From shelter to stroke: A mediation-proxy analysis of homelessness-associated stroke risk in dialysis patients
Sanat Subhash1, Haziq Rabbani1, Nanak Dhillon1, Noah Glass1, Flora Gallegos1, Shyam Bihari Bansal2, Prakash Gudsoorkar3, Rupesh Raina1
1 Akron Nephrology Associates/Cleveland Clinic Akron General Medical Center, Akron, OH, USA, 2 Department of Nephrology, Medanta-Medicity, Gurgaon, India, and 3 Department of Internal Medicine, University of Cincinnati, Cincinnati, OH, USA
DOI 10.5414/CN111835
Abstract
Background: Dialysis patients face high ischemic stroke risk, only partly explained by cardiometabolic and genetic factors. Homelessness, race, and other social determinants may further elevate risk, but the pathways affecting outcomes in dialysis patients remain uncharacterized. Materials and methods: We conducted a retrospective cohort study using the TriNetX electronic health record network, identifying 8,821 adults with at least one homelessness code and 320,432 housed adults who initiated dialysis between 2014 – 2024 with ≥ 5 years of prior data. The primary outcome was time to first ischemic stroke within 5 years. A multivariable Cox model adjusted for age, sex, race/ethnicity, and dialysis modality estimated the total homelessness effect (HR<sub>0</sub>). We then ran 3 hazard‐attenuation models, each adding 1 composite mediator block, clinical (hypertension, diabetes, hyperlipidemia, obesity), psychosocial (nicotine dependence, depression), and social-adversity (ICD-10 Z56–Z65 codes), to calculate the percent reduction in the homelessness log‐hazard. Sensitivity analyses used a composite stroke‐or‐death endpoint and a landmark survivor cohort excluding early deaths. Results: Homelessness conferred a 4.61‐fold higher 5‐year stroke hazard. The clinical block attenuated this hazard by 12.1%, and the psychosocial block by 15.2%. In contrast, the social-adversity block amplified the effect by 27.4%, reflecting a suppressor phenomenon from under-documented social determinants of health codes. Sensitivity checks confirmed robustness, with a stroke‐or‐death HR 1.35 and death‐free cohort HR 4.50. Conclusion: Pre-dialysis homelessness and Black race independently predict stroke risk in dialysis patients. This reaffirms the need for standardized social‐needs screening, area‐level deprivation measures, and advanced data‐extraction methods to accurately capture social adversity and guide targeted interventions.
Author Details
Authors
Departments
- 1 Akron Nephrology Associates/Cleveland Clinic Akron General Medical Center, Akron, OH, USA,
- 2 Department of Nephrology, Medanta-Medicity, Gurgaon, India, and
- 3 Department of Internal Medicine, University of Cincinnati, Cincinnati, OH, USA
Address
Rupesh Raina, MD, FAAP, FACP, FASN, FNKF, Consultant Nephrologist, Adult-Pediatric Kidney Disease/Hypertension, Department of Nephrology, Cleveland Clinic Akron General and Akron Children’s Hospital, 214 West Bowery Street, Akron, Ohio 44308, USA
Email:
[email protected]
Citation
Sanat Subhash, Haziq Rabbani, Nanak Dhillon, Noah Glass, Flora Gallegos, Shyam Bihari Bansal, Prakash Gudsoorkar, and Rupesh Raina.From shelter to stroke: A mediation-proxy analysis of homelessness-associated stroke risk in dialysis patients
. ; : 0-8. doi: 10.5414/CN111835.