1
How to read evidence
First identify the question and the type of source. Then check the study design, population, comparison, outcome, risk of bias, uncertainty and whether the conclusion agrees with the wider evidence.
Give more weight to evidence that is well designed, reproducible and directly relevant. Do not treat a headline, abstract, report count or isolated study as the whole evidence base.
Read the full guide →
2
What counts as a claim
A claim is a statement that can be checked against evidence. It might say that something causes an outcome, changes a risk, is effective or ineffective, or that a report or study proves a conclusion.
Before judging it, rewrite the claim precisely: which vaccine, which population, which outcome, what time period and what comparison? Broad wording can hide several different questions.
See claims in context →
3
How safety signals work
A safety signal is information suggesting a possible new or changed risk that should be investigated. Signals can come from spontaneous reports, studies, health records or other surveillance.
A signal is not the same as a confirmed side effect. Investigators assess timing, patterns, exposure, background rates, alternative explanations and other evidence before deciding whether causation is supported.
See how monitoring works →
4
How scientific consensus forms
Consensus is the broad conclusion that develops when multiple relevant lines of evidence are repeatedly assessed by researchers, reviewers, regulators and professional or public-health bodies.
It does not require every expert to agree, and it is not a vote that replaces evidence. A consensus is stronger when findings are consistent across high-quality studies and reviews, and it can change when important new evidence changes the balance.
Learn how studies are assessed →
Three distinctions used throughout this site
After does not automatically mean because of. Timing matters, but it is not sufficient on its own to establish causation.
A signal is not a verdict. It is a reason to investigate.
Uncertainty is not ignorance. Evidence can support a conclusion with a stated degree of confidence while leaving some questions unresolved.
Reading rule: match the strength of your conclusion to the strength, relevance and consistency of the evidence.