|
Activity Number:
|
491
- Causal Inference Within Reach: Pragmatic Approaches to Model Construction and Validation
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 1, 2018 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #326882
|
Presentation
|
|
Title:
|
Weighting-Based Sensitivity Analysis in Causal Mediation Studies: Interactive Tools for Analysts
|
|
Author(s):
|
Guanglei Hong* and Xu Qin and Fan Yang
|
|
Companies:
|
University of Chicago and University of Chicago and University of Colorado Denver
|
|
Keywords:
|
Causal inference;
direct effect;
indirect effect;
propensity score;
RMPW;
selection bias
|
|
Abstract:
|
A general weighting-based approach to sensitivity analysis extends the ratio-of-mediator-probability weighting (RMPW) method for identifying natural indirect effect and natural direct effect. The new strategy assesses bias associated with one or more omitted pretreatment or posttreatment confounders. In its essence, the discrepancy between a new weight that adjusts for the omitted confounders and an initial weight that omits them captures the role of the confounders. The effect size of the bias is a product of two sensitivity parameters, one associated with the degree to which the omitted confounders predict the mediator and the other associated with the degree to which they predict the outcome. This approach reduces the reliance on functional form assumptions and removes constraints on measurement scales. A graphical display indicates the threshold values of bias great enough to alter the inference; a sensitivity analysis table lists the estimated actual or potential bias. We discuss considerations in developing a software package that interacts with the user, guides analytic decision-making, and supplies information necessary for evaluating an initial causal conclusion.
|
Authors who are presenting talks have a * after their name.
