Christian et al. BMC Nutrition (2023) 9:9 BMC Nutrition https://doi.org/10.1186/s40795-022-00667-9 RESEARCH Open Access Infant and young child feeding practices are associated with childhood anaemia and stunting in sub-Saharan Africa Aaron Kobina Christian1*, Eric Afful‑Dadzie2 and Grace S. Marquis3 Abstract Background The co‑occurrence of anaemia and stunting (CAS) presents acute development and morbidity chal‑ lenges to children particularly in sub‑Saharan Africa (SSA). Evidence on the effect of child feeding recommendations on CAS is scarce. Methods We used data from 22 recent Demographic and Health Surveys in SSA countries to examine the associa‑ tion between caregivers’ implementation of recommendations on infant and young child feeding and the CAS in their 6‑ to 23‑mo‑old children. Results Overall, in multiple logistic regression models, child feed index score, high wealth of household, increas‑ ing household size, household head with at least secondary school education, improved sanitation of household, an increase in caregiver’s age and caregiver’s with at least secondary education were associated with lower odds of CAS (i.e., AOR: 0.86; 95% CI; 0.84 – 0.88: 0.75; 0.69 – 0.82: 0.98, 0.98 – 0.99: 0.76, 0.70 – 0.83: 0.81, 0.74 – 0.87: 0.87, 0.81 – 0.94: 0.69, 0.62 – 0.77 respectively). Having a diarrhoea in the past 2 weeks and having fever in the past month were associated with higher odds of CAS (AOR:1.1, 95% CI; 1.0 – 1.2: 1.1, 1.0 – 1.2, respectively). Results from the decision tree analysis showed that the educational level of women was the most important predictor of CAS, followed by child feeding score, the level of education of the family head and state of drinking water. Conclusion The results buttress the importance of interventions aimed at improving feeding practices and parental educational as a vehicle to improve children’s nutritional status. Keywords Anaemia, Stunting, Co‑occurrence of anaemia and stunting: feeding practice, Decision tree Introduction vitamin B12  and vitamin A),  infections (e.g., soil-trans- Anaemia is defined as having low haemoglobin concen- mitted helminths, malaria, tuberculosis and HIV/AIDS) tration or red blood cell mass as resulting from either and/or due to genetic factors (e.g., thalassemia) [1].The nutritional factors  (e.g., deficiencies of iron, folic acid, global prevalence of anaemia in 2010 was estimated at 32.9% with sub-Saharan Africa (SSA) among the hard- est hit regions alongside South Asia [2]. According to the *Correspondence: Aaron Kobina Christian World Health Organization (WHO), an estimated 273.2 akchristian@ug.edu.gh million children aged 6–59 months were anaemic in 2011 1 Regional Institute for Population Studies, University of Ghana, [3]. Stunting on the other hand is assessed by measuring Legon‑Accra, Accra, Ghana 2 Department of Operations and Management Information Systems, child’s length or height and interpretating the measure- University of Ghana Business School, Accra, Ghana ment by comparing them with an acceptable set of stand- 3 School of Human Nutrition, McGill University, Montreal, QC, Canada ard values per the child’s sex and age category. Stunting © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4.0 /. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Christian et al. BMC Nutrition (2023) 9:9 Page 2 of 13 is considered as the best indicator for assessing children’s and childhood nutrition outcomes (including anaemia, well-being. Over the past two decades whereas global stunting, and the co-occurrence of anaemia and stunting numbers of stunted children have decreased in most (CAS)), which was the primary reason for which the indi- regions, SSA has witnessed an increase [4]. cators were generated. The aim of the current study is to Child growth was captured as a key indicator of devel- use the most recent nationally representative data across