Component- and dimension-level network associations between sleep quality and health-related quality of life in older adults with hypertension.
Older adults with hypertension commonly experience poor sleep quality and reduced health-related quality of life (HRQoL). Previous studies have mainly examined these associations using total scores, providing limited insight into interactions between specific symptom dimensions. This study aimed to construct a component- and dimension-level network linking sleep quality and HRQoL and to compare network differences across depressive-symptom status.
This study included 2,257 older adults. Sleep quality, HRQoL, and depressive symptoms were assessed using the B-PSQI, EQ-5D-5L, and PHQ-9, respectively. We estimated a component- and dimension-level network based on the EBICGlasso-regularized Gaussian graph model, calculated centrality and bridge centrality metrics, validated stability using the bootstrap method, and compared network differences between the depressive and non-depressive-symptom groups.
The final network included 10 nodes and 32 non-zero edges among 45 possible edges, with a density of 0.711. The strongest associations were between sleep efficiency and sleep duration (SE-ST, 0.793), sleep interruption and subjective sleep quality (SW-SQ, 0.583), and self-care and usual activities (SC-UA, 0.560). Usual activities (UA) showed the highest strength centrality and expected influence, followed by subjective sleep quality (SQ), self-care (SC), and sleep efficiency (SE). Sleep efficiency had the highest bridge expected influence, followed by subjective sleep quality. Centrality indices demonstrated good stability (CS = 0.75). No significant differences were found in network structure (M = 0.224, P > 0.05) or global strength (S = 0.119, P > 0.05); repeated 1:1 subsampling broadly supported these findings.
Sleep components and HRQoL dimensions formed a closely connected network in older adults with hypertension. Usual activities showed the highest centrality, while sleep efficiency and subjective sleep quality showed prominent cross-community. These domains may be useful for screening or hypothesis generation but should not be interpreted as confirmed intervention targets.
This study included 2,257 older adults. Sleep quality, HRQoL, and depressive symptoms were assessed using the B-PSQI, EQ-5D-5L, and PHQ-9, respectively. We estimated a component- and dimension-level network based on the EBICGlasso-regularized Gaussian graph model, calculated centrality and bridge centrality metrics, validated stability using the bootstrap method, and compared network differences between the depressive and non-depressive-symptom groups.
The final network included 10 nodes and 32 non-zero edges among 45 possible edges, with a density of 0.711. The strongest associations were between sleep efficiency and sleep duration (SE-ST, 0.793), sleep interruption and subjective sleep quality (SW-SQ, 0.583), and self-care and usual activities (SC-UA, 0.560). Usual activities (UA) showed the highest strength centrality and expected influence, followed by subjective sleep quality (SQ), self-care (SC), and sleep efficiency (SE). Sleep efficiency had the highest bridge expected influence, followed by subjective sleep quality. Centrality indices demonstrated good stability (CS = 0.75). No significant differences were found in network structure (M = 0.224, P > 0.05) or global strength (S = 0.119, P > 0.05); repeated 1:1 subsampling broadly supported these findings.
Sleep components and HRQoL dimensions formed a closely connected network in older adults with hypertension. Usual activities showed the highest centrality, while sleep efficiency and subjective sleep quality showed prominent cross-community. These domains may be useful for screening or hypothesis generation but should not be interpreted as confirmed intervention targets.