Evidence-appraisal glossary
Inconsistency (Certainty of Evidence)
Inconsistency is unexplained variation in results across the studies pooled for an outcome. When effects point in different directions or differ widely in size without a clear reason, certainty in the combined estimate is rated down.
Also called: unexplained heterogeneity.
Some spread across studies is expected, but when the estimates scatter widely and subgroup or sensitivity analyses cannot explain why, a single pooled number becomes hard to trust. In the GRADE framework this domain draws on signals such as non-overlapping confidence intervals, a high I-squared, and a significant heterogeneity test, interpreted together rather than mechanically. It concerns unexplained heterogeneity specifically: variation that is understood and accounted for does not lower certainty in the same way.
Read the full Reading the Evidence blog.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.