Activity Number:
|
337
- Environmental Epidemiology and Analysis of Large Database
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 9, 2022 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #322530
|
|
Title:
|
Minimizing Estimation Bias in Analyses Utilizing Electronic Health Records Data
|
Author(s):
|
Zhibao Mi* and Ellen J Dematt and Eileen M Stock and Min Zhan and Kousick Biswas
|
Companies:
|
VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center
|
Keywords:
|
Estimation Bias;
Electrical Health Records;
Auxiliary Data;
Clinical Trial;
Confounding ;
Propensity Scores
|
Abstract:
|
Electrical health records (EHR) are increasingly utilized as auxiliary data to enhance clinical trial performance and as real-world data to derive real-world evidence (RWE). Some of the challenges of using EHR data are obtaining all data points necessary to define a study cohort, missing data, and covariate imbalances between comparison groups. We developed an approach to analyze EHR data from VA Corporate Data Warehouse (CDW) to obtain reliable estimators to support our VA clinical trial designs via building analysis cohort to emulate trial study groups, handling missing data, controlling confounding factors by adjusting them using a propensity score approach and stratifying prognostic factors with interested subgroup information, and further assessing the estimation bias by sensitivity analyses. The detailed methods are discussed using a recent clinical trial design to illustrate the proposed analytical approach. Numerical studies are also performed to evaluate estimation robustness on various analytical scenarios.
|
Authors who are presenting talks have a * after their name.