pyhealth

Build clinical deep-learning pipelines for electronic health records and physiological signals.

2|Updated Jun 4, 2026

Search Tags:#deep learning #healthcare #clinical #ehr #mimic #pyhealth

Ready to run

npx skills add https://github.com/Lord1Egypt/scientific-agent-toolkit --skill pyhealth-lord1egypt

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: pyhealth
Source: https://github.com/Lord1Egypt/scientific-agent-toolkit/tree/main/scientific-skills/pyhealth
Command: npx skills add https://github.com/Lord1Egypt/scientific-agent-toolkit --skill pyhealth-lord1egypt

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyhealth, torch, and includes references (resource) and assets (resource) components.

What problem does it solve?

This skill streamlines the complex, multi-stage process of building clinical machine learning pipelines, ensuring that researchers can move from raw electronic health records to trained models without writing repetitive boilerplate code.

Core Features & Use Cases

Unified Pipeline: Standardizes the Dataset, Task, Model, Trainer, and Metrics workflow for clinical data.
Medical Code Utilities: Provides built-in support for cross-mapping and looking up medical coding systems like ICD-9/10, ATC, NDC, and RxNorm.
Use Case: A researcher can use this skill to quickly build a mortality prediction model on MIMIC-III data by defining the dataset, selecting a Transformer model, and using the built-in trainer to handle training and evaluation.

Quick Start

Use the pyhealth skill to build a mortality prediction pipeline using the MIMIC-III dataset and a Transformer model.

Frequently Asked Questions about pyhealth

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build a clinical deep learning pipeline for MIMIC-III data?▼

To build a clinical deep learning pipeline for MIMIC-III, you can use this skill to standardize the dataset loading, model instantiation, and training workflow. It provides modular components to process EHR data and train models like Transformers for mortality prediction.

Can I use PyTorch to develop drug recommendation models from electronic health records?▼

Yes, you can use PyTorch to develop drug recommendation models from electronic health records. This skill integrates PyTorch-based architectures and supports cross-mapping medical coding systems like ATC, NDC, and RxNorm for drug recommendation tasks.

What is the best way to map medical codes like ICD-9 to ICD-10 for EHR analysis?▼

The best way to map medical codes like ICD-9 to ICD-10 for EHR analysis is using built-in medical code utilities. This skill provides native support for cross-mapping and looking up ICD-9/10, ATC, NDC, and RxNorm coding systems.

Does this skill support mortality prediction across both MIMIC-IV and eICU datasets?▼

Yes, this skill supports mortality prediction across both MIMIC-IV and eICU datasets. It standardizes the dataset, task, model, and trainer workflow, allowing researchers to train and evaluate clinical models across diverse EHR and physiological signal datasets.

Why should I use a unified pipeline for clinical machine learning instead of custom scripts?▼

You should use a unified pipeline for clinical machine learning to avoid writing repetitive boilerplate code. This skill streamlines the complex multi-stage process, allowing researchers to move from raw EHR data to trained models using standardized data loading, training, and clinical metric evaluation.