SCI Literature Structured Reading

Problem

Clinical guidelines often run to 50 or more pages. An abstract captures conclusions but hides recommendation strength, evidence grade, applicability conditions, and differences from prior or similar guidelines. This skill targets that kind of long, normative document and turns the original PDF into a structured, evidence-based reading.

How It Works

• Input validation: It expects the original guideline PDF; titles, abstracts, or links alone stop the workflow.
• Segmented full-text reading: It follows the table of contents to read methods, recommendations, and supporting evidence, reducing truncation and missed items.
• Structured extraction: It extracts recommendation number, strength, evidence quality, and applicable population, then flags high-risk, workflow-changing, and immediately actionable items.
• Comparative analysis: It searches earlier versions and related authoritative guidelines, then marks agreements, conflicts, additions, and deletions.
• Personalized reporting: It combines the user's specialty to generate departmental implementation checks, resource needs, quality metrics, and evidence gaps, and can push an IMA note.

Boundaries

It suits clinical guidelines, expert consensuses, position statements, technical norms, and related evidence documents. It is not a substitute for summaries without full source. If comparison guidelines cannot be fully retrieved, the limitations are stated explicitly, and the output does not replace clinical judgment.