Anxiety Symptoms as the Central Node : A Network Analysis of Mental Health Sequelae in a Post-COVID-19 National Sample

Zhitao Yuan^1,2*Xianyang Wang²Mengyuan Yang²Lixin Zhao²Chaoxian Wang³Xiangyu Ren⁴Lei Ren²Na Ni^1*Shengjun Wu^2*

• ¹Shaanxi University of Chinese Medicine, Xianyang, China
• ²Air Force Medical University, Xi'an, China
• ³Naval Medical University, Shanghai, China
• ⁴People's Liberation Army Unit 96607, baoji, China

Background: The COVID-19 pandemic has precipitated a global mental health crisis, with the long-term effects of Post-COVID-19 syndrome presenting a co mplex interplay of physical and psychological symptoms. However, the dynami c network interrelationships between social support, perceived stress, and genera l mental health and how these networks vary across demographic groups remai n inadequately explored, hindering the development of targeted interventions. Methods: To address this gap, we constructed an interactive network model usi ng national survey data (N=7,997). A Gaussian Graphical Model (GGM) was e mployed to estimate partial correlations among variables, and network comparis on tests were conducted to examine structural differences across gender, income, residence, and marital status. Results: Anxiety and insomnia symptoms (GHA, from the General Health Ques tionnaire) exhibited the highest strength centrality in the network, indicating the ir position as the most statistically interconnected node. A strong negative corre lation was identified between social support (SSS) and feelings of helplessness (PLC, from the Perceived Stress Scale), underscoring the potential protective ro le of social support. Notably, network invariance tests revealed significant struct ural variations across demographics with meaningful effect sizes. Women showe d stronger stress-depression connections, whereas men exhibited a stronger nega tive link between social support and helplessness. Low-income groups demonstr ated tighter anxiety-depression connectivity; distinct network topologies were ob served between urban and rural residents, and marital status differentially influe nced the prominence of tension-related versus helplessness-support connections. Conclusion: Mental health issues in the post-pandemic era exhibit distinct netw orked characteristics, with social support serving as a key buffer against stress. The substantial variation in symptom networks across demographic lines under scores the necessity of developing tailored, precision interventions for specific populations.