Statistical Methods for Demonstrating Equivalence in Crossover Trials Based on the Ratio of Two Location Parameters
- PDF / 329,325 Bytes
- 6 Pages / 504 x 719.759 pts Page_size
- 2 Downloads / 159 Views
0092-86 15/2000
Copyright 0 2000 Drug Information Association Inc.
STATISTICAL METHODS FOR DEMONSTRATING EQUIVALENCE IN CROSSOVER TRIALS BASED ON THE RATIO OF TWO LOCATION PARAMETERS MEINHARD KIESER,PHD Head Biometrics, Dr. Willmar Schwabe. Pharmaceuticals, Karlsruhe, Germany
DIETER HAUSCHKE, PHD Head Department of Preclinical Biometry, Byk Gulden Pharmaceuticals, Konstanz, Germany
Equivalence trials have the objective of demonstrating rhar an investigational treatment, for example, a test drug under development, is not different from a reference treatment by more than a prespecified clinically irrelevant amount. The purpose of this paper is to investigate for the crossover design the situation when equivalence is defined in terms of the ratio of location parameters. An approximate formula for sample size calculation is presented for the case of normally distributed endpoints, and nonparametric methods for testing and calculation of confidence intervals are provided. Key Words: Equivalence; Ratio of location parameters; Confidence intervals; Sample size
INTRODUCTION
starting the clinical trial and will depend on medical grounds. Furthermore, this Note for THE NOTE FOR GUIDANCE “Sratisrical Guidance requires that the statistical evaluaPrinciples for Clinical Trials, published in tion should be based on the use of confidence 1998 by the European Agency for the Evalu- intervals. Only a limited number of publicaation of Medicinal Products (I), addresses tions, however, address the situation for an the statistical analysis of trials to show equiv- underlying location model and when the alence. In such studies the issue is not to specification of equivalence limits is done by show a difference between two active treat- expressing the term relative to the unknown ments, but to demonstrate that an investiga- location parameter of the reference. In a tional treatment is not different from the ac- recent paper (2), the corresponding two cepted reference treatment by more than an one-sided tests procedure and the two-sided irrelevant amount. Obviously, the extent of (1 - 209 100% confidence interval for twothe accepted difference must be fixed before sided equivalence studies assuming normally distributed data were presented. In addition, the authors addressed the exact sample size determination for the crossover. The calculaReprint address: Meinhard Kieser. PhD. Dr. Willtion of the power is mathematically complex mar Schwabe Pharmaceuticals, P.O. Box 41 09 25, D-76209 Karlsruhe, Germany. E-mail: meinhard.kieser and not accessible to practitioners, therefore, an approximate formula for sample size [email protected]. 563 Downloaded from dij.sagepub.com at NANYANG TECH UNIV LIBRARY on June 1, 2015 ”
Meinhurd Kieser and Dieter Hauschke
564
termination is presented in this paper. A further purpose of this paper is to derive an analogous nonparametric decision procedure without assuming the normal distribution.
the acceptable difference between pT and p R is expressed as a proportion of p R (2): PT H~: I el or k 2 e2
PR
C
Data Loading...
