Complex intervention trials should be able to answer both pragmatic and explanatory questions in order to test the theories motivating the intervention and help understand the underlying nature of the clinical problem being tested.  Key to this is the estimation of direct effects of treatment and indirect effects acting through intermediate variables, such as mediators. Using psychological treatment trials as an example of complex interventions, this talk explains statistical methods which evaluate both direct and indirect effects in the presence of hidden confounding between mediator and outcome. We introduce principal stratification and structural mean models, and discuss approaches for attaining identifiability of key parameters of the basic causal model. Assuming that there is no direct effect of treatment leads to the use of instrumental variable methods, using randomisation and its interactions with baseline covariates as instruments. The new methodology is illustrated with motivating examples of randomised trials from the mental health literature.

Richard's slides

Meeting contact: