Global dietary quality in 185 countries from 1990 to 2018 show wide differences by nation, age, education, and urbanicity

The study addresses a major global health concern: poor diet is a leading cause of disease worldwide and is responsible for a large share of preventable mortality. While optimal dietary patterns are well characterized and validated, their global distributions—particularly among children and adolescents—are poorly described. Prior global assessments often relied on national per capita food availability or sales data, which can misestimate individual intakes and typically excluded younger populations. Moreover, few studies have assessed disparities by age, sex, education, and urbanicity, and no prior global analyses jointly evaluated multiple validated diet quality indices (AHEI, DASH, MED). To fill these gaps, the authors quantified global, regional, and national dietary patterns and trends among children and adults across 185 countries in 1990 and 2018, stratified by age, sex, education, and urbanicity, using harmonized individual-level dietary data from the Global Dietary Database (GDD).

Previous studies were limited in scope (small subsets of countries), often used aggregate food availability or sales data as direct inputs—which substantially misestimate individual intake—and typically did not include children, adolescents, or young adults (<25 years). Evidence on global disparities by sociodemographic factors (age, sex, education, urbanicity) has been sparse. Additionally, no prior global study has jointly assessed several validated diet quality metrics including AHEI, DASH, and MED. Earlier analyses suggested differences between healthy and unhealthy diet components across income levels and regions, but comprehensive, up-to-date, individual-level global estimates—especially for younger age groups—were lacking.

Data were drawn from the Global Dietary Database (GDD) 2018, a collaborative compilation and standardization of individual-level dietary intake data on up to 53 foods, beverages, and nutrients from 1,139 dietary surveys in 175 countries (covering 7.46 billion people in 2018), with additional surveys compiled overall from 188 countries. Most surveys were nationally or subnationally representative and used 24-hour recalls or food-frequency questionnaires; data were jointly stratified by age, sex, education, and urban/rural residence. Dietary intakes were standardized to age-specific energy intakes to assess composition independent of quantity and reduce measurement error. A Bayesian hierarchical model estimated log-means (and standard deviations) of dietary intakes within a nested structure with random effects at global, regional, and country levels; included fixed effects for sex, education, urbanicity, and nonlinear age; survey-level indicators (dietary assessment method, metric type); and national year-specific covariates. Overdispersion accounted for non-national representativeness or limited stratification. Uncertainty was quantified using 4,000 posterior draws for each stratum (country, year, age, sex, education, urbanicity), providing medians and 95% uncertainty intervals (UIs). Time trends were strengthened via a second Bayesian varying-slopes model incorporating FAO Food Balance Sheets and Global Expanded Nutrient Supply data. Validity checks included five-fold cross-validation, plausibility assessments, and visual inspection via heat maps. Dietary pattern characterization: Primary analysis used the Alternative Healthy Eating Index (AHEI) components available in GDD: fruit, non-starchy vegetables, whole grains, sugar-sweetened beverages (SSBs), legumes/nuts, unprocessed red/processed meats, seafood omega-3 fat, PUFAs, and sodium (alcohol and trans fat not available). Each component was scored 0–10, summed to 0–90, and scaled to 0–100. Secondary analyses calculated DASH (eight components, scored 1–5 by sex-specific quintiles; total 8–40) and MED (eight components excluding alcohol, scored 0/1 by sex-specific medians; total 0–8). For consistency, DASH and MED distributions were derived from 2018 and applied to other years. All intakes were standardized to 2,000 kcal/day to derive pattern scores. Statistical analysis used population-weighted averages (UN Population Division for populations; Barro-Lee for education distributions; UN/World Bank for urbanicity) to compute global, regional, and national scores and subgroup differences. Spearman correlations assessed inter-relationships among AHEI, DASH, and MED. Changes from 1990 to 2018 were computed using all posterior draws and standardized to 2018 demographic distributions. Given the Bayesian framework, 95% UIs guide interpretation rather than formal significance tests.

