Global Dietary Database Research Articles

Global, Regional and National Consumption of Major Beverages in 2015: Systematic Analysis of Country-Specific Nutrition Surveys Worldwide (P10-038-19)

ObjectivesAssessing the impact of beverage intake on health has been limited by lack of reliable and standardized individual-level data on a global scale. We aimed to quantify global distributions of intakes of sugar-sweetened beverages (SSBs), fruit juices, milk, coffee, and tea by country (n = 188), super-region (n = 6), key population characteristics (all ages, sex, education, urban/rural location) and time (1990, 2015). MethodsPrimary national and subnational survey data on beverage intake in the Global Dietary Database (GDD) were identified through systematic searches of electronic databases and communications with data owners worldwide. This includes 1137 survey-years representing 6.78B people from 185 countries worldwide (97.5% of the world’s population). To address missing data and estimate distributions of intake for each diet factor and by country-, age-, sex-, education-, urban/rural- and year-specific stratum we developed a Bayesian hierarchical imputation model. The model combined primary intake data with time-varying country-level and survey-level covariates and further accounted for differences in intakes vs availability from FAO food disappearance data. ResultsIn 2015, global SSB intake in adults (20 + y) was 0.37 servings (8-oz)/day (Fig 1). The Latin-American region had the highest SSB intake (1.15 servings/day) and Asia the lowest (0.14 servings/day). For fruit juices, global intake across all ages was 0.19 servings (8-oz)/day; highest intakes were seen in Latin America (0.64 servings/day) and lowest in Asia (0.06 servings/day) (Fig 2). Global milk intake was 0.39 servings (8-oz)/day, ranging from 0.64 servings/day in the High-Income region to 0.20 servings/day in Asia (Fig 3). Intakes were generally similar by sex and higher at younger ages, urban areas, and higher education levels. Findings for coffee and tea, and over time, will be presented at the meeting. ConclusionsThese novel global dietary data highlight substantial variation in beverage intake worldwide, further informing global diet surveillance, priority setting and nutrition strategies. Funding SourcesBill & Melinda Gates Foundation. Supporting Tables, Images and/or Graphs▪▪▪

Global, Regional and National Consumption of Plant-Source Foods in 2015: Systematic Analysis of Country-Specific Nutrition Surveys Worldwide (OR21-01-19)

ObjectivesEstimating plant-source food (PSF) intake is crucial for evaluating associated health and environmental impacts; yet, worldwide consumption remains largely uncharacterized. Our aim was to quantify global distributions of daily intakes of fruits, vegetables (non-starchy, potatoes, other starchy), legumes, grains (refined, whole), and nuts/seeds by country (n = 188), super-region (n = 6), key population characteristics (all ages, sex, education, urban/rural location), and time (1990, 2015). MethodsThe primary data sources of PSF intake in the Global Dietary Database were from national and subnational surveys identified by systematic searches of electronic databases and communications with data owners from 1980 through 2015, including 1137 survey-years representing 185 countries (97.5% of the world’s population). To address missingness and estimate national and stratum-specific intake distributions worldwide, we used a Bayesian hierarchical imputation model. The model combined stratum-specific individual-level intakes with time-varying country- and survey-level covariates, and accounted for differences in intakes vs per capita availability from FAO food disappearance data. ResultsIn 2015, global fruit intake was 114 g/d (1.1 servings (100 g)/d), with highest intakes reported in Latin America (146 g/d) and lowest in the Middle East and North Africa (MENA) region (97 g/d) (Figure). Non-starchy vegetable global intake was 203 g/d (2 servings (100 g)/d); highest intakes were seen in Asia (269 g/d) and lowest in the Former Soviet Union (FSU) (135 g/d) and High-Income (136 g/d) regions. Global legume intake was 17 g/d (0.2 servings (100 g)/d), highest in Latin America (52 g/d) and lowest in FSU (3 g/d). For refined grains global average was 215 g/d (4.3 servings (50 g)/d), varying from 254 g/d in MENA to 117 g/d in Sub-Saharan Africa. Intakes were generally similar by sex and urban/rural location, and higher at older ages and higher education levels. Findings for potatoes, other starchy vegetables, whole grains and nuts/seeds, and over time, will be presented at the meeting. ConclusionsSuch global data, showing substantial variation in PSF consumption by nation and population worldwide, inform and highlight the need for country-specific priority setting and nutrition policies. Funding SourcesGates Foundation. Supporting Tables, Images and/or Graphs▪