opment by the Millennium Development Goals [5] and the SSA to examine the relationship between IYCF prac- subsequently within the Sustainable Development Goals tices and CAS in 6- to 23-months-old children. The key (SDGs) [6]. The SDG Target 2.2. aims to end all forms of research questions addressed in this analysis are (1) Does malnutrition, including stunting in children. Conversely, IYCF predict CAS and (2) what is the relative importance childhood anaemia influences under five years mortal- of the association of other covariates to CAS in sub-Saha- ity, delayed motor development, reduced mental capac- ran Africa? ity and poor educational attainment [7] and substantial economic burden [8, 9]. The consequence of stunting Methods includes low educational performance, suboptimal func- Data sources and procedures tion later in life and, for pregnant women, adverse effects We used data from the most recent Demographic and such as restricted uterine blood flow and growth of the Health Surveys (DHS) of selected sub-Saharan Africa uterus, placenta, and foetus [10]. countries (SSA) provided by the Integrated Public Use Most undernutrition problems share overlapping risk Microdata Series (IPUMS) (Minnesota Population Cen- factors, especially with respect to their proximate deter- tre). The DHS employs a stratified two-stage cluster minants [15, 16]. Factors such as poor socioeconomic sampling technique to select census enumeration based status, inadequate health service utilization, suboptimal on probability (proportional to area size), then a ran- childcare practices and poor water and sanitation often dom selection of households from a listing of households serve as correlates of multiple health conditions includ- within the selected enumeration area. These surveys ing malnutrition outcomes [11, 12]. Given the overlap- collect household and individual information on a wide ping causes of anaemia and stunting, their co-occurrence range of themes, using comparable questionnaires across would be expected to occur particularly in poor childcare study countries. For selected households, a member who settings [13]. is usually the household head answers general questions There have been several adaptations of the oft-cited on household socioeconomic and demographic charac- United Nations Children’s Fund conceptual framework teristics through a face-to face interview. that speaks to the correlates of childhood nutritional sta- All consenting women within the reproductive age tus. An example of such a framework is the WHO frame- bracket (i.e., 15 to 49  years old) are also interviewed work that gives importance to inadequate complimentary on questions such as their reproductive health, their child feeding practices -poor quality food, low dietary children’s dietary intake and their child feeding prac- diversity and intake of food, infrequent and inadequate tices. Caregivers also provide information on the index feeding, and insufficient frequency of feeding[14, 15]. It is children (i.e., children between 6 to 59  months) in the generally agreed that appropriate infant and young child selected households. Children’s length/height and weight, feeding (IYCF) is likely to increase child survival and pro- and haemoglobin concentration are also assessed. The mote optimal growth and development and has the most surveys also gather information on the household mem- impact on children between six months and two years. In bers and characteristics of the household. 