Instrumental Variable Estimation of Marginal Structural Mean Models for Time-Varying Treatment

Robins introduced Marginal Structural Models (MSMs), a general class of counterfactual models for the joint effects time-varying treatment regimes in complex longitudinal studies subject to confounding. In his work, identification MSM parameters is established under Sequential Randomization Assumption (SRA), which rules out unmeasured confounding assignment over time. We consider sufficient conditions subclass, Mean (MSMMs), when sequential randomization fails hold due confounding, using instead instrumental variable. Our require that no unobserved confounder predicts compliance type treatment. describe simple weighted estimator and examine its finite-sample properties simulation study. apply proposed effect delivery hospital on neonatal survival probability. Supplementary materials this article are available online.