Global and National Consumption of Animal Source Foods for Children and Adults in 2015: Systematic Analysis of Country-Specific Nutrition Surveys Worldwide (P10-077-19)

ObjectivesGlobal intakes of animal source foods (ASFs) are important for several health outcomes─ broadly positive for maternal and child health outcomes (MCH) and broadly negative for some non-communicable diseases (NCDs). While ASFs represent an important source of key nutrients for many (particularly lower-income) consumers, the distributions of ASF intake is not well established. Our objective was to generate comprehensive estimates of daily intakes of milk, eggs, unprocessed red meat, and additional ASFs by country, age, sex, education and urban/rural location in 2015. MethodsNational and subnational surveys on dietary intake were identified through systematic searches of multiple electronic databases, and through personal communication with experts worldwide. Presently, the Global Dietary Database (GDD) includes 1137 survey-years of global data representing participants from 185 countries (97.5% of the world’s population). Bayesian hierarchical modeling methods were used to combine individual-level intake data with country-level availability data and estimate mean intake and its statistical uncertainty for each country-, age-, sex-, education-, urban/rural- and year-specific stratum. ResultsIn 2015, global milk intake was 95.0 g/day (95% uncertainty interval [UI] 70.1–134.7 g/day) with region-specific intakes ranging from 49.2 g/day in Asia to 155.9 g/day in high-income countries (Figure 1). Mean national milk intake was highest in Sweden (299.9 g/day) and lowest in China (23.4 g/day). Global egg intake was 25.0 g/day (13.5–54.1 g/day). Region-specific intake was highest in Latin American and the Caribbean (45.4 g/day) and lowest in Sub-Saharan Africa (13.9 g/day; Figure 2). Mean national egg intake was ≥1 serving per day (55 g) in 3 of the 185 countries. Global unprocessed red meat intake was 57.9 g/day (40.1–89.2 g/day) and intake was greatest in the Former Soviet Union (75.3 g) and Latin America and the Caribbean (69.4 g; Figure 3). National red meat intake was ≥1 serving per day (100 g) in 7 of the 185 countries. Additional findings for cheese, yogurt, reduced fat milk, whole milk, seafood and processed meat will be presented at the meeting. ConclusionsOur findings highlight the substantial range of ASF intake worldwide. These data are valuable for dietary monitoring and the development of policies to prevent MCH and NCDs associated with both ASF insufficiency and surplus. Funding SourcesGates Foundation. Supporting Tables, Images and/or Graphs▪▪▪

Global Intakes of Select Micronutrients; Findings from the 2015 Global Dietary Database Project (OR07-08-19)

ObjectivesAdequate micronutrient intake is crucial for a variety of health outcomes. Micronutrient deficiency can lead to severe consequences for maternal child health (MCH) while excessive intakes of micronutrients such as sodium have been linked to non-communicable disease (NCD) outcomes. Having comparable estimates of micronutrient intake around the world is crucial for surveillance, intervention, and nutrition policy. Our objective was to estimate individual level daily nutrient intake for 18 vitamins and minerals from infancy through adulthood across 188 countries for 35 years from 1980 to 2015. MethodsWe collected and analyzed data from nationally representative or sub-national dietary surveys of individual-level intakes. In total, our model incorporated data from 1137 survey years of data representing 185 countries and 97.5% of the world’s population. We used a Bayesian hierarchical model, informed by additional covariates such as FAO food balance sheets and GENuS nutrient estimates, to estimate intakes of 18 micronutrients (Table 1) selected based on known or suspected causal links to MCH or NCDs. Estimates of mean intakes and uncertainty were generated jointly by country, time, age (15 groups), sex, urban or rural residence, and education (three levels). ResultsIn 2015, important global variation was identified in dietary intakes of these micronutrients. Here we present findings for three key nutrients. For calcium, the global mean was 630 mg/d (Figure 1). Regionally, only high-income countries met the adult EAR of 800 mg/d with a mean intake of 902 mg/d. For iron, we estimated a global mean of 13.4 mg/d (Figure 2). For Vitamin A, we estimated a global mean of 821 μg/d (Figure 3). Globally, notable differences include variations in intake between men & women, urban vs. rural residency, and high vs. low education. Additional detailed findings on other nutrients and population strata will be presented at the meeting. ConclusionsBy combining extensive global data, significant and specific variation was identified in optimal individual dietary intakes of micronutrients, including by region and population subgroups. These novel findings provide the best available global evidence to inform disease estimation, policy, and prevention efforts. Funding SourcesBill & Melinda Gates Foundation. Supporting Tables, Images and/or Graphs▪▪▪▪

