medspaCy — Clinical NLP Toolkit

medspaCy — Clinical NLP Toolkit

medspaCy extends spaCy with clinical NLP components tailored for EHR text. It handles the quirks of clinical documentation — section headers, templated notes, negated findings, family history — that general NLP tools miss entirely.

Key Features

• Clinical section detector: History of Present Illness, Assessment/Plan, Medications, Social History, etc.
• ConText algorithm: negation, hypothetical, historical, family-member modifiers
• Concept extraction via custom rule sets and machine learning models
• UMLS entity linking via scispaCy
• Targets clinical note types: discharge summaries, progress notes, radiology reports
• Interoperable with standard spaCy pipelines

Quick Start

import spacy
import medspacy

nlp = medspacy.load()
doc = nlp("Patient has no history of diabetes. Mother has hypertension.")
for ent in doc.ents:
    print(ent.text, ent.label_, ent._.is_negated, ent._.is_family)

npx ai-supply add medspacy-clinical-nlp

Curated mirror of the open-source medspaCy (MIT). Get it from the source.