One Table, Explained: A Relative Risk and Its Interval

Read the labels before the number

The following row is entirely hypothetical. It is not a patient dataset, study result, or claim about a treatment.

Comparison	Outcome window	Relative risk	95% confidence interval
Group A versus Group B	By 30 days	0.80	0.55 to 1.16

Before interpreting it, confirm which group is the reference, what the outcome means, how follow-up was handled, and whether the estimate is adjusted.

What 0.80 means

For a risk ratio, 0.80 corresponds to an estimated risk in Group A that is 80% of the risk in Group B. It is not automatically a 20-percentage-point reduction. Absolute differences depend on the underlying risks and the context of the comparison.

The no-difference value for a risk ratio is 1. The displayed interval includes 1 and spans both lower and higher risk. Reporting only the point estimate would hide that uncertainty. Reporting only “not significant” would hide its direction and range.

What the row cannot tell you

It does not establish causality, show whether follow-up was complete, reveal how confounding was handled, or tell you whether the outcome was measured consistently. Those questions need the methods and wider results.

Do not interpret a hazard ratio or odds ratio as though it were this risk ratio. They are different measures. Also do not put a raw risk ratio into a calculator intended for normally distributed differences; ratio analyses often use a log scale.

A reusable reading routine

Identify the measure, reference, units, time window, analysis population, adjustment, and interval method. Then ask what effect sizes remain compatible with the result and which limitations affect the interpretation.

Download the interpretation worksheet. Explore differences with the Results Interpretation Playground, or discuss your own tables through results consultation.

Reference: Cochrane Handbook on relative and absolute effects.