Global Intakes of Total Protein and Sub-types; Findings from the 2015 Global Dietary Database (P10-050-19)

ObjectivesAdequate macronutrient intake is critical for child growth and health and the prevention of non-communicable diseases. Presently, the global distributions of macronutrient intake are not well established. Our objective was to comprehensively estimate daily intakes of total, animal, and plant protein, and additional macronutrients for individuals of all ages from 185 countries in 2015. MethodsWe collected and analyzed individual-level dietary intake data from nationally or sub-nationally-representative surveys, comprising 1137 survey-years of data representing 185 countries (97% of the world’s population). Using a Bayesian hierarchical model, we estimated mean macronutrient intake by country, time (35 years), age (15 groups), sex, education (3 levels), urban or rural residence, and pregnancy status (when applicable). The primary model incorporated the number of subjects in each stratum, sampling representativeness, dietary assessment method, type of dietary metric, availability from the United Nations FAO, and country, and superregion (6 groupings of regions) random effects as covariates. ResultsIn 2015, global total protein intake was 78.2 g/day, ranging from 61.0 g in the Middle East and North Africa to 92.5 g/day in Asia. Mean national total protein intake was ≥46 g/day in all 185 countries. Global animal protein intake was 33.3 g/day and intake was highest in Latin American and the Caribbean (42.8 g/day) and lowest in Sub-Saharan Africa (38.3 g/day). The countries with the lowest mean national animal protein intake were Bangladesh (7.2 g/day) and Nepal (10.1 g/day). Global plant protein intake was 29.1 g/day. Region-specific intake was highest in the Middle East and North Africa (32.1 g/day) and lowest in the Former Soviet Union (25.5 g/day). Plant protein intake was usually greater in males than females. Additional findings for dairy protein, carbohydrate, added sugar, fiber, total fat, saturated fat, polyunsaturated fat, seafood omega-3 fat, plant omega-3 fat, and trans-fat will be presented at the meeting. ConclusionsOur results show trends and variation in protein worldwide, by region, and by country. These data provide the best available global evidence on macronutrient intakes and are valuable for informing public health policy and disease prevention efforts. Funding SourcesBill & Melinda Gates Foundation. Supporting Tables, Images and/or Graphs▪▪▪

Global Improvement in Dietary Quality Could Lead to Substantial Reduction in Premature Death

The preventable premature mortality achievable by improvement in dietary quality at a global level is unclear. The aim of this study was to assess dietary quality globally, and to quantify the potential global impact of improving dietary quality on population health. We applied the Alternate Healthy Eating Index (AHEI, potential range 0-100) to a global dietary database to assess dietary quality among adults in 190 countries/territories. The relation of AHEI score to risks of major chronic disease was estimated from 2 large cohorts of men and women for whom many repeated dietary assessments during up to 30 years were available. We calculated the preventable premature deaths achievable by shifting from current national diets to a reference healthy diet. The global mean AHEI score in 2017 was 49.5 for males and 50.5 for females. Large differences between current and target intakes existed for whole grains, sodium, long-chain n-3 polyunsaturated fats, polyunsaturated fats, and fruits. From 1990 to 2017, the global mean AHEI score increased modestly from 45.4 to 50.0. Diet quality varied substantially across the world. Coastal Mediterranean nations, the Caribbean region, and Eastern Asia (except China and Mongolia) had a higher AHEI score, whereas Central Asia, the South Pacific, and Eastern and Northern Europe had a lower score. An improvement in dietary quality from the current global diet to the reference healthy diet could prevent >11 million premature deaths, ∼24% of total deaths in 2017. These included 1.6 million cancer deaths, 3.9 million coronary artery disease deaths, 1.0 million stroke deaths, 1.7 million respiratory disease deaths, 0.4 million neurodegenerative disease deaths, 0.5 million kidney disease deaths, 0.6 million diabetes deaths, and 1.2 million digestive disease deaths. Global dietary quality is slowly improving, but remains far from optimal and varies across countries. Improvements in dietary quality have the potential to reduce mortality rates substantially.