2002, the WHO and UNICEF adopted a  global strategy IPUMS DHS was established in 2012 to support for IYCF aimed at bringing global attention to the impact comparative over-time and international research with of feeding practices. The WHO has since introduced a the DHS [18]. IPUMS provides added value to nation- set of indicators for complementary feeding among chil- ally represented survey and census data by coding dren aged 6 to 23 months to track and encourage optimal variables consistently across different countries and feeding practices [16]. over time. IPUMS also provides a systems integration Although research on the effect of single feeding prac- and documentation which enables researchers to eas- tice on nutritional outcomes is critical, it does not permit ily merge information from different countries and the examination of child feeding practices on children’s across different data types [19]. Interested variables health and nutrition outcome [17]. Apart from a study and samples (i.e., countries) were selected on IPUMS by Gebremedhin (2019), pooled indicators for IYCF data request portal and submitted. The pooled data us practices across the entire SSA have not been published. then downloaded automatically. The following were Additionally, critically missing is the adequate scholar- the inclusion criteria for selecting surveys for the cur- ship examining the relationship between IYCF practices rent analysis: (i) only sub-Saharan African countries, C hristian et al. BMC Nutrition (2023) 9:9 Page 3 of 13 (ii) survey period between 2008 and 2017, and (iii) 2nd November 2022). Written informed consent was included children’s anaemia and stunting status. The obtained from all study participants. DHS is a nationally representative cross-sectional household survey that is conducted approximately Measures every 5 years. All DHS follow standardised procedures Main outcome in the survey design and data collection across study Childhood anaemia was defined as having a haemoglobin countries [20]. (Hb) level of < 11.0  mmol/L in children 6 to 59  months old, based on WHO classification [3]. The length/height- for-age Z-scores (LAZ/HAZ) for children were calculated Ethical considerations based on child sex, age, and length (if aged < 2 years) or Ethical clearances for the DHS studies are sought from height (if aged ≥ 2  years) using WHO Anthro software the Ethics Committee of ORC Macro Inc. and from version 3.2 [21]. Children with a LAZ/HAZ less than –2 the Institutional Review Boards and Ethics Boards of were considered stunted. Thus, CAS was considered as partner institutions (e.g., Ministries of Health) of the having both a Hb level < 11.0 mmol/L and LAZ/HAZ < -2. studied countries. Protocols adopted for the DHS ensure that standards for the protection of respond- Main independent variable‑IYCF practice ents’ privacy and confidentiality are strictly adhered An IYCF index was created for this study based on age- to. The Inner-City Fund International ensures that the specific (6 to 8 months, 9 to 23 months) feeding recom- survey conforms with the regulations for the respect mendations (Table  1). Breastfeeding is recommended of human subjects as stipulated by the United States for at least 24 months [22]; 2 points were given for chil- Department of Health and Human Services. All proce- dren in all age categories who were breastfed and 0 if not dures were performed in accordance with guidelines in breastfed. Use of bottles are discouraged [23]; 1 point was accordance with the ethical standards as laid down in given for not using a bottle and 0 if a bottle was used. the 1964 Declaration of Helsinki and its later amend- Dietary diversity was evaluated based on the number ments or comparable ethical standards. As this study of food groups a child ate in the past 24 h using a list of used data from the DHS provided through IPUMS, no seven food groups: grains, roots and white tubers; leg- further ethical approval or consideration was required. umes; nuts and seeds; dairy products; flesh foods (meat, Additional information concerning protocols and poultry, fish, and offal); eggs; pro-vitamin A-rich foods ethical standards of the DHS can be found at https:// (yellow and orange-fleshed roots and tubers, orange- dhspr ogram. com/ metho