Statistical Approaches to Improving Trial Efficiency and Conduct

Given the trend towards increasing complexity and cost of clinical trials today, future trials may end up being small, complex, and under-powered to detect clinically meaningful treatment effects. In order to continue to perform important research in the future, we need to make clinical trial designs more efficient. Through the retrospective statistical analysis of variation in the design of past trials and the prospective comparisons of clinical trials methods, we can determine which trial procedures truly influence the bias and precision of treatment estimates and where complexity and costs can be reduced. We provide two examples of the retrospective study of clinical trials methods that could change the conduct of future trials. First, an overview of the effect of outcome adjudication on treatment estimates for cardiovascular trials is presented. Second, a prognostic model to detect fraud within multi-centre trials is developed as part of a system of central statistical monitoring. There are many more unanswered questions about efficiencies in clinical trials methodology that need to be examined by statisticians and researchers.