Gomberg-Maitland Mardi, Frison Lars, Halperin Jonathan L
Rush Heart Institute for Pulmonary Heart Disease, Rush School of Medicine, Chicago, Ill 60612-3824, USA.
Am Heart J. 2003 Sep;146(3):398-403. doi: 10.1016/S0002-8703(03)00324-7.
The randomized, double-blind, placebo-controlled trial is the optimum method for clinical evaluation of new treatments, as assessed by clinicians and statisticians. However, if a known standard of therapy exists, it may be difficult to prove that a new therapy is superior. Equivalence and noninferiority clinical trial designs are now frequently utilized in clinical medical research. This article reviews the statistical differences between superiority, equivalence, and noninferiority design schemes, which pose specific ethical questions and have important implications for interpretation and clinical application of trial results. A guideline is proposed as a standard approach for reporting to facilitate qualitative assessment of the methodology of these trials.
经临床医生和统计学家评估，随机双盲安慰剂对照试验是临床评估新疗法的最佳方法。然而，如果存在已知的治疗标准，可能难以证明新疗法更优。等效性和非劣效性临床试验设计目前在临床医学研究中经常被采用。本文回顾了优效性、等效性和非劣效性设计方案之间的统计学差异，这些差异引发了特定的伦理问题，并对试验结果的解释和临床应用具有重要意义。本文提出了一项指南作为报告的标准方法，以促进对这些试验方法的定性评估。
