Last update:
01-Oct-2009

Randomized clinical trials are the most widely accepted approach for comparing the benefits of alternative treatments. In designing such studies, formal estimations of sample size are required in order to ensure that the probability of missing an important difference is small, to reduce unnecessary cost and to reduce wastage. An appropriate sample size generally depends on five study design parameters: the level of statistical significance considered appropriate (alpha level), the statistical power of the test, the minimum expected difference (effect size) between the compared groups that would be regarded as clinically significant, the estimated measurement variability and whether a oneor two-tailed statistical analysis is planned. The ability of a test to identify correctly that there is a difference between the compared groups in a trial is called the power of the test, that is, the ability of a test to reject the null hypothesis when it should be rejected. In randomized clinical trials, the statistical power is customarily set to a number greater than or equal to 0.8. If the study finds a difference between treatments, when in actuality there is no difference, then a type I error is present. If the study fails to find a difference between treatments when in actuality there is a difference, then a type II error is present. Sample size is best considered early in the planning of a study, when modifications in study design can still be made.

Key words: Effect size, Randomized clinical trials, Statistical power, Type I error, Type II error.