISN, 91400, Orsay, France
The development of clinical natural language processing (NLP) systems is severely hampered by the sensitive nature of medical records, which restricts data sharing under stringent privacy regulations, particularly in France and the broader European Union. To address this gap, we introduce PARHAF, a large open-source corpus of clinical documents in French. PARHAF comprises expert-authored clinical reports describing realistic yet entirely fictitious patient cases, making it anonymous and freely shareable by design. The corpus was developed using a structured protocol that combined clinician expertise with epidemiological guidance from the French National Health Data System (SNDS), ensuring broad clinical coverage. A total of 104 medical residents across 18 specialties authored and peer-reviewed the reports following predefined clinical scenarios and document templates. The corpus contains 7394 clinical reports covering 5009 patient cases across a wide range of medical and surgical specialties. It includes a general-purpose component designed to approximate real-world hospitalization distributions, and four specialized subsets that support information-extraction use cases in oncology, infectious diseases, and diagnostic coding. Documents are released under a CC-BY open license, with a portion temporarily embargoed to enable future benchmarking under controlled conditions. PARHAF provides a valuable resource for training and evaluating French clinical language models in a fully privacy-preserving setting, and establishes a replicable methodology for building shareable synthetic clinical corpora in other languages and health systems.
Figures & tables
Path
Type
Description
Root level
name
string
Name of the corpus
version
string
Version
licenses
string
Licenses of the dataset
license_urls
string
URL to license descriptions
description
string
Short description of the corpus
Table 1: Path-based schema description
Figure 2: Patient count per specialty. Specific documents, off distribution, were written for four use cases. A subset of the dataset is temporarily embargoed to enable future evaluations under controlled conditions, thereby limiting the risk of large language model contamination through prior exposure to the data. In total, we release 4,254 patients ( 6,185 documents) and keep 755 patients ( 1,209 documents) under embargo for future release.
Figure 3: PARHAF population pyramid chart.
Figure 4: Document and word count by document type. The number of documents to write for a patient was imposed by the initial scenario. Word count is computed by the library edsnlp (v0.20).
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
Figure S1: Change in distribution for each specialty, after the adjustment described in Section 2.3.1 .
Background: Dutch medical corpora are scarce, limiting NLP development. \ Methods: We translated English datasets, identified medical text in generic corpora, and extracted open Dutch medical resources. \ Results: The resulting corpus comprises ± 35 billion tokens across the medical domain in about 100 million documents, freely available on Hugging Face. \ Conclusion: This work establishes the first large-scale Dutch medical language corpus for pre-training and downstream NLP tasks.
B. van Es
Central Diagnostic Laboratory, University Medical Center Utrecht, Utrecht, The Netherlands. · R&D, B-lab, Castricum, The Netherlands.
Synthetic data is increasingly used to enable the development and evaluation of AI systems in domains where access to real-world data is restricted. In healthcare, clinical documentation presents particular challenges due to its sensitivity. This work introduces a synthetic clinical notes pipeline and dataset designed to support the development of clinical AI tools while avoiding the privacy risks associated with real patient data. The dataset is generated using a modular pipeline that combines structured patient generation, semi-structured patient journey simulation, and unstructured clinical note generation using large language models. The pipeline is designed to prioritise internal consistency across longitudinal patient records, while also capturing variation in writing style, note structure, and clinical detail. Additional mechanisms, including LLM-based validation and augmentation steps, are used to improve faithfulness, realism, and diversity of the generated notes. We release a dataset of 70 synthetic patients, each associated with 20-50 clinical notes spanning a full hospital journey. The dataset is provided at multiple levels of validation, enabling users to balance realism and scalability depending on their use case. This dataset supports the development, testing, and evaluation of clinical AI systems, including summarisation tools, coding models, and decision support systems, without reliance on real patient data.
We present eCREAM-MedCorpus, a new and unique large-scale dataset of clinical notes produced in Emergency Departments of Italian hospitals. The corpus, in its current version, is composed of approximately 4 million clinical notes fully anonymized, covering diverse phases of patient care during the stay in the emergency department. In addition, a subset of about six thousand notes has been manually annotated by clinical experts through a structured Case Report Form (CRF) containing 132 items relevant for two patient situations in emergency departments, dyspnea and loss of consciousness. Items may assume numerical values (e.g., for blood saturation), categorical (e.g., for level of consciousness ), binary (e.g., for presence of traumas), and mixed value types. The annotation process involved multiple clinicians and underwent iterative revision to resolve ambiguities in item formulation, resulting in a richly structured (although high imbalanced) resource. The dataset aims to fill a relevant gap of data able to support both the development and the use of Large Language Models in concrete medical applications. We describe the data collection protocol, the on-site anonymisation pipeline, corpus statistics, and the annotation scheme. Finally, we propose CRF-filling as a novel structured information extraction benchmark, and provide zero-shot baseline resulting from Gemma-27B and MedGemma-27B. To the best of our knowledge, eCREAM-MedCorpus is the largest freely available dataset of clinical notes existing for the Italian language.
Tiziano Labruna, Guido Bertolini, Pietro Ferrazzi +1
Fondazione Bruno Kessler, Trento, Italy · Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Milan, Italy · University of Padua, Padua, Italy