Development of a Tool to Predict Risk of Behavioral Health Evacuation From Combat

In the present study, we examined administrative data on 667,437 deployments of at least 30 days duration to Iraq and Afghanistan from 2011 through the end of 2016 to determine risk factors for evacuation from the combat zone for behavioral health reasons. Demographic data, military-specific data, responses on predeployment mental health assessments, and presence of previous treatment for psychiatric conditions were entered into a logistic regression based on expert determination, distinguishing the 2,133 behavioral health evacuations from those deployments that either did not end in evacuation or included evacuations for reasons other than behavioral health. The model, derived from a random half of the sample (training set), was verified on the other half (validation set). Predictor variables used in the model were calendar year; gender; age; rank; marital status; parental status; number of prior war zone deployments; branch of service; screens for symptoms of posttraumatic stress disorder, depression, and hazardous alcohol use on the predeployment mental health assessment; and prior substance- and non-substance-related behavioral health diagnoses. Odds ratios (range: 1.05-3.85) for selected variables that contributed to the model were used to assign risk scores in the Behavioral Health Evacuation Risk Tool, which can aid predicting which service members are more likely to be evacuated from combat for behavioral health reasons, thus indicating where resources can be allocated for behavioral health referrals and war zone care.