Toward the Narrow Approach to Posttraumatic Stress Disorder Diagnostic Criteria: An Item Response Theory Analysis

Abstract

A crucial debate in the field of traumatic stress involves the question of whether posttraumatic stress disorder (PTSD) is better represented by a broad or a narrow approach to establishing the set of symptom criteria (Stein et al., 2014). The broad approach, exemplified by the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5; APA, 2013), includes a wide range of trauma-related symptoms, regardless of whether they overlap with other disorders. Conversely, the narrow approach, exemplified by the International Classification of Diseases, 11th Edition (ICD-11; WHO, expected release 2018) retains only a limited set of what are argued to be core symptoms specific to PTSD (Brewin, 2013; Maercker et al., 2013; Resick & Miller, 2009).. Although the ICD-11 workgroup narrowed symptom criteria using theory, empirical research, and clinical judgment, it remains empirically unclear whether the retained symptoms are truly the core PTSD symptoms. Item response theory (IRT), a statistical technique that examines each symptom’s relative contribution to a construct, is a powerful tool that can inform PTSD symptom selection for the narrow approach. Although IRT studies on PTSD measures exist, no firm conclusions can be drawn about the core symptoms due to the over-restrictiveness of the models employed, variations in measures and populations examined, and the change from DSM-IV-TR (APA, 2000) to DSM-5 (APA, 2013) criteria. To empirically address the question of which items represent the core PTSD symptoms, IRT was employed to examine item difficulty and item discrimination parameters. Undergraduates who experienced a DSM-5 Criterion A event completed the LEC-5 and PCL-5. Physiological reactivity, internal avoidance, persistent negative emotional state, detachment from others, and concentration and sleep difficulties emerged as the most discriminating symptoms within each DSM-5 symptom cluster. Importantly, this list only has one symptom in common with ICD-11 PTSD criteria, suggesting that, in general, the symptoms retained for ICD-11 are not in fact the most discriminating. Future research should employ IRT in a clinical population.