Clinical Guidelines Tools

ToolUniverse provides 8 clinical guideline tools for searching and extracting authoritative medical guidelines.

Tool List

Search Tools (6):

  • NICE_Clinical_Guidelines_Search - UK NICE official guidelines

  • WHO_Guidelines_Search - WHO international guidelines

  • PubMed_Guidelines_Search - PubMed peer-reviewed guidelines

  • EuropePMC_Guidelines_Search - Europe PMC guidelines

  • TRIP_Database_Guidelines_Search - TRIP evidence-based database

  • OpenAlex_Guidelines_Search - OpenAlex scholarly database

Full-Text Extraction (2):

  • NICE_Guideline_Full_Text - Extract complete NICE guidelines (5,000-20,000+ chars)

  • WHO_Guideline_Full_Text - Extract WHO content + PDF links (1,500+ chars)

Quick Start

from tooluniverse import ToolUniverse

tu = ToolUniverse()
tu.load_tools()

# Search guidelines
results = tu.run({
    "name": "NICE_Clinical_Guidelines_Search",
    "arguments": {
        "query": "diabetes",
        "limit": 5
    }
})

# Extract full text
full_text = tu.run({
    "name": "NICE_Guideline_Full_Text",
    "arguments": {
        "url": results[0]['url']
    }
})

Tool Comparison

Tool

Data Source

Features

NICE_Clinical_Guidelines

UK NICE official website

Official UK guidelines with summaries

WHO_Guidelines

WHO publications database

International health guidelines

PubMed_Guidelines

NCBI PubMed database

Peer-reviewed with abstracts & PMIDs

EuropePMC_Guidelines

Europe PMC database

European & international research

TRIP_Database

TRIP evidence database

Evidence-based medicine focus

OpenAlex_Guidelines

OpenAlex scholarly database

Comprehensive with citation counts

Usage Examples

Search NICE Guidelines

results = tu.run({
    "name": "NICE_Clinical_Guidelines_Search",
    "arguments": {
        "query": "type 2 diabetes",
        "limit": 5
    }
})

for guideline in results:
    print(f"{guideline['title']}")
    print(f"URL: {guideline['url']}")
    print(f"Summary: {guideline['summary'][:200]}...")

Search Multiple Sources

query = "hypertension"

# Search NICE
nice = tu.run({
    "name": "NICE_Clinical_Guidelines_Search",
    "arguments": {"query": query, "limit": 3}
})

# Search PubMed
pubmed = tu.run({
    "name": "PubMed_Guidelines_Search",
    "arguments": {"query": query, "limit": 3}
})

# Search WHO
who = tu.run({
    "name": "WHO_Guidelines_Search",
    "arguments": {"query": query, "limit": 3}
})

print(f"Found: NICE {len(nice)}, PubMed {len(pubmed)}, WHO {len(who)}")

Extract NICE Full Text

# Step 1: Search
search = tu.run({
    "name": "NICE_Clinical_Guidelines_Search",
    "arguments": {"query": "diabetes", "limit": 1}
})

# Step 2: Extract full text
full_text = tu.run({
    "name": "NICE_Guideline_Full_Text",
    "arguments": {"url": search[0]['url']}
})

print(f"Length: {full_text['full_text_length']:,} characters")
print(f"Sections: {full_text['sections_count']}")
print(f"Recommendations: {full_text['recommendations_count']}")
print(f"\nContent:\n{full_text['full_text'][:500]}...")

Extract WHO Content + PDF

# Step 1: Search
search = tu.run({
    "name": "WHO_Guidelines_Search",
    "arguments": {"query": "HIV", "limit": 1}
})

# Step 2: Extract content
content = tu.run({
    "name": "WHO_Guideline_Full_Text",
    "arguments": {"url": search[0]['url']}
})

print(f"Content length: {content['content_length']:,} characters")
print(f"Has PDF: {content['has_pdf']}")
if content['has_pdf']:
    print(f"PDF URL: {content['pdf_download_url']}")

Complete Example

See examples/clinical_guidelines_demo.py for a complete demonstration:

python examples/clinical_guidelines_demo.py

FAQ

Q: What’s the difference between search tools and full-text tools?

  • Search tools: Return multiple results with title, URL, summary (200-2,500 chars each)

  • Full-text tools: Extract complete guideline content (5,000-20,000+ chars) from a single URL

Q: When should I use full-text tools?

Use full-text tools when you need:

  • Complete guideline content for detailed analysis

  • All clinical recommendations extracted

  • Content to feed into LLMs

  • PDF download links for offline reading

Q: Which tools return the most complete content?

  • NICE search: 300-2,500 char summaries

  • NICE full-text: 5,000-20,000+ char complete guidelines

  • PubMed: 200-2,000 char abstracts

  • WHO full-text: 1,500+ chars + PDF links

Q: How to filter by year?

# OpenAlex supports year filtering
results = tu.run({
    "name": "OpenAlex_Guidelines_Search",
    "arguments": {
        "query": "cancer screening",
        "year_from": 2023,
        "limit": 10
    }
})

Return Fields

Search Tools Return:

  • title: Guideline title

  • url: Direct link to guideline

  • summary/abstract/description: Content summary (200-2,500 chars)

  • Tool-specific: pmid, doi, authors, date, cited_by_count, etc.

NICE Full-Text Returns:

  • full_text: Complete guideline (5,000-20,000+ chars)

  • full_text_length: Character count

  • sections_count: Number of sections

  • recommendations: List of recommendations

  • recommendations_count: Count

  • metadata: Publication info

  • success: Boolean status

WHO Full-Text Returns:

  • overview: Overview section

  • main_content: Main content

  • content_length: Character count

  • has_pdf: Boolean

  • pdf_download_url: PDF link (if available)

  • key_facts: List of key facts

  • metadata: Publication info

  • success: Boolean status

Summary

8 Tools Covering:

  • ✅ 6 search tools - Quick search across multiple sources

  • ✅ 2 full-text tools - Deep content extraction

  • ✅ Official sources - NICE (UK), WHO (International)

  • ✅ Academic databases - PubMed, Europe PMC, OpenAlex

  • ✅ Evidence-based - TRIP Database

  • ✅ Complete extraction - 5,000-20,000+ characters

  • ✅ PDF downloads - WHO guidelines

  • ✅ 100% test coverage - 23/23 tests passing

For More Information:

  • Complete demo: examples/clinical_guidelines_demo.py

  • Source code: src/tooluniverse/unified_guideline_tools.py

  • Tests: tests/test_guideline_tools.py