Scientific Interpretation
Evidence Hierarchy
Evidence hierarchies are useful starting frameworks for matching a study design to the question it can answer. They are not automatic rankings of truth: review quality, risk of bias, directness, precision, consistency, and applicability can raise or lower confidence within any design category.
Evidence Snapshot
Evidence: StrongerSystematic reviews and meta-analyses
- Human evidence
- Can provide high-confidence summaries when the question is well defined, eligible studies are appropriate, methods are transparent, and the underlying evidence is sufficiently trustworthy and comparable.
- Research signal
- Can place clinical and mechanistic findings in broader context, but pooled estimates do not repair bias or indirectness in the included studies.
- Safety profile
- Confidence still depends on the included populations, follow-up, adverse-event reporting, heterogeneity, publication bias, and review methods.
Evidence Snapshot
Evidence: StrongerRandomized controlled trials
- Human evidence
- Can provide direct causal evidence for the population, intervention, comparator, and outcomes actually studied when randomization and trial conduct are credible.
- Research signal
- Can test whether a mechanistically plausible intervention produces measurable human outcomes, but the mechanism itself is not proven by a positive clinical result.
- Safety profile
- Sample-size adequacy, duration, missing data, participant diversity, outcome selection, adherence, and adverse-event capture all affect confidence.
Evidence Snapshot
Evidence: ModerateObservational human evidence
- Human evidence
- Can identify associations, longer-term patterns, uncommon exposures, and real-world signals that trials may not capture.
- Research signal
- May help connect biological hypotheses with human patterns, while residual confounding and selection bias can limit causal interpretation.
- Safety profile
- Especially useful for generating and investigating safety signals, but associations still require careful control of confounding and measurement error.
Evidence Snapshot
Evidence: TheoreticalMechanistic, animal, and theoretical evidence
- Human evidence
- Mechanistic or preclinical evidence alone does not establish a clinically meaningful human benefit, an effective oral dose, or long-term safety.
- Research signal
- Useful for understanding receptors, signaling pathways, metabolism, pharmacology, and biological plausibility.
- Safety profile
- A plausible pathway does not guarantee effectiveness or safety in humans.
Use the hierarchy with the research question
Different designs can be strongest for different questions. Randomized trials are often preferred for causal treatment effects, while observational evidence may be essential for uncommon harms or long-term real-world patterns. A well-conducted study that directly answers the question can be more informative than a nominally “higher” design that is biased, indirect, or built from poor underlying studies.
Referenced Research
Research references
Source ledger
References
1 source
Learning context
How this concept connects to supplement decisions
Educational overview of evidence hierarchy, clinical trials, mechanistic evidence, ethnobotanical context, and scientific interpretation systems. Learning pages explain the reasoning layer behind the herb and compound library. They are designed to make mechanisms, evidence quality, safety tradeoffs, and product claims easier to interpret.
Use Evidence Hierarchy to build better questions before choosing a supplement: what outcome is being targeted, what mechanism is claimed, what human evidence exists, what dose was studied, and what risks could change the answer for a specific person?
Mechanistic plausibility is useful, but it should be weighed against trial design, safety history, product quality, and the possibility that a simpler intervention may be more appropriate.