PyHealth

Deep learning for healthcare: from raw EHR data to a trained, evaluated model in about 15 lines.

PyHealth is an open-source PyTorch toolkit for building, benchmarking and reproducing clinical AI models. It works with electronic health records (MIMIC-III, MIMIC-IV, eICU, OMOP), clinical text, biosignals (EEG, ECG) and medical images.

.. image:: https://img.shields.io/pypi/v/pyhealth.svg?color=brightgreen :target: https://pypi.org/project/pyhealth/ :alt: PyPI version

.. image:: https://static.pepy.tech/badge/pyhealth :target: https://pepy.tech/project/pyhealth :alt: Downloads

.. image:: https://img.shields.io/github/stars/sunlabuiuc/pyhealth.svg :target: https://github.com/sunlabuiuc/pyhealth/stargazers :alt: GitHub stars

.. image:: https://img.shields.io/readthedocs/pyhealth?logo=readthedocs&label=docs&version=latest :target: https://pyhealth.readthedocs.io/en/latest/ :alt: Docs

.. image:: https://github.com/sunlabuiuc/PyHealth/workflows/CI/badge.svg :target: https://github.com/sunlabuiuc/PyHealth/actions :alt: CI status

.. image:: https://img.shields.io/badge/Discord-Join-5865F2?logo=discord&logoColor=white :target: https://discord.gg/mpb835EHaX :alt: Discord

.. image:: https://img.shields.io/badge/Tutorials-Google%20Colab-orange?logo=googlecolab :target: https://pyhealth.readthedocs.io/en/latest/tutorials.html :alt: Tutorials

.. image:: https://img.shields.io/badge/YouTube-16%20Videos-red?logo=youtube :target: https://www.youtube.com/playlist?list=PLR3CNIF8DDHJUl8RLhyOVpX_kT4bxulEV :alt: YouTube

Documentation <https://pyhealth.readthedocs.io/en/latest/>_ · Tutorials <https://pyhealth.readthedocs.io/en/latest/tutorials.html>_ · Paper (ICML 2026) <https://openreview.net/forum?id=gMLVFN9hl8>_ · Discord <https://discord.gg/mpb835EHaX>_ · Mailing list <https://docs.google.com/forms/d/e/1FAIpQLSfpJB5tdkI7BccTCReoszV9cyyX2rF99SgznzwlOepi5v-xLw/viewform?usp=header>_

.. image:: figure/poster.png :width: 810

Why PyHealth

• Healthcare-first data layer. Loaders for MIMIC-III, MIMIC-IV, eICU, OMOP-CDM, EHRShot, sleep and EEG datasets, chest X-rays and clinical notes. Includes medical code lookup and mapping across ICD, CCS, ATC, NDC and RxNorm.
• Ready-made clinical tasks. Mortality, readmission, length of stay, drug recommendation, medical coding, sleep staging and more. A custom task is a plain Python function.
• 33+ models, one interface. RNN, Transformer and Hugging Face models through RETAIN, SafeDrug, GAMENet, StageNet, ConCare and EHRMamba. Swap a model by changing one line.
• Reproducible evaluation. Patient-level splits, a trainer that keeps the best checkpoint, calibration metrics and bootstrap confidence intervals.
• Fast on large datasets. Polars and Dask processing with an on-disk cache, about 3x faster than pandas.

Quickstart

.. code-block:: sh

pip install pyhealth

This example trains and evaluates an in-hospital mortality model on synthetic MIMIC-III data. You don't need credentials to run it.

.. code-block:: python

from pyhealth.datasets import MIMIC3Dataset, get_dataloader, split_by_patient
from pyhealth.models import RNN
from pyhealth.tasks import MortalityPredictionMIMIC3
from pyhealth.trainer import Trainer

if __name__ == "__main__":  # needed in scripts; set_task uses worker processes
    dataset = MIMIC3Dataset(
        root="https://storage.googleapis.com/pyhealth/Synthetic_MIMIC-III/",
        tables=["DIAGNOSES_ICD", "PROCEDURES_ICD", "PRESCRIPTIONS"],
    )
    samples = dataset.set_task(MortalityPredictionMIMIC3())
    train, val, test = split_by_patient(samples, [0.8, 0.1, 0.1])

    model = RNN(dataset=samples)
    trainer = Trainer(model=model, metrics=["roc_auc", "pr_auc"])
    trainer.train(
        train_dataloader=get_dataloader(train, batch_size=32, shuffle=True),
        val_dataloader=get_dataloader(val, batch_size=32),
        epochs=3,
    )
    print(trainer.evaluate(get_dataloader(test, batch_size=32)))

Pointing root at your own MIMIC copy uses real data. You can also swap in MIMIC4Dataset, another task from pyhealth.tasks or another model from pyhealth.models, and the rest of the code stays the same. The tutorials <https://pyhealth.readthedocs.io/en/latest/tutorials.html>_ walk through each stage.

If PyHealth helps your work, please star the repo and cite the paper <#citing-pyhealth>_. Stars and citations help us keep the project funded and maintained.

1. Installation :rocket:

---

Python version: PyHealth 2.x requires Python 3.12 or 3.13 (>=3.12,<3.14).

Install the latest release from PyPI. This also installs PyTorch and the other deep learning dependencies:

.. code-block:: sh

pip install pyhealth

For the latest development version, or to contribute, install from source:

.. code-block:: sh

git clone https://github.com/sunlabuiuc/PyHealth.git
cd PyHealth
pip install -e .

Legacy version: PyHealth 1.1.6 supports Python 3.9+ but lacks the 2.x features:

.. code-block:: sh

pip install pyhealth==1.1.6

News: We keep adding influential papers and benchmarks to PyHealth. See the planned list <https://docs.google.com/spreadsheets/d/1PNMgDe-llOm1SM5ZyGLkmPysjC4wwaVblPLAHLxejTw/edit#gid=159213380>_. You're welcome to pick one and send a PR, or to suggest a paper.

2. Introduction :book:

---

pyhealth provides these functionalities (we are still enriching some modules):

.. image:: figure/overview.png :width: 770

You can use the following functions independently:

This is the beginning of the README. Read the rest at the original. Read the full README ↗