|
Activity Number:
|
129
- Quantile and Nonparametric Regression Models
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 31, 2017 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #323376
|
|
|
Title:
|
A Multiple Imputation Method Based on Bayesian Quantile Regression Model for Handling Intermittent Missing Data
|
|
Author(s):
|
MinJae Lee* and Mohammad Rahbar and John Reveille and Michael Weisman and Michael Ward and Lianne Gensler and Matthew Brown
|
|
Companies:
|
The University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and The University of Texas Health Science Center at Houston and Cedars-Sinai Medical Center and National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)/NIH and University of California, San Francisco and Queensland University of Technology
|
|
Keywords:
|
Multiple Imputation ;
Intermittent Missing ;
Bayesian Quantile Regression ;
Longitudinal Model ;
NSAID ;
Prospective Study of Outcomes in AS (PSOAS)
|
|
Abstract:
|
In longitudinal studies, the data collected during the follow up time points can be intermittently missing. Although longitudinal models can deal with specific types of missing data, inappropriate handling of this issue could lead to biased estimation of regression parameters when missing data mechanisms are complex and depend on multiple sources of variation. We propose a multiple imputation (MI) approach based on a Bayesian quantile regression model that not only accounts for longitudinal missing data, but also deals with data that are not normally distributed. Findings from our simulation study indicate that the proposed method performs better than traditional MI methods under certain scenarios of data distribution. We also demonstrate application of our method to data from the Prospective Study of Outcomes in ankylosing spondylitis (AS) (PSOAS) when examining an association between longitudinal Nonsteroidal anti-inflammatory drug (NSAID) usage and radiographic progression in AS while the longitudinal NSAID indices data were intermittently missing.
|
Authors who are presenting talks have a * after their name.
