Children's and adolescents' rising animal-source food intakes in 1990–2018 were impacted by age, region, parental education and urbanicity

The study investigates how children’s and adolescents’ consumption of animal-source foods (ASF) varies globally and over time, and how it relates to age, sex, parental education, and urbanicity. Child malnutrition, including micronutrient deficiencies, stunting, wasting, and rising overweight/obesity, remains a global concern, particularly in Asia and Africa. ASF supply key nutrients (amino acids, fatty acids, iron, zinc, iodine, vitamin A), yet environmental and adult health impacts vary across ASF types. Prior data are limited for children aged 5–14 years and for sociodemographic disparities. The authors aim to fill these gaps by estimating ASF intakes (milk, cheese, yogurt, eggs, seafood, unprocessed red meat, processed meat) among 0–19-year-olds across 185 countries for 1990 and 2018, and by quantifying differences by age, sex, parental education, urban/rural residence, country, and region.

The paper highlights scarce nationally representative dietary data for adolescents and limited global/regional analyses of ASF intakes for children aged 5–14 years. Prior systematic assessments often focused on ≤5 and ≥15 years in mainly low-income settings, without granularity by age or sociodemographics. Evidence links ASF consumption in early life to improved growth outcomes (e.g., reduced stunting) but also notes varied long-term health and environmental impacts across ASF types (fish, dairy vs red/processed meats). Existing datasets (e.g., Demographic and Health Surveys, Gallup World Poll) indicate that while many infants/young children and most adolescents consume some ASF, there are substantial cross-country disparities, with lowest intakes in parts of sub-Saharan Africa and South Asia. The study builds on and extends previous GDD work by providing detailed age-specific and subgroup estimates for children and adolescents worldwide.

Data sources: The Global Dietary Database (GDD) compiled 1,248 dietary surveys from 188 countries (99% of global population). For ASF outcomes (milk, cheese, yogurt, eggs, seafood, unprocessed red meat, processed meat), 498 surveys reported ASF, including 429 with data on ages 0–19 years, covering 3.3 million children in 125 countries (93.1% of global child population). Most surveys were nationally representative (88.1%), individual-level (78.1%), and included urbanicity (71.1%) and parental education (53.4%). Surveys included 24-hour recalls, food frequency questionnaires, or standardized short questionnaires; large cohort or household budget surveys were used when necessary. Special populations were excluded. Data extraction and harmonization: Using standardized protocols, survey data were extracted on means/SDs, units, and stratified by age, sex, parental education, and urbanicity; quality checks assessed selection bias, representativeness, response rate, and diet method validity. Intakes were harmonized to mean individual intakes, standardized definitions/units, and adjusted for total energy using age-specific intakes: <1y: 700 kcal/d; 1–<2y: 1,000; 2–5y: 1,300; 6–10y: 1,700; 11–19y: 2,000 kcal/d. Analyses for harmonization used SAS v9.4, Stata v14.0, RStudio v1.1.453. Modelling and uncertainty: A Bayesian hierarchical model with nested random effects by country and region estimated intakes jointly stratified by age groups (<1, 1–2, 3–4, 5–9, 10–14, 15–19 years), sex, parental education (<6, 6–<12, ≥12 years), and urbanicity (urban/rural). Inputs included stratified survey intakes, survey characteristics, and country-year covariates. Overdispersion accounted for non-nationally representative surveys or those lacking finer strata. Uncertainty was quantified via 4,000 posterior draws per stratum; medians and 95% uncertainty intervals (UIs) were from the 50th, 2.5th, 97.5th percentiles. Fivefold cross-validation assessed predictive validity; plausibility was evaluated via heat maps. A second time-series Bayesian model incorporated FAO Food Balance Sheets and GENUS data to inform temporal changes. Analyses used RStudio v3.3 and Stan v2.29. Statistical analysis: Mean intakes were estimated for each of 72 population strata (age×sex×education×urbanicity) in 185 countries for 1990 and 2018. Global, regional, national, and subgroup intakes were computed using population-weighted averages (UN Population Division; Barro-Lee for education; UN for urbanicity). Units were grams/day and servings/day or per week, using standardized serving sizes: milk 245 g; cheese 42 g; yogurt 245 g; egg 55 g; seafood 100 g; unprocessed red meat 100 g; processed meat 50 g. Spearman correlations assessed relationships between ASF intakes. Differences between 1990 and 2018 used all posterior draws and were standardized to 2018 population weights. Ethics: Exempt from IRB as modelling on published, de-identified data; individual surveys had local IRB approvals.

