Marginal Structural Models for Sufficient Cause Interactions

摘要

Sufficient cause interactions concern cases in which there is a particular causal mechanism for some outcomenthat requires the presence of 2 or more specific causes to operate. Empirical conditions have been derived to testnfor sufficient cause interactions. However, when regression outcome models are used to control for confoundingnvariables in tests for sufficient cause interactions, the outcome models impose restrictions on the relation betweennthe confounding variables and certain unidentified background causes within the sufficient cause framework; often,nthese assumptions are implausible. By using marginal structural models, rather than outcome regression models,nto test for sufficient cause interactions, modeling assumptions are instead made on the relation between thencauses of interest and the confounding variables; these assumptions will often be more plausible. The usenof marginal structural models also allows for testing for sufficient cause interactions in the presence of timedependentnconfounding. Such time-dependent confounding may arise in cases in which one factor of interestnaffects both the second factor of interest and the outcome. It is furthermore shown that marginal structural modelsncan be used not only to test for sufficient cause interactions but also to give lower bounds on the prevalence of suchnsufficient cause interactions.