Ethical and Statistical Norms in the Justification of Randomised Clinical Trials

Abstract

This paper analyses how to combine the statistical justification of the results obtained in a randomised clinical trial and the ethical and social arguments that support its implementation. It aims at showing how this justification varies according to the view on probability we adopt and the different ethical cases that can be made to sustain it. Both the frequentist and the Bayesian approaches to the design of clinical trials will be seen as equally defensible on these grounds. To be socially acceptable both approaches must meet an impartiality requirement, which is best served by the technique of randomisation.