(PS10-34) Reevaluating PTSD in Dimensional Models of Psychopathology: Differential Prediction of PTSD Symptom Domains by Latent Negative Affect and Detachment Factors

Abstract Text:

Background: Hierarchical dimensional models of psychopathology (e.g., HiTOP, Kotov et al., 2017; DSM-5 AMPD) organize psychiatric phenotypes into higher-order factors including Internalizing/Negative-Affect, Detachment, Antagonism, Disinhibition, and Psychoticism, allowing for more targeted assessment and intervention. To date, PTSD has largely been classified within the Internalizing dimension alone. However, supporting research typically involves diagnosis-level analysis despite significant heterogeneity of PTSD symptom presentation, spurring recent calls for incorporating symptom-level data into dimensional nosology research. The current study evaluates whether latent dimensions of four PTSD symptom domains –Intrusion, Avoidance, Cognition/Mood Alterations, and Hyperarousal – are differentially predicted by latent dimensions of negative affect (NA) and detachment (DET).

Method: Participants were 583 trauma-exposed individuals recruited via MTurk (61.4% female; 76.8% White). In Mplus, confirmatory factor analysis (CFA) was used to model the four PTSD symptom domains using PCL-5 items as indicators. Exploratory structural equation modeling (ESEM) was used to model latent NA and DET using existing scale measures (NA indicators: PANAS-NA, PROMIS-Anxiety, and ASI-3; DET indicators: INQR Perceived Burdensomeness and Thwarted Belongingness subscales, PROMIS-Depression, and PMBS Low Self-Worth subscale). Bivariate correlations between all latent variables were evaluated within the hybrid ESEM-CFA model, and subsequently, each of the latent PTSD symptom domains were regressed simultaneously onto latent NA and DET factors.

Results: The hybrid ESEM-CFA measurement model fit the data well; CFI=.920, TLI=.907, RMSEA=.08, SRMR=.04. At the bivariate level, both latent NA and DET were positively associated with all four latent PTSD domains. In multiple regression models including both latent NA and DET as predictors, latent NA similarly predicted all four latent PTSD symptom domains (Bs Intrusion=.59, Avoidance=.52, Cognition/Mood Alterations=.34, Hyperarousal=.62, ps < .001). In contrast, latent DET predicted Cognition/Mood Alterations (B=.46, p < .001) and Hyperarousal (B=.14, p < .05), though it was no longer a significant predictor of Intrusion or Avoidance.

Conclusions: The latent Negative-Affect factor broadly predicted all latent PTSD symptom domains, consistent with traditional classifications that place PTSD exclusively within the Internalizing/Negative-Affect dimension. However, the current findings suggest that higher-order Detachment also uniquely predicts alterations in cognition and mood, as well as hyperarousal, highlighting the importance of examining PTSD at the symptom level within dimensional models of psychopathology. Future research should explore the degree to which PTSD symptoms differentially map onto other higher-order factors, such as Antagonism, Disinhibition, and Psychoticism. In doing so, dimensional models may better account for the heterogeneity of PTSD symptom expression and, in clinical practice, better inform both case conceptualization and treatment planning.