key: cord-0319053-du404slv authors: Gatica Dominguez, G.; Neves, P. A. R.; Barros, A. J. D.; Victora, C. G. title: Complementary feeding practices in 80 low- and middle-income countries: prevalence and socioeconomic inequalities in dietary diversity, meal frequency and dietary adequacy date: 2020-12-02 journal: nan DOI: 10.1101/2020.12.01.20241372 sha: 7dc9c75af587285cc11ae717b969e210feee6a1e doc_id: 319053 cord_uid: du404slv Objective. To describe patterns and socioeconomic inequalities in complementary feeding practices among children aged 6-23 months in 80 low and middle-income countries (LMICs). Methods. We analyzed national surveys carried out since 2010. Complementary feeding indicators for children aged 6-23 months included minimum dietary diversity (MDD), minimum meal frequency (MMF) and minimum acceptable diet (MAD). Between- and within-country inequalities were documented using relative (wealth deciles) and absolute (estimated household income) socioeconomic indicators. Results. Only 21.3%, 56.2% and 10.1% of the 80 countries showed prevalence levels above 50% for MDD, MMF and MAD, respectively. Western & Central Africa showed the lowest prevalence for all indicators, whereas the highest for MDD and MAD was Latin America & Caribbean, and for MMF in East Asia & the Pacific. Log per capita gross domestic product was positively associated with MDD (R2 = 48.5%), MMF (28.2%) and MAD (41.4%). Pro-rich within-country inequalities were observed in most countries for the three indicators; pro-poor inequalities were observed in two countries for MMF, and in none for the other two indicators. Breastmilk was the only type of food with a pro-poor distribution, whereas animal-source foods (dairy products, flesh foods and eggs) showed the most pronounced pro-rich inequality. Dietary diversity improved sharply when absolute annual household incomes exceeded about US$20,000. There were no consistent differences among boys and girls for any of the indicators studied. Conclusion. Monitoring complementary feeding indicators in the world and implementing policies and programs to reduce wealth related inequalities are essential to achieve optimal child nutrition. During the first two years of life, all children must be optimally breastfed and receive an appropriate and diverse diet from six months of age in order to achieve optimal growth and development [1] [2] [3] [4] . Departures from optimal growth vary in different groups of countries. In most high-income countries (HIC), childhood overweight/obesity is a major concern, partially caused by high-energy-dense diets 5 . In low-and middle-income countries (LMIC), stunting (low height-for-age) and micronutrient deficiencies are more prevalent due to poor-quality diets 6 . The introduction of a healthy diet in early childhood contributes to better food preferences and health outcomes throughout the life course 6 . In 2007, the World Health Organization (WHO) proposed a set of complementary feeding indicators for monitoring infant and young child feeding (IYCF) practices among children aged 6-23 months 7, 8 . The core indicators address the diversity (minimum dietary diversity or MDD) and frequency (minimum meal frequency or MMF) of child diets. A third indicator -minimum acceptable diet or MADrelates to child diets that met both diversity and frequency requirements. Analyses conducted in South Asia using national surveys found that children whose diets complied with the IYCF recommendations were less likely to be ill or malnourished 9, 10 . Socioeconomic inequalities represent a major threat to optimal feeding practices. 11, 12 Using the 2007 definitions, a UNICEF report analyzed data from up to 87 national surveys in LMICs. The study found low overall prevalence levels: 29.4% for MDD, 52.2% for MMF, and only 16% for MAD. The report also found that prevalence of the three indicators increased with household wealth within countries 13 . In 2018, UNICEF and WHO updated the definitions of the three indicators, mainly in order to refine analyses of the diets of breastfed children 14 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint three complementary feeding indicators requirements were Sub-Saharan Africa and Latin America & Caribbean, respectively. MDD prevalence ranged from 18% to 54%, MMF from 41% to 72%, and MAD from 9% to 40%. Stark disparities by wealth quintile were observed in most LMICs studied, particularly for MDD, which was also positively associated with GNI PPP at country level 15 . As far as we are aware, there are no multi-country analyses using the 2018 definitions of the three indicators. The Sustainable Development Goals, part of the 2030 Agenda for Sustainable Development 16 call for action towards a better future for all, which includes appropriate diets for children, addressing goals 2 (zero hunger) and 3 (good health and well-being). Disaggregated analysis by socioeconomic indicators, including recent nationally representative surveys carried out in LMICs, are essential to track progress and identify challenges regarding complementary feeding practices. In the present analyses, we describe wealth-related inequalities in complementary feeding practices among children aged 6-23 months in 80 LMICs. We used the 2018 definitions and provide breakdowns by wealth deciles, to allow greater granularity than wealth quintiles, and also by estimated absolute income of households in international dollars. The database of the International Center for Equity in Health (www.equidade.org) includes over 400 national surveys with information on child health and nutrition in LMICs. We selected the most recent survey in each country, carried out since 2010, that included information on the three complementary feeding indicators described below, and sample sizes of at least 25 children aged [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; https://doi.org/10.1101/2020.12.01.20241372 doi: medRxiv preprint surveys rely on multistage sampling procedures, selecting regions within countries, administrative units within each region (e.g., municipalities), census tracts within each administrative unit, and households within each tract. All women aged 15-49 years from selected households are invited for an interview on the nutrition and health of their under-five children. Further information on survey methodology is available in each survey's published national reports. Three complementary feeding indicators were estimated for children aged 6-23 months, based on 24-hour dietary recall. 