Editorial policy

Sourcing

• Claims are sourced to primary records: the published study, the trial registry entry, the chemical database record, or the regulator's own document. Where a secondary source is used it is labelled as such in the reference list.
• Bibliographic data is pulled from PubMed programmatically, never retyped. An unknown citation fails the build.
• Regulatory facts are read from the regulator's own page or document. Vote tallies, dates and quoted concerns come from the minutes and briefing documents themselves.
• Where a primary source could not be retrieved, the site says so and names the gap rather than substituting a secondary account. The WADA listing on the regulatory page is a current example.

How findings are described

• The study is the subject of the sentence. Pages say what a study measured, not what the compound does.
• The species and model are named wherever a finding appears.
• Results are reported as the authors reported them, including null results, negative results and results that contradict popular claims.
• A mechanism is never presented as demonstrating a clinical benefit. Receptor activity is receptor activity.
• Findings for an analogue, a related compound or market product are never presented as findings for ipamorelin.

What is never published here

• Doses, schedules, routes, reconstitution instructions or administration guidance of any kind.
• Combination, stacking or cycling protocols.
• Pooled effect estimates across heterogeneous studies, or any combined efficacy score.
• Before-and-after claims, testimonials, or any content presenting individual experience as evidence.
• Product links, affiliate links, advertising or sponsored content.
• A fabricated citation, count, date or regulatory fact, under any circumstances.

Safety and adverse events

• Adverse events are reported as the source reported them, with the source named, including where two sources describe the same trial differently.
• Causality is never asserted beyond what the source asserts. Where a regulator states that a causal link is unclear, the site states that it is unclear.
• Serious adverse events and deaths are never omitted for being uncomfortable, and never presented as established consequences when the source does not support that.

Numbers

Study counts, trial counts, participant totals and database statistics are computed at build time from the underlying data. No count is typed into a page, so a page cannot drift out of step with the database behind it, and a record added or removed updates every page that reports a total.

Publication workflow

Records move through five states:

1. Draft — annotated but not yet checked.
2. Source Verified — bibliographic data confirmed against the primary record by automated check.
3. Scientific Review — a named human reviewer is assessing interpretation and grading.
4. Approved — signed off by that reviewer.
5. Published — live on the site.

Approval states live in a git-tracked file, so every change to one is attributable and reversible. A production build can be configured to refuse to render any record below Approved. Nothing is auto-approved by the pipeline: passing the automated checks reaches Source Verified and no further. The current status of every record is shown on the methodology page and on each record in the research database.

Corrections

Corrections are made to the underlying data, which rebuilds every page that draws on it. Material factual corrections are noted on the affected page. Each page carries its own last-reviewed date, and the site footer carries the dates the research corpus and regulatory records were last checked.

Independence

This site has no commercial relationship with any seller, manufacturer, compounding pharmacy or clinic. It carries no advertising and no affiliate links, and no commercial relationship influences what appears here. Where research was conducted by the company that developed the compound, that is noted in the record's limitations.