Reducing racial/ethnic disparities in diabetes: the Coached Care (R2D2C2) project

Abstract

Background: Despite numerous efforts to change healthcare delivery, the profile of disparities in diabetes care and outcomes has not changed substantially over the past decade.

Objective: To understand potential contributors to disparities in diabetes care and glycemic control.

Design: Cross sectional analysis.

Setting: Seven outpatient clinics affiliated with an academic medical center.

Patients: Adult patients with type 2 diabetes who were Mexican American, Vietnamese American or non-Hispanic white (n = 1,484).

Measurements: Glycemic control was measured as hemoglobin A1c (HbA1c) level. Patient, provider and system characteristics included demographic characteristics; access to care; quality of process of care including clinical inertia; quality of interpersonal care; illness burden; mastery (diabetes management confidence, passivity); and adherence to treatment.

Results: Unadjusted HbA1c values were significantly higher for Mexican American patients (n = 782) (mean = 8.3 % [SD:2.1]) compared with non-Hispanic whites (n = 389) (mean = 7.1 % [SD:1.4]). There were no significant differences in HbA1c values between Vietnamese American and non-Hispanic white patients. There were no statistically significant group differences in glycemic control after adjustment for multiple measures of access, and quality of process and interpersonal care. Disease management mastery and adherence to treatment were related to glycemic control for all patients, independent of race/ethnicity.

Limitations: Generalizability to other minorities or to patients with poorer access to care may be limited.

Conclusions: The complex interplay among patient, physician and system characteristics contributed to disparities in HbA1c between Mexican American and non-Hispanic white patients. In contrast, Vietnamese American patients achieved HbA1c levels comparable to non-Hispanic whites and adjustment for numerous characteristics failed to identify confounders that could have masked disparities in this subgroup. Disease management mastery appeared to be an important contributor to glycemic control for all patient subgroups.

Figure 1.

Comprehensive conceptual framework for evaluating diabetes disparities.

Figure 2.

Results of sequential regression models assessing contributions of patient characteristics, access to care, quality of care and medication adherence to glycemic control. *p < 0.05, **p < 0.01, ***p < 0.001. Model 1: Race/ethnicity only. Model 2: Model 1 and gender; age; duration of diabetes; born in the U.S.; education level; income (< $20–$40 k, $40–$60 k, > $60 k). Model 3: Model 2 and insurance status (uninsured, Medicaid, Medicare vs. commercial); barriers to access. Model 4: Model 3 and regimen intensification, all five processes of care, number of diabetes medications, currently on insulin. Model 5: Model 4 and physician–patient language discordance; physician–patient communication; participatory decision-making style; trust in physician. Model 6: Model 5 and Total Illness Burden Index; SF-36 physical function index; depression symptomology; diabetes burden scale. Model 7: Model 6 and management confidence; passive orientation toward health care. Model 8: Model 7 and adherence to treatment scale.