medkit

medkit facilitates the development, evaluation, and reproducibility of phenotyping pipelines from clinical text data in Electronic Health Records (EHRs).

Key Features:

• Pipeline Composition: medkit enables composing data processing pipelines made of reusable software components called medkit operations.
• Reusable Operations: The library provides a core framework with shareable operations and pipelines for the phenotyping community to reuse and enrich.
• Open-Source Implementation: medkit is implemented as an open-source Python library.

Scientific Applications:

• Clinical Text Phenotyping: medkit supports identifying individuals associated with specific traits or conditions from unstructured clinical text in EHRs for observational studies.
• Secondary Use of EHRs: medkit enables phenotyping pipelines that leverage clinical text data for research studies relying on secondary EHR use.

Methodology:

medkit implements a modular architecture based on composable operations for building phenotyping pipelines.