Independent · non-commercial · publishes on a quarterly cycle|Current cycle 2026 Q3
Compound Evidence InstituteEvidence synthesis · established 2023Graded assessments of compounds, trials, methods and supply
Document set current to 30 July 2026
Finding aid

Certainty index

Every outcome the Institute grades, organised by certainty level rather than by compound.

§1The four levels

A certainty rating describes confidence in an effect estimate for a stated population, comparator and outcome. It is not a recommendation, not an endorsement, and does not transfer to a different population, comparator or outcome.

Table 1. The four certainty levels and their meanings.

LevelGraded outcomesMeaning
High62The Institute is confident that the true effect lies close to the estimate. Further research is very unlikely to change confidence in the estimate.
Moderate114The true effect is likely to be close to the estimate, but there is a possibility that it is substantially different. Further research is likely to have an important impact on confidence in the estimate.
Low52Confidence in the effect estimate is limited. The true effect may be substantially different from the estimate.
Very low53The Institute has very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate. In this series a very low rating most often reflects an absence of controlled human evidence rather than conflicting evidence.
Distribution of certainty ratingsShare of assessed outcomes at each certainty level.281graded
HighModerateLowVery low
Figure 1. Distribution of certainty across everything the Institute grades. Very low ratings are listed with equal prominence throughout this index, on the reasoning that a reader looking for what is not known is as legitimate as one looking for what is.

§2How a rating is arrived at

Randomised evidence starts at high certainty. Each serious concern in one of five domains reduces the rating by one level and each very serious concern by two. Observational evidence starts at low and may be upgraded, rarely, where the effect is large, a dose–response gradient is present, or a plausible confounder would have reduced the observed effect.

Table 2. The five domains.

DomainWhat it assesses
Risk of biasLimitations in the design and conduct of the contributing studies.
InconsistencyUnexplained heterogeneity of results across contributing studies.
IndirectnessDifferences between the population, intervention, comparator or outcome studied and those of the assessment question.
ImprecisionWidth of the confidence interval relative to the decision threshold, and the number of events.
Publication biasRisk that results were selectively reported or that unpublished studies exist.
Nothing published by the Institute is medical advice, a diagnosis, a prescription, a treatment recommendation or a purchasing recommendation. Compounds supplied for research use are not approved for human or veterinary use in any jurisdiction, and a favourable analytical assessment of a supplier is not a statement that any product is safe or effective. The Institute publishes certainty ratings and never recommendations. No telephone number, messaging handle or ordering channel appears anywhere on this site.