A network analysis and simulated intervention study of perceived stress in trauma-exposed adults with autism spectrum disorder

This study employed network analysis to construct a perceived stress network among Chinese adults with autism spectrum disorder (ASD) who had experienced interpersonal trauma, identify core nodes, compare network structures between high and low cumulative trauma exposure groups, and explore potential intervention targets through simulated interventions. A convenience sampling method was used. Electronic questionnaires were distributed through a WeChat public platform between April 2025 and May 2026. A total of 459 Chinese adults with ASD who reported at least one type of interpersonal trauma were included. All participants completed the 14-item Perceived Stress Scale (PSS-14) and a self-reported trauma exposure checklist. After controlling for demographic covariates, a regularized partial correlation network was estimated. Expected Influence (EI) was used to assess node centrality. Network stability was evaluated using a case-dropping bootstrap procedure. The Network Comparison Test (NCT) was used to compare the global strength and edge weights between high- and low-trauma exposure groups. Simulated intervention analyses were conducted on binarized Ising network models to identify potential aggravation and alleviation targets. The network demonstrated high stability (CS-coefficient = 0.749) and density (0.604). The three core nodes with the highest EI were “inability to control important things” (P2, EI = 1.168), “confidence in handling personal problems” (P6, EI = 1.091), and “unable to overcome difficulties” (P14, EI = 1.006).Network comparison analyses detected no significant differences in global strength or network structure between the high- and low-trauma groups; however, the study was underpowered to detect small-to-moderate differences, and these null findings should be interpreted as inconclusive rather than as evidence of structural invariance. Exploratory simulated intervention analyses suggested that, in the primary analysis, P4 (“successful handling of hassles”, Δ = +0.737) and P10 (“sense of mastery”, Δ = −1.806) emerged as potential aggravation and alleviation targets, respectively; however, target identification was not stable across perturbation magnitudes and binarization cut-offs. In the strict analysis, the aggravation target was P6 (“confidence in handling personal problems”, Δ = +1.870) and the alleviation target was P2 (“inability to control important things”, Δ = −1.454), which differed from the primary analysis. These targets should therefore be regarded as exploratory statistical predictions rather than stable clinical intervention targets. Nodes related to perceived control and self-efficacy were central in the perceived stress network of trauma-exposed adults with ASD. Exploratory simulated intervention analyses tentatively pointed to problem-solving capacity and sense of control as potential themes; however, the identified targets varied across analytic parameters, and the network comparison was underpowered to detect small-to-moderate differences. These findings should be considered preliminary and require validation through longitudinal designs and more refined trauma measurements before any clinical inference can be made.