|
Activity Number:
|
345
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #301096 |
|
Title:
|
Correcting for Measurement Error in Diagnoses of Post-Traumatic Stress Disorder
|
|
Author(s):
|
Juned Siddique*+ and Robert Gibbons and Bonnie Green
|
|
Companies:
|
The University of Chicago and University of Illinois at Chicago and Georgetown University Medical School
|
|
Address:
|
Department of Health Studies, Chicago, IL, 60637,
|
|
Keywords:
|
imputation ; measurement error ; post-traumatic stress disorder
|
|
Abstract:
|
An accurate diagnosis of Post-Traumatic Stress Disorder (PTSD) requires an expert clinician who is able to obtain a detailed trauma history of the patient and then perform a fine-grained analysis of symptom severity. In mental health research it is often unfeasible for participants to be assessed by expert clinicians. Instead, studies must use lay interviewers or rely on questionnaires that measure self-reported symptoms. Both approaches tend to focus on the number of PTSD symptoms rather than symptom severity and can result in a high number of false positives. We present a multiple imputation approach to correct for measurement error in diagnoses of PTSD using a Bayesian ordinal probit model. The method is applied to a depression treatment study where nurse practitioners were twice as likely as psychologists to diagnose PTSD.
|
