Exact Critical Values for Farrington-Manning Noninferiority Exact Test
- PDF / 4,705,226 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 12 Downloads / 221 Views
David Sotres-Ramos, PhD Progmrna de Estadistica. Colegio de Postgraduados. carretera Mexico-Texcoco. Mexico
Felix Almeidra-Arao, PhD UPIITA del Instituto Politecnico Nacional. Mexico, DF, Mexico
Cecilia Ramirez-Figieroa, PhD Program de Estadistica, Colegio de Postgraduados, carretera Mhico-Texcoco. Mexico
Key Words Noninferiority; Size ofa test; Fam'ngton-Manning; Exact test
Correspoidmce Address David Sotrm-Ramos, Sierra Apaneca ## 43. Col. Jardines en la Montafia. Mexico, DF, CP 14210, Mexico (email:
[email protected]).
159
Exact Critical Values for Farrington-Manning Noninferiority Exact Test Noninfzriority tests are fiequently used in clinical trials to demonstrate that a new theram, with minimum side effects or low cost, is not significantly inferior in efficacy to the standard thenam. Ihe exact version of the Fam'ngtonManning noninfm*on.tytest for two independent binomial proportions has been studied and recommended ty several authors. Howeva this
INTRODUCTION As more and more effective drugs become available and fewer and fewer breakthrough new drugs emerge, the objective of clinical investigation changes. Oftentimes, it seeks noninferiority of the new drug in active controlled trials. Active controlled clinical trials, with the intention of demonstrating efficacy by showing the new test treatment to be not inferior to the standard treatment by more than a small predefined margin, have been called noninferiority trials. Refer to the ICH E9 guideline, "Statistical Principles for Clinical Trials" (l),for important statistical considerations. The use of an active control in noninferiority trials is playing an increasingly important role in global new drug development. The Farrington-Manning (FM) noninferiority exact test for two independent binomial proportions has been studied and recommended by several authors (2-10). In a recent article Almendra-Arao and Sotres-Ramos (11) compared the FM asymptotic test with six other asymptotic noninferiority tests. They have shown that the significance level and the power of the FM asymptotic test have the best behavior among these asymptotic tests for sample sizes of 30 In I100, as well as for the three most frequently used noninferiority margins and for the nominal levels 0.01 and 0.05. In addition, AlmendraArao et al. (12) compared the significance level of the FM exact and asymptotic tests. They found, for almost all sample sizes considered in
exact test is difficult to apply because the ctitical constant and the size of the test are complicated to calculate. To facilitate the use of the Fmhgton-Manning exact test, a computer progmm has been written in PLUS 2000, which calculates the critical constant and the size of this test. Also, it has been veriped that this test is not very conservative.
the range 5 5 n I 100,that the significance level of the FM exact test is closer to the nominal level than the FM asymptotic test, and that the FM asymptotic test is much more conservative than the FM exact test. The main goal of this work is to provide a compute
Data Loading...
