B0259
Title: The robust inverse-dispersion weighted estimator in Mendelian randomization
Authors: Alfonso Garcia-Perez - Universidad Nacional de Educación a Distancia (UNED) (Spain) [presenting]
Abstract: First, a new robust estimator is defined for the effect of exposure X on outcome Y, in the context of Mendelian randomization (MR), a method that uses a genetic variation to avoid possible biases in the regression of Y on X, due to lack of complete randomization, or reverse causation, or confounders. MR uses instrumental variables Z for this purpose being the two-stage least squares estimator classical estimator of this effect. In the second stage, the combination of these classical estimators, for different instrumental variables, is done with the classical inverse-variance weighted (IVW) estimator, which has a 0 breakdown point. A new robust version of the IVW is also included in which these effects are combined. This new robust estimator is called the robust inverse-dispersion weighted estimator.
