Non-inferiority Study Designs
As the medical literature moves forward and standards of care in the treatment of many conditions are established, newer therapies which offer additional appealing features, such as ease of administration or fewer side effects, will often be studied with non-inferiority trial designs. The idea is that the newer therapies need not prove superiority to existing standards of care if they indeed offer other benefits to patients or health systems. As a result, non-inferiority designs are becoming increasingly common. Non-inferiority designs bring a different set of implications than standard superiority trials with regard to sources of bias and understanding the numeric results. In this chapter, we review how non-inferiority margins are calculated (i.e., how much worse can you be and still be considered “not worse than”?), interpreting power in non-inferiority designs, how the risk of bias needs to be interpreted differently in non-inferiority designs, and tips for teaching this fascinating topic to learners.
