2025 Southeastern Residency Conference

Title: Impact Assessment of Food Deserts and Food Insecurity on Diabetes Mellitus


Authors: 
Allison Eppenauer


Background: 
The worldwide epidemic and challenges of managing diabetes mellitus (DM) is partially driven by the intake of nutrient-poor and highly processed foods, however access to nutrient dense foods can be limited by financial restrictions and proximity to supermarkets. Areas of low access have been identified by the United States Department of Agriculture (USDA) Economic Research Service (ERS) Food Access Research Atlas as ‘food deserts’. As negative health outcomes of residing in food deserts or experiencing food insecurity have been previously demonstrated in other studies, this study aims to evaluate if these factors are an area that requires additional focus within the current population of patients with DM at the Ralph H. Johnson Veterans Affairs Healthcare System (RHJVAHS). Identification or focus on patients residing in food deserts or experiencing insecurity may allow clinicians to streamline preventative interventions during diabetes management to improve outcomes.


Methods: 
This project is a retrospective observational longitudinal cohort study for quality improvement comparing DM outcomes in veterans residing in a food desert/experiencing food insecurity and those not residing in a food desert/experiencing food insecurity. Eligible patients in the study included all patients enrolled within the RHJVAHS catchment area who had an active ICD 9/10 code diagnosis of DM or A1c > 6.5% between January 1, 2016 and December 31, 2018 with at least one annual visit with a VA provider in a primary location. Patients were then targeted over a follow up period to include lab values, hospitalizations, and deaths during the follow up period. 

The primary endpoint for this project was to determine if there is a significant difference in the frequency of uncontrolled A1c or urine albumin-creatinine ratio (UACR) based on a composite status of residing in a food desert or experiencing food insecurity and food insecurity alone. The secondary endpoint analyzed the difference in the frequency of all-cause hospitalizations, DM-related hospitalizations and mortality based on a composite status of residing in a food desert or experiencing food insecurity and food insecurity alone.


Results: Of the 4,209 veterans included in this study, 2054 (49%) patients were found to be residing in a food desert. 18 (0.004%) of patients screened positive for food insecurity. The primary outcome of uncontrolled A1c and UACR did not differ significantly between between the composite group of non-food desert/non-food insecurity (NFD/NFI) vs food-desert/food insecurity (FD/FI).  A1c >7% occurred in 45.7% of NFD/NFI vs. 44.2% FD/FI; p = 0.338.) A1c >8% occurred in 16.8% of NFD/NFI vs. 15.1% FD/FI; p = 0.133.) UACR >30 occurred in 39.1% of NFD/NFI vs. 39.5% FD/FI; p = 0.791. UACR >300 occurred in 4.2% of NFD/NFI vs. 5.4% FD/FI; p = 0.167. 

The primary outcome did not differ in patients with food insecurity alone with overall similar findings to the composite group. Although not statistically significant, patients with food insecurity had higher UACR >30 (63.6% in those experiencing food insecurity vs. 39.2% in those without, p = 0.098.)

The secondary outcome of all-cause hospitalization rate, DM-related hospitalizations, and mortality did not differ between NFD/NFI vs. FD/FI - about 20% of patients were hospitalized for any reason, 4% hospitalized due to complications of DM, and~20% of patients passed away. In food insecurity alone, all-cause hospitalizations was higher in those experiencing food insecurity vs. those not (44.4% vs 20.8%, respectively, OR 3.0 (1.2 - 7.7); p = 0.014.) DM-related hospitalizations did not differ between groups (3.7% in food insecurity vs. 5.6% in non-food insecurity; p = 0.864.) Mortality was also higher in patients with food insecurity (44.4% vs 20.2%, respectively, p = 0.018)

Additionally, patients with food insecurity had higher odds of missing a UACR lab value within the reporting period (19.7% vs 38.9%; OR 2.6 (1.0 - 6.7) ; p = 0.042.)


Conclusion: 
In summary, there was no difference in DM outcomes (UACR, A1c, hospitalizations, or mortality) between patients residing in a food desert/experiencing food insecurity versus those who are not. Patients with food insecurity were less likely to be evaluated for UACR despite the fact their UACR >30 trended higher than patients without food insecurity. Additionally, patients with food insecurity were more likely to be hospitalized for any reason or have higher mortality rate than those without food insecurity. The project was limited by the small patient population included in the food insecurity group which likely contributed to underreported findings. Data was not collected from other facilities, limiting the information available for patients seen at other medical centers which is an especially important consideration for patients residing in rural areas that may not seek care at the VA facility a significant distance away. Patients were also excluded if their address changed during the follow-up period.