Confounding adjustment through front-door blocking in longitudinal studies

A common aim of epidemiological research is to estimate the causal effect of a particular exposure on a particular outcome. Towards this end, observed associations are often 'adjusted' for potential confounding variables. When the potential confounders are unmeasured, explicit adjustment becomes unfeasible. It has been demonstrated that causal effects can be estimated even in the presence of umeasured confounding, utilizing a method called 'front-door blocking'. In this paper we generalize this method to longitudinal studies. We demonstrate that the method of front-door blocking poses a number of challenging statistical problems, analogous to the famous problems associated with the method of 'back-door blocking'.