What is a non-inferiority trial?


New drugs can be tested in trials that are designed to establish that their effect on a particular outcome measure is better than a placebo. But if an established treatment already exists, it is more meaningful to compare the new treatment with an established one. Trials can be designed to test if the new drug is better than the standard therapy (so-called superiority trials). But superiority trials using an active control need to be much larger than placebo-controlled trials because the expected difference between the treatments is small and so a large sample will be needed to show a significant difference.

So very often trials are designed to show only that the new drug is not inferior to a standard drug (in a non-inferiority trial).  This means the trial needs to show the new drug is no worse, within a specified margin, than the standard drug. Non-inferiority is assumed if the lower confidence limit for the difference between the two drugs is not more negative than a pre-specified amount, known as the non-inferiority margin. A non-inferiority trial does not actually rule out the possibility that the test drug is more effective than the standard drug. And it is possible, once non-inferiority has been demonstrated, to go on to do a statistical test to show the drug is statistically superior to the standard treatment. 

Non-inferiority trials usually need a much smaller sample size than trials looking for superiority and so they are quicker and cheaper to perform, and also less likely to produce disappointing results for a new drug.This type of trial is often used to test new drugs.