|
Activity Number:
|
183
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 1, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Mental Health Statistics Section
|
|
Abstract #320928
|
View Presentation
|
|
Title:
|
A Model for Estimating Missing Items Score on Self-Reported Psychometric Scales
|
|
Author(s):
|
Balasubramani G.K.* and Stephen R. Wisniewski
|
|
Companies:
|
University of Pittsburgh and University of Pittsburgh
|
|
Keywords:
|
Imputation ;
Estimation ;
Major Depressive Disorder ;
Longitudinal data ;
Weighted mean ;
Psychometric scales
|
|
Abstract:
|
In almost all clinical studies of patients with major depressive disorder, the data on the psychometric rating scales were collected by the self-reported methods. The problem of missing items is deceptive to self-report measure. At any one visit a study participant might fail to complete specific measures or items resulting in missing data for those self-administered scales. Treating unobservable data simply as missing can potentially lead to estimation bias of an unquantifiable magnitude. In this paper we developed a statistical weighted means model to estimate efficiently the missing items scores on the self-administered psychometric scales of longitudinal data. To estimate the model we used the assumption that the patient response is non-uniform across items. We also explored different estimating techniques such as mean model, using modal score and compare it with the proposed approach. We evaluated the models overall performance using simulation techniques. We demonstrated its use by application data from STAR*D study. The developed method efficiently estimates scores of psychometric rating scales with partially missing data which would be valuable to clinical researchers.
|
Authors who are presenting talks have a * after their name.
