Duke S P, Garrett S
Department of Clinical Research, Atrix Laboratories, Inc., Fort Collins, CO, USA.
J Periodontol. 1998 Jun;69(6):650-4. doi: 10.1902/jop.1998.69.6.650.
There has been a tendency by clinicians in periodontal research to confuse lack of significance in a clinical trial designed for superiority for equivalence between the treatment arms. This paper is intended to define for the clinician what equivalency means in statistical terms. To demonstrate equivalence, the hypotheses and analysis specific to equivalence must be carried out; the information cannot be extracted directly from trials designed with superiority hypotheses and analysis. Further, acceptable mean differences in equivalency trials should be determined based not only on statistical considerations, but also on the clinical relevance of the proposed differences. The relationship between superiority and equivalency design and analysis is discussed and an example using a hypothetical clinical study is given.
在牙周病研究中，临床医生往往会混淆为证明治疗组间优效性而设计的临床试验中缺乏显著性与等效性之间的区别。本文旨在为临床医生从统计学角度定义等效性的含义。为了证明等效性，必须进行特定于等效性的假设和分析；不能直接从基于优效性假设和分析设计的试验中提取相关信息。此外，等效性试验中可接受的均值差异不仅应基于统计学考虑来确定，还应考虑所提出差异的临床相关性。本文讨论了优效性与等效性设计及分析之间的关系，并给出了一个使用假设性临床研究的示例。
