Where a finding sits in the hierarchy
In vitro results describe behaviour in cells. Animal models add a whole organism but not human physiology. Early human trials measure safety and pharmacology. Randomised controlled trials test whether an intervention changes outcomes that matter.
Systematic reviews and meta-analyses summarise the trial base. A single positive rodent study and a large randomised trial are not comparable forms of knowledge, even when the headline sounds similar.
- Phase 1: safety, tolerability, pharmacokinetics — usually small.
- Phase 2: dose finding and early efficacy signals.
- Phase 3: confirmatory efficacy and safety in larger populations.
- Phase 4: post-marketing surveillance after approval.
Fig. 02 / Molecular pathwaySurrogate endpoints and effect size
Many peptide trials measure surrogates — a biomarker, a scan, a weight change — rather than long-term clinical events. Surrogates are useful, but they can move without the outcome moving.
Always look for the absolute effect, the comparator, the duration, and the dropout rate. A large relative improvement in a short, small trial with high attrition is a weak signal.
Language that signals uncertainty
Responsible sources write 'associated with', 'in animal models', 'in a phase 2 trial', or 'evidence is limited'. Promotional sources write 'clinically proven', 'heals', or 'optimises'. The vocabulary is diagnostic.
Fig. 03 / Evidence landscapeEditorial disclaimer
This site is educational and does not provide medical advice, dosing protocols, or treatment recommendations. Regulatory status and evidence strength differ sharply between compounds. A “research use only” label does not indicate that a product is appropriate for human use. Decisions about medicines belong with a qualified clinician.