Healthy eating as self-care: commitment, agency, structural constraints and fruit and vegetable intake: a cross-sectional survey of UK adults

AbstractBackgroundHealthy eating is a domain of self-care, but its relationship with commitment, agency and material food-access constraints requires examination. We investigated these associations with dietary behaviours among UK adults.MethodsAdults were recruited through Prolific for a cross-sectional online survey conducted from 29 July to 27 August 2026. Among 2,087 complete responses, 1,811 had complete data for the main model. Modified Poisson regression estimated adjusted prevalence ratios (PRs) for meeting a brief self-reported five-a-day indicator (≥5 fruit and vegetable portions/day). Exposures were study-specific commitment and agency scores, four perceived structural conditions and an adapted Food Insecurity Experience Scale (FIES) count. Secondary outcomes were fruit and vegetable portion intake and the consumption frequency of six selected processed-food categories.ResultsOf the 2,087 respondents, 936 (44.8%) met the survey’s five-a-day indicator. Among 2,014 respondents with scoreable FIES responses, 29.6% reported at least one experience. In the main model, commitment (PR 1.53 per SD, 95% CI 1.40–1.66) and agency (PR 1.10 per SD, 95% CI 1.01–1.19) were positively associated with five-a-day, while FIES was inversely associated (PR 0.90 per experience, 95% CI 0.86–0.95). For the five-a-day indicator, the joint test of the four structural items gave a Holm-adjusted p value of 0.76. In exploratory analyses, the corresponding p values were 0.011 for portion count and <0.001 for processed-food frequency, before adjustment for multiple testing.ConclusionsCommitment, agency and food-access experiences were associated with dietary behaviours, with structural associations differing by outcome. The findings support evaluating self-care support alongside enabling material conditions. Cross-sectional, non-probability data do not establish causality or population prevalence.