Abstract:
|
When estimating causal effects, unmeasured confounding and model misspecification are potential sources of bias. To manage these, sensitivity of a study to potential unmeasured confounders can be evaluated in terms of the strength of confounding need to invalidate the result, while flexible model fitting techniques reduce parametric assumptions and thus minimize the chance of model misspecification. We propose a two-parameter sensitivity analysis strategy that incorporates Bayesian Additive Regression Trees (BART) into a larger framework to assess sensitivity of posterior distributions of treatment effects to choices of sensitivity parameters. This results in an easily interpretable framework for testing for the impact of an unmeasured confounder that also limits the number of modeling assumptions. We evaluate our approach in a large-scale simulation setting and with high blood pressure data taken from the Third National Health and Nutrition Examination Survey. The model is implemented as open-source software, integrated into the treatSens package for the R statistical programming language.
|