Food insecurity and the risk of diabetes: understanding the role of BMI as a mediator and moderator

Objective: This retrospective cohort study examined BMI as a mediator and moderator of the association between food insecurity and diabetes. Methods: Data came from the electronic health records of 74,174 primary care patients at a large academic medical center. We used multivariate Poisson regression models to examine the association between food insecurity and diabetes. We used causal mediation analysis to evaluate the direct and indirect effects by which BMI mediates and moderates this association and the extent to which these effects varied by age, sex, and race and ethnicity. Results: During the 5-year period, 6.2% of patients were newly diagnosed with diabetes. Food insecurity was associated with a higher risk of diabetes (relative risk [RR] 1.19, 95% CI: 1.03-1.36) after multivariate adjustment. In mediation analysis, BMI was a significant mediator of the association between food insecurity and diabetes (natural indirect effect: RR 1.10, 95% CI: 1.06-1.14). When accounting for an interaction with BMI, BMI remained a significant mediator (natural indirect effect: RR 1.12, 95% CI: 1.0-1.17). In subgroup analyses, the mediator-moderator effect of BMI was stronger among adults <45 years old and female patients; there were no differences by race or ethnicity. Conclusions: Our findings support BMI as a mediator and moderator of the association between food insecurity and diabetes.