A PROSPECTIVELY PLANNED CUMULATIVE METAANALYSIS APPLIED TO A SERIES OF CONCURRENT CLINICAL-TRIALS

Sequential designs are now a familiar part of clinical trial methodolo gy. In particular, the triangular test has been used in several indivi dual studies. Methods of combining studies are also well-known from th e literature on meta-analysis. However, the combination of the two app roaches is new. Consider the situation where a series of studies is to be conducted, following broadly similar protocols comparing a new tre atment with a control treatment. In order to obtain an answer as quick ly as possible to an efficacy or safety question it may be desirable t o perform a cumulative meta-analysis on one particular variable. This could, for example, be the primary efficacy variable, an expensive ass essment conducted in only a subgroup of patients, or a serious side-ef fect. To allow for the size of the treatment difference varying from s tudy to study we might wish to provide a global estimate. Hence a rand om effects combined analysis, within a sequential framework, would app ear to be appropriate. A methodology which utilizes efficient score st atistics and Fisher's information is presented. Simulations show that the proposed methodology will achieve the specified error probabilitie s with reasonable accuracy provided that any random effect is relative ly small. Ignoring random effects when they are present can lead to in accuracies. A simulated example illustrates a number of practical issu es. (C) 1997 by John Wiley & Sons, Ltd.