Longevity Atlas

METHODOLOGY · VERSION 1.2

How we decide what counts as evidence.

A transparent framework for grading human evidence, claim relevance, replication, safety and uncertainty.

§ 1 — Evidence Pipeline

From source to decision in four stages

01
INDEX
  • ClinicalTrials.gov
  • PubMed
  • Primary sources
02
SCREEN
  • Human relevance
  • Study design
  • Endpoint relevance
03
GRADE
  • T–1 to T–5
  • 7-factor score
  • Uncertainty review
04
DECIDE
  • Include
  • Caveat
  • Early stage
  • Caution
WHAT THE ENGINE SEESNMN · SAMPLE RUN
SOURCE RECORDS
ClinicalTrials.gov31
PubMed RCT papers16
Meta-analyses11
EXCLUDED
Non-human42
Wrong endpoint18
Duplicate7

§ 2 — Evidence Tiers

Five levels of human evidence maturity

T–5

Extensive replicated human evidence

Multiple independent RCTs and consistent synthesis across populations

T–4

Strong human evidence

Several relevant RCTs with reasonable consistency across outcomes

T–3

Limited or mixed human evidence

Small studies, conflicting outcomes or limited replication

T–2

Early human evidence

Pilot trials or observational evidence only

T–1

Preclinical evidence

Mechanistic, animal or traditional evidence only

Evidence tier reflects maturity, not effect size or medical recommendation.

On the Evidence Coordinates map, high-volume evidence is defined operationally as ≥100 indexed PubMed RCT publications, and the confidence divider is set at 75/100. These are navigational thresholds, not efficacy claims.

§ 3 — Confidence Score

Seven factors. Transparent weights.

Human RCT quality
35%
Independent replication
20%
Meta-analysis consistency
15%
Sample size & duration
10%
Endpoint relevance
10%
Safety evidence
5%
Publication bias
5%
TECHNICAL NOTE
Confidence score is bounded by relevance, quality and uncertainty checks.
Score = Σ(weight_i × factor_i)
NMN · SCORE 79 · EXAMPLE
RCT quality
28 / 35
Replication
15 / 20
Meta-analysis
11 / 15
Sample & duration
8 / 10
Endpoint relevance
8 / 10
Safety
5 / 5
Publication bias
4 / 5
TOTAL 79 / 100

§ 4 — Source Hierarchy

Different evidence answers different questions

Primary Sources
Trial registries and original studies
ClinicalTrials.gov registrations, peer-reviewed RCT papers, protocol documents
Peer-Reviewed Synthesis
Systematic reviews and meta-analyses
Pooled effect estimates across populations and study designs
Contextual Evidence
Mechanistic, historical and traditional literature
Pathway research, ethnobotany, pharmacokinetic data — informs but does not prove efficacy
Commercial Context
Claims, formulation, supplier and market readiness
Not evidence of efficacy — relevant to decision-making for formulators and buyers

Not all evidence proves efficacy — different sources answer different questions in the decision process.

Boundary

We are an evidence intelligence layer, not a medical provider.

Longevity Atlas does not diagnose, treat, prescribe, or replace licensed clinical care. We help brands and formulation teams ask better questions before they commit money or trust.

Submit an ingredient →