nejm_dataset_v1.jsonl: A dataset related to the New England Journal of Medicine (NEJM) clinical cases.
medqa_extended_v1.jsonl: Extended dataset for medical question-answering tasks with comprehensive coverage.
medqa_v1.jsonl: Dataset focused on general medical question-answering.
mimiciv_v1.jsonl: Dataset based on the MIMIC-IV medical database with patient trajectories.
nejm_extended_v1.jsonl: Extended version of the NEJM dataset with additional clinical scenarios.

Usage

To load the datasets, simply use the following code:

python

import json

# Example for loading a dataset
with open("dataset_filename.jsonl", "r") as f:
    data = [json.loads(line) for line in f]

License

This repository is under the MIT License. See the LICENSE file for more details.

Acknowledgments

This work was supported by the MedAgentSim project.
The MIMIC-IV dataset is publicly available and was used for medical data simulations.
Citation for AgentClinic:

@misc{schmidgall2024agentclinic,
      title={AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments}, 
      author={Samuel Schmidgall and Rojin Ziaei and Carl Harris and Eduardo Reis and Jeffrey Jopling and Michael Moor},
      year={2024},
      eprint={2405.07960},
      archivePrefix={arXiv},
      primaryClass={cs.HC}
}

Citation for Self-Evolving Multi-Agent Simulations:

@misc{almansoori2025selfevolvingmultiagentsimulationsrealistic,
      title={Self-Evolving Multi-Agent Simulations for Realistic Clinical Interactions}, 
      author={Mohammad Almansoori and Komal Kumar and Hisham Cholakkal},
      year={2025},
      eprint={2503.22678},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2503.22678}, 
}

Contact

For any questions or inquiries, please reach out to Komal Kumar.

ItsMaxNorm/MedAgentSim-datasets

MedAgentSim Datasets

Datasets Included

Usage

License

Acknowledgments

Contact

ItsMaxNorm/
MedAgentSim-datasets