dology/ Prote cting- the- Priva fleshed fruits, and dark green leafy vegetables); and other cy- of-D HS- Survey- Respon dents.c fm ( accessed on fruits and vegetables [24]. A published scoring scheme Table 1 Variables and scoring system used to create the child feeding index for 6 to 23 months old children, by age group Age category Variables 6 to 8 months 9 to 23 months Rationale for scoring Is breastfed No = 0; Yes = 2 No = 0; Yes = 2 Breast milk provides at least half of the energy requirements for children 6 to 12 months and one third for those between 12 to 24 months. Reduces risk of acute disease and chronic diseases provides long‑term coronary health (Brown et al., 1990; Kathyrn Dewey, 2008) Uses bottle No = 1; Yes = 0 No = 1; Yes = 0 Bottle feeding is linked to unhygienic conditions increasing the risk of illness and mortality (Boone et al., 2016; Muhammad et al., 2018) For past 24 h Dietary diversity scorea 0 = 0 0 = 0 Having a more diverse diet is a proxy for adequate micronutrient‑dense foods. Con‑ 1 to 3 = 1 1 to 3 = 1 sumption from at least 4 food groups over 24 h. is considered adequate and likely ≥ 4 = 2 ≥ 4 = 2 to contain at least one fruit or vegetable and one animal source food in addition to a staple food, usually cereals or other starchy staples (Working Group on Infant and Young Child Feeding Indicators, 2006) Meal frequency 0 meals/d = 0 0 meals/d = 0 The required meal frequency for children is dependent on their energy requirement 1 meal/d = 1 1 to 2 meals/d = 1 (Kathyrn Dewey, 2008) ≥ 2 meals/d = 2 ≥ 3 meals/d = 2 Maximum score 7 points 7 points a Dietary diversity from 24-h recall; maximum score of 7 when a child consumes food items from all 7 food groups: grains and tubers, legumes, dairy, flesh (red meat, fish, offal), eggs, vitamin A rich fruits, fruits and vegetables Christian et al. BMC Nutrition (2023) 9:9 Page 4 of 13 for diet diversity was modified and used as part of a feed- For individual countries, binary logistic regression was ing index as follows: for all age categories, a score of 2 used to test whether the magnitude and statistical signifi- points for eating from 4 or more food groups, 1 point was cance of the association between the child feeding index awarded to those who ate from 1- 3 food groups, and 0 and CAS remained after controlling for other covariates was assigned to children who did not eat from any food noted to influence anaemia and stunting. group over the past 24  h. Based on the recommended We further used decision tree analysis to comprehen- meal frequency, 6- to 8-mo-old and 9- to 23-mo-old chil- sively understand the weight of the factors that influ- dren who were fed at least twice and thrice times, respec- enced CAS observed, that is to examine correlates with tively, were given 2 points. One point was given for fewer respect to the extent of their importance in predicting meals and 0 for no meals. Taking cue from previous stud- the outcome of interest. A decision tree is a non-para- ies a final IYCF index was generated as a summation of metric supervised learning method used for both clas- the scores obtained for each variable described [25]. The sification and predictive analysis problems. The goal is feeding scores were categorized as ‘Low’ and ‘High’ based to create a model that concurrently predicts and classi- on computation of the 50th percentile. fies the value of a dependent variable (in this case CAS) by learning simple decision rules inferred from the data Other covariates attributes [29]. Visually, it is an inverted tree where each Selection of covariates was guided by the UNICEF frame- node represents a given data attribute, each branch a work for childhood malnutrition and the WHO frame- decision rule (condition) and each leaf represent an out- work that gives importance to inadequate complimentary come (categorical or continuous value). In each node child