Marginal structural models
link the expected value of a counterfactual outcome to covariates; unlike in regression models, the expected value is not conditional on the covariates
link the expected value of a counterfactual outcome to covariates; unlike in regression models, the expected value is not conditional on the covariates