4025 — Modified PROMIS Anger Measure for Veterans with Deployment-Related PTSD

Lead/Presenter: Vanessa Panaite, COIN - North Florida/South Georgia and Tampa
All Authors: Panaite V (James A. Haley Veterans' Hospital) Luther SL (James A. Haley Veterans' Hospital) Toyinbo PA (James A. Haley Veterans' Hospital) Marx B (VA Boston Healthcare System) Campbell T (Hunter Holmes McGuire VA Medical Center) Erbes C (Minneapolis VA Health Care System) Kisala P (University of Delaware) Tulsky D (University of Delaware)

Objectives:
To develop an item bank and short form to better measure anger in Veterans with deployment-related post-traumatic stress disorder (PTSD).

Methods:
Health-related quality of life (HRQOL) domains and constructs relevant to Veterans with deployment-related PTSD were identified using methods outlined by the NIH Patient-Reported Outcomes Measurement Information System (PROMIS) and Neuro-QOL programs. The process included input from clinicians and patients through focus groups, comparison with existing PROMIS/NeuroQOL measures, development of item pools, cognitive testing, field testing, and psychometric analyses of items. One domain/construct related to anger was identified through this process. A total of 7 new anger items were field tested along with 5 items from the existing PROMIS Anger short form. Responses to these 12 items were evaluated with Item Response Theory (IRT) methodology to calibrate and create a final short form for use with Veterans with deployment-related PTSD.

Results:
Veterans who completed the field test of Anger items (n = 538), had a mean (SD) age of 57.1 (13.4), were primarily male (n = 489, 91.0%), white (n = 400, 74.0%), with 216 (40.0%) having served in the OEF/OIF/OND conflicts or Gulf War. Dimensionality statistics from confirmatory bifactor analysis including explained common variance of the general factor (ECV = 74%; recommended minimum is 60%) and the degree of a common dimension underlying the raw scores (Omega-H = .84; recommended minimum is .7) support a unidimensional construct. Evaluation of test scale and individual item fit to graded response model identified four items for exclusion. Our final short scale is composed of 8 items (6 new and 2 PROMIS items). Compared to the existing PROMIS anger short form, the new short form provides greater information among the respondents, and a higher percentage of the total information for high Anger levels.

Implications:
The newly developed short form can be better suited to measure anger in Veterans with deployment-related PTSD, than the existing PROMIS anger short form.

Impacts:
Improved Anger measure for Veterans with deployment-related PTSD can facilitate higher quality research, program evaluation, and clinical practice related to a highly relevant domain for improved quality of life among Veterans.