CRITICAL REVIEW OF THE ARMITAGE AND BROSS SEQUENTIAL PROCEDURES IN CLINICAL REHABILITATION TRIALS

Abstract

The use of sequential analysis in clinical rehabilitation research allows a spectrum of analyses in the comparison of two or more treatment protocols. Researchers in rehabilitation medicine are increasingly making use of sequential designs in their clinical investigations. This review serves to highlight the optimal use of the classical sequential designs presented by Armitage and Bross in the mid-twentieth century. A discussion of the limitations of this most basic sequential analysis is presented for the information of clinical researchers considering this study design. Examples of the use of classical and group sequential designs addressing both continuous and dichotomous outcomes are provided, and the advantages and disadvantages of classical and group sequential procedures as compared to fixed sample designs are illustrated with rehabilitation examples. As little literature has been published regarding the application of sequential analysis in clinical rehabilitation trials, clinical pharmacological, and medical trials in addition to medical statistical sources were used in this review.