- Global/regional status in 2018: The global mean AHEI score was 40.3 (95% UI 39.4, 41.3) on a 0–100 scale. Regional means ranged from 30.3 (28.7, 32.2) in Latin America and the Caribbean to 45.7 (43.8, 49.3) in South Asia. Among healthier components, highest global component scores were legumes/nuts 5.0 (4.8, 5.3), whole grains 4.7 (4.5, 5.0), seafood omega-3 fat 4.2 (3.8, 5.1), and non-starchy vegetables 3.9 (3.8, 4.0). Among unhealthier components, highest scores (indicating lower intake) were SSBs 5.8 (5.7, 5.9) and red/processed meat 4.8 (4.5, 5.1). Component profiles varied widely by region (e.g., South Asia scored high on whole grains and low red/processed meat and SSBs; Latin America/Caribbean scored high on legumes/nuts and lower sodium). - National variation (2018): Only ten countries (<1% of world population) had AHEI ≥50. Among the 25 most populous countries, highest mean AHEI scores were in Vietnam, Iran, Indonesia, and India (54.5–48.2), and lowest in Brazil, Mexico, the United States, and Egypt (27.1–33.5). Across populous countries, component variation was substantial: ~100-fold for sodium, ~90-fold for red/processed meat, ~23-fold for SSBs; least variation for PUFA (~2-fold) and non-starchy vegetables (~3-fold). - Demographic differences (2018): Globally, children’s mean AHEI (39.2; 38.2, 40.3) was similar to adults’ (40.8; 39.8, 42.0), but children scored lower than adults in Central/Eastern Europe and Central Asia, high-income countries, and the Middle East/North Africa. By age, J- or U-shaped relationships were common, with highest scores in the youngest (≤5 years) and/or oldest (≥75 years) groups. Component differences (children vs adults) included lower scores for fruit (2.2 vs 2.5), non-starchy vegetables (3.1 vs 4.3), SSBs (5.3 vs 6.1), and seafood omega-3 (3.3 vs 4.7) in children, and higher in children for PUFAs (2.1 vs 1.4) and sodium (4.6 vs 3.2). By sex, women had higher AHEI globally and regionally, with largest differences in high-income countries (+4.4; 3.8, 5.0) and Central/Eastern Europe & Central Asia (+3.6; 2.1, 5.3). Women modestly outscored men for fruit (+0.2), non-starchy vegetables (+0.3), and whole grains (+0.4). By education, higher attainment corresponded to higher AHEI globally and in most regions (largest differences: Central/Eastern Europe & Central Asia +3.6; Latin America/Caribbean +3.5; South Asia +2.9), with no clear differences in Middle East/North Africa or Sub-Saharan Africa. More educated individuals had higher scores for fruit (+0.8), sodium (+0.7), whole grains (+0.6), and non-starchy vegetables (+0.5), but lower scores for red/processed meat (-0.6), SSBs (-0.6), and legumes/nuts (-0.1). By urbanicity, global AHEI did not differ significantly between urban and rural, but was higher in urban areas in Central/Eastern Europe & Central Asia (+2.2; 0.9, 3.5) and Southeast/East Asia (+1.4; 0.6, 2.4), and lower in urban areas in Middle East/North Africa (-3.8; -5.5, -2.2). Urban residents had higher fruit (+0.2) and whole grains (+0.2) scores, but lower SSBs (-0.5), red/processed meat (-0.4), and legumes/nuts (-0.1). - Time trends (1990–2018): Globally, AHEI increased modestly by +1.5 (1.0, 2.0), with increases in Central/Eastern Europe & Central Asia (+4.6), high-income countries (+3.2), Southeast/East Asia (+2.7), Middle East/North Africa (+2.2), and Latin America/Caribbean (+1.3); no significant change in South Asia (0; -0.9, 1.1); and a decline in Sub-Saharan Africa (-1.1; -1.8, -0.4). Component trends globally: increases in non-starchy vegetables (+1.1), legumes/nuts (+1.1), and fruit (+0.1); decreases (worsening) in red/processed meat (-1.4), SSBs (-0.6), and sodium (-0.4); stable for whole grains (+0.1), PUFAs (0), and seafood omega-3 (0). Among populous countries, largest improvements: Iran (+12.0; 9.9, 13.9), United States (+4.6; 4.1, 5.1), Vietnam (+4.5; 2.4, 7.2), China (+4.3; 2.8, 5.9); largest declines: Tanzania (-3.7; -5.8, -1.5), Nigeria (-3.0; -5.3, -0.7), Japan (-2.7; -3.1, -2.3), Philippines (-1.8; -2.7, -0.9). - Secondary diet indices (2018): Global mean DASH was 22.9 (22.6, 23.2) and MED 4.1 (3.9, 4.2). Regionally, both were higher in South Asia and lower in Latin America/Caribbean. Adults scored higher than children (DASH: 23.2 vs 22.3; MED: 4.3 vs 3.7); little difference by sex. Scores were higher among more educated (+2.6 DASH; +0.3 MED) and, for DASH only, among urban vs rural (+0.4; 0.2, 0.7). From 1990 to 2018, global DASH increased by +1.0 (0.8, 1.1) and MED by +0.3 (0.2, 0.4). Inter-correlations across strata: AHEI–DASH 0.8; AHEI–MED 0.5; DASH–MED 0.6.

