Pubmed Database

API Rate Limits: 3 req/s without key, 10 req/s with key. Register at https://www.ncbi.nlm.nih.gov/account/

Quick Start

import requests, time

BASE_URL = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/"
API_KEY = "YOUR_API_KEY"  # Optional but recommended

def pubmed_request(endpoint, params):
    params.setdefault("api_key", API_KEY)
    response = requests.get(f"{BASE_URL}{endpoint}", params=params)
    response.raise_for_status()
    time.sleep(0.1 if API_KEY != "YOUR_API_KEY" else 0.34)
    return response

# Search → Fetch workflow
search_resp = pubmed_request("esearch.fcgi", {
    "db": "pubmed", "term": "CRISPR[tiab] AND 2024[dp]",
    "retmax": 5, "retmode": "json"
})
pmids = search_resp.json()["esearchresult"]["idlist"]
print(f"Found: {pmids}")

fetch_resp = pubmed_request("efetch.fcgi", {
    "db": "pubmed", "id": ",".join(pmids),
    "rettype": "abstract", "retmode": "text"
})
print(fetch_resp.text[:500])

Core API Summary

For complete code examples across all 6 endpoints and search patterns, see references/api_reference.md.

For systematic review workflow and literature monitoring pipeline, see references/workflows.md.

For complete field tag reference and query templates, see references/search_syntax.md and references/common_queries.md.

Key Concepts

E-utilities Endpoint Summary

History Server Pattern

For result sets >500 articles, use the history server to avoid URL length limits:

1. ESearch with usehistory=y → returns WebEnv + QueryKey
2. EFetch in batches using WebEnv + QueryKey + retstart/retmax
3. EPost to upload additional PMIDs to the same WebEnv

Automatic Term Mapping (ATM)

When no field tag is specified, PubMed maps terms through the MeSH Translation Table → Journals Table → Author Index → Full Text. Bypass ATM with explicit field tags (e.g., term[tiab]) or quoted phrases.

Common MeSH Subheadings

Key Parameters

Best Practices

1. Always use an API key — register at NCBI for 10 req/sec instead of 3
2. Use history server for >500 results — avoids URL length limits
3. Include rate limiting — time.sleep(0.1) with API key, time.sleep(0.34) without
4. Cache results locally — PubMed data changes slowly
5. Combine MeSH + free text — (diabetes mellitus[mh] OR diabetes[tiab]) for comprehensive coverage
6. Document search strategies — record exact queries, dates, result counts for systematic reviews
7. Parse XML for structured data — text is human-readable but XML preserves field structure

Troubleshooting

Bundled Resources

• references/search_syntax.md — Complete field tag reference, Boolean/wildcard/proximity syntax, automatic term mapping rules, all filter types
• references/common_queries.md — Ready-to-use query templates organized by domain (~40 example patterns)
• references/api_reference.md — Complete ESearch/EFetch/ESummary/ELink/ECitMatch code examples
• references/workflows.md — Systematic review workflow and literature monitoring pipeline

References

• PubMed Help: https://pubmed.ncbi.nlm.nih.gov/help/
• E-utilities Documentation: https://www.ncbi.nlm.nih.gov/books/NBK25501/
• NCBI API Key: https://www.ncbi.nlm.nih.gov/account/

Related Skills

• openalex-database — broader academic literature beyond biomedical
• literature-review — systematic review methodology and PRISMA framework