Abstract P065: Odds Of Heart Disease And Arrhythmia Associated With Exposure Dose To Homelessness

Abstract

Background: Adults experiencing homelessness are at 40-50% greater mortality risk from cardiovascular disease (CVD) than the general population, however little is understood about how the duration or degree of homelessness increases the risk of CVD. The Vulnerability Index-Service Prioritization Decision Assistance Tool (VI-SPDAT v1, 2012) is a 50-item measure intended to proscribe appropriate levels of intervention and prioritize the waiting list for housing. The measure also includes an item to assess history of CVD. The objective of this study is to examine the relation between duration and frequency of homelessness and CVD diagnosis, while controlling for potential confounders. Method: Version 1 of the VI-SPDAT was used to assess a cohort of individuals experiencing homelessness (n=4,739 unique participants; median age=47, {17-88}) and seeking subsidized housing services in a single county in Texas between 2014 and 2017. Medical questions were validated through electronic, deterministic linkage to the Community’s Health Information Exchange (HIE) for indigent populations. Logistic regression models identified independent covariates and potential effect modifiers of a theoretical dose-response relationship between homelessness and CVD. Results: A history of heart disease or arrhythmia was reported in 24.98% of the sample. For participants with repeated assessments (n=713; median interval of homelessness: 306 days; IQR: 216-441 days) the prevalence increased from 25.6% to 30.6% (p<0.05). Criterion validation of the tool using HIE data supported that this prevalence estimate was under-reported (38.67%; 26.44% adjusted for full sample). Odds of CVD increased with age (OR=1.03; 95%CI:1.02-1.04) and spending more than two years in homelessness (OR=1.45; 95%CI:1.27-1.67). Proportion of the sample with CVD was also higher for females, non-Hispanic participants (both p<0.05), those reporting medication adherence issues, problems with concentration, history of traumatic brain injury, and mental health-related ED visits (all p<0.001). Age at assessment (p<0.001), total time spent homeless (p<0.001), and recent episodes (p<0.05) were all independently associated with CVD. Cognitive deficits and medication adherence issues mediated the relationship between multiple demographics, homeless episodes, and reported CVD diagnosis, but fit independently into a model with age, gender, and 2 years or more spent in homelessness. Discussion: This study presents a rare window into the complex interrelationships of the time-dose of exposure to homelessness and the concomitant social and medical vulnerabilities. Time spent in homelessness and the number of recent homelessness periods were independently associated with CVD, even while controlling for age, suggesting a complex dose-response relationship.