Comparative Analysis of Comorbidity Networks of Anxiety and Depression in Elderly Individuals with and without a History of Falls

Traditional psychopathological views symptoms as manifestations of latent disorders, while recent perspectives reconceptualize mental health as multidimensional continuums. The network analysis model elucidates dynamic symptom interactions via “comorbidity networks”, where co-occurring symptom clusters form nodes driving disorders. Although falls are associated with exacerbated anxiety and depression in older adults, prior research lacks network analysis examining associations between fall history and anxiety-depression comorbidity patterns. This study compared anxiety-depression comorbidity network structures between older adults with and without fall history to identify potential differences in symptom interactions. 371 elderly participants (169 with fall history, 202 without; mean age 81.00 ± 8.12, $60.4 \%$ male) were recruited from 10 Anhui elderly care institutions via convenience sampling. Data were collected via one-on-one interviews using GAD-7 (Cronbach’s $\alpha=0.930$) and PHQ$9(\alpha=0.849)$. Network analysis via EBICglasso in R assessed global structure, edge weights, node strength centrality, and robustness. Results showed fall history correlated with higher anxiety and depression scores $(p\lt 0.05)$. Global network structure and connectivity did not differ, but specific edges and node centrality varied. In the fall group, “uncontrollable worries” and “inability to relax” (anxiety) were core nodes; in the non-fall group, “excessive worries” (anxiety) and “inferiority” (depression) were central. Significant edge differences included “inferiority-suicidal symptoms” ($p=0.023$) and “excessive worries-irritability” ($p=$ 0.007). Network robustness was confirmed (CS-C $\gt 0.5$). Findings demonstrate distinct symptom interaction patterns associated with fall history despite preserved global architecture. Symptom-specific intervention targets “uncontrollable worries” for the fall group and “inferiority” for the non-fall group may optimize network-based approaches to elderly mental health management.