Social Determinants of Health, Processes, & Outcomes in T2D
Social Determinants of Health, Processes, & Outcomes in T2D
Demographic characteristics for this sample of 615 adults with type 2 diabetes are shown in Table 1. The mean age was 61 years, with the majority being men (61.6%), non-Hispanic black (64.9%), and employed (65.3%). 13% had less than a high school diploma, and 41.6% earned less than $20,000 annually. Descriptive information on self-care and psychological measures included in the model are also presented in Table 1.
The estimated model demonstrated good data fit, chi2 (15) = 17.68, p = 0.28; RMSEA = 0.02 (90% CI 0.00, 0.04) and CFI = 0.99. Standardized direct, indirect and total effects of the path analysis are shown in Table 2 and Table 3. While all measures hypothesized as being part of the model in Figure 1 were included in the analysis, for the sake of parsimony and interpretation indicator variables for each mediator/moderator were chosen as shown in Figure 2. In addition, non-significant paths were retained in the model rather than presenting results of a trimmed model to provide information on the full conceptual framework.
(Enlarge Image)
Figure 2.
Path model of social determinants of health on glycemic control, adjusting for age, gender, race and health literacy. Overall model fit chi2 (15) = 17.68, p = 0.28; RMSEA = 0.02 (90% CI 0.00, 0.04), CFI = 0.99. p = 0.06, *p < 0.05, **p < 0.01, ***p < 0.001. Note: coefficient for path between access to care and glycemic control is based on visits to primary care rather than patient centered care.
As indicated in Table 2 and Figure 2, there were significant total effects of socioeconomic and psychosocial factors on glycemic control for employment (r = 0.13, p = 0.002), fatalism (r = −0.09, p = 0.03), self-efficacy(r = −0.30, p < 0.001), and diabetes distress (r = 0.12, p = 0.03), such that less hours worked, more fatalistic attitudes, more self-efficacy, and less diabetes distress were associated with lower HbA1c. The majority of the total effects were direct: 92% for employment, 88% for fatalism, 93% for self-efficacy, and 92% for diabetes distress.
While the direct and total effects for social support were not significant, the indirect effect was (r = 0.02, p = 0.03). Table 3 shows that this indirect effect is mediated by access to care and process of care, where there was a significant total effect of social support on access to care (r = 0.08, p = 0.03) and processes of care (r = 0.01, p = 0.04). 100% of these effects are direct.
As indicated in Table 3 and Figure 2, there were also significant total effects of socioeconomic and psychosocial factors on self-care (medication adherence) for diabetes distress (r = −0.14, p = 0.01) and perceived stress (r = −0.15, p = 0.001), such that lower diabetes distress and perceived stress is associated with higher self-care. There were significant total effects of socioeconomic and psychosocial factors on access to care (patient centered care) for income (r = 0.08, p = 0.03), diabetes distress (r = −0.21, p < 0.001) and social support (r = 0.08, p = 0.03), such that higher income, lower diabetes distress, and higher social support were associated with higher access. Lastly, there significant total effects of socioeconomic and psychosocial factors on processes of care (diabetes education in past 12 months) for income (r = −0.11, p = 0.03), social support (r = 0.10, p = 0.04), and perceived stress (r = 0.10, p = 0.04) such that lower income, higher social support and higher perceived stress were associated with higher processes of care. The majority of all these paths were direct effects (93–100% for each).
Overall, the model explained 29% of the variance in HbA1c, 30% of the variance in self-care (medication adherence), 53% of the variance in access to care (patient centered care), 16% of processes of care (diabetes education in the past 12 months), and 76% of the variance overall.
Results
Sample Demographics
Demographic characteristics for this sample of 615 adults with type 2 diabetes are shown in Table 1. The mean age was 61 years, with the majority being men (61.6%), non-Hispanic black (64.9%), and employed (65.3%). 13% had less than a high school diploma, and 41.6% earned less than $20,000 annually. Descriptive information on self-care and psychological measures included in the model are also presented in Table 1.
Validation of the Conceptual Framework
The estimated model demonstrated good data fit, chi2 (15) = 17.68, p = 0.28; RMSEA = 0.02 (90% CI 0.00, 0.04) and CFI = 0.99. Standardized direct, indirect and total effects of the path analysis are shown in Table 2 and Table 3. While all measures hypothesized as being part of the model in Figure 1 were included in the analysis, for the sake of parsimony and interpretation indicator variables for each mediator/moderator were chosen as shown in Figure 2. In addition, non-significant paths were retained in the model rather than presenting results of a trimmed model to provide information on the full conceptual framework.
(Enlarge Image)
Figure 2.
Path model of social determinants of health on glycemic control, adjusting for age, gender, race and health literacy. Overall model fit chi2 (15) = 17.68, p = 0.28; RMSEA = 0.02 (90% CI 0.00, 0.04), CFI = 0.99. p = 0.06, *p < 0.05, **p < 0.01, ***p < 0.001. Note: coefficient for path between access to care and glycemic control is based on visits to primary care rather than patient centered care.
As indicated in Table 2 and Figure 2, there were significant total effects of socioeconomic and psychosocial factors on glycemic control for employment (r = 0.13, p = 0.002), fatalism (r = −0.09, p = 0.03), self-efficacy(r = −0.30, p < 0.001), and diabetes distress (r = 0.12, p = 0.03), such that less hours worked, more fatalistic attitudes, more self-efficacy, and less diabetes distress were associated with lower HbA1c. The majority of the total effects were direct: 92% for employment, 88% for fatalism, 93% for self-efficacy, and 92% for diabetes distress.
While the direct and total effects for social support were not significant, the indirect effect was (r = 0.02, p = 0.03). Table 3 shows that this indirect effect is mediated by access to care and process of care, where there was a significant total effect of social support on access to care (r = 0.08, p = 0.03) and processes of care (r = 0.01, p = 0.04). 100% of these effects are direct.
As indicated in Table 3 and Figure 2, there were also significant total effects of socioeconomic and psychosocial factors on self-care (medication adherence) for diabetes distress (r = −0.14, p = 0.01) and perceived stress (r = −0.15, p = 0.001), such that lower diabetes distress and perceived stress is associated with higher self-care. There were significant total effects of socioeconomic and psychosocial factors on access to care (patient centered care) for income (r = 0.08, p = 0.03), diabetes distress (r = −0.21, p < 0.001) and social support (r = 0.08, p = 0.03), such that higher income, lower diabetes distress, and higher social support were associated with higher access. Lastly, there significant total effects of socioeconomic and psychosocial factors on processes of care (diabetes education in past 12 months) for income (r = −0.11, p = 0.03), social support (r = 0.10, p = 0.04), and perceived stress (r = 0.10, p = 0.04) such that lower income, higher social support and higher perceived stress were associated with higher processes of care. The majority of all these paths were direct effects (93–100% for each).
Overall, the model explained 29% of the variance in HbA1c, 30% of the variance in self-care (medication adherence), 53% of the variance in access to care (patient centered care), 16% of processes of care (diabetes education in the past 12 months), and 76% of the variance overall.