The plausible range around a study's result.
A confidence interval is the range of plausible values for a true effect, calculated from the study's data. A 95% CI means that if the study were repeated many times using the same methods, 95% of the intervals calculated would contain the true effect. It turns a single point estimate into a range that reflects how precisely that effect has been measured.
When appraising a study, the CI shows how much confidence to place in the reported effect beyond the p-value alone. A narrow CI suggests a precise estimate, usually from a larger sample or lower variability, while a wide CI signals imprecision that may stem from a small sample, high variability in responses, or both. This guards against overinterpreting a single point estimate as more certain or generalisable than the data actually support.
CIs are reported alongside effect estimates such as mean differences, odds ratios, or risk ratios, usually as '[lower, upper]' next to the point estimate. When appraising, check whether the interval is narrow enough to be clinically informative, and whether it crosses the line of no effect (zero for a mean difference, one for a ratio) — if it does, the result is not statistically clear at that alpha level. Systematic reviews and forest plots display CIs visually as horizontal lines through each point estimate, making precision and overlap between studies easy to compare.
A CI is commonly misread as 'a 95% probability the true value lies in this range' — strictly, it describes the long-run behaviour of the method across repeated sampling, not a probability statement about this one interval. A statistically significant result with a very wide CI may still be clinically uninformative, and conversely, a CI that excludes no effect does not guarantee the effect is large enough to matter to patients. CI width and p-values are related but not interchangeable — always look at the interval itself rather than relying on whether it crosses the null.
This guide was auto-drafted and is pending editorial review.