Abstract
BACKGROUND: High levels of psychosocial distress are correlated with worse glycemic control as measured by glycosylated hemoglobin levels (HbA(1c)). Some interventions specifically targeting diabetes distress have been shown to lead to lower HbA(1c) values, but the underlying mechanisms mediating this improvement are unknown. In addition, while type 2 diabetes mellitus (T2D) disproportionately affects low-income racial and ethnic minority populations, it is unclear whether interventions targeting distress are differentially effective depending on participants' baseline characteristics. OBJECTIVE: Our objective was to evaluate the mediators and moderators that would inform interventions for improvements in both glycemic control and diabetes distress. METHODS: Our target population included 290 Veterans Affairs patients with T2D enrolled in a comparative effectiveness trial of peer support alone versus technology-enhanced peer support with primary and secondary outcomes including HbA(1c) and diabetes distress at 6 months. Participants in both arms had significant improvements in both HbA(1c) and diabetes distress at 6 months, so the arms were pooled for all analyses. Goal setting, perceived competence, intrinsic motivation, and decisional conflict were evaluated as possible mediators of improvements in both diabetes distress and HbA(1c). Baseline patient characteristics evaluated as potential moderators included age, race, highest level of education attained, employment status, income, health literacy, duration of diabetes, insulin use, baseline HbA(1c), diabetes-specific social support, and depression. RESULTS: Among the primarily African American male veterans with T2D, the median age was 63 (SD 10.2) years with a baseline mean HbA(1c) of 9.1% (SD 1.7%). Improvements in diabetes distress were correlated with improvements in HbA(1c) in both bivariate and multivariable models adjusted for age, race, health literacy, duration of diabetes, and baseline HbA(1c). Improved goal setting and perceived competence were found to mediate both the improvements in diabetes distress and in HbA(1c), together accounting for 20% of the effect of diabetes distress on change in HbA(1c). Race and insulin use were found to be significant moderators of improvements in diabetes distress and improved HbA(1c). CONCLUSIONS: Prior studies have demonstrated that some but not all interventions that improve diabetes distress can lead to improved glycemic control. This study found that both improved goal setting and perceived competence over the course of the peer support intervention mediated both improved diabetes distress and improved HbA(1c). This suggests that future interventions targeting diabetes distress should also incorporate elements to increase goal setting and perceived competence. The intervention effect of improvements in diabetes distress on glycemic control in peer support may be more pronounced among White and insulin-dependent veterans. Additional research is needed to understand how to better target diabetes distress and glycemic control in other vulnerable populations.