feeding practices [26] and other socio-demographic of the tree, a splitting rule (referred to as a condition) is characteristics that known to have an influence on child- applied to determine which class a datum should belong hood malnutrition. Household characteristics such as for a specific attribute that has a property of data asso- wealth status, total number of people (i.e., household ciated with the node [30]. The algorithm has an added size), access to improved sanitation and drinking water power of inferring the dominant attributes most impor- were included [27]. The following were considered as tant to the prediction. In this instance, the condition of improved water sources: piped water, boreholes, or tube the node which is closest to the root of the tree is consid- wells, protected dug wells, protected springs, rainwater, ered the most influential or the most important as far as and packaged or delivered water and the unimproved the prediction is concerned. sources were unprotected dug well, unprotected spring, To obtain the tree and nodes, we employed the gain river, dam, lake, pond, stream, canal, and irrigation canal. ratio methodology as the criterion on which attributes Improved sanitation facilities were flush/pour flush to were selected for splitting. The gain ratio criterion is a piped sewer systems, septic tanks, or pit latrines; ven- variant of information gain used in measuring how well a tilated improved pit latrines, composting toilets, or pit given attribute separates the training examples based on latrines with slabs with pit latrines without a slab or their target classification. The goal for the decision tree platform, hanging latrines or bucket latrines and open analysis was to identify relatively important factors that defecation considered as unimproved facilities. Child predict CAS. Automatic feature selection was performed characteristics that were selected included having expe- in the RapidMiner 2018 software (RapidMiner, Boston) rienced diarrhoea and fever of a specified period [28]. to select only attributes that contributed significantly to Whether or not a child was given either vitamin A or iron the prediction of CAS. Criteria for selection included (1) supplements within the past 6 months was also assessed. attributes (variables and the categories of that variable) that had a strong relationship with the outcome variable Statistical analysis (CAS) and (2) independent variables that had very low Factoring in the sampling design of the DHS data, sam- correlation between themselves. pling weights were calculated to account for differential probabilities of selection and participation of households Results across countries. Weighted proportions and 95% confi- Characteristics of the study sample dence intervals (CI) were calculated with STATA version The final analytic sample for assessing the association 14.2 (Stata, College Station, TX, USA). The svyset com- of child feeding index with CAS was 33,846 caregiver- mand in STATA was used to describe the sample design. child dyads across all the selected 22 SSA countries. The The svy: prefix is used before commands for running heads of the households were mostly men (79%), and models such as regress or logistic. Chi-square was used to only about 28% of the household heads had more than a test the association between child feeding index terciles secondary level of education. In terms of household con- (low, medium, and high) and CAS in a bivariate analysis. ditions, slightly more than half of the households (56%) C hristian et al. BMC Nutrition (2023) 9:9 Page 5 of 13 had improved water source, and 45% had improved toi- (Fig.  2). A further breakdown of the tree through the let facilities. The mean household size was 7.2 (4.6 SD). nodes revealed that, where caregivers had no educa- Over half of the countries (12/22, 55%) had less than 20% tion at all or