Data-Driven Identification of Social Determinants Driving Health Disparities in Rural Appalachian West Virginia: A Population-Level Analysis

Abstract

West Virginia is the only U.S. state lying entirely within the Appalachian Regional Commission's 423-county Appalachian Region, and it carries a well documented, disproportionate burden of chronic disease and "diseases of despair." Yet county-level, statewide, data-driven analyses that connect specific social determinants of health (SDOH) indicators to specific chronic disease outcomes, at the resolution local health departments and state planners actually operate at, remain comparatively scarce. We assembled a population-level dataset covering all 55 West Virginia counties from two real, public sources: the CDC PLACES 2023 release, a model-based small-area estimation product built on the Behavioral Risk Factor Surveillance System, and the Appalachian Regional Commission's fiscal year 2027 county economic status classification. We examined six chronic health outcomes (diabetes, obesity, hypertension, chronic obstructive pulmonary disease, current asthma, and frequent mental distress) against four county-level social-need indicators (lack of health insurance, food insecurity, housing insecurity, and transportation barriers) and current cigarette smoking prevalence. The four social-need indicators proved to be extremely intercorrelated at the county level (pairwise r = 0.84 to 0.98; a joint regression model showed variance inflation factors of 11 to 112), and a principal component analysis confirmed they collapse onto a single latent hardship dimension explaining 96.6 percent of their shared variance. We therefore constructed a composite SDOH Hardship Index and an analogous composite Health Disparity Index (the mean z-score across the six outcomes) and analyzed county-level relationships between them. The two composite indices correlate at r equals 0.924 (p less than 0.001) across all 55 counties. A two-predictor model (Hardship Index plus smoking prevalence) explains 90.9 percent of county-level variance in the Health Disparity Index (p less than 0.001), though its explanatory power varies sharply by individual outcome, from 98 percent for COPD to 23 percent for obesity. County economic distress status, independently assigned by the Appalachian Regional Commission, stratifies the Health Disparity Index cleanly and monotonically (one-way ANOVA F equals 32.8, p less than 0.001): economically Distressed counties average 1.21 standard deviations above the state mean on the composite disparity measure, while the state's sole Competitive-status county averages 2.05 standard deviations below it. We identify the ten counties carrying the state's highest social and health hardship burden, nearly all officially Distressed, as the most concrete candidates for targeted intervention. As an ecological, cross-sectional analysis of publicly available aggregate data, this study cannot support individual-level or causal claims, a limitation we discuss explicitly alongside the practical, county-targetable implications of the findings for Appalachian public health planning.