Something for Nothing in Noninferiority/Superiority Testing: A Caution
- PDF / 402,739 Bytes
- 5 Pages / 504 x 719.759 pts Page_size
- 20 Downloads / 183 Views
Drug Information Journal, Vol. 35, pp. 241-245. 2001 Printed in the USA. All rights reserved.
SOMETHING FOR NOTHING IN NONINFERIORITY/SUPERIORITY TESTING: A CAUTION BRIANL. WIENS,PHD* Associate Director, Department of Biornetrics and Data Management, Quintiles, Inc., Kansas City, Missouri
Clinical trials comparing two active treatments often are conducted to show that an investigational treatment is not inferior to a standard therapy. When noninferiority is shown, a conditional analysis can be used to establish whether the investigational treatment is superior to the standard. Alternatively when superiority is desired but not shown, a conditional analysis can be used to establish noninferiority. Such analyses can be used without adjustment to the type I error rate, under certain assumptions. This paper presents cautions in implementing such an analysis strategy. Equivalence and noninferiority analyses often use the per protocol group of subjects in the primary analysis, while analyses that want to show a difference generally use a wider group of subjects from the study. If the two analyses use different groups of subjects, the type I error rate may be inflated unless a stepwise procedure in the proper direction is instituted. Strategies for planning and interpreting such analyses are discussed. Key Words: Equivalence; Intersection-union test; Intent-to-treat; Per protocol
INTRODUCTION CLINICAL TRIALS THAT compare an investigational therapy to an active standard therapy are common. Such studies may be conducted to gain advantageous labeling from a regulatory agency, or to support publication of results. One way to make such a claim is through an equivalence or noninferiority analysis. Such an analysis is used to show that two therapies are sufficiently close in efficacy that it does not matter which therapy is used. This is done by demonstrating that if the investigational therapy is not superior to the standard therapy, the difference is
Reprint address: Brian Wiens, WS24-3-C. Arngen, Inc., One Amgen Center Drive, Thousand Oaks, CA 91320. E-mail: [email protected]. *Dr. Wiens is now Manager 11, Biostatistics, at Amgen, Inc., Thousand Oaks, CA.
less than some prespecified value 6. Of course, showing that an investigational therapy is superior to a standard therapy will generally be much more valuable. Dunnett and Gent (1) proposed that both a noninferiority and a superiority comparison can be made without adjustment for the multiple comparisons. If 0 is the parameter of interest, the hypotheses in testing noninferiority are Hol: 0 2-6 versus HI,:0 > -6, assuming that positive values of 0 occur when the investigational therapy is superior to the standard. If noninferiority is demonstrated by rejecting Hal, then superiority can be tested via H,: 0 5 0 versus HI2:8 > 0. The tests can be conducted by constructing a confidence interval for the true value of 0. If the lower bound exceeds -6, noninferiority can be concluded; if the lower bound also exceeds 0, superiority can be concluded. No adjustment to the type I er
Data Loading...