This comprehensive global assessment of dietary quality in 185 countries shows that overall diet quality is modest, with substantial heterogeneity by region, age, sex, education, urbanicity, and dietary components. The findings address the lack of rigorous, individual-level global evidence—particularly for children and adolescents—by leveraging standardized data from over 1,100 surveys and applying validated metrics (AHEI, DASH, MED). Regional patterns highlight different priority areas: in South Asia and Sub-Saharan Africa, low intakes of unhealthy items (SSBs, red/processed meats) coexist with suboptimal intakes of healthy foods (fruit, non-starchy vegetables, legumes/nuts, seafood omega-3, plant oils), suggesting policies to increase access and affordability of produce, seafood, and plant oils. In high-income regions, Central/Eastern Europe & Central Asia, and the Middle East/North Africa, improvements in healthy components have been offset by minor reductions or stability in red/processed meat, SSBs, and sodium, implying a dual strategy to promote healthy foods while reducing harmful components. The observed associations of AHEI with reduced risks of cardiovascular disease, diabetes, cancer, and mortality in multiple cohorts suggest that the modest global diet quality documented here likely contributes to substantial preventable chronic disease burden. Age-related patterns (highest scores in very young and very old, with declines during adolescence) underscore the need for interventions targeting older children and adolescents and for fostering healthy habits early in life. Educational disparities, generally favoring higher education levels, and mixed urban–rural patterns by region indicate that socio-economic and environmental contexts shape diet quality differently across settings, informing tailored national and subnational policies to improve nutrition security and equity.

Global dietary quality in 2018 was modest and improved only slightly since 1990, with considerable variation by region and demographic factors. The study provides comprehensive, contemporary, individual-level estimates of diet quality across 185 countries, including children and adolescents, and evaluates three validated dietary patterns (AHEI, DASH, MED). The results can guide targeted policies: increasing healthy food intake where it is low (for example, fruits, vegetables, legumes, whole grains, seafood, plant oils) and reducing harmful components (red/processed meat, sodium, SSBs), with strategies tailored to regional and sociodemographic contexts. Future research should develop and validate diet quality metrics that capture micronutrient adequacy across all age groups, especially for children and in low- and middle-income countries, and further validate or adapt indices (including components like alcohol and trans fat) for diverse global populations.

Individual-level dietary data are subject to measurement error; survey availability was limited or incomplete for certain nations, dietary factors, demographic strata, and years (fewer than one quarter of surveys included children aged 3–9 and adults ≥85). Although Bayesian hierarchical models incorporated additional uncertainty and covariates to address heterogeneity, sampling and information bias cannot be ruled out. Standardizing intakes to 2,000 kcal/day for comparability may misrepresent groups with lower or higher energy needs (e.g., infants, young children, seniors, high-energy consumers). Trans fat (AHEI) and alcohol (AHEI, MED) were not available, so scores reflect remaining components. The AHEI, MED, and DASH were originally developed/validated in adults in high-income countries; while applied here to children and older adults, their correlation with nutrient adequacy—particularly micronutrients—may be limited in some populations. Other, less validated indices were not assessed and warrant evaluation once better validated for diverse settings.