Abstract
AIMS: To examine the relationship between food insecurity and emergency department (ED) visits, hospitalizations, A1c, and diabetes medication adherence over one year of follow-up among individuals >65 years with diabetes mellitus. METHODS: We conducted a longitudinal cohort study of adults >65 years with diabetes who did (n = 742) or did not (n = 2226) report food insecurity at baseline. We used conditional logistic regression for the ED visits or hospitalization outcomes, and mixed effects models for A1c and non-insulin diabetes medication adherence. RESULTS: In bivariate analyses, individuals with food insecurity were more likely to have an ED visit (OR = 1.40, 95% CI 1.15-1.72) or hospitalization (OR = 1.41, 95% CI 1.11-1.78) in the year after the food security assessment. In addition, A1c was higher (7.5% vs. 7.2%, p < 0.001). There was no difference in medication adherence. These differences persisted with adjustment for basic demographic and clinical characteristics, but were attenuated with further adjustment for socioeconomic status. CONCLUSIONS: Differences in diabetes outcomes by food insecurity status were attenuated by adjustment for socioeconomic status. Adverse outcomes in individuals with diabetes and food insecurity may be driven by effects of food insecurity per se or be mediated by a constellation of basic resource needs or lower socioeconomic status.