Home > FAQ > How do I interpret an absolute risk reduction alongside its confidence interval?

How do I interpret an absolute risk reduction alongside its confidence interval?

November 2, 2026
absolute risk reductionconfidence intervalsbaseline riskevidence interpretationeffect measures

Read the absolute risk reduction against the baseline risk, then use the confidence interval endpoints to state the smallest and largest plausible benefit. The Cochrane Handbook's guidance on interpreting results treats confidence intervals as the range of effects compatible with the data, so the interval carries the uncertainty that the point estimate alone hides.

Anchor the reduction to baseline risk

Absolute risk reduction (ARR) is the difference in event risk between groups, such as 20% in the control group versus 15% in the treated group, an ARR of 5 percentage points. That number only means something next to the baseline risk it came from: the same 5-point reduction is a large proportional change on a 10% baseline and a modest one on a 50% baseline. Report the baseline risk and the ARR together.

Read the interval endpoints as the plausible range

  1. Identify the lower and upper confidence limits of the ARR and confirm whether the interval crosses zero. An interval spanning zero includes the possibility of no benefit, and the width shows how precisely the data locate the effect.
  2. Translate each endpoint into the same units as the point estimate, for example an ARR of 5 percentage points with a 95% CI of 1 to 9 points. The lower limit is the smallest plausible benefit and the upper limit the largest.
  3. Compare the lower limit with the baseline risk to judge whether even the smallest plausible benefit would matter in the population you are considering, and note whether the trial population matches that group.

State the range in plain language

Write one sentence naming the baseline risk, the point estimate, and both endpoints, such as: "In a group with a 20% event risk, treatment reduced the risk by about 5 percentage points, with a plausible range from 1 to 9 points." Keep the statement descriptive. The Cochrane Handbook's interpretation chapter places applicability judgments and recommendations in a separate step that weighs certainty, indirectness, and the decision context.

Watch for common misreads

A relative measure such as a risk ratio can look large while the absolute reduction stays small, so pair the two. Pooled estimates from meta-analysis carry additional uncertainty from heterogeneity and missing data, which the Cochrane Handbook's meta-analysis chapter addresses through sensitivity analyses and careful reading of forest plots.

←
PreviousHow do I interpret a prediction interval alongside a pooled estimate in a meta-analysis?
NextHow do I interpret an odds ratio with a wide confidence interval in a winter-health study?
→