Examining the DSM-5 latent structures of posttraumatic stress disorder in a national sample of student veterans

• Published in: Journal of anxiety disorders 2020-08, Vol.74, p.102262-102262, Article 102262
• Main Authors: Drake-Brooks, Malisa M., Hinkson, Kent D., Osteen, Philip, Bryan, Craig J.
• Format: Article
• Language: English
• Subjects: Adult; Confirmatory factor analysis; Diagnostic and Statistical Manual of Mental Disorders; DSM-5; Factor Analysis, Statistical; Female; Humans; Male; Middle Aged; PCL-5; PTSD; Stress Disorders, Post-Traumatic - diagnosis; Stress Disorders, Post-Traumatic - psychology; Student veterans; Students - psychology; Students - statistics & numerical data; Veterans - psychology; Veterans - statistics & numerical data; Young Adult
• ISSN: 0887-6185; 1873-7897
• DOI: 10.1016/j.janxdis.2020.102262

•Student veterans as a population are different than students or veterans alone.•1-factor comparison model is missing from most factor analysis studies.•Multi-factor models did not show significant improvement over a 4-factor model.•In the multi-factor comparative analyses, no significant improvement was found. To date, no studies have examined the latent structures of posttraumatic stress disorder (PTSD) within a sample of student veterans. To examine these constructs in a student veteran sample (n = 297), confirmatory factor analysis (CFA) was conducted on six different models of PTSD, including a one-factor model, based on the 20 symptoms found in the DSM-5; PTSD was assessed using the PCL-5. Global fit statistics suggest that fit across all models, including the 1-factor model, were good [RMSEAs(0.054−0.056); CFIs(0.928−0.940); SRMRs(0.043−0.045)], and the AIC was lowest for the seven-factor hybrid model. Statistical tests and fit guidelines for nested models suggest there is no quantitative advantage of a five, six, or seven-factor model over the existing DSM-5 four-factor model. Given the high percentage of student veterans that screened positive for a probable PTSD diagnosis (53 %) in this study compared to non-student veterans (11–20 %) and the general student population (11–15 %) found in other studies, further research is needed to assess the clinical utility of these symptoms and model structures.