T. D. Klepp1,2, J. P. Herrera-Escobar1, E. Stanek1, J. Barrett1,2, H. M. Kaafarani1, S. E. Sanchez1,2, A. Salim1, D. Nehra1, N. Levy-Carrick1,3 1Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 2Boston University School of Medicine,Boston, MA, USA 3Brigham And Women’s Hospital,Ambulatory Clinical Services, Dept Psychiatry,Boston, MA, USA
Background: Screening tools and prevalence rates of Post Traumatic Stress Disorder (PTSD) across study populations vary significantly. The transition from DSM-IV to DSM-V criteria has, moreover, expanded symptom clusters. We aim to clarify salient PTSD symptoms identified by two screening tools and assess their relative utility in identifying symptoms implicated in long-term functional impairment following traumatic injury.
Methods: Trauma patients (≥18 years) with moderate-to-severe injuries (Injury Severity Score ≥9) admitted to three Level-I trauma centers were contacted 6- and 12-months post-hospitalization. PTSD screening was administered using either the Breslau et al DSM-IV screen during the first study phase (2015-2018; n = 1388), or the abbreviated 8-item PTSD Checklist for DSM-V (PCL-5) during the second study phase (2018-2019; n = 428). Chi-squared tests and multivariate analysis was used to compare PTSD prevalence between the two screening methods. Diagnostic criteria was tested for internal consistency using Chronbach’s alpha.
Results: Cohort demographics revealed a greater percentage of male participants in the second phase (p = 0.012), with no significant difference in other patient characteristics (age, insurance, injury severity score, injury type, length of stay or previous psychiatric illness). Probable PTSD was identified in 19.2% of respondents using the Breslau PTSD compared with 8.2% of those using the PCL-5 screen (adjusted p < 0.001). Subsequent analysis separated questions by symptom criteria comparing Likert-based responses from PCL-5 with binary (yes-no) responses from the Breslau screen (Table 1). Across each symptomatic category, responses of “moderately agree” significantly differed from Breslau outcomes although “somewhat agree” did not. Preliminary analysis of each symptomatic category showed low internal consistency for both scales when measuring negative alterations in mood and hyperarousal symptoms. Within negative alterations in mood, respondents were more likely to endorse having lost interest and isolation but less likely to endorse having negative beliefs or future-oriented thinking in both the Breslau scale and PCL-5.
Conclusion: We observed discrepant rates of positive PTSD identification between two commonly used measures. Variability is likely multifactorial, including both a shift from binary responses to Likert-based responses and differences in scope of symptoms identified. We also observed specific items within the negative alterations in mood that more likely to be present. This has important implications for optimizing both screening and treatment planning to address the most impairing symptom clusters impacting survivors of traumatic injury.