terminated at the primary level, there was of women attaining secondary or higher education. Fif- 73.1% of CAS and 26.9% of no CAS. However, where the teen out of the twenty-two countries had 20% or more of level of education was secondary school and above, the mothers indicating that their children had diarrhoea or most important predictor towards CAS was the child fever over the two weeks prior to the survey. There was feeding index score. A feeding index score greater than no vitamin A supplementation in Angola, no iron supple- five resulted in 84.4% chance of no CAS. However, if the mentation in Zimbabwe, South Africa, Rwanda, Namibia, child feeding score was less than five, the level of educa- and Lesotho. The household, caregiver, and child charac- tion of the family head was the most important predic- teristics that may influence the nutritional outcome of tor of CAS. If the household head’s level of education was the children varied widely across the countries (Table 2). below primary school, this resulted in CAS. However, if The prevalence of households with improved water the level of education was secondary and above, state of source and access to improved sanitation facility was drinking water (i.e., having an improved water source) lowest in Madagascar (31%) and highest in Malawi (80%). was the predictor to focus on. Where there was improved Both maternal and head of household education varied water source, there was no CAS, however, all households greatly across the SSA countries. Burkina Faso and Niger without improved water source had CAS. had the lowest and South Africa had the highest educa- tional achievement. Other covariates associated CAS In the pooled analysis, a lower odds of CAS was also Nutritional status of children and child feeding practices: associated with greater household wealth, larger family Prevalence of anaemia, stunting and CAS size, higher level of education of the household head and Overall, the prevalence of anaemia, stunting and the the mother, improved sanitation, and older mothers (See CAS among the sample children were 76%, 32% and 25%, Table  4). Children with recent diarrhoea or fever had a respectively. More than half of the study countries had higher odds of CAS. There were country-specific differ- over 80% of mothers breastfeeding their 6- to 23-months- ences in how sociodemographic variables were related to old children (Table 3). South Africa was the only country CAS. For example, after controlling for other predictors, with less than half of mothers breastfeeding. In Burkina more caregiver formal education was associated with Faso, breastfeeding was close to universal (94%) and bot- a lower odd of a child being both anaemic and stunted tle feeding almost non-existent (1%) for this age group. only in six countries. Not all countries showed a negative Overall, only 16% of the children consumed from at least association between wealth and CAS. Improved sanita- four food groups. tion was protective against CAS (decreased odds) in Bur- kina Faso, Burundi, Niger, and Senegal Faso. Having had Child feeding practices and its association with nutritional childhood diarrhoea recently was a predictor of CAS in outcome Angola, Cameroon, and Zimbabwe. Figure  1 shows bivariate associations between feeding practices and CAS for each country. Overall, a lower pro- Discussion portion of children with CAS was associated with having This study aimed to examine the relationship between a high compared to a low score on the IYCF score. There IYCF recommend practices and the co-occurrence of were however some country variations in this pattern. anaemia and stunting. Arguably, this study provides the After adjusting for other covariates in the pooled data most comprehensive up-to-date evidence of the associa- set, child feeding index significantly predicted CAS (See tion of IYCF practice on CAS in children 