Refining Posttraumatic Stress Disorder Diagnosis: Evaluation of Symptom Criteria With the National Survey of Adolescents

Objective: To compare the prevalence estimates, comorbidity rates, and structural validity of a revised symptom criteria set for the diagnosis of posttraumatic stress disorder (PTSD) with those of the DSM-IV criteria in a representative community sample of adolescents.

Method: Cross-sectional data from the National Survey of Adolescents, a 1995 household probability sample of 4,023 adolescents aged 12-17 years, were examined. DSM-IV PTSD symptoms were assessed with a modification of the National Women’s Study PTSD module. Three- and 4-factor DSM-IV models were compared to a 2-factor PTSD model that deleted symptoms potentially overlapping with depression or other anxiety disorders. Comorbidity was assessed using DSM-IV criteria for major depressive episodes and substance use disorders.

Results: PTSD prevalence varied across models (ie, 5.2%-8.8%, lifetime; 3.2%-5.7%, past 6 months). When the 2-factor model was used with a proportionate symptom threshold, lifetime PTSD prevalence was comparable to that with the 3-factor DSM-IV model, and major depressive episode comorbidity was reduced by 9%-14%. Comorbidity with substance use disorders was comparable across models. Structural validity, tested with confirmatory factor analyses, showed that the 2-factor model and a 4-factor DSM-IV model were superior to the DSM-IV 3-factor model.

Conclusion: Compared to the DSM-IV 3-factor PTSD model, a 2-factor model that removed depression and anxiety symptoms and used a proportionate symptom threshold may produce comparable lifetime PTSD prevalence estimates, reduced PTSD-depression comorbidity, and superior structural validity (comparable to a 4-factor PTSD model) when applied to community samples of adolescents. Further research on PTSD structure and diagnosis with adolescents is warranted.