A Tale of Confusion From Overlapping Confidence Intervals

Abstract

In clinical research presentations, study results are commonly reported in the form of P values and confidence intervals as an estimate of association and treatment effect. The interpretation of confidence intervals that overlap can be confusing and difficult for the reader to draw clinically meaningful conclusions. In this brief report, we describe the basics of confidence intervals and present an example from a recently published randomized control trial to illustrate a common confusion that overlapping confidence intervals between the means of two independent groups may not necessarily reject the true significant difference of effect. It is recommended that investigators use the direct difference of means between groups for confidence interval estimation to reduce type II errors. Clinicians should interpret overlapping confidence intervals with caution and avoid the assumption that overlapping confidence intervals always implies a lack of difference of treatment effect to decide application of treatment.