Is food security associated with diet and health? A cross-sectional online panel of adults in the UK

Abstract

BackgroundPrevalence of adult food insecurity is estimated to be 8–10% in the UK, but the impact on population diet and health in unknown. We aimed to investigate the associations between food security and diet and health outcomes in adults in the UK. MethodsWe used cross-sectional UK data from the International Food Policy Study, which recruited 4047 UK adults (aged 18–64 years) through quota sampling from an online panel. 767 participants were excluded for incomplete adult food security status (measured with the USDA Adult Food Security Survey Module). 748 further participants were excluded for incomplete outcome data. Adjusted linear regression models were used to explore associations between food security and frequency of fruit and vegetable intake. Adjusted logistic regression models were used to explore associations between food security and diet quality, health, and stress. Sample weights were used to improve population representativeness and our sample was similar to the UK population in terms of sex, age, and region of residence. Findings2551 participants were included in our analysis. When weighted, 24·3% of participants were food insecure. Food insecure and food secure adults did not differ in their fruit or vegetable intake frequency, but food insecure adults drank fruit juice 0·35 times more per day (95% CI 0·20–0·49) than food secure adults. Odds of poor self-rated diet quality were higher in food insecure than in food secure adults (odds ratio 1·65, 95% CI 1·31–2·09), as were odds of poor general health (1·90, 1·50–2·41), poor mental health (2·10, 1·65–2·69), and high stress levels (3·15, 2·42–4·11). InterpretationThis study is the first, to our knowledge, to examine associations between food security and diet and health in the general adult population in the UK. Food insecurity was associated with some markers of poor diet and health. Strong associations with mental health and stress, especially, point to possible psychological consequences of food insecurity. This study was cross-sectional, so further research into these associations is needed. Self-reported data are subject to social desirability bias. Quota sampling from an online panel may have introduced selection bias. Exclusions from the initial sample could limit representativeness. FundingCentre for Diet and Activity Research (CEDAR) (for AY via a CEDAR PhD studentship, JA, MW), University of Cambridge