Date of Award

2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Program

Health Outcomes and Policy Research

Track

Health Services Research

Research Advisor

Satya Surbhi

Committee

Asos Mahmood; Chi-Yang Chiu; Esra Ozdenerol; James Bailey

Keywords

Cardiometabolic Disease, Social Determinants of Health, Vulnerable Populations

Abstract

Background: Social determinants of health (SDoH) are non-medical risk factors that play a major role in shaping outcomes and can contribute to health disparities. It is critical to understand the effects of multiple domains of SDoH and their associations with cardiometabolic disease outcomes, especially among vulnerable populations. Objectives: This dissertation includes three research studies with the following aims: (1) to investigate the association of multiple domains of individual-level SDoH risk factors with uncontrolled HbA1c, blood pressure, and lipid levels in West Tennessee, (2) to identify and map census tract Hot Spots and Cold Spots of uncontrolled diabetes, hypertension, and hyperlipidemia in Shelby County, TN, and characterize these Hot and Cold spots based on neighborhood-level SDoH risk factors at the census tract level, and (3) to examine the mediating effect of nonadherence to hypoglycemic medications on the association between SDoH and uncontrolled diabetes among Medicaid patients in West Tennessee. Methods: The first study was retrospective and cross-sectional. We conducted unadjusted logistic regression analysis to examine the association between each SDoH domain and uncontrolled HbA1c, blood pressure, and lipid levels. Next, multivariable logistic regression models were used to examine the association between the cumulative PRAPARE tool score, categorized into risk tiers, and uncontrolled HbA1c, blood pressure, and lipid levels. The second study used a combination of spatial epidemiology and statistical analysis to identify and characterize census tract-level Hot and Cold spots of uncontrolled cardiometabolic conditions. Patient residence addresses were geocoded to census tracts using Esri ArcGIS Pro. Hot Spot analysis was conducted using the Getis-Ord Gi* statistic to identify significant census-level clusters for uncontrolled cardiometabolic conditions. Logistic regression models were then fit to examine the association between census-level SDoH risk factors and the likelihood of a census tract being categorized as a Hot Spot. The third study followed a retrospective cohort design and used causal mediation analysis to examine whether nonadherence to hypoglycemic medications mediates the association between SDoH and uncontrolled HbA1c. Results:The first study included three exclusive subsamples of 3,534 patients with diabetes and complete A1c measurements, 10,418 patients with hypertension and recorded blood pressure measurements, and 2,205 patients with hyperlipidemia and lipid level assessments. Across the three cardiometabolic conditions, SDoH risk factors, including Black race, other race, Hispanic ethnicity, lack of housing, lack of insurance, lack of access to basic needs, stress, and transportation issues, were strongly associated with uncontrolled HbA1c, blood pressure, and lipid levels in univariate models. Our adjusted models showed a dose-response relationship between cumulative SDoH risk categories and uncontrolled diabetes. Patients in the moderate-risk (AOR = 1.13, 95% CI = 0.94–1.36), high-risk (AOR = 1.37,95% CI = 1.12–1.69), and very high-risk SDoH categories (AOR = 1.53, 95% CI =1.13–2.06) were more likely to have uncontrolled diabetes (HbA1c>8%) compared to those in the lowest-risk category. In the second study, we identified 40 significant Hot Spot clusters of uncontrolled diabetes, 67 significant Hot Spot clusters of uncontrolled hypertension, and 22 significant Hot Spot clusters of uncontrolled lipid levels. Results from the multivariable logistic regression analysis showed that census tracts with higher proportions of African American and uninsured residents were more likely to be Hot Spots for uncontrolled diabetes and hypertension. Finally, in the third study, our results revealed that medication nonadherence was a significant mediator of the association between SDoH risk factors and uncontrolled HbA1c. Medication nonadherence mediated approximately 27% of the association between SDoH risk factors and poor glycemic control. Conclusions: The study findings show that patient-reported SDoH risk factors were strongly associated with uncontrolled cardiometabolic outcomes, especially with uncontrolled diabetes. In addition, our study identified distinct geographic clustering of uncontrolled cardiometabolic conditions across Shelby County, with Hot Spots associated with adverse SDoH. Furthermore, adverse health behaviors, such as medication nonadherence, significantly mediated the association between SDoH risk factors and uncontrolled diabetes, suggesting that medication nonadherence is one of the mechanisms by which SDoH affects uncontrolled diabetes.

ORCID

https://orcid.org/0009-0008-5129-1051

DOI

10.21007/etd.cghs.2025.0691

Download

Available for download on Tuesday, June 30, 2026

Included in

Epidemiology Commons

Share

COinS