- In 2018, global mean total ASF intake among 0–19-year-olds was 1.9 servings/day (95% UI: 1.9, 2.1); 16% of children consumed ≥3 daily ASF servings. - Regional variation: 0.8 servings/day in South Asia vs 4.2 in Central/Eastern Europe and Central Asia. Intakes varied 4.5-fold by age: 0.6 servings/day at <1 year vs 2.5 at 15–19 years. - Among 20 most populous countries, total ASF highest in Russia, Brazil, Mexico, Turkey (3.0–5.0 servings/day), lowest in Uganda, India, Kenya, Bangladesh (0.7–0.8). - Dairy (global means, 2018): milk 103 g/d (~0.4 servings/day); yogurt 18 g/d (~0.1 servings/day); cheese ~6 g/d. Milk highest in ages 10–14 and 5–9 years (~105–112 g/d); lowest <1 year (~63 g/d). Regional milk range: 46 g/d (sub-Saharan Africa) to 252 g/d (high-income countries). Highest national milk: Mexico, USA, Brazil, Turkey (222–257 g/d); lowest: Nigeria, DRC, Bangladesh, Tanzania (32–43 g/d). - Eggs: global 17 g/d (~0.3 servings/day); higher in Central/Eastern Europe and Central Asia, East/Southeast Asia, MENA, LAC; lowest in sub-Saharan Africa and South Asia. Highest national: Vietnam, China, Mexico, Russia (33–44 g/d); lowest: Nigeria (4 g/d). Egg intake correlated with unprocessed red meat (r=0.6) and processed meat (r=0.5), not seafood (r=0.03). - Seafood: global 21 g/d (~0.22 servings/day); 2 g/d at <1 year to 31 g/d at 15–19 years. Highest regions: Southeast/East Asia (~32 g/d); lower in high-income and South Asia (≤10 g/d). Highest national: Vietnam, Indonesia, DRC, Bangladesh (37–43 g/d); lowest: Pakistan, Ethiopia, Turkey, Bangladesh (≤6 g/d). Seafood negatively correlated with unprocessed red meat (r=−0.2), not with processed meat (r≈0.008). - Unprocessed red meat: global 40 g/d (~0.4 servings/day); by age: <1y 3 g/d to 15–19y 59 g/d. Highest regions: Central/Eastern Europe and Central Asia; Southeast/East Asia; lowest: South Asia, sub-Saharan Africa. Among populous countries, highest in Russia, Brazil, Vietnam; lowest in India, Bangladesh, Ethiopia, China, Uganda. - Processed meat: global 18 g/d (~0.4 servings/day); 15-fold regional range: 3 g/d (South Asia) to 44 g/d (Central/Eastern Europe and Central Asia). Highest national among populous: Philippines, Kenya, India, Brazil, Russia (40–71 g/d); lowest: Bangladesh, Indonesia, Tanzania (≤4 g/d). - Unprocessed vs processed meat moderately correlated across nations (r=0.5). Typically unprocessed > processed; exceptions include Sierra Leone (9 vs 67 g/d), Mongolia (58 vs 114 g/d), Philippines (24 vs 71 g/d), Armenia (52 vs 95 g/d), Georgia (13 vs 56 g/d). - Sex differences: minimal globally; girls slightly higher seafood (+0.04 servings/week; 0.01, 0.07) and milk (+0.03 servings/week; 0.01, 0.06). - Parental education: higher education associated with higher intakes of all ASF. Largest absolute global differences (high vs low education) in servings/week: milk +1.7 (1.6, 1.9); eggs +0.7 (0.6, 0.9); seafood +0.5 (0.4, 0.6); processed meat +0.4 (0.08, 0.8); yogurt +0.3 (0.3, 0.4); cheese +0.3 (0.2, 0.3); unprocessed red meat +0.2 (0.2, 0.3). Regionally, largest gaps varied by ASF and region. - Urbanicity: urban > rural for all ASF. Largest global absolute differences (urban − rural), servings/week: processed meat +0.6 (0.06, 1.3); milk +0.5 (0.4, 0.7); eggs +0.4 (0.3, 0.6); yogurt +0.3 (0.2, 0.6); seafood +0.2 (0.09, 0.3); cheese +0.2 (0.02, 0.3); unprocessed red meat +0.2 (0.1, 0.2). Largest regional gaps differed by ASF. - Time trends (1990–2018): global total ASF increased +0.5 servings/week (0.4, 0.6); increases in all regions except sub-Saharan Africa. Largest national increases among populous: Brazil +1.8, China +1.8, Vietnam +1.6, Mexico +1.1 servings/week; largest decreases: Tanzania −1.1, Iran −0.3, Kenya −0.2. Increases were larger at older ages (e.g., <1y +0.3; 15–19y +0.7 servings/week). - Specific ASF trends: milk increased in LAC (+2.4/week) and high-income (+0.3), no change MENA, decreased in sub-Saharan Africa (−0.03). Largest national milk increases: Brazil +4.5, Mexico +4.1, Turkey +2.3, Russia +2.3; decreases: Philippines −2.7, Iran −1.7, Kenya −0.9. Cheese increased globally (+0.1/week), notably in high-income and LAC; yogurt largely stable. Eggs doubled globally (+1.0/week), with largest increases in Southeast/East Asia and countries like Vietnam (+4.8) and China (+3.7). Seafood modestly rose globally (+0.2/week), with Southeast/East Asia up (+0.9) and sub-Saharan Africa down (−0.5); Tanzania had a large national decline (−7.6). Unprocessed red meat meaningfully increased in Southeast/East Asia (+3.9) and LAC (+1.2), declined in Central/Eastern Europe and Central Asia (−0.7). Processed meat increased in four regions (up to +1.5), stable in MENA, South Asia, sub-Saharan Africa; largest national increase in the Philippines (+6.2).

