Statistical considerations in the design, analysis, and interpretation of comparative clinical studies.

Comparative clinical studies are extensively used as a basis of support for the potential approval of treatments by the Food and Drug Administration and also for other general health policy decisions. The purpose of this paper is to review statistical principles that are relevant to the design, analysis, and interpretation of comparative clinical studies. Three general criteria are emphasized: clarity, comparability, and generalizability. These criteria are discussed in terms of their implications to such considerations as scope of patient population, sample size, randomization, and covariance analysis. Also, an extensive reference list further addressing these topics is presented.