alterlab-pyhealth

What problem does it solve?

PyHealth provides an integrated toolkit for developing, testing, and deploying machine learning models on clinical data, addressing the need for standardized pipelines in healthcare AI research.

Core Features & Use Cases

• Comprehensive healthcare AI toolkit with datasets, models, preprocessing, and evaluation workflows.
• Supports EHR data, medical coding, physiological signals, and imaging for end-to-end clinical prediction tasks.
• Use cases include mortality prediction, readmission risk, LOS estimation, drug recommendation, and code translation across coding systems.

Quick Start

Load a MIMIC-4 dataset, apply the mortality prediction task, initialize a Transformer model, and train/evaluate end-to-end.