01
Source order
- Current FDA records and product labels
- Posted ClinicalTrials.gov results and completed registry records
- Randomized human outcomes and human pharmacology
- Systematic reviews, guidelines, and official regulatory bodies
- Preclinical work for mechanism and translation questions
- Promotional claims only as objects to verify, never as proof
02
Scheduled review
Approved-product briefs are checked at least every 12 months. Investigational and high-controversy briefs are checked at least every six months because status, trial results, and safety communications can change quickly.
The automated freshness check flags any brief beyond its cadence, future-dated reviews, and source links that no longer resolve.
URL reachability is only transport evidence. It does not prove that a source supports a claim. Claim records separately identify the responsible internal editor, the clinical-review boundary, and whether each cited source supports the scoped claim, defines an evidence limit, or supports a corrective verdict.
03
Current-label control
Products under active label review are keyed by application and product in a public current-label manifest. Product records and claim citations must match the reviewed manifest URL.
A newer label revision is a review trigger even when the headline claim does not change. The editor must compare indication, population, warning, precaution, and safety changes before retaining prior wording.
04
Immediate review triggers
- Approval, withdrawal, label change, safety communication, or enforcement action
- Posted result from a pivotal or claim-relevant trial
- Retraction, expression of concern, major correction, or invalid source
- New anti-doping classification
- Evidence that a molecule name or marketed product identity has changed
05
What a review must preserve
Positive evidence is acknowledged within its real scope. New uncertainty is not used to erase a supported result, and a new signal is not inflated into a settled outcome.
Material changes are written into page-level review history and, when corrective, the corrections record.
06
Long-form education depth
Substantial learning guides and dedicated molecule deep dives are maintained at 2,000 to 3,000 rendered words. The range is an editorial depth check, not proof of quality: every section must add distinct definitions, evidence interpretation, applied examples, source limits, or safety context.
Evidence briefs, individual claim records, glossary definitions, topic hubs, source indexes, and policy pages use purpose-based lengths. They remain concise when structured coverage and fast verification serve the reader better than an essay.
07
Clinical review boundary
The current edition has not been independently medically reviewed. Label interpretation, safety synthesis, and clinical-context pages are candidates for qualified physician or pharmacist review before any future expansion into decision support.