Investigating stroke risk factors: A cross-cultural network analysis

Abstract

Stroke is the second leading cause of death and a major source of disability, with well-documented, complex risk factors. While stroke risk factors have been studied in their respective fields, research on how biological, psychological, and social factors interact remains limited, and no studies have explored whether patterns vary across ethnic groups. This study used network analysis to explore the unique associations and directional probabilities among known stroke risk factors across different ethnic groups. Data from 15,460 participants across 134 countries were collected via the Stroke Riskometer mobile application, classified into six ethnic groups (White/European, African, Asian, Indian, Latin American/Hispanic, Other). Bayesian Gaussian Graphical Models (BGGM) showed that 78-84% of associations were stable across ethnicities, indicating that most stroke risk relationships are consistent across cultures, with some unique differences. Associations between cardiovascular and non-modifiable risk factors of age, TBI and stroke history were identified cross-culturally, consistent with literature and confirming the viability of the networks. The Directed Acyclic Graph (DAG) revealed probable risk pathways stemming from non-modifiable risk factors such as age, stroke history, parental stroke, TBI, and sex, leading to psychological distress, with cardiovascular and lifestyle factors acting as powerful mediators, and resulting in poor memory and diabetes identified as terminal outcomes. These findings suggest that stroke risk develops through multiple biopsychosocial pathways. While non-modifiable risk factors have substantial upstream influence, the DAG identified modifiable factors, such as psychological distress, cardiovascular conditions, and lifestyle factors, that can be targeted for prevention. Stroke prevention should be multifaceted, reflecting the network’s dynamic interconnections and hierarchical nature. Overall, these results highlight a diverse but largely consistent cross-cultural network of stroke risk factors.