Quantifying direct effects of social determinants of health on glycemic control in adults with type 2 diabetes

Abstract

Background: The aim of this study was to investigate if self-care is the pathway through which social determinants of health impact diabetes outcomes by analyzing the direct and indirect effects of socioeconomic and psychosocial factors on self-care and glycemic control.

Subjects and methods: Six hundred fifteen adults were recruited from two primary care clinics in the southeastern United States. A series of confirmatory factor analyses identified the latent factors underlying social status, psychosocial determinants (psychological distress, self-efficacy, and social support), and self-care (diet, exercise, foot care, glucose testing, and medication adherence). Structured equation modeling investigated the relationship among social determinants, self-care and glycemic control.

Results: Latent variables were created for diabetes self-care, psychological distress, self-efficacy, social support, and social status. The final model [χ(2)(275)=450.07, P<0.001, R(2)=99, root mean square error of approximation=0.03, comparative fit index=0.98] showed lower psychological distress (r=-0.13, P=0.012), higher social support (r=0.14, P=0.01), and higher self-efficacy (r=0.47, P<0.001) were significantly related to diabetes self-care. Lower psychological distress (r=0.10, P=0.03), lower social support (r=0.10, P=0.02), and higher self-efficacy (r=-0.37, P<0.001) were significantly related to lower glycemic control. When social determinants of health variables were included in the model, self-care was no longer significantly associated with glycemic control (r=0.01, P=0.83).

Conclusions: This study suggests a direct relationship between psychosocial determinants of health and glycemic control. Although associated with self-care, the relationship between social determinants of health and glycemic control is not mediated by self-care. Development of interventions should take psychosocial factors into account as independent influences on diabetes outcomes, rather than as indirect influences via self-care behavior.

Figures

FIG. 1.

Hypothesized model of the relationship between social determinants of health and glycemic control. HbA1c, hemoglobin A1c.

FIG. 2.

Structured equation modeling of social determinants of health on diabetes outcomes. Note that coefficients are standardized path coefficients. The fit of the final model was as follows: χ2(275), n=615)=450.07, P>0.000, R2=99, root mean square error of approximation=0.03, comparative fit index=0.98. Statistically significant difference are indicated: *P<0.05, **P<0.001. All loadings onto latent variables were significant at the P<0.001 level. HbA1c, hemoglobin A1c; MOS, Medical Outcomes Study; PDS, Perceived Diabetes Self-Management; SPD, Serious Psychological Distress.