Editorial Policy

MediSum summaries should stay tied to visible source records. A public card should not invent titles, journals, dates, PMIDs, authors, study details, limitations, or confidence values.

When exact specialty supply is limited, the product should broaden to related real records or show an availability message rather than filling the page with synthetic examples.

Summaries may describe study design, population, measured outcomes, direction of findings, and stated limitations. They should not convert a paper into diagnosis, treatment guidance, or patient-specific advice.

When evidence is preliminary, observational, mixed, or limited, wording should preserve that uncertainty and avoid implying clinical efficacy beyond what the source supports.

The public demo, sample issue, and specialty research update pages should keep PMIDs and PubMed links visible so clinicians can verify provenance quickly.

Placeholder PMIDs, unsupported confidence scores, loading-only output, and synthetic article cards are not acceptable public demo states.