One Large, Well-Designed, Multicenter Study as an Alternative to the Usual Fda Paradigm
- PDF / 533,211 Bytes
- 7 Pages / 504 x 720.24 pts Page_size
- 43 Downloads / 159 Views
0092-86 15/99 Copyright 0 1999 Drug Information Association Inc.
ONE LARGE, WELL-DESIGNED, MULTICENTER STUDY AS AN ALTERNATIVE TO THE USUAL FDA PARADIGM LLOYDD. FISHER,PHD* Professor of Biostatistics. Department of Biostatistics, University of Washmgton, Seattle, Washington
This article considers when a single study presents both: 1. The same strength of evidence for a drug or biologic effects as the usual Food and Drug Administration (FDA) twostudy paradigm. and 2. Also gives sufficient evidence of replicability as in two studies done at disjoint sets of institutions. It is argued that 1. holds if the two-sided signi’jicance level is 0.00125 and 2. holdr ifthe clinical sites are divided into two subsets that maximize the smaller sample size that each is statistically significant at the 0.05 significance level. Other approaches to satisfying 2. are also addressed. The approach is illustrated with data from the Condylox@Gel 0.5% (podojilox gel) versus vehicle control. Key Words: One large study; FDA two-study paradigm; Strength of evidence; Significance level; Replicating results
INTRODUCTION IN 1995 THE UNITED States Food and Drug Administration published the following: “FDA believes good science dictates that a showing of effectiveness must be methodologically sound and provide a high level of confidence in the validity of the result. For human drug products, this ordinarily is achieved by independently replicating the result in a second study, to constitute an adequate demonstration of effectiveness for a new product. While a second study may well be needed to replicate results demonstrated in a first study, in some instances, it is possible to replicate
*The author was a consultant for Oclassen Pharmaceuticals. Oclassen also supported the preparation of this manuscript, as well as kindly allowing the use of its data. Reprint address: Lloyd D. Fisher, PhD, Professor of Biostatistics. Department of Biostatistics, Box 357232, University of Washington, Seattle, WA 98155-7232.
results within one large, welldesigned, multicenter study. FDA emphasizes that this approach can be successful only when results are strong. The agency has, in the past, approved new human drug products on the basis of a single, multicenter study. Examples include domase alfa for the treatment of cystic fibrosis, timolol for treatment of people after a heart attack, and zidovudine for AIDS. A statistically marginal result, even in a very large study, cannot provide convincing evidence without replication” (1).
More recently, the agency has circulated a draft guidance document-not for implementation-ntitled “Guidance for Industry: Providing Clinical Evidence of Effectiveness for Human Drug and Biologic Products” (2). This document, in part, expands upon the 1995 document; it gives advice about when one large trial may be adequate evidence of efficacy. The draft guidance addresses situations where one trial may suffice but with
265 Downloaded from dij.sagepub.com at UNIVERSITY OF SASKATCHEWAN LIBRARY on March 18, 2015
Uoyd D. Fisher
266
Data Loading...
