Alliance to Reduce Disparities in Diabetes

Chicago Site Team at AcademyHealth

Integrating multi-level interventions to improve diabetes care and outcomes

2014 AcademyHealth Annual Research Meeting
2:30 PM - 4:00 PM, Sunday, June 8
San Diego, CA

Peek, M.E., Geary, N., Gao, Y., Rowell, D.J., O’Neal, Y., Roberson, T.S., Whyte, J.M., Bergeron, N., & Chin, M.H.

Research Objective
The drivers of health care disparities in diabetes are multifactorial, including limited access to medical care, health insurance, and community-based resources for diabetes self-management. Few studies have integrated interventions at multiple levels (e.g. person, organization) across both the health care system and community to address these diverse causes.  

Study Design
Our multi-level intervention included four components that were carried out in two university clinics and four federally-qualified health centers on the South Side of Chicago from 2008 to 2012.  Patients were invited to participate in culturally-tailored diabetes empowerment classes, which focused on shared decision-making and diabetes education.  Health care providers were trained in cultural competency and patient-centered communication. Clinics implemented a range of quality improvement projects, including integrating care management, customizing electronic medical records to foster population management, and referral tracking.  And finally, patients were linked to community-based resources that support healthy lifestyles.

We collected cross-sectional diabetes-related care processes (e.g., influenza vaccinations, diabetes education, and HbA1c testing) and health outcomes (e.g., HbA1c levels, blood pressure, and LDL levels) from each clinic site from 2008 to 2012. For each year, a random sample of 100 patients with diabetes were selected from each clinic. Generalized estimating equations were used to examine changes in diabetes-related processes and health outcomes over time, controlling for clinic and patient characteristics (i.e., age group, gender, race, insurance, treatment type, and comorbidities).

Population Studied
A total of 2527 medical charts were reviewed across 5 years. The majority of patients were African American (69 pct), female (64 pct), and had type 2 diabetes (92 pct).  The mean age was 57 years old, and 44 percent were insured through Medicaid or uninsured.  Treatment types for patients varied, with most using oral medications (60 pct), followed by oral medications and insulin (19 pct), insulin only (15 pct), and diet only (6 pct).         

Principal Findings
Compared to 2008, there were significant improvements in several diabetes-related process and outcome measures over time.  In 2012, patients were more likely to be offered an influenza vaccination (24 to 53 pct , adjusted OR, aOR=2.67, CI 1.63-4.38), receive diabetes education during the appointment (22 to 60 pct, aOR=6.81, CI 2.72-17.02), and receive an HbA1c test (77 to 94 pct, aOR=2.99, CI 1.55- 5.78). In 2011, patients were more likely to have HbA1c values less than 8 percent (62 to 71 pct, aOR=1.57, CI 1.06-2.33), and LDL levels less than 100 (49 to 61 pct, aOR=1.66, CI 1.19-2.30). An unexpected small decrease in blood pressure control (i.e. less than 130/80) was observed (34 to 33 pct, aOR=0.73, CI 0.54-0.97).

Some diabetes related process measures and patients’ health outcomes in the six clinics are improving. While outcomes did not improve consistently across years, there is an overall upward trend. In general, diabetes-related process measures improved faster and with greater magnitude than health outcomes, which is expected since health outcomes take longer to change.  

Implications for Policy and Practice
Our results suggest that policies and practices that integrate interventions at multiple levels have the potential to improve comprehensive diabetes care and outcomes over time, and potentially reduce diabetes disparities.