Abstract:
|
Confounding and exposure measurement error are common barriers to drawing causal inference. While there are broad methodologies for addressing each phenomenon individually, confounding and exposure measurement biases frequently co-occur and there is a paucity of methods that address them simultaneously when targeting marginal causal effects. In this paper, methods are derived which leverage the likelihood structure under classical additive measurement error to draw inference using only measured variables. Three estimators are proposed based on g-computation, inverse probability weighting, and doubly-robust estimation techniques. The estimators are shown to be consistent and asymptotically normal, and the doubly-robust estimator is shown to exhibit its namesake property. The methods perform well in finite samples under both confounding and measurement error as demonstrated by simulation studies. The proposed doubly-robust estimator is applied to study the effects of two biomarkers on HIV-1 infection using data from the HVTN 505 vaccine trial.
|