14 For each of the eight food groups used to calculate the MDD indicator, we also reported the percentage of children who during the previous day consumed foods or beverages from each of the eight food groups: 1) breastmilk; 2) cereals and grains (grains, white/pale starchy roots, tubers, and plantains); 3) legumes and nuts (beans, peas, lentils, nuts, and seeds); 4) dairy products (milk, infant formula, yogurt, cheese); 5) flesh foods (meat, fish, poultry, organ meats); is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; Wealth deciles. Household asset scores, generated through principal component analysis (PCA), were available in the DHS and MICS datasets. The PCA includes variables on household assets, building materials, and utilities like water and electricity, which are adjusted for the place of residence. 18 The first component of the PCA, a continuous variable, was used to classify households into wealth deciles, with the first decile (D1) representing the poorest 10% of all families and the tenth decile (D10) representing the wealthiest 10% of all families. Per capita gross domestic product (GDP). This indicator is expressed in current international dollars converted by purchasing power parity (PPP), is the sum of the gross value added by all resident producers in the country plus any product taxes and minus any subsidies not included in the value of the products. The PPP conversion factor is a spatial price deflator and currency converter that eliminates the effects of the differences in price levels between countries. 19 We obtained the GDP data through the wbopendata module 16.3, which draws from the core World Bank development indicators. It presents the most current and accurate global development data available, compiled from officially-recognized international sources 20 . Absolute income for each wealth decile. This was calculated based on the national income levels obtained from the World Bank database (http://databank.worldbank.org/data/reports.aspx?source=world-developmentindicators), and national income inequality data collected from the Standardized World Income Inequality Database (https://www.wider.unu.edu/project/wiidworld-income-inequality-database). Dollar values (2011 purchase power parity adjusted international dollars) were then assigned to each household wealth decile, accounting for income's log-normal distribution 21 . The SII is a summary measure of absolute inequality, which is calculated through logistic regression models with the natural logarithms of the odds of the complementary feeding variables as the outcomes, and the wealth deciles as the independent . CC-BY 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; variable. The SII represents the difference in the fitted value of the outcome between the highest and the lowest values of the wealth index scale 22 , and is interpreted as percentage points (p.p.). For each country included in the analyses, we estimated the prevalence of the three indicators at national level, by sex of the child and by wealth decile, and calculated the SII and its 95% confidence interval (CI). Next, we grouped the countries according to UNICEF world regions and World Bank income group classifications for the year of the survey 23,24 . Regional and income group estimates were weighted by the size of the population of children aged 6 to 23 months in the year when the survey was conducted 25 . Equiplot graphs were used to illustrate how weighted mean prevalence varied by wealth deciles. We fitted polynomial equations for each complementary feeding indicator according to log GDP at national level, but this procedure did not improve the fit of the model compared to a linear equation. We present R 2 values for the linear models, which express the proportion of the is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; https://doi.org/10.1101/2020.12.01.20241372 doi: medRxiv preprint the institutions that conducted the surveys in each country handled the respective ethical clearance. The most recent surveys were analyzed for 80 countries, with dates ranging from 2010 to 2019 Figure 1 shows the ecological analyses with countries as the units. There were direct linear associations between log GDP per capita and MDD (R 2 = 48.5%), MMF (R 2 = 28.2%) and MAD (R 2 = 41.4%). MDD was also correlated to MMF (r = 0.68; p<0.001; data not shown). Table 1 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. Inequalities in the consumption of cereal and grains, legumes and nuts, and vitamin A rich fruits and vegetables were small, but for the other four food groups, particularly dairy products, tended to be wide (Table 3) . Additional results of food groups by world regions are presented in Supplementary table 6. In the last set of analyses, the three complementary feeding indicators were plotted against absolute income. In Figure 2 , the points represent the 800 deciles in all countries included in the analyses, and the lines are fractional polynomials for the three World Bank country income . CC-BY 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. Our analyses add to the literature by presenting the first report on the three internationally 15 had also shown that minimum dietary diversity had lower prevalence than minimum feeding frequency. