Online Program Home
My Program

Abstract Details

Activity Number: 164 - SPEED: Causal Inference and Related Methodology
Type: Contributed
Date/Time: Monday, July 30, 2018 : 10:30 AM to 12:20 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #330452 Presentation
Title: Accounting for Variation in Instrumental Effect Estimates Leads to More Precise Estimates of Causal Effects in MR Studies
Author(s): Richard Barfield* and Li Hsu
Companies: Fred Hutchinson Cancer Research Center and Fred Hutchinson Cancer Research Center, USA
Keywords: Summary Statistics; GWAS; Empirical Bayes; Attenuation Bias ; Mendelian Randomization
Abstract:

Mendelian randomization (MR) estimates the causal effect of a mediating variable on an outcome using genetic variants, as genetic variants are independent of unmeasured confounders. Summary statistics let researchers pool together large studies without sharing individual level data. Given this, there has been development of MR methods that use summary statistics to estimate the mediating variable's causal effect. Most methods however do not account for the variation in the instrument effects, leading to attenuation bias in the causal effect estimate. We developed an Empirical Bayes approach using multivariate summary MR for estimation of the causal effect of the mediating variable on the outcome. This approach accounts for the variation of the instrumental effect estimates, and corrects for the downward bias often seen in summary MR methods. The proposed method leads to a more precise estimate of the causal effect of interest. We performed extensive simulation scenarios to demonstrate the estimator's empirical performance. We applied the approach to summary statistics from two publically available GWAS of height and BMI to assess their causal effect on various disease phenotypes.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program