6 to 23 months Table 4). That is, overall higher child feeding index score of age, using nationally representative samples across was associated with a lower odds of children being both sub-Saharan African countries. Based on the WHO stunted and anaemic (p ≤ 0.001). There were country- classification for assessing severity of malnutrition [31], specific differences in the models results after adjusting approximately 64% of countries in the current sample for other covariates (Supplementary Table 1). had very high prevalence of stunting (≥ 30%). Conversely, The decision tree analysis offered an alternative means the prevalence of childhood anaemia was severe [3] for all of identifying the level of relevance of attributes predict- the countries in the study (i.e., above 40%). A quarter of ing CAS. Overall, the caregiver’s educational level was the children (25%) were anaemic and stunted at the same the most important predictor of CAS, followed by child time. Overall, the child feeding index was associated feeding index score, the household head’s educational with CAS, and the decision tree analysis showed mater- level, and access to improved water source for drinking nal education as the most important predictor of CAS in Christian et al. BMC Nutrition (2023) 9:9 Page 6 of 13 Table 2 Sociodemographic characteristics and health and nutrition indicators for the study sample population, by c ountrya Angola Benin Burkina Faso Burundi Cameron Cote d’Ivoire DRC Ethiopia Ghana Guinea Lesotho Year of survey 2015 2011 2010 2016 2011 2011 2013 2016 2014 2012 2014 Household Household size (#) 6.2 ± 2.8 6.5 ± 3.2 7.5 ± 3.8 5.7 ± 2.1 7.9 ± 4.2 7.8 ± 4.5 6.8 ± 3.0 5.9 ± 2.1 5.5 ± 2.6 8.5 ± 4.5 6.0 ± 2.5 Household head Male (Yes) 71.2 85.5 91.3 79.2 85.9 84.4 78.6 87.0 75.8 85.5 70.8 Attained at least secondary education (Yes) 53.9 21.8 5.9 12.9 42.1 21.2 70.6 13.5 66.6 20.3 36.3 Access to improved water source (Yes)b 38.5 67.2 72.2 78.3 60.4 59.0 40.5 50.2 63.3 69.8 62.3 Access to improved sanitation (Yes)c 61.7 31.9 23.5 52.2 53.1 42.0 38.8 9.8 66.2 36.8 69.9 Wealth tertilesd Low 38.5 45.7 40.8 42.4 48.1 43.0 45.4 43.9 43.3 46.9 41.5 Middle 18.0 21.7 23.4 22.0 19.0 21.2 18.7 23.3 19.7 19.6 25.2 Highest 43.5 32.5 35.8 35.6 32.9 35.8 35.9 32.9 37.1 33.5 33.3 Female caregiver Age (y) 15 to 24 41.7 24.6 33.2 25.6 39.7 37.7 32.1 27.9 25.2 35.3 51.5 24 to 49 58.3 75.4 66.8 74.4 60.3 62.3 67.9 72.1 74.8 64.7 48.6 Attained at least secondary education (Yes) 33.2 14.2 5.5 11.4 34.0 9.7 39.1 7.6 53.0 10.9 56.6 Child Had diarrhea in past 2 wk (Yes) 26.1 13.3 23.1 37.8 33.8 27.2 31.2 19.4 16.1 23.0 23.5 Had fever in past 2 wk (Yes) 20.1 16.3 30.5 47.9 35.3 34.6 37.7 20.1 15.5 37.5 19.9 Vitamin A in past 6 mo (Yes) 0.0 50.2 67.4 72.4 51.7 61.7 68.7 42.7 70.1 38.1 72.5 Iron supplement in past 7 d (Yes) 11.1 39.3 8.3 7.7 9.9 16.5 15.6 7.8 22.9 10.1 0.0 Anaemia (Yes)e 78.7 67.4 94.2 69.3 73.6 86.5 68.1 72.0 77.7 85.1 61.0 Stunted (Yes)f 37.6 37.8 30.6 52.9 27.5 27.2 34.0 32.9 14.6 22.5 29.2 CAS (Yes)g 29.5 25.0 29.1 36.8 21.3 23.8 22.6 24.1 12.0 19.9 20.8 Madagascar Malawi Mali Mozambique Namibia Niger Rwanda Senegal South Africa Tanzania Zimbabwe Year of survey 2008 2016 2012 2011 2013 2012 2014 2017 2016 2015 2015 Household Household size (#) 6.0 ± 2.5 5.3 ± 2.0 7.2 ± 3.5 5.8 ± 2.4 6.7 ± 3.3 7.6 ± 3.7 5.1 ± 1.8 14.3 ± 8.5 6.3 ± 3.0 7.2 ± 4.4 5.5 ± 2.4 Household head Male (Yes) 85.5 72.5 92.1 72.3 47.3 85.7 80.6 74.0 45.7 82.9 62.7 Attained at least secondary education (Yes) 22.7 35.7 10.5 19.3 62.5 6.8 11.3 17.1 86.0 18.2 75.6 Access to improved water source for drinking 31.3 80.0 57.8 40.1 71.0 63.1 65.1 46.2 60.9 47.0 64.4 (Yes)b Access to improved sanitation (Yes)c 5.3 81.0 37.4 19.6 35.9 19.0 69.6 66.4 68.9 69.4 58.1 C hristian et al. BMC Nutrition (2023) 9:9 Page 7 of 13 Table 2 (continued) Madagascar Malawi Mali Mozambique Namibia Niger Rwanda Senegal