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; https://doi.org/10.1101/2020.12.01.20241372 doi: medRxiv preprint We found that East Asia and Pacific was the region with the highest mean MMF, while the Latin America and Caribbean region had the best performance for MDD and MAD. In contrast, the earlier analyses by White et al 13 showed that East Asia and Pacific had the highest values for the three indicators. These differences may be due to changes in the definition of the indicators, and to the fact that the number of countries in the analyses varied between the two studies. In both studies the lowest mean of the weighted prevalence for all three complementary feeding indicators were observed in the two Sub-Saharan Africa regions and in South Asia. There were striking wealth related inequalities, with pro-rich patterns present in betweencountry and within-country analyses, in all seven world regions. Dietary diversity started to improve when absolute household income exceeded about US$20,000. The analyses by Baye and colleagues also found a direct association between GDP and diversity. 15 Similar patterns were observed for within-country inequalities, which tended to be wider for diversity than for frequency in five of the seven regions of the world, the exceptions being Latin America & Caribbean and Middle East & North Africa. In contrast to earlier analyses relying on wealth quintiles, our results by wealth decile were able to document socioeconomic gradients with greater granularity, while also confirming earlier reports of wider inequalities for dietary diversity than for frequency at global and regional level. 13, 15 When we analyzed each of the eight food groups, the widest inequalities were observed for consumption of animal-source foods, mainly caused by dairy products consumption, followed by flesh foods and eggs, and for consumption of fruits and vegetables other than those rich in vitamin A. Such inequalities were likely due to the high cost of these foodstuffs. There were virtually no socioeconomic differences in consumption of cereals and grains, legumes and nuts, and vitamin A rich fruits and vegetables. The only food group with higher consumption among children from poor families was breastmilk, a finding that is in accordance with the literature. 26 The limitations of our study include the fact that indicators are derived from 24-recall by the survey respondents rather than actual observation and measurement of feeding patterns. . CC-BY 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; https://doi.org/10.1101/2020.12.01.20241372 doi: medRxiv preprint Although it is possible that respondents may have overreported on the types of foodstuffs and frequencies of feeds, the fact that prevalence of adequate complementary feeding was low suggests that poor diets are a major problem, particularly in low-income countries. Another limitation is that no data were available for 60 of all 140 LMICs 24 ; these included several upper-middle income countries such as China and Brazil, where standardized surveys have not been conducted in recent years. Among the strengths of our analyses, this is the first comprehensive study that complied with the new definitions of complementary feeding indicators in LMICs, while also addressing both relative and absolute inequalities using different socioeconomic indicators and with greater granularitywealth deciles rather than quintilesthan previously reported studies. In addition, we explored socioeconomic inequalities according to food groups. Inadequate complementary feeding practices are a major determinants of poor child growth and development [1] [2] [3] [4] . Regular monitoring of dietary adequacy is an essential component for tracking progress towards the health and nutrition-related Sustainable Development Goals. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted December 2, 2020. ; https://doi.org/10.1101/2020.12.01.20241372 doi: medRxiv preprint Maternal and child undernutrition: global and regional exposures and health consequences Maternal and child undernutrition and overweight in low-income and middle-income countries Maternal and child nutrition: building momentum for impact World Health Organization. Infant and young child feeding Worldwide trends in children's and adolescents' body mass index, underweight, overweight and obesity, in comparison with adults, from 1975 to 2016: A pooled analysis of 2,416 population-based measurement studies with 128 World Health Organization. Indicators for assessing infant and young child feeding practices: part 1: definitions: conclusions of a consensus meeting held 6-8 World Health Organization. Indicators for assessing infant and young child feeding in Bangladesh Prevalence and factors associated with complementary feeding practices among children aged 6-23 months in India: a regional analysis Infant feeding practices and dietary consumption of US infants and toddlers: National Health and Nutrition Examination Survey (NHANES) Predictors of complementary feeding practices in Afghanistan: Analysis of the 2015 Demographic and Health Survey Complementary feeding practices: Current global and regional estimates United Nations Fund for Children. Meeting report -Interagency technical consultation on infant and young child feeding indicators 11-13 Estimates of dietary quality in infants and young children (6-23 months): evidence from demographic and health surveys of 49 low-and middle-income countries' United Nations. The Sustainable Development Goals Measuring coverage in MNCH: tracking progress in health for women and children using DHS and MICS household surveys The DHS Wealth Index: Approaches for rural and urban areas