Social Determinants of Health and Diabetes

Introduction: Social determinants of health (SDOH) research demonstrates that poverty, access to healthcare, discrimination, and other environmental factors as influencing health outcomes. There are 4 SDOH models (World Health Organization, Healthy Peoples, County Health Rankings, and Kaiser Permanente) commonly used to operationalize research, policy, and clinical applications of SDOH, yet there is limited understanding of how these SDOH models compare in predicting diabetes risk. Social workers are frequently engaged in the provision of care for SDOH and are well positioned to lead research agendas analyzing SDOH on health outcomes. The current study empirically tested SDOH for diabetes outcomes while identifying implications for social work research and policy.

Methods: The National Longitudinal Study of Adolescent to Adult Health (Add Health) was used to perform a logistic regression for Hemoglobin A1C (HbA1c) and a linear regression model for (body mass index) BMI; two commonly used dependent variables used in diabetes research. Previous literature has identified the World Health Organization (WHO), Healthy People, County Health Rankings, and Kaiser Family Foundation as the conventional SDOH models, and we operationalized these using variables in the Add Health dataset. Each SODH model includes overarching categories including social context, education, and physical environment and operationalizes these categories with a distinct variables. Common SDOH variables include income, discrimination, and healthcare access. Four linear regression models were run to analyze the SDOH models impact on BMI, and four logistic regression models were run to analyze the influence of the SDOH models on HbA1c.

Results: 1555 participants were included in the final model. The majority identified as White (68%), followed by participants identifying as Black (22%). Gender was dichotomized within the Add Health dataset with 57% identifying as male and 43% identifying as female. Each of the SDOH models was statistically significant. For logistic regression, the Healthy People model had the best overall fit (Wald χ2(4) = 89.21, p < .0001) and produced the most significant predictors of HBA1C risk including poverty, unstable housing, lack of community engagement, and incarceration history. For the linear regression models, the Kaiser model was most significant for BMI [F (28, 1448) = 5.96, p < .0001], accounting for 10.7% of the variance in BMI.

Discussion: Each of the SDOH models was statistically significant in predicting BMI and HbA1c. Each of the SDOH models demonstrated that participants who have lower income were more at risk for increased BMI and HbA1c. In addition, each model demonstrated that individuals who identify as people of color were at increased risk for elevated HbA1c. Last, education was significant across the four SDOH models. Social Work researchers are well equipped to identify assessment and interventions to address these risks within healthcare settings. Future research may implement interventions within clinics which target these health concerns while identifying their health impact. In addition, social work researchers should be using SDOH that best predict healthcare risk. This study has provided evidence for using the Health Peoples Model for HbA1c and the Kaiser Model for BMI.