LAUSR

To the Editor: C linical trial results that are statistically significant are routinely considered to be clinically important, and conversely, those with statistically insignificant results are interpreted as being clinically unimportant. However, statistical and clinical significance are distinct concepts. Statistical significance implies that the difference seen between groups is larger than 0 but does not convey importance. In contrast, clinical significance indicates that the difference in effectiveness between treatments is clinically meaningful (ie, is important to patients and is worth changing practice for). Unfortunately, clinical and statistical significance are often confused, and the terms are often considered synonymous. In the context of clinical trial design, clinical significance, in the form of the minimal clinically important difference (MCID), is used in the sample size calculation to ensure the proposed trial will have enough power, and; therefore, yield a statistically significant result, to detect the smallest difference that would lead to a change in patient management. The role played by the MCID, unfortunately, usually ends there. The MCID remains restricted to the description of the sample size calculation and clinical trials results are interpreted in light of their statistical significance only. Under perfect circumstances, a clinical trial that accrues it target sample size calculation should not produce statistically significant results for differences smaller than the MCID (and; therefore, not clinically meaningful). Unfortunately, sample size calculation is an imprecise process based on the pre-experiment probability of hypothetical findings meaning that clinical trials that reach their