Nonparametric Analysis of Covariance and Its Role in Noninferiority Clinical Trials
- PDF / 1,114,161 Bytes
- 15 Pages / 504 x 720 pts Page_size
- 57 Downloads / 180 Views
Drug Information JoumI. Vol. 33. pp. 1145-1159. 1999
Printed in the USA. All rights reserved.
NONPARAMETRIC ANALYSIS OF COVARIANCE AND ITS ROLE IN NONINFERIORITY CLINICAL TRIALS GARYG. KOCH, PHD Department of Biostatistics, University of North Carolina, Chapel Hill, North Carolina
CATHERINE M. TANGEN, DRPH Southwest Oncology Group, Fred Hutchinson Cancer Research Center, Seattle, Washington
Some randomized clinical trials are conducted to compare test treatment, an active reference control, and placebo with the objectives of demonstrating that test treatment is not inferior to reference control, and to show that both test treatment and reference control are better than placebo. We typically want to adjust for covariables that are strongly associated with the response of interest in order to gain variance reduction, to adjust for random imbalances of covariables, and to clarify the degree to which differences between randomized groups are due to treatments rather than other factors associated with response. Parametric modeling is often used to evaluate the relationship between covariables and the conditional distributions of response given the covariables. There can be concerns, however; about model assumptions, and they are not always straightforward to assess. An alternative is to use a nonparametric method for the primary evaluation of treatment comparisons. The nonparametric method is performed through linear models for (unconditional) differences between treatment groups for means of response criteria (orfunctions of such means) and covariables jointly with specifications that adjust random differences for means of covariables to zero. This paper discusses the role of nonparametric analysis of covariance in clinical trials to compare test treatment, an active reference control, and placebo with three examples. Key Words: Reference control; Population average treatment parameter; Weighted least squares; Random imbalance of covariables; Variance reduction from covariance adjustment
INTRODUCTION IN there is interest in comparing test treatment Resented by b.Koch at the 5’ Annual DIA Biostatistics Workshop: “Antidiabetics, Antibiotics, and ICH E10 (Choice of Control Group in Clinical Trials),” S e g tember 4-5, 1998. Tokyo, Japan. Reprint address: Dr. G q Koch, Department of Biostatistics, CB 7400, School of Public Health, University of ~ 0 1 t hCarolina, Chapel KII, NC 275%7400.
(T), an active reference control (R), and placebo (P). For trials with this design, there are usually two objectives. One is to show that both test treatment and the active reference control are better than placebo, and the other is to show that test treatment is not inferior to the active reference control (1.2). When a study convincingly demonstrates both of these objectives, its evidence for the Of test treatment is ‘Omparable to that from two studies (ie, T SUPrior to P in one study, and T noninferior to
1145
Downloaded from dij.sagepub.com at Bobst Library, New York University on April 11, 2015
Gary G. Koch and Catherine M .
Data Loading...
