Evaluating measures of combat deployment for U.S. Army personnel using various sources of administrative data

This study's purpose is to inform future research decisions about optimal measures for identifying combat deployments. We aim to evaluate four commonly utilized measures available in population-level administrative data to identify combat deployments in recent military operations among active duty Army personnel.

We compare these measures in three ways: (1) agreement (assessing the extent to which soldiers were differentially identified as combat deployed via each measure); (2) validity (calculating the sensitivity of each measure against a criterion measure); and (3) corroboration (examining how each measure predicted subsequent incidence of traumatic brain injury and post-traumatic stress disorder).

We found that using personnel records to identify deployments to Iraq, Afghanistan, and/or Kuwait captured over 98% of combat-related deployments identified via self-reported measures. The addition of Kuwait allowed for detection of nearly 100% of battle injuries, improving sensitivity from 94.5% to 99.8%. However, self-reported combat exposure measures showed the largest differential in subsequent incidence of traumatic brain injury and post-traumatic stress disorder. Completeness and accuracy of different combat deployment measures varied significantly.

Using personnel records to identify deployment to Iraq, Afghanistan, and/or Kuwait was the most valid and comprehensive measure of combat deployment. However, self-reported combat exposure measures were more predictive of combat-related outcomes.