Talk to the Veterans Crisis Line now
U.S. flag
An official website of the United States government

VA Health Systems Research

Go to the VA ORD website
Go to the QUERI website
2015 Conference Logo

2015 HSR&D/QUERI National Conference Abstract

1041 — Cross Validation of the Veterans Mental Health Risk Screen

Carlson EB, National Center for PTSD; Palmieri PA, Summa Health System; Eisen S, Boston University; Vogt D, National Center for PTSD;

Objectives: Many military veterans who have PTSD and depression symptoms following military service do not seek or accept care. A screen for risk of persisting mental health problems following military service would allow providers to direct their efforts to those most in need and may persuade some reluctant veterans to engage in mental health care.

Methods: To identify risk factors for chronic PTSD and depression, we conducted analyses on data from a prospective study of military personnel recently returned from deployment, a sample of Gulf War veterans studied 10 years after discharge, and a prospective study of adults exposed to traumatic injury. We also collected new data from veterans seeking VA primary care services on trauma exposure and on stress in relationship related to transition from military to civilian life. The top seven risk factors that were strongly related to PTSD in the datasets studied (r > .50), relevant to all veterans, and measurable with only a few items included: current PTSD symptoms, relationship stress, social constraints on disclosure, posttraumatic negative cognitions, social support, psychotic symptoms, and dissociation. To cross-validate a screen assessing these risks, we then collected new data from 223 veterans coming to VHA for primary health care on these risk factors (using 15 items) and on symptoms of PTSD and depression 6 months later.

Results: The prototype screen showed sensitivity of .85 and specificity of .81 in predicting elevated PTSD. In a subsample of 64 veterans who served in Iraq or Afghanistan, sensitivity was a bit higher (SE = .89), and specificity was lower (SP = .70). In both groups, sensitivity to predict later PTSD was similar for 10 current PTSD symptoms, but specificity was lower (.75 in all veterans and .62 in Iraq and Afghanistan veterans).

Conclusions: Accurate prediction of chronic mental illness in veterans new to VA appears to be possible, and risk factor screening may produce fewer false positives than screening for current symptoms.

Impact: Information about risk for chronic mental illness of veterans seeking primary care services at VA may be helpful to veterans who are not inclined to seek mental health care.