How income and food prices influence global dietary intakes by age and sex: evidence from 164 countries

BackgroundWhile income and prices are key drivers of dietary choices, how their influence varies by food category, nation, and demographics is not well established. Based on intake data for 164...

Global Dietary Data Availability: Results of a Systematic Worldwide Search for Data on Dietary Intakes of 55 Foods and Nutrients

IntroductionQuantification of dietary intakes of key foods and nutrients across the lifecourse in all regions of the world is foundational to developing/targeting population‐based dietary policies and interventions to improve both maternal/child health and reduce chronic disease burdens. Yet currently, comprehensive/comparable data on dietary intakes around the world remains sparse in low/ middle‐income countries (LMIC), many of which currently bear dual burdens of under‐ and over‐nutrition characteristic of rapid nutrition transition.ObjectiveTo systematically search for and assess the availability/quality of nationally or subnationally‐representative dietary intake data in 188 countries around the world, with particular focus on children and pregnant women in LMIC.MethodsTo identify, assess, and harmonize dietary data from around the world, we established the Global Dietary Database (GDD) Consortium, a global collaborative effort to compile individual‐level data on dietary intakes of 55 foods and nutrients between 1990 and 2015 stratified by age, sex, urban vs. rural residence, and level of education. We systematically searched databases of published literature to identify contacts for national or subnational data, as well as databases of publically available microdata. Our current data searches focus on eight world regions that encompass most LMIC : Sub‐Saharan Africa, North Africa and the Middle East, South Asia, Southeast Asia, Central Asia, Oceania, Latin America, and the Caribbean. We contacted the corresponding authors of relevant publications, inviting them to become members of the GDD Consortium and contribute dietary intake data. For countries in which no data source was identified through either PubMed searches, personal contact with experts, or searches for public data, we have launched searches of region‐specific databases. To identify publically‐available data, we have searched large databases such as WHO Global Health Observatory data repository, the Harvard Dataverse, the Inter‐university Consortium for Political and Social Research (ICPSR), and the Demographic and Health Surveys (DHS), as well as country ministry of health website. Identified data is currently being graded for quality of both dietary assessment and survey sampling methodology.ResultsTo‐date, we have completed systematic searches of PubMed, screening a total of 2,878 articles and identifying ~600 unique data owners around the world. Thus far, we have identified 139 country‐specific surveys through personal contact with data owners. Overall, 36% of these data are nationally representative, 33% are based on multiple 24 hr recall, 53% are based on food‐frequency questionnaires, 38% include data on pregnant/nursing women, and 66% include data on children and adolescents. Additionally, we have identified over 800 publically‐available country‐year‐specific surveys that may have relevant data on dietary intakes or biomarkers, and we are in the process of assessing these surveys for relevance and quality (Figure).ConclusionsThrough systematic global data searches, we have identified numerous surveys on dietary intakes of key foods and nutrients. These data, once harmonized, will provide comprehensive estimates of dietary intakes of key foods and nutrients in countries around the world, crucial for understanding diet in poor and vulnerable populations and for characterizing epidemiologic transition in LMIC.Support or Funding InformationThis work was funded by the Bill & Melinda Gates Foundation.

