The confidence interval (CI) gives a range of values which are likely to cover the true but unknown value.
- CIs provide more useful information than the p-value alone
- CI around a result obtained from a study sample indicates the range of values within which there is a specific level of certainty (usually 95%) that the true population value for that result lies
- if a CI around the difference in mean effects of two interventions contains the value zero, then one cannot rule out the possibility that there is no difference in effect between the interventions and therefore the differencebetween the interventions would not be statistically significant. If however, if the CI excludes the value zero, one can be reasonably (95%) certain that there is a difference between the interventions and therefore the difference would be statistically significant
- if a CI
is around a statistic that is a ratio (e.g. relative risk, odds ratio, hazard
- if the CI does not contain the value 1.0, this would indicate that a statistically significant difference exists
- CIs can also
be constructed around the measure of effect of each separate intervention
- if the CIs around each measure of effect do not overlap, this would indicate that the effects of each separate intervention are significantly different. If the CIs do overlap then it is less certain whether there is or is not a significant difference between the effects of each intervention
- width of
the CI (i.e. the range of values of the CI) also provides useful information
- a CI that is tight around the point estimate indicates that the study has sufficient power to be relatively precise
- a CI that is very wide indicates that the study could be underpowered and the point estimate imprecise
- MeReC Briefing (2005);30:1-7.
- Best Pract Res Clin Obstet Gynaecol. 2005;19(1):15-26.
- Wiebe S. The principles of evidence-based medicine.Cephalalgia. 2000;20 Suppl 2:10-3.