Design and Analysis of Dose-Response Studies: Reality Versus Regulatory Requirements
- PDF / 418,329 Bytes
- 8 Pages / 504 x 719.759 pts Page_size
- 48 Downloads / 153 Views
0092-8615197 Copyright 0 195’7 Drug Information Association Inc.
DESIGN AND ANALYSIS OF DOSERESPONSE STUDIES: REALITY VERSUS REGULATORY REQUIREMENTS ALANPHILLIPS,BSc, PHD Director Biostatistics, Europe, Wyeth Research, Berkshire, United Kingdom
Within the pharmaceutical industry the assessment of dose-response is a critical part of a drug development program. This paper reviews how dose-response studies are
designed and analyzed, and compares and contrasts this against regulatory requirements. The paper aims to show that current practices for processing dose-response data do not meet the expectations of regulatory bodies. Alternative methods will be suggested. Key Words: Dose-response; Regulatory guidelines
INTRODUCTION
patients are allocated to one of several (4-5) fixed dose groups. Whenever ethical, one of WITHIN THE PHARMACEUTICAL inthe groups is usually a placebo group since, dustry the assessment of dose-response is a as illustrated by Ruberg (l), a significant critical part of a drug development program. trend in response with increasing dose is not Dose-response studies are undertaken to esevidence of a drug effect. A positive control tablish the effectiveness and safety of a new is also sometimes included to establish the drug, and the information obtained is used “sensitivity” of the study. The test drug dose to prepare dosage and administration instruclevels are often determined by clinical judgetions. This paper describes how dose-rement. Patients may be placed immediately on sponse studies are designed and analyzed the selected dosage or may require a “forced” within the pharmaceutical industry. This patitration to it. Typically, three dose levels are per aims to show that current methods of included in each study. analysis are unsatisfactory and do not meet Earlier on in a drug development program, regulatory requirements. Alternative methassignment to dose group may be ordered, ods for processing dose-response data will not random. The advantage of such nonranbe suggested. domized studies is that they provide a quick approximation of the dose-response relationSTUDY DESIGNS ship, while ensuring that doses well onto the plateau of the dose-response curve which The most commonly used study design to may be associated with undesired effects are assess dose-response within the pharmaceutinever used. For confirmatory dose-response cal industry is the parallel group study, where studies, however, patients are randomized to dose groups which are typically selected to be the minimal effective dose (low dose), the Reprint address: Alan Phillips, BSc, PhD, Director Biostatistics, Europe, Wyeth Research, Huntercombe Lane expected therapeutic dose (medium dose), South, Taplow. Maidenhead, Berkshire S M OPH. and the highest tolerated dose (high dose). United Kingdom. When drug effects develop rapidly and 737 Downloaded from dij.sagepub.com at CARLETON UNIV on May 10, 2015
738
Alan Phillips
patients return to baseline conditions quickly, dose-response is sometimes assessed using a multiple period crossover
Data Loading...
