Home > FAQ > How do I read a risk ratio and its confidence interval in a winter-health meta-analysis?

How do I read a risk ratio and its confidence interval in a winter-health meta-analysis?

November 2, 2026
risk ratio interpretationconfidence interval meaningmeta-analysis forest plotpooled effect estimatewinter health evidence

Read the pooled risk ratio as the average effect across included studies, compare its confidence interval with the null value of 1, and treat the interval as the range of effects compatible with the data. The Cochrane Handbook covers effect measures, forest plots, and interpretation of pooled results in Chapter 10 and confidence intervals and relative effects in Chapter 15.

Locate the three numbers

In the forest plot or summary table, find the pooled estimate (often a diamond), its 95% confidence interval, and the null line. For a risk ratio, the null value is 1: a ratio of 1 means the risk is the same in both groups. A pooled risk ratio of 0.80 means the outcome was less frequent in the intervention group across the pooled studies.

Translate the interval

  1. Check whether the confidence interval crosses 1. If the entire interval lies below 1, the pooled data are compatible with lower risk in the intervention group; if it lies above 1, with higher risk; if it spans 1, the direction of the pooled effect is unresolved.
  2. Read the interval endpoints as the range of effects compatible with the data. An interval from 0.70 to 0.92 describes a smaller reduction than one from 0.70 to 0.99, even though both exclude 1.
  3. Compare the magnitude with the question's context. A ratio near 1 represents a small relative difference; a ratio farther from 1 represents a larger one. Report the ratio and interval together so readers see both size and uncertainty.

Write the plain-language sentence

State direction, magnitude, and uncertainty in one sentence, for example: "Across the pooled studies, the intervention was associated with a lower risk of the outcome (risk ratio 0.80, 95% CI 0.70 to 0.92)." If the interval crosses 1, say the data are compatible with both lower and higher risk and that the direction remains uncertain. Keep the sentence about the pooled association; clinical recommendations belong in a separate discussion of applicability and certainty.

←
PreviousHow do I read a forest plot to see which winter-health trials drive the pooled effect?
NextRelative or absolute effects: which should a clinical summary report?
→