Abstract P265: Macrosocial Disparity, Diet, and Cardiovascular Mortality in Global Populations Undergoing Transition

Background: Economic and social disparities are linked to poor health outcomes in diverse countries worldwide. Differences in diet may mediate many of these effects, yet to-date, the impact of social, economic, and demographic inequality on the quality and diversity of diet has not been examined in a comparable manner globally. Objective: To comparably quantify the relationships of macroeconomic indicators of disparity with diet quality in 84 countries worldwide, and to assess their associations with CVD mortality. Methods: Data on country-specific per-capita GDP, urbanization, income inequality (income Gini coefficient and share of income held by top 1%), education, and educational inequality (educational Gini); on dietary quality and diversity; and on CVD mortality rates were obtained from our Global Dietary Database and other sources (Figure). Relationships between income inequality, diet quality, and CVD mortality were examined through multivariate regression controlling for country-level per-capita GDP, including country-level fixed effects, as well as through non-parametric Loess regression. Analyses were performed separately for 1990, 2010, and for the percent change between 1990 and 2010 for each variable. Results: From 1990-2010 income inequality increased by >5% in 51 of 84 countries and decreased by > 5% in 15/84 countries. Of countries with increasing income inequality, 61% had worsening scores for unhealthy diet; of those with decreasing income inequality, only 27% had worsening unhealthy diet scores. CVD mortality increased in 43% of countries with increasing income inequality, primarily in Eastern Europe/Asia; while it declined in ½ of countries with declining income inequality (Figure). Results for educational inequality and urbanization are in progress. Conclusions: This first cross-country analysis of macrosocial disparities, diet quality and CVD mortality highlights regional heterogeneity and time variance the impacts of these inequities on diet and subsequent risk of CVD.

Global Expanded Nutrient Supply (GENuS) Model: A New Method for Estimating the Global Dietary Supply of Nutrients.

Insufficient data exist for accurate estimation of global nutrient supplies. Commonly used global datasets contain key weaknesses: 1) data with global coverage, such as the FAO food balance sheets, lack specific information about many individual foods and no information on micronutrient supplies nor heterogeneity among subnational populations, while 2) household surveys provide a closer approximation of consumption, but are often not nationally representative, do not commonly capture many foods consumed outside of the home, and only provide adequate information for a few select populations. Here, we attempt to improve upon these datasets by constructing a new model—the Global Expanded Nutrient Supply (GENuS) model—to estimate nutrient availabilities for 23 individual nutrients across 225 food categories for thirty-four age-sex groups in nearly all countries. Furthermore, the model provides historical trends in dietary nutritional supplies at the national level using data from 1961–2011. We determine supplies of edible food by expanding the food balance sheet data using FAO production and trade data to increase food supply estimates from 98 to 221 food groups, and then estimate the proportion of major cereals being processed to flours to increase to 225. Next, we estimate intake among twenty-six demographic groups (ages 20+, both sexes) in each country by using data taken from the Global Dietary Database, which uses nationally representative surveys to relate national averages of food consumption to individual age and sex-groups; for children and adolescents where GDD data does not yet exist, average calorie-adjusted amounts are assumed. Finally, we match food supplies with nutrient densities from regional food composition tables to estimate nutrient supplies, running Monte Carlo simulations to find the range of potential nutrient supplies provided by the diet. To validate our new method, we compare the GENuS estimates of nutrient supplies against independent estimates by the USDA for historical US nutrition and find very good agreement for 21 of 23 nutrients, though sodium and dietary fiber will require further improvement.

Assessing global dietary habits: a comparison of national estimatesfrom the FAO and the Global Dietary Database

Background: Accurate data on dietary habits are crucial for understanding impacts on disease and informing policy priorities. Nation-specific food balance sheets from the United Nations FAO provided the only available global dietary estimates but with uncertain validity.Objectives: We investigated how FAO estimates compared with nationally representative, individual-based dietary surveys from the Global Dietary Database (GDD) and developed calibration equations to improve the validity of FAO data to estimate dietary intakes.Design: FAO estimates were matched to GDD data for 113 countries across the following 9 major dietary metrics for 30 y of data (1980–2009): fruit, vegetables, beans and legumes, nuts and seeds, whole grains, red and processed meats, fish and seafood, milk, and total energy. Both absolute and percentage differences in FAO and GDD mean estimates were evaluated. Linear regression was used to evaluate whether FAO estimates predicted GDD dietary intakes and whether this prediction varied according to age, sex, region, and time. Calibration equations were developed to adjust FAO estimates to approximate national dietary surveys validated by using randomly split data sets.Results: For most food groups, FAO estimates substantially overestimated individual-based dietary intakes by 74.5% (vegetables) and 270% (whole grains) while underestimating beans and legumes (−50%) and nuts and seeds (−29%) (P < 0.05 for each). In multivariate regressions, these overestimations and underestimations for each dietary factor further varied by age, sex, region, and time (P < 0.001 for each). Split–data set calibration models, which accounted for country-level covariates and other sources of heterogeneity, effectively adjusted FAO estimates to approximate estimates from national survey data (r = 0.47–0.80) with small SEs of prediction (generally 1–5 g/d).Conclusions: For all food groups and total energy, FAO estimates substantially exceeded or underestimated individual-based national surveys of individual intakes with significant variation depending on age, sex, region, and time. Calibration models effectively adjusted the comprehensive, widely accessible FAO data to facilitate a more-accurate estimation of individual-level dietary intakes nationally and by age and sex.

