Online Program Home
My Program

Abstract Details

Activity Number: 374
Type: Contributed
Date/Time: Tuesday, August 2, 2016 : 10:30 AM to 12:20 PM
Sponsor: Mental Health Statistics Section
Abstract #320364
Title: On Choosing Optimal Diagnostic Short-Form Scales from Unidimensional Criteria Sets: Application to Alcohol Use Disorder Diagnosis
Author(s): Cheryl Raffo* and Melanie Wall, PhD
Companies: Columbia University and Columbia University
Keywords: short-form ; DSM-5 ; AUD ; scales
Abstract:

In psychiatry, a disorder is defined by a set of symptom criteria and a diagnosis can be made when a minimum number of these criteria are met. The DSM-5 provides a fixed-length set of 11 unidimensional criteria associated with Alcohol Use Disorder (AUD) and a diagnosis is made when two or more criteria are met. It is desirable to create diagnostic short-forms that can provide a simplified measure of a disorder. We describe an empirical framework that starts with an exhaustive search to obtain subscales of all lengths and narrows down the subscales based on optimization rules performed sequentially. We choose top subscales based on bootstrapped correlations of sum scores of all subscales with the full form and among these top performers, a threshold component is incorporated to create diagnostic short-forms. Top performers are then chosen based on bootstrapped sensitivity/specificity analyses when compared to the current DSM-5 AUD rule and among these, top diagnostic short-forms are chosen based on how similarly they compare to DSM-5 AUD prevalence. Lastly, bootstrapped concordance analyses are conducted with AUD severity to highlight top performers. The NESARC used as data source.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

 
 
Copyright © American Statistical Association