Longitudinal PTSD network structure: Measuring PTSD symptom networks over a five year period

Background: Network modeling has been applied in a range of trauma exposed samples, yet results are limited by an over reliance on cross-sectional data. The current analyses used posttraumatic stress disorder (PTSD) symptom data collected over a five-year period to estimate a more robust between-subject network and an associated symptom change network. Methods: A PTSD symptom network is measured in a sample of military veterans across four time points (Ns = 1254, 1231, 1106, 925). The repeated measures permits isolating between-subject associations by limiting effects of within-subject variability. The result is a highly reliable PTSD symptom network. A symptom slope network depicting covariation of symptom change over time is also estimated. Results: Negative trauma-related emotions had particularly strong associations with the network. Trauma-related amnesia, sleep disturbance, and self-destructive behavior had weaker overall associations with other PTSD symptoms. Conclusions: PTSD’s network structure appears stable over time. There is no single “most important” node or node cluster. The relevance of self-destructive behavior, sleep disturbance, and trauma-related amnesia to the PTSD construct may deserve additional consideration.