This study addresses a major data gap by providing comprehensive, age-specific, and subgroup-specific estimates of ASF consumption among children and adolescents across 185 countries in 1990 and 2018. It establishes that while global ASF intake approaches 2 servings/day, substantial heterogeneity exists by region, age, and sociodemographic factors. Younger age groups consume fewer ASF, with milk forming a larger share of total ASF in early childhood and more diverse ASF profiles in adolescence. The findings link higher ASF intake with greater parental education and urban residence, indicating socioeconomic and environmental determinants of diet that intensify with age. The results have critical implications for public health and nutrition policy. In South Asia and sub-Saharan Africa—regions with the lowest ASF intakes and high burdens of micronutrient deficiencies—very low intakes of seafood, eggs, and unprocessed red meat highlight the need for targeted interventions to improve diet quality among children and adolescents. Conversely, in regions with high ASF intakes (Central/Eastern Europe and Central Asia, LAC, high-income countries), the predominance of red and processed meats among adolescents suggests opportunities to shift toward healthier and more environmentally sustainable ASF (e.g., fish and dairy) without compromising nutrient adequacy. The study’s nuanced quantification of disparities by parental education and urbanicity can guide tailored programs and policies to reduce inequities and improve child and adolescent nutrition while considering planetary health goals.

Global ASF consumption among children and adolescents in 2018 was approximately 1.9 servings/day, with wide variation by age, region, country, parental education, and urbanicity. Total ASF intake rose modestly between 1990 and 2018, except in sub-Saharan Africa, and ASF subtype trends differed substantially. These comprehensive estimates can inform policies and programs to improve nutrient adequacy in low-intake settings and to encourage healthier, more sustainable ASF choices where intakes are high. Future work should expand dietary components (e.g., poultry), incorporate infant feeding sources (breastfeeding/formula), and explore interventions addressing socioeconomic and urban–rural disparities, particularly as differences grow with age.

- Survey availability was limited for some ASF subtypes (e.g., cheese, yogurt), age groups, countries, and years; heterogeneity in survey design and dietary assessment necessitated assumptions about serving sizes, food group definitions, energy adjustment, and disaggregation of household data. - The GDD framework did not include poultry; omission likely underestimates total ASF, particularly in high-income countries, LAC, and China; poultry may still constitute a meaningful share even in South Asia and sub-Saharan Africa. - No data were collected on breastfeeding/formula, so infant energy intake from these sources was not accounted for. - Individual-level dietary surveys are subject to sampling and measurement bias; while the Bayesian models incorporated uncertainty and overdispersion, residual bias cannot be excluded. - Despite cross-validation and plausibility checks, model-based estimates depend on input data quality and covariates used for temporal modeling.