pubmed_skill

Overview

This skill helps you search and evaluate biomedical literature on PubMed using well-constructed queries, targeted filters, and critical appraisal rules. It emphasizes using MeSH, Boolean logic, study hierarchy, and practical search strategies to find high-quality evidence quickly. The goal is to turn literature search time into actionable clinical or research insights.

How this skill works

The skill guides construction of precise queries (MeSH terms, phrase searching, field tags, truncation) and recommends essential filters (study type, date, species, free full text). It inspects retrieved records for study design, sample size, outcomes, CI/p-values, funding, and methodological red flags. It also suggests follow-up steps: citation tracking, Related Articles, saving searches, and exporting references.

When to use it

• When you need an evidence-based answer for clinical decisions (treatment, diagnosis, prognosis).
• When conducting a literature review or building the background for a research project.
• When monitoring new publications on a focused topic via saved searches or alerts.
• When appraising the trustworthiness of a paper before citing or applying its results.
• When selecting high-quality studies for guideline development or systematic reviews.

Best practices

• Start with MeSH terms and broaden with keywords; use PICO to structure clinical questions.
• Apply study-design and date filters early to prioritize high-level evidence (systematic reviews, RCTs).
• Read methods and results, not just abstracts; check sample size, confidence intervals, and primary outcomes.
• Watch for funding sources, conflicts of interest, and journal reputation to assess bias risk.
• Save searches, use Related Articles and citation tracking, and export results to a reference manager.

Example use cases

• Find recent systematic reviews and RCTs on a new antihypertensive drug using MeSH + publication date filter.
• Assess diagnostic accuracy studies for a screening test by filtering for sensitivity/specificity designs.
• Monitor emerging COVID-19 literature with saved queries and free full-text filter for rapid access.
• Compile cohort studies for long-term prognosis after myocardial infarction using human-only and date filters.
• Verify trial results and detect potential conflicts by checking author affiliations, funding, and disclosures.