Hyunjoo Joo, Jonghyuk Choi
Background/Objectives: Diet quality is a leading modifiable determinant of non-communicable disease, yet its upstream, person-level determinants are seldom studied. We examined whether health literacy is associated with overall diet quality in a nationally representative Korean sample. Methods: This cross-sectional study analyzed 5353 adults aged ≥19 years from the 2024 Korea National Health and Nutrition Examination Survey. Health literacy, measured with a validated 10-item instrument, was analyzed as score-based categories, continuously, and as the official binary indicator; diet quality was the Korean Healthy Eating Index (KHEI, 0-100) from a single 24 h recall. Survey-weighted linear and logistic regression was used, with component, subgroup, sensitivity, and multiple imputation analyses. Results: The crude difference between the lowest (≤28 points) and highest (≥33 points) categories was positive (β = +1.41); adjustment for age alone inverted it (β = -2.76), and the fully adjusted difference was β = -1.70 (95% CI, -2.66 to -0.75; -0.13 SD), with higher odds of lowest tertile diet quality (OR = 1.37; 95% CI, 1.09 to 1.72; p = 0.007). Each 1-SD increment in health literacy corresponded to β = +0.77 (95% CI, 0.42 to 1.12). Component differences were negative for 12 of 14, but none survived false discovery rate correction. The association was more evident in men (women vs. men, interaction β = -1.03 per 1-SD increment; 95% CI, -1.67 to -0.40). Multiple imputation (m = 20; n = 5678) gave consistent estimates. Conclusions: Lower health literacy was associated with modestly poorer diet quality, an association masked and reversed by age until adjustment. The cross-sectional design precludes causal inference, the single-day assessment may not represent habitual intake, and component and subgroup findings are exploratory.