Abstract P358: Assessing Global Dietary Habits: A Comparison of National Estimates from the Food and Agricultural Organization and the Global Dietary Database

Background: Accurate data on dietary habits are crucial for understanding impacts on disease and informing policy priorities. National food availability estimates from the UN Food &amp; Agricultural Organization (FAO) have been widely used as the only global data available, but with uncertain validity. Objective: To investigate the validity of FAO data compared to nationally representative data on individual-based dietary intakes from the Global Dietary Database (GDD). Methods: We compared 30 y (1980-2010) of FAO food supply data, comprising 101 food items/y, matched to corresponding GDD survey data for113 countries across major food groups, including fruits &amp; fruit juices, vegetables, beans &amp; legumes, nuts &amp; seeds, whole grains, red &amp; processed meats, fish &amp; seafood, milk, and total energy. Absolute and percent differences in FAO-GDD country-year pairs were compared using t-tests. Statistical significance of heterogeneity was tested using likelihood ratio tests comparing nested models with and without interaction terms. Correction equations were developed and validated using randomly split datasets, using multivariate regression for FAO estimates as predictors of GDD results. Results: FAO estimates substantially overestimated individual-based dietary intakes for all food groups except beans &amp; legumes ( p &lt;0.0001) and nuts &amp; seeds ( p &lt;0.05). Modest heterogeneity in estimation was evident by age (Figure),sex, diet assessment method, and survey representativeness; substantial heterogeneity was observed by world region for all food groups ( p &lt;0.01). Results for split-data correction equations to be presented. Conclusion: For most food groups, FAO estimates were substantially different than individual-based national surveys, with significant heterogeneity according to several factors. Future national and global dietary analyses should rely on individual-based national survey data or multivariable-corrected FAO data for accurate assessment of undernutrition, disease burdens, and relevant policy.

Abstract MP75: The Influence of Income on Global Dietary Habits by Country, Age, Sex, and Time

Background: Income may influence diet, especially in low and middle-income countries, and particularly in the setting of relatively high recent food prices and the global financial crisis. Yet, the influence of income on dietary habits in most countries is not well established, nor how this influence varies by age, sex, or time across different countries. Aim: To assess the relation of income with dietary habits globally, including from least developed to highly industrialized countries, and by age, sex, and time. Methods: We estimated the income elasticities of consumption of 11 food categories in 180 countries using the Global Dietary Database, based on 325 national surveys on individual-level dietary intakes worldwide. Per-capita income by country in real terms and on purchasing power parity basis were obtained from the World Bank. Elasticities were modelled using a modified Working’s expenditure share model by country, age (20-80+ in 5-year intervals), sex, and time (1990, 2005, 2010). Results: Significant income elasticities were identified. Food intake in low-income countries was generally more responsive to income than in high-income countries ( Figure ). In addition, income elasticity of dietary intake varied for different items (fruit, fruit juices, vegetables, beans/legumes, nuts/seeds, whole grains, unprocessed red meat, processed meat, milk, sugar sweetened beverages, fish). Additional heterogeneity was evident by age and sex, that also varied by food item. Conclusions: Intakes of different foods are variably responsive to income, and diets in low-income countries are more sensitive to income changes than in high-income countries, even after accounting for purchasing power. Additional food-specific patterns of variation in income elasticity were seen by age and sex. These first-ever global findings of dietary income elasticity based on individual-reported dietary intakes and by age and sex inform the relevance and impact of underlying societal determinants of dietary habits across the world.