Although network structure did not differ significantly between groups, the severe dependence group showed significantly higher global strength, suggesting a more densely connected HRQoL network, and the fundamental architecture of post-stroke HRQoL differs across levels of functional dependence.
Key Findings
Results
The global strength of the HRQoL network was significantly higher in stroke survivors with severe functional dependence compared to those with mild-to-moderate dependence.
Network Comparison Test yielded p = 0.021 for the difference in global strength between groups.
Higher global strength in the severe group indicates a more densely connected HRQoL network.
Mild-to-Moderate group: Barthel Index > 40 (n = 259); Severe group: Barthel Index ≤ 40 (n = 192).
Total sample: 451 inpatients assessed in a multi-center cross-sectional design.
Results
Network structure did not differ significantly between the mild-to-moderate and severe functional dependence groups.
Network Comparison Test for network structure yielded p = 0.928.
This indicates the overall pattern of connections among HRQoL domains was similar regardless of functional status.
Polychoric correlations were used to account for the ordinal nature of SS-QOL items.
Results
In the mild-to-moderate group, Self-Care and Social Roles were the most central nodes in the HRQoL network.
Self-Care (S7) had the highest strength centrality at 1.209 in the Mild-Moderate group.
Social Roles (S8) had the second highest strength centrality at 1.014 in the Mild-Moderate group.
Strength centrality identifies nodes with the greatest total connection strength to all other nodes in the network.
Results
In the severe functional dependence group, Upper Extremity Function and Thinking were the most central nodes in the HRQoL network.
Upper Extremity Function (S10) had the highest strength centrality at 1.124 in the Severe group.
Thinking (S9) had the second highest strength centrality at 1.031 in the Severe group.
This contrasts with the Mild-Moderate group where physical self-care and social participation dominated centrality.
Results
Thinking was the highest bridge node in the mild-to-moderate group, while Language and Thinking were the highest bridge nodes in the severe group.
Bridge expected influence for Thinking (S9) was 0.694 in the Mild-Moderate group.
In the Severe group, Language (S3) had the highest bridge expected influence at 0.722, followed by Thinking (S9) at 0.701.
Bridge expected influence was calculated based on a priori communities: Physical, Psychological, Social, and General.
Bridge nodes represent domains that most strongly connect different communities within the HRQoL network.
Results
Network stability was acceptable in both functional dependence groups as assessed by bootstrap analysis.
CS-coefficients for strength centrality were 0.595 for the Mild-Moderate group and 0.438 for the Severe group.
Both values met acceptable thresholds for network stability, supporting confidence in the centrality estimates.
Bootstrap analysis was used to assess the stability of network parameters.
Methods
The study used the Stroke-Specific Quality of Life (SS-QOL) scale across 12 domains to assess HRQoL in 451 inpatients stratified by functional dependence.
Participants were stratified using the Barthel Index: Mild-Moderate group (BI > 40; n = 259) and Severe group (BI ≤ 40; n = 192).
The study was a multi-center cross-sectional design.
Polychoric correlations were used to model relationships among the ordinal SS-QOL items.
Strength centrality and bridge expected influence were the primary network metrics calculated.
Discussion
The authors propose that findings support a precision rehabilitation framework differentiated by functional severity.
For milder cases, the framework would target the synergy between physical functions, particularly self-care and social roles.
For severe disability, the framework would focus on supporting cognitive integrity and social adaptation.
The identification of Thinking (S9) as a bridge node in both groups highlights cognition as a cross-community connector in post-stroke HRQoL.
What This Means
This research suggests that the way different aspects of quality of life are connected to each other after stroke differs depending on how severely a person is physically affected. Researchers studied 451 stroke patients in hospitals and mapped out how 12 different quality-of-life domains—such as mobility, thinking, language, mood, and social roles—relate to each other, forming a kind of web or network. They found that among patients with severe physical disability, these domains were more tightly interconnected with each other compared to patients with milder disability, meaning that changes in one area of life were more likely to ripple through and affect other areas.
The research also identified which quality-of-life domains act as the most influential 'hubs' and 'bridges' in each group. For patients with milder impairment, self-care ability and participation in social roles were the most central concerns, while for those with severe impairment, arm and hand function and cognitive abilities like thinking and language took on greater importance as connectors across different life domains. Notably, thinking ability appeared as a critical bridge between the physical, psychological, and social aspects of quality of life in both groups.
This research suggests that rehabilitation programs might benefit from being tailored to a patient's level of physical dependence rather than using a one-size-fits-all approach. For those with milder stroke effects, focusing on restoring physical self-care and social participation may have the broadest impact on overall quality of life. For those more severely affected, supporting cognitive functions—particularly thinking and language—and facilitating social adaptation may be especially important given their role as bridges connecting multiple life domains.
Ma J, Zhang Y, Zhou T, Tong L, Shi L. (2026). The architecture of quality of life after stroke: a network comparison of mild-to-moderate versus severe functional impairment.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1808859