RESUMO
OBJECTIVE: Food insecurity (FI) affects approximately 11·1 % of US households and is related to worsened infant outcomes. Evidence in lower income countries links FI and infant mortality rates (IMR), but there are limited data in the USA. This study examines the relationship between FI and IMR in North Carolina (NC). DESIGN: NC county-level health data were used from the 2019 Robert Woods Johnson Foundation County Health Rankings. The dependent variable was county-level IMR. Eighteen county-level independent variables were selected and a multivariable linear regression was performed. The independent variable, FI, was based on the United States Department of Agriculture's Food Security Supplement to the Current Population Survey. SETTING: NC counties. PARTICIPANTS: Residents of NC, county-level data. RESULTS: The mean NC county-level IMR was 7·9 per 1000 live births compared with 5·8 nationally. The average percentage of county population reporting FI was 15·4 % in the state v. 11·8 % nationally. Three variables statistically significantly predicted county IMR: percent of county population reporting FI; county population and percent population with diabetes (P values, respectively, < 0·04; < 0·05; < 0·03). These variables explained 42·4 % of the variance of county-level IMR. With the largest standardised coefficient (0·247), FI was the strongest predictor of IMR. CONCLUSIONS: FI, low birth weight and diabetes are positively correlated with infant mortality. While correlation is not causation, addressing FI as part of multifaceted social determinants of health might improve county-level IMR in NC.