Network structure of depression and anxiety in Bangladeshi women.

PURPOSE: Depression and anxiety are highly prevalent and frequently co-occurring mental health conditions, yet their symptom-level interrelationships remain poorly understood in low-and middle-income countries. Conceptualizing mental disorders as networks of interacting symptoms, this study aimed to examine the network structure of depression and anxiety symptoms among Bangladeshi women. METHODS: This study used cross-sectional data from the 2022 Bangladesh Demographic and Health Survey, comprising 19,930 women aged 15–49 years. Sixteen symptoms from the Generalized Anxiety Disorder-7 (GAD-7) and Patient Health Questionnaire-9 (PHQ-9) were analyzed. A Gaussian graphical model was estimated using the Extended Bayesian Information Criterion graphical lasso (EBICglasso). Centrality and bridge centrality indices were computed to identify highly connected symptoms and cross-domain connectivity, respectively. Network stability and edge weight accuracy were evaluated using bootstrapping procedures. RESULTS: The network demonstrated dense within-domain connectivity for both anxiety and depression symptoms, alongside multiple cross-domain conditional associations. Trouble relaxing and nervousness were among the most central anxiety symptoms, while sad mood, psychomotor symptoms, and suicidal ideation were highly central within the depression subnetwork. Bridge centrality identified feeling afraid, irritability, and nervousness as prominent anxiety-related bridge symptoms, and sad mood, anhedonia, psychomotor symptoms, and suicidal ideation as key depressive symptoms linking the two domains. The network showed excellent stability (CS-coefficient = 0.75). CONCLUSION: This study provides the first nationally representative network analysis of anxiety and depression symptoms among Bangladeshi women. The findings highlight distinct roles of symptoms in overall connectivity and cross-domain linkage, offering a descriptive framework for understanding comorbidity patterns in resource-limited settings.