Novel Network Analysis of County- and Individual-Level Factors Associated With Functional Outcomes After Stroke

Abstract

Background: Social determinants are known to impact stroke risk and poststroke outcomes. Using complexity science, we examined interrelations between county- and individual-level social and clinical determinants influencing stroke functional outcomes.

Methods: We examined a retrospective cohort of 2 961 664 patients diagnosed with acute ischemic or hemorrhagic stroke from 2218 US hospitals participating in the Get With The Guidelines-Stroke Registry from 2013 to 2019, linked by ZIP code with the county-level institute for health metrics and evaluation data. We constructed multilayer networks, estimating mixed graphical models of 32 nodes representing social and clinical factors. Networks included 4 layers of factors: (1) county-level social, (2) individual-level social, (3) clinical comorbidities, and (4) hospital encounters. Networks were estimated for patients with less favorable (modified Rankin Scale score 3-6) versus favorable (modified Rankin Scale score 0-2) outcomes. We compared network structure and node centrality measures between groups using bootstrap permutation analyses, identifying influential (hub) nodes.

Results: The overall influence of social determinants (global connectivity) was greater in patients with less favorable outcomes (P<0.001). Homelessness and Black race were hub nodes, indicating their role in mediating relationships between social and downstream clinical factors in patients with less favorable outcomes. Being uninsured had greater influence (closeness centrality; P<0.001) in patients with less favorable outcomes, indicating its role in amplifying the effects of social determinants. Greater county-level high school completion (P<0.001) and a lower proportion of the population living below the US poverty line (P=0.030) were directly associated with faster onset-to-arrival time in patients with less favorable functional outcomes. The clinical-social determinant network explained 34% of the variance of modified Rankin Scale scores.

Conclusions: Social determinants have a substantial influence on functional outcomes after stroke. County-level poverty directly affected onset-to-arrival time and quality of care. Health insurance status and homelessness were influential and modifiable patient-level factors that may serve as critical leverage points for future interventions aimed at improving outcomes.

Keywords: cause of death; longevity; poverty; social determinants of health; stroke.