Finite mixture models to look at the causal effects of process variables

State of the art talk by Richard Emsley, University of Manchester, UK

Chair: Tim Croudace, Wednesday 22nd July, 11.45 - 12.30, Palmeston Lecture Theatre, Fisher Building.

Using psychological treatment trials as an example of complex interventions, we review statistical methods which evaluate both direct and indirect effects, in the presence of hidden confounding between the process variables and outcome. We review the historical literature on mediation and moderation of treatment effects before introducing two methods from within the causal inference literature, principal stratification (finite mixture models) and structural mean models. We demonstrate how these can be applied in the context of mediation before discussing approaches and assumptions necessary for attaining identifiability of key parameters of the basic causal model. We describe how moderation can occur through post-randomisation variables, and extend the principal stratification approach to multiple group methods with explanatory models nested within the principal strata. We illustrate the new methodology with motivating examples of randomised trials from the mental health literature.