Abstract Details
Activity Number:
|
68
|
Type:
|
Contributed
|
Date/Time:
|
Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
|
Sponsor:
|
ENAR
|
Abstract #312677
|
View Presentation
|
Title:
|
Causal Inference with Truncation by Death in Observational Study
|
Author(s):
|
Linbo Wang*+ and Xiao-Hua Andrew Zhou
|
Companies:
|
University of Washington and University of Washington
|
Keywords:
|
Causal Inference ;
Observational Study ;
Truncation by Death ;
Propensity Score
|
Abstract:
|
It is common that in observational studies, some subjects die before the primary outcomes are measured. However, the primary outcomes (such as depression) may only be defined for alive people. In this case, the restricted analysis in people who survive does not lead to a meaningful causal estimate. We instead consider estimating the Survivor Average Causal Effect (SACE), i.e., the average causal effect in a subpopulation that would survive regardless of treatment received. However, SACE is not identifiable from observational studies for two reasons: members of the subpopulation are not observed directly, and confounding can present without randomization. In this paper, we show identifiability of SACE under additional assumptions, including existence of a baseline covariate that is similar to an instrumental variable, and a modified strongly ignorable treatment assignment assumption. We then propose a consistent non-parametric estimator for SACE based on propensity score stratification, as well as a consistent variance estimator. Finally, we apply our method to estimate effect of statins use on cognitive decline among subjects with normal cognition in an observational study.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.