Assessing posttraumatic stress in military service members: improving efficiency and accuracy

Posttraumatic stress disorder (PTSD) is assessed across many different populations and assessment contexts. However, measures of PTSD symptomatology often are not tailored to meet the needs and demands of these different populations and settings. In order to develop population- and context-specific measures of PTSD it is useful first to examine the item-level functioning of existing assessment methods. One such assessment measure is the 17-item PTSD Checklist-Military version (PCL-M; Weathers, Litz, Herman, Huska, & Keane, 1993). Although the PCL-M is widely used in both military and veteran health-care settings, it is limited by interpretations based on aggregate scores that ignore variability in item endorsement rates and relatedness to PTSD. Based on item response theory, this study conducted 2-parameter logistic analyses of the PCL-M in a sample of 196 service members returning from a yearlong, high-risk deployment to Iraq. Results confirmed substantial variability across items both in terms of their relatedness to PTSD and their likelihood of endorsement at any given level of PTSD. The test information curve for the full 17-item PCL-M peaked sharply at a value of θ = 0.71, reflecting greatest information at approximately the 76th percentile level of underlying PTSD symptom levels in this sample. Implications of findings are discussed as they relate to identifying more efficient, accurate subsets of items tailored to military service members as well as other specific populations and evaluation contexts.