|
Abstract:
|
In longitudinal studies with time-varying treatment, the standard approach to the estimation of the effect of treatment on an outcome of interest is to model the outcome at time t as a function of past treatment history. This approach may be biased, whether or not one further adjusts for the past history of time-dependent confounding covariates, when these covariates predict subsequent outcome and treatment history and are, themselves, influenced by past treatment. In this roundtable, I will discuss several methods that can provide, under certain assumptions, valid estimates of the causal effect of a time varying treatment in the presence of time varying confounding factors. These include inverse-probability-of-treatment-weighted (IPTW) estimation of a marginal structural model, g-estimation of a structural nested model, and the parametric g-computation algorithm. The three methods will be compared.
|
