Using network analysis to characterize clinical improvement during cognitive processing therapy

Cross-sectional network studies find mixed results regarding changes in network structure as a response to treatment across disorders. This study characterized improvement in mental health following Cognitive Processing Therapy (CPT) for PTSD in veterans from the perspective of network psychometrics and explored how cross-sectional networks inform our understanding of PTSD recovery.

Veterans with PTSD participated in CPT-based intensive treatment programs (ITPs), offered in two-week (N = 635) or three-week (N = 457) formats. PTSD symptoms were self-reported on the PTSD Checklist for DSM-5 (PCL-5) at pre-, mid-, and post-treatment. Cross-sectional networks for each time point were compared using network comparison tests. Linear regression tested if the relationship of initial treatment gains from admission to mid-treatment with overall outcomes was associated with the expected influence centrality of a node.

Substantial improvement in PTSD symptoms were found, but network structure remained largely unaffected, with global edge strength increasing from pre-to post-treatment. Initial treatment gains in nodes with high expected influence were associated with overall treatment outcomes. A post-hoc simulation based on a common-cause model produced similar regression results, indicating that while our findings align with spreading activation, they are not exclusive to this mechanism.

The indiscernibility of cross-sectional networks between pre- and post-treatment raises questions about whether cross-sectional networks can illuminate PTSD recovery beyond traditional measures of treatment response.