scientific-clinical-nlp

Extract clinical entities with negation detection and UMLS concept linking.

3|1|Updated Feb 11, 2026

Search Tags:#de-identification #ner #negation #umls #clinical-nlp #medical-linking

Ready to run

npx skills add https://github.com/nahisaho/satori --skill scientific-clinical-nlp

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: scientific-clinical-nlp
Source: https://github.com/nahisaho/satori/tree/main/src/.github/skills/scientific-clinical-nlp
Command: npx skills add https://github.com/nahisaho/satori --skill scientific-clinical-nlp

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Clinical notes and biomedical texts are unstructured and rich in terminology, making it hard to extract actionable information. This skill provides a scalable NLP pipeline to identify clinical entities, detect negation, segment sections, and map concepts to standardized vocabularies for downstream analysis.

Core Features & Use Cases

Clinical NER and concept linking using MedSpaCy and scispaCy
Negation detection (NegEx/ConText) and automatic section classification
UMLS/SNOMED-CT ICD-10 linkage for terminology bridging
De-identification (PHI redaction) to support data privacy
TU-compatible integration for broader biomedical data workflows

Quick Start

Run the clinical NLP pipeline on a sample set of clinical notes to extract entities, detect sections, and anonymize PHI.

Frequently Asked Questions about scientific-clinical-nlp

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

FAQPage Schema

How do I extract clinical entities and detect negation from unstructured electronic health records?▼

To extract clinical entities and detect negation from electronic health records, you can use a pipeline built with MedSpaCy and scispaCy. It identifies medical entities, applies NegEx or ConText for negation detection, and segments clinical note sections automatically.

Can I de-identify PHI in clinical notes before running NLP extraction?▼

Yes, you can de-identify Protected Health Information (PHI) in clinical notes for data privacy. The pipeline includes PHI redaction capabilities to anonymize patient data before or during the clinical entity extraction and terminology linking process.

How do I map extracted clinical concepts to standardized medical vocabularies like UMLS or SNOMED-CT?▼

To map extracted clinical concepts to standardized vocabularies, the pipeline uses a medical terminology linker. It bridges recognized entities to UMLS, SNOMED-CT, and ICD-10 codes, enabling accurate terminology matching for downstream biomedical analysis.

Does this clinical NLP pipeline support section segmentation for medical notes?▼

Yes, the clinical NLP pipeline supports automatic section classification for medical notes. It segments unstructured clinical text into relevant sections, improving the context accuracy of entity recognition and negation detection.

What is the best way to process biomedical literature for accurate medical entity recognition?▼

The best way to process biomedical literature for medical entity recognition is using scispaCy and MedSpaCy models. They provide specialized NLP pipelines to accurately identify clinical entities, handle negation, and link biomedical terms to standardized vocabularies.