Network analysis of symptom severity and symptom distress in hospitalized patients with decompensated cirrhosis

This study aimed to explore the network characteristics of symptom severity and symptom distress in hospitalized patients with decompensated cirrhosis, conduct an exploratory comparison of the two networks, and analyze factors associated with the occurrence of the symptom ranked highest in centrality within each network to inform clinical symptom management. Using convenience sampling, we recruited 459 hospitalized patients with decompensated cirrhosis from a tertiary Grade A hospital in China between July 2025 and January 2026. Symptoms were assessed using the Chinese version of the Memorial Symptom Assessment Scale, and the correlation between symptom severity and symptom distress was assessed using Spearman’s rank correlation analysis. Symptom severity and symptom distress networks were constructed separately, with node centrality evaluated using expected influence. Network differences were assessed using the Network Comparison Test, and factors associated with the occurrence of the symptom ranked highest in centrality within each network were examined using binary logistic regression. Patient-level mean severity and distress scores across the 16 symptoms included in the primary networks were strongly positively correlated (Spearman’s ρ = 0.957). No statistically significant differences were detected between the two networks in global strength or overall network structure. “Feeling sad” had the highest expected influence in the symptom severity network, while “I don’t look like myself” had the strongest expected influence in the symptom distress network. The occurrence of “Feeling sad” was associated with the Model for End-Stage Liver Disease score category and per capita monthly household income. The occurrence of “I don’t look like myself” was associated with the Model for End-Stage Liver Disease score category, per capita monthly household income, and primary caregiver. “Feeling sad” and “I don’t look like myself” ranked first in centrality within their respective networks. When considered alongside patients’ actual symptom burden and the patient characteristics associated with the occurrence of these two symptoms, the findings of this study can provide a reference for clinical symptom management.