Sleep components and HRQoL dimensions formed a closely connected network in older adults with hypertension, with usual activities showing the highest centrality and sleep efficiency and subjective sleep quality showing prominent cross-community bridge roles.
Key Findings
Results
The component- and dimension-level network between sleep quality and HRQoL included 10 nodes and 32 non-zero edges among 45 possible edges, with a density of 0.711.
The network was estimated using EBICGlasso-regularized Gaussian graph model.
The sample comprised 2,257 older adults with hypertension.
Sleep quality was assessed with the B-PSQI, HRQoL with the EQ-5D-5L, and depressive symptoms with the PHQ-9.
Network density of 0.711 indicates a high proportion of non-zero edges among all possible connections.
Results
The strongest pairwise associations in the network were between sleep efficiency and sleep duration, sleep interruption and subjective sleep quality, and self-care and usual activities.
Sleep efficiency and sleep duration (SE-ST) had the strongest edge weight at 0.793.
Sleep interruption and subjective sleep quality (SW-SQ) had an edge weight of 0.583.
Self-care and usual activities (SC-UA) had an edge weight of 0.560.
These were the top three strongest associations among the 32 non-zero edges in the network.
Results
Usual activities (UA) showed the highest strength centrality and expected influence among all nodes in the network.
Usual activities was the EQ-5D-5L dimension with the highest centrality metrics.
The centrality ranking after usual activities was: subjective sleep quality (SQ), self-care (SC), and sleep efficiency (SE).
Centrality indices demonstrated good stability with a case-dropping subset bootstrap stability coefficient (CS) of 0.75.
Higher centrality suggests these nodes have stronger and more numerous connections to other nodes in the network.
Results
Sleep efficiency had the highest bridge expected influence in the network, followed by subjective sleep quality.
Bridge centrality measures a node's role in connecting different communities (sleep components vs. HRQoL dimensions) within the network.
Sleep efficiency and subjective sleep quality showed prominent cross-community connections.
These findings suggest these sleep components may be particularly relevant to HRQoL dimensions in older adults with hypertension.
The authors note these domains 'may be useful for screening or hypothesis generation but should not be interpreted as confirmed intervention targets.'
Results
No significant differences were found in network structure or global strength between the depressive-symptom and non-depressive-symptom groups.
Network structure comparison yielded M = 0.224, P > 0.05.
Global strength comparison yielded S = 0.119, P > 0.05.
Repeated 1:1 subsampling broadly supported these findings of no significant group differences.
Depressive symptoms were assessed using the PHQ-9.
Methods
The study used a network analysis approach rather than traditional total-score comparisons to examine interactions between specific sleep quality components and HRQoL dimensions.
Previous studies had mainly examined sleep-HRQoL associations using total scores, providing limited insight into interactions between specific symptom dimensions.
The EBICGlasso regularization method was applied to estimate the Gaussian graphical model.
Bootstrap methods were used to validate stability of centrality indices.
The network included nodes from both the B-PSQI sleep quality components and EQ-5D-5L health dimensions.
What This Means
This research suggests that in older adults with high blood pressure, poor sleep and reduced quality of life are not just loosely related — they form a tightly interconnected web of specific symptoms and experiences. Using a statistical technique called network analysis, the researchers mapped out how specific aspects of sleep (like sleep efficiency, sleep duration, and subjective sleep quality) connect to specific aspects of health-related quality of life (like ability to care for oneself and perform usual daily activities) in a group of 2,257 older adults. Rather than simply asking whether 'sleep' and 'quality of life' are related in general, this approach reveals which particular features are most central or act as bridges between the two domains.
The study found that 'usual activities' — a person's ability to carry out their normal daily tasks — was the most central node in the network, meaning it had the strongest and most connections to other symptoms. Sleep efficiency (how much of time in bed is actually spent sleeping) and subjective sleep quality (how a person feels about their own sleep) acted as key bridge nodes, linking the sleep-related side of the network to the health-quality-of-life side. Interestingly, the network looked similar regardless of whether participants had depressive symptoms or not, suggesting these sleep-health connections may be consistent across this population.
This research suggests that targeting specific sleep components — particularly sleep efficiency and subjective sleep quality — might be relevant to understanding how sleep problems relate to broader health and functioning in older adults with hypertension. The authors caution that these findings, while useful for identifying areas of interest and for generating new research questions, should not yet be interpreted as confirmed targets for clinical intervention. The network approach itself is relatively new, and more research would be needed before drawing firm clinical conclusions.
Sun X, Guo Y, Ding B, Hu J. (2026). Component- and dimension-level network associations between sleep quality and health-related quality of life in older adults with hypertension.. Frontiers in public health. https://doi.org/10.3389/fpubh.2026.1890897