From Group Sequential to Adaptive Designs [chapter]

Christopher Jennison, Bruce Turnbull
2010 Handbook of Adaptive Designs in Pharmaceutical and Clinical Development  
increased probability of seeing spurious results and making a premature and erroneous decision. To overcome this danger of over-interpretation of interim results, special statistical analysis methods are required. To address this need, the first classic books on sequential analysis were published by Wald (1947) , motivated primarily by quality control applications, and by Armitage (1960) for medical trials. In this chapter, we shall be concerned with the latter application. The benefits of
more » ... oring data in clinical trials are obvious: Administrative. One can check on accrual, eligibility and compliance, and generally ensure the trial is being carried out as per protocol. Economic. Savings in time and money can result if the answers to the research questions become evident early -before the planned conclusion of the trial. Ethical. In a trial comparing a new treatment with a control, it may be unethical to continue subjects on the control (or placebo) arm once it is clear that the new treatment is effective. Likewise if it becomes apparent that the treatment is ineffective, inferior or unsafe, then the trial should not continue. It is now standard practice for larger Phase III clinical trials to have a Data Monitoring Committee (DMC) to oversee the study and consider the option of early termination. Note that many of the same considerations apply to animal and epidemiologic studies as well. It was soon recognized by researchers that fully sequential procedures, with continuous monitoring of the accumulating data, were often impractical and, besides that, much of the economic savings could be achieved by procedures that examined the data on a limited number of occasions throughout the trial -at six month intervals, for example, in a multi-year trial. The corresponding body of statistical analysis and design techniques has become known as group sequential methodology because the accumulated data are examined after observing each successive group of new observations. There is a large body of literature in the biostatistical and medical journals and there have been several comprehensive books published. These include Whitehead (1997) , Jennison and Turnbull (2000) and Proschan, Lan and Wittes (2006) . Of related interest are books on the practical considerations for the operation of DMCs by Ellenberg, Fleming
doi:10.1201/b10279-6 fatcat:d2j64sk6rje4nfj2uml7n25am4