Diabetes risk among US adults with different socioeconomic status and behavioral lifestyles: evidence from the National Health and Nutrition Examination Survey

Abstract

Background: Diabetes disproportionately affects minorities and those with low socioeconomic status (SES) in the United States, and differences in behavioral lifestyles are largely responsible for the unequal distribution of diabetes among different groups.

Methods: With data of 9,969 participants collected in the 2007-2008 and 2009-2010 cycles of the US National Health and Nutrition Examination Survey (NHANES), this study examined several mediators and their mediating effects in the connection between SES and the risk of diabetes. The SES is assessed by the income-to-poverty ratio (IPR), education level, and employment status. For the mediation analysis, we used health-related behaviors as mediators (smoking, alcohol use, consumption of green vegetables and fruits, physical activity and sedentary time, health insurance, and healthcare). In this study, the structural equation model was utilized to evaluate the mediating effects of behavioral lifestyle as a mediator in the relationship between SES and diabetes.

Results: A total of 9,969 participants were included in this study. We found a negative nonlinear association between IPR and diabetes risk (P_overall < 0.001; P_non-linear = 0.46), which was independent of the majority of known or suspected risk factors and confounding variables (gender, age, race). Participants with lower SES had higher risk of diabetes compared with those with higher SES. In mediating analysis, we found alcohol intake (OR = 0.996), physical activity (OR = 0.993), health insurance (OR = 0.998), and healthcare (OR = 1.002) mediated the IPR-diabetes association. But in the relationship between education status and diabetes, the mediation effect of alcohol intake (OR = 0.995), physical activity (OR = 0.991), and health care (OR = 1.008) were obvious. Likewise, alcohol intake (OR = 0.996), fruit intake (OR = 0.998), and health care (OR = 0.975) were important mediators in the association between employment status and diabetes.

Conclusion: This study provides critical insights on the link between SES and diabetes. Our results highlight that poor health-related behaviors and limited access to healthcare are important pathways for increased diabetes risk related to those with low SES, particularly among Mexican Americans and males. They should be top priorities for agencies and healthcare providers to develop behavior-related interventions to reduce inequalities in diabetes risk.

Keywords: Mexican American; access to healthcare; behavioral factors; diabetes mellitus; socioeconomic status.

Figure 1

Restricted cubic spline regression model fitting the relationship between income-to-poverty ratio and diabetes risk. Confounding variables such as age, gender, and race were adjusted in the model. The reference value of ORs was the median of income-to- poverty ratio (2.06); the solid red line represents the OR, and the black dashed line represents the 95% CI of the OR; P_overall < 0.001; P_non-linear = 0.46.

Figure 2

Mediating and direct effects of SES on diabetes. The proportion of the direct effect and the proportion of the sum of the mediation effect is shown in white, the proportion of the mediation effect with OR < 1 is shown in blue, and the proportion of the mediation effect with OR > 1 is shown in red. OR = exp. (a_i*b_i). OR < 1 means that improving SES can reduce the risk of diabetes through this mediating factor pathway; conversely, OR > 1 means that improving SES can increase the risk of developing diabetes through this mediating factor pathway. ^*p < 0 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 3

Mediation and direct effects of SES on diabetes in Mexican Americans, non-Hispanic white, non-Hispanic black, and Hispanics. The proportion of the direct effect and the proportion of the sum of the mediation effect is shown in white, the proportion of the mediation effect with OR < 1 is shown in blue, and the proportion of the mediation effect with OR > 1 is shown in red. OR = exp. (a_i*b_i). OR < 1 means that improving SES can reduce the risk of diabetes through this mediating factor pathway; conversely, OR > 1 means that improving SES can increase the risk of developing diabetes through this mediating factor pathway. ^*p < 0 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 4

Mediation and direct effects of SES on diabetes in males and females. The proportion of the direct effect and the proportion of the sum of the mediation effect is shown in white, the proportion of the mediation effect with OR < 1 is shown in blue, and the proportion of the mediation effect with OR > 1 is shown in red. OR = exp. (a_i*b_i). OR < 1 means that improving SES can reduce the risk of diabetes through this mediating factor pathway; conversely, OR > 1 means that improving SES can increase the risk of developing diabetes through this mediating factor pathway. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 5

Mediation and direct effects of SES on diabetes in people aged 20–30, 30–40, 40–50, 50–60, 60–70, 70–80. The proportion of the direct effect and the proportion of the sum of the mediation effect is shown in white, the proportion of the mediation effect with OR < 1 is shown in blue, and the proportion of the mediation effect with OR > 1 is shown in red. OR = exp. (a_i*b_i). OR < 1 means that improving SES can reduce the risk of diabetes through this mediating factor pathway; conversely, OR > 1 means that improving SES can increase the risk of developing diabetes through this mediating factor pathway. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.