Comparing Care Coordination and Food Assistance Referrals to Improve Diabetes Outcomes: A Cohort Study

Background: Care coordination or food assistance referrals can both improve diabetes outcomes by addressing health-related social needs such as food insecurity. However, the evidence comparing these approaches is limited. Objective: To examine whether care coordination and food assistance referrals are differentially associated with diabetes outcomes. Design: Longitudinal, observational cohort study. Participants: Adults with type 2 diabetes seen in community-based health centers referred for care coordination or food assistance. Main Measures: The primary outcome was hemoglobin A1c (HbA1c). Secondary outcomes were systolic and diastolic blood pressure (SBP and DBP) and low-density lipoprotein cholesterol (LDL). The primary timepoint was 6 months after referral; 12, 18, and 24 months were secondary timepoints. Targeted minimum loss estimation was used to adjust for age, sex, race and ethnicity, language, income, health insurance, comorbidities, body mass index, census tract social vulnerability index, food insecurity, housing instability, transportation barriers, and baseline outcome levels. Key Results: 2108 individuals were included (food assistance = 960; care coordination = 1148). Mean age was 55.2 years (SD: 12.4), 61.9% were female, and mean baseline HbA1c was 7.6% (SD: 2.1%). Results did not suggest clinically meaningful differences in outcomes when comparing food assistance referrals to care coordination referrals. For example, the estimated 6-month difference comparing food assistance to care coordination was 0.01% (95%CI −0.15 to 0.18) for HbA1c; −0.20 mm Hg (95%CI −1.62 to 1.22) for SBP; 0.78 mm Hg (95%CI −0.14 to 1.69) for DBP; and 2.78 mg/dL (95%CI −2.30 to 7.86) for LDL. Results were similar at other time points. Conclusions: This study did not find clinically meaningful differences in diabetes outcomes between adults who received a referral for food assistance versus care coordination. Both interventions have been proven effective in other studies, so clinics might decide which intervention to offer based on factors such as feasibility and patient preference.