A Pragmatic Approach to Psychometric Comparisons between the DSM-IV and DSM-5 Posttraumatic Stress Disorder (PTSD) Checklists in Acutely Injured Trauma Patients

Psychiatry. 2020 Aug 6;1-12. doi: 10.1080/00332747.2020.1762396. Online ahead of print.

Abstract

Objective: Prior investigations suggest the relative equivalence of the DSM-IV and DSM-5 versions of the Posttraumatic Stress Disorders (PTSD) Checklist, yet no investigations have assessed the psychometric properties of the scales in pragmatic trials of acutely injured trauma survivors.

Method: DSM-IV and DSM-5 versions of the PTSD Checklist were included in follow-up interviews of physically injured patients enrolled longitudinally in a pragmatic clinical trial; pragmatic trials aim to efficiently implement research procedures to inform healthcare system policy changes. Psychometric comparisons of the DSM-IV, DSM-5, and a 20-item blended version of the scale included evaluations of internal consistency, correlational assessments, evaluation of item level agreements, and estimation of DSM-5 cutoffs that optimize electronic health record screening protocols.

Results: 128 patients were included in the pragmatic psychometric study. Cronbach's alphas for the 3 versions of the PTSD Checklist ranged from 0.93 to 0.95. Correlations between the 3 scales ranged from 0.79 to 0.99. All 3 measures demonstrated excellent convergent and discriminant properties. Item level agreement ranged from 70-80%. For the DSM-5 and blended versions of the scale, a score of 30 and 24, respectively, best approximated the DSM-IV cutoff of ≥35 that had previously optimized PTSD detection in conjunction with EHR screening.

Conclusions: Among injured trauma survivors, the psychometric performance of the DSM-IV PTSD Checklist with the addition of the 4 new DSM-5 PTSD Checklist items is nearly equivalent to the DSM-5 PTSD Checklist. The investigation also suggests that pragmatic psychometric methods can catalyze the rapid translation of research findings into real-world practice settings.