Abstract 15849: Studying the Effects of Race, Community Health, and Vascular Inequities on Outcomes for Peripheral Artery Disease: Principal Components Analysis of Community Health Metrics Using the Agency for Healthcare Research and Quality Social Determinants of Health Database

Abstract

Introduction: Healthcare disparities associated with peripheral artery disease (PAD) are influenced by differences in healthcare quality and social determinants of health (SDOH). Hypothesis: We used principal components analysis (PCA) to summarize key themes of community-level SDOH variables. Methods: This retrospective analysis used data from a single health system quality improvement database of patients diagnosed with PAD between 2016-2019. Subjects were geocoded to census tracts and variables from the Agency for Healthcare Quality and Research SDOH database were obtained to examine small-area variability in social, economic, educational, physical infrastructure, and healthcare contexts. PCA was used to identify major themes in SDOH heterogeneity across communities. Relevant SDOH variables were normalized within a data matrix and those with a loading value >0.2 were retained in a principal component (PC). PCs with an eigenvalue > 1.0 were identified to parsimoniously exemplify key themes. Results: Among 7,187 subjects with PAD, the mean age was 72.2 11.1 years, 45.9% were female, 9.4% self-reported as non-Hispanic Black, and 86.6% as non-Hispanic White. PCA was applied to 29 SDOH variables relevant to vascular health. PCA retained 6 PCs to an SDOH matrix describing 73.8% of variance in census tract level data. The greatest variance in data was attributed to socioeconomic advantage (PC1) and urbanicity/migration context (PC2) with some overlap in Black neighborhood context (PC3), income inequality (PC4), multi-race cohesion (PC5), and underrepresented minority context (PC6). Conclusions: PCA transforms high-dimensional SDOH data into distinct typologies of community health while retaining complex features of the data. Future analysis will weight these results within a hierarchical model that accounts for patient-level characteristics and quality of care to inform a more complete social-ecological model of whole-person vascular care.