Sample Size Determination

Home Significance tests Meta-analysis Puzzles About Contact

Clinical Trial: Non-inferiority Design

This utility can be used to calculate required sample size for a clinical trial with non-inferiority design when the outcome variable is in ratio scale.

Formula used and other details
Your guesstimate of Mean in test arm (M1) Approxiate estimate of mean in test arm based on previous studies / pilot studies etc.
Your guesstimate of Standard Deviation of mean in test arm (SD1) Approxiate estimate of Standard Deviation of mean in test arm, based on previous studies / pilot studies etc.
Your guesstimate of mean in control arm (M2) Approxiate estimate of mean in control arm based on previous studies / pilot studies etc.
Your guesstimate of Standard Deviation of mean in control arm (SD2) Approxiate estimate of Standard Deviation of mean in control arm, based on previous studies / pilot studies etc.
Ratio of Sample Size in two arms (Groups) (N2 / N1) Sample Size in Group 2 (control arm) / Sample Size in Group 1 (test arm)
Confidence Level % In percentage. Commonly used values are 95, 99 and 90. Should be between 0.01 to 99.99.
Power ( 1 - β) In Percentage. Should be between 0.01 to 99.99. Common values are 80 % and 90 %
Non-Inferiority Margin (d) It is the difference between control arm mean and test arm mean which can be tolerated to define non-inferiority. Can be positive or negative depending on the direction of non-inferiority. It is positive where inferiorty of test arm is defined, if test arm mean is less than control arm mean. In rare situaions, non-inferiority limit is negative, where inferiority of test arm is defined when test arm mean is more than control arm mean.
One sided test Being a non-inferiority design, hypothesis test is always one sided
Sample Size Required =