Category: Military and Veterans Psychology

PS5- #A13 - A Factor Analytic Evaluation of the WHO Disability Assessment Schedule 2.0 With Veterans Seeking Mental Health Services

Friday, Nov 17
1:30 PM – 2:30 PM
Location: Indigo Ballroom CDGH, Level 2, Indigo Level

Keywords: Veterans | Psychometrics | Assessment

Functional impairments are critical components for the establishment of most mental health diagnoses. The Diagnotic and Statistical Manual of Mental Disorders (DSM-5) removed the relatively subjective Global Assessment of Functioning score and currently recommends a dimensional approach to evaluating the functional impact of symptoms. As such, it has advocated for the use of scales such as the World Health Organization Disability Assessment Schedule 2.0 (WHODAS 2.0) to aid this purpose. However, the factor analytic properties of the 36-item WHODAS 2.0 vary in robustness when applied to populations with significant mental health symptoms and the survey has yet to be evaluated for use with Veterans. To this end, we conducted an investigation of the factor structure based on data provided by 464 Veterans seeking mental health services at a generalist mental health outpatient service at a Midwestern VA. Participant data was split into two datasets of 232 cases (Samples 1 and 2). Responses from our first sample were submitted to an exploratory factor analysis (EFA). In the second analytic step, our EFA-derived model was submitted to Confirmatory Factor Analysis (CFA). An additional CFA was conducted with data from the second sample using the six-factor model structure previously investigated with mental health populations. Results of our EFA revealed that a seven-factor solution provided best fit to the data. Of note, the accepted six-factor model did not result in close fit based on evaluation of chi-square alone (χ2= 5821.816, p < .001). Following recommended modification indices permitting correlating pairs of error terms, the CFA found the seven-factor model derived from the EFA demonstrated adequate fit (χ2 = 742.07, p < .001; RMSEA = .055). Zero-order correlations among latent variables in this model ranged from .433 to .795. Next, the six-factor model based on research with non-Veteran mental health participants was evaluated with CFA using identical procedures. However, neither the initial (χ2 = 1962.80, p <.001; RMSEA = .102 or modfied (χ2 = 1165.93, p < .001; RMSEA = .067) version of this model represented adequate fit to the data. Because our developed model was not nested within the original model, chi-square difference testing could not be conducted to compare fit. Rather, model AIC values were compared with a difference of 10 or greater considered indicative of superior fit. This comparison (EFA-derived AIC = 922.07; original model AIC = 1351.93) confirmed that the seven-factor model represented a significantly better fit. This research offers an empirical contrast of factor analytic approaches with mental healthcare utilizers and yields important information regarding using the WHODAS 2.0 with Veterans.

Minden B. Sexton

Women Veterans Mental Health Coordinator
Ann Arbor Veterans Healthcare System
Ann Arbor, Michigan

Margaret T. Davis

Postdoctoral Fellow
Yale School of Medicine

Rebecca K. Lusk

Staff Psychologist / Assistant Chief of Mental Health Clinic
Ann Arbor Veterans Healthcare System
Ann Arbor, Michigan

Mark Lyubkin

Ann Arbor Veterans Healthcare System

Stephen T. Chermack

Ann Arbor Veterans Healthcare System