Open Skills Logo
Open Skills

pyhealth

Build clinical/healthcare deep-learning pipelines with PyHealth — loading EHR/signal/imaging datasets (MIMIC-III/IV, eICU, OMOP, SleepEDF, ChestXray14, EHRShot), defining tasks (mortality, readmission, length-of-stay, drug recommendation, sleep staging, ICD coding, EEG events), instantiating models (Transformer, RETAIN, GAMENet, SafeDrug, MICRON, StageNet, AdaCare, CNN/RNN/MLP), training with the PyHealth Trainer, computing clinical metrics, and using medical code utilities (ICD/ATC/NDC/RxNorm lookup and cross-mapping). Use this skill whenever the user mentions PyHealth, MIMIC, eICU, OMOP, EHR modeling, clinical prediction, drug recommendation, sleep staging, medical code mapping, ICD/ATC codes, or any healthcare ML pipeline that fits the dataset → task → model → trainer → metrics pattern, even if "PyHealth" isn't named explicitly.

Author

@K-Dense-AI

Category

Development Tools

Install

Hot:17

Download and extract to your skills directory

Copy command and send to AI Agent for auto-install:

Download and install this skill https://openskills.cc/api/download?slug=k-dense-ai-skills-pyhealth&locale=en&source=copy

PyHealth - Clinical Deep Learning Toolkit

Skills Overview


PyHealth is a Python toolkit specifically designed for clinical and medical deep learning. It provides a unified 5-stage pipeline (Dataset → Task → Model → Trainer → Metrics) to handle electronic health records, physiological signals, and medical imaging data.

Suitable Scenarios


  • MIMIC Data Analysis: Works with standard medical datasets such as MIMIC-III/IV, eICU, and OMOP-CDM to perform patient-level predictive modeling and clinical analysis.

  • Clinical Predictive Modeling: Builds deep learning models for clinical tasks including mortality prediction, readmission prediction, length-of-stay prediction, drug recommendation, sleep staging, ICD coding, and more.

  • Medical Code Processing: Provides lookup and cross-mapping for medical codes such as ICD-9-CM, ICD-10-CM, ATC, NDC, RxNorm, and CCS.

    • Core Features


  • Unified 5-Stage Pipeline: Offers a modular workflow from Dataset → Task → Model → Trainer → Metrics. Each stage can be independently replaced, with stable interfaces—making it easier to build and maintain clinical deep learning pipelines.

  • Rich Dataset Support: Natively supports medical datasets including MIMIC-III/IV, eICU, OMOP-CDM, EHRShot, SleepEDF, SHHS, ISRUC, COVID19-CXR, ChestX-ray14, etc., simplifying data loading and preprocessing.

  • Diverse Models and Tasks: Includes model architectures such as Transformer, RETAIN, GAMENet, SafeDrug, MICRON, StageNet, AdaCare, and more. Supports clinical prediction tasks such as binary classification, multiclass classification, and multilabel prediction, and provides evaluation metrics tailored for clinical use.

    • FAQ

    How does PyHealth handle privacy and access issues for MIMIC data?


    PyHealth provides a synthetic MIMIC-III dataset for demonstrations and learning, without requiring access credentials. For real MIMIC data, users must apply for access through PhysioNet, and then provide the local CSV data path to PyHealth. The toolkit also supports caching to avoid repeated parsing.

    How is PyHealth different from other medical AI frameworks?


    PyHealth focuses on the clinical deep learning 5-stage pipeline and provides stable interfaces from datasets to tasks. Compared with general-purpose frameworks, it includes built-in medical dataset loaders, clinical task definitions, and medical code utilities, reducing redundant work in medical data preprocessing. Unlike FHIR tools, PyHealth is oriented toward modeling rather than data interoperability.

    Is PyHealth suitable for developers without a medical background?


    PyHealth is more suitable for developers who understand medical data or collaborate with medical experts. Although the toolkit simplifies technical implementation, the correctness of clinical prediction tasks depends on understanding the medical domain (e.g., meanings of ICD codes, patient grouping logic, and choice of clinical metrics). It’s recommended to start with the synthetic datasets and the examples in the documentation.

      PyHealth - Clinical Deep Learning Toolkit | EHR Modeling and Medical Prediction - Open Skills