CLINICAL-TRIALS AND CAUSATION - BAYESIAN PERSPECTIVES

In addition to the safety, it is essential to establish the causal eff icacy of extant and new treatments, and well-designed clinical trials are thought by most to be the 'gold standard' to accomplish this. Cont rary to most statisticians' and regulators' views, however, I will arg ue that the concept of causation involved in clinical trials is not al l that clear. I discuss the manipulability approach to causation, inte rpreted counterfactually, which seems to fit causation as it is found in such sciences as physiology, but it has unclear relations to a conc ept of causation proposed by a number of epidemiologists. I characteri ze 'epidemiological causation' as probabilistic and formulated at a po pulation level, and dependent on certain general criteria for causatio n as well as study-design considerations. I then attempt to clarify th e connections between these concepts of causation and Cartwright's vie ws on complexity and causality, a 'Bayesian' framework proposed by Rub in and further elaborated by Holland, and Glymour and his colleagues' recent directed graphical causal modelling approach.