South Africa Tanzania Zimbabwe Year of survey 2008 2016 2012 2011 2013 2012 2014 2017 2016 2015 2015 Wealth tertilesd Low 49.1 44.9 48.3 44.0 47.3 47.6 36.9 46.4 47.0 45.4 47.4 Middle 19.4 20.5 19.9 18.4 19.6 21.4 21.9 19.3 20.7 18.8 17.7 Highest 31.5 34.7 31.8 37.6 33.1 31.0 41.2 34.3 32.3 35.8 34.9 Female caregiver Age (y) 15 to 24 42.2 45.3 32.3 39.6 35.8 27.6 25.0 29.7 34.6 37.7 36.9 24 to 49 57.9 54.7 67.7 60.4 64.2 72.4 75.0 70.3 65.4 62.3 63.1 Attained at least secondary education (Yes) 20.6 23.1 9.1 12.0 72.2 6.4 13.7 18.1 90.0 17.3 65.7 Child Had diarrhoea in past 2 wk (Yes) 16.2 37.3 13.4 18.3 35.1 28.0 21.9 27.8 18.8 21.5 30.9 Had fever in past 2 wk (Yes) 14.5 36.9 13.6 17.2 33.2 22.8 25.1 28.5 26.1 22.6 17.5 Vitamin A in past 6 mo (Yes) 43.2 67.6 53.0 69.6 86.5 60.4 82.4 57.4 78.9 43.1 76.5 Iron supplement in past 7 d (Yes) 1.6 11.4 18.7 21.0 0.0 10.5 0.0 2.9 0.0 1.4 0.0 Anaemia (Yes)e 62.4 80.9 88.6 79.3 63.8 86.4 52.2 83.6 68.7 75.3 55.7 Stunted (Yes)e 48.2 32.9 30.5 41.5 17.9 38.9 37.0 16.3 33.3 32.2 26.4 CAS (Yes)g 31.2 27.3 27.5 34.3 11.3 33.6 18.6 14.4 25.7 24.6 14.9 DRC Democratic Republic of the Congo, CAS co-occurrence of anaemia and stunting a Sample includes only households with complete data on children’s height, age, and haemoglobin level. Values represent mean ± standard deviation or percentage b Improved water source: piped household water connection, public standpipe, borehole, protected dug well, protected spring and rainwater c Improved sanitation: ventilated improved pit latrines and pit latrines with a slab or covered pit. d Wealth quintiles were calculated from an asset-based wealth index using assigned asset weights from a principal components analysis to create standardized asset scores. e Anemia:hemoglobin concentration of < 110 mmol/L. f Stunted: length/height-for-age Z-score < –2 g CAS: haemoglobin level < 11.0 mmol/L and length/height-for-age Z-score < -2 Christian et al. BMC Nutrition (2023) 9:9 Page 8 of 13 Table 3 Feeding practices reported by caregivers, by countrya Angola Benin Burkina Faso Burundi Cameron Cote d’Ivoire DRC Ethiopia Ghana Guinea Lesotho Currently... Breastfeeding 77.6 72.9 93.7 91.8 67.7 75.1 87.4 88.3 83.2 89.1 64.4 Uses bottle 15.8 14.9 1.3 11.5 7.7 4.5 5.6 14.8 12.1 4.9 22.5 Dietary diversity for past 24 h Consumed from no food group 16.2 23.5 28.4 13.2 10.7 16.0 19.7 24.3 11.6 24.7 8.2 Consumed from 1—3 food groups 57.8 52.0 68.7 75.4 67.2 74.3 70.3 64.8 72.4 68.8 80.3 Consumed from 4—7 food groups 26.0 24.6 2.9 11.4 22.1 9.7 10.0 10.9 16.1 6.6 11.5 Meal frequency 6 to 8 mo old children 1.8 ± 1.5 1.3 ± 1.5 1.0 ± 1.3 1.8 ± 1.0 1.3 ± 1.5 1.4 ± 1.4 1.7 ± 1.2 1.5 ± 1.5 1.7 ± 1.4 1.0 ± 1.3 2.7 ± 1.7 9 to 11 mo old children 2.0 ± 1.4 1.9 ± 1.7 1.9 ± 1.3 2.2 ± 1.0 1.4 ± 1.6 1.8 ± 1.3 2.0 ± 1.3 2.1 ± 1.4 1.9 ± 1.2 1.5 ± 1.3 2.6 ± 1.7 12 to 23 mo old children 2.2 ± 1.4 2.5 ± 1.6 2.5 ± 1.3 2.3 ± 1.1 1.3 ± 1.7 2.5 ± 1.2 2.3 ± 1.2 2.4 ± 1.5 2.5 ± 1.2 2.3 ± 1.6 3.2 ± 1.7 Child feeding indexb 3.9 ± 1.4 3.7 ± 1.4 3.9 ± 1.0 4.2 ± 1.1 3.6 ± 1.2 3.9 ± 1.2 4.0 ± 1.2 3.9 ± 1.2 4.1 ± 1.2 3.8 ± 1.1 4.0 ± 1.3 Madagascar Malawi Mali Mozambique Namibia Niger Rwanda Senegal South Africa Zimbabwe Tanzania Currently... Breastfeeding 85.6 86.2 85.1 83.0 59.1 86.5 93.3 80.3 43.1 68.7 81.3 Uses bottle 2.1 5.6 4.9 77.1 35.4 2.4 5.1 4.8 48.1 7.7 4.4 Dietary diversity for past 24 h Consumed from no food group 8.8 10.7 25.9 13.7 18.1 22.9 15.6 15.1 7.8 4.4 5.9 Consumed from 1—3 food groups 80.7 69.7 59.5 59.4 57.3 69.7 72.9 63.9 54.0 75.7 77.5 Consumed from 4—7 food groups 10.4 19.6 14.6 26.9 24.6 7.4 11.5 21.0 38.2 19.9 16.6 Meal frequency 6 to 8 mo old children 0.2 ± 0.8 1.6 ± 1.1 1.0 ± 1.2 2.2 ± 1.7 1.7 ± 1.4 1.7 ± 1.6 1.3 ± 1.2 1.2 ± 1.2 1.8 ± 1.3 1.8 ± 1.2 1.9 ± 1.0 9 to 11 mo old children 0.2 ± 0.7 2.1 ± 1.1 1.4 ± 1.1 2.5 ± 1.5 2.0 ± 1.3 2.5 ± 1.7 2.3 ± 1.0 1.6 ± 1.2 2.0 ± 1.1 2.2 ± 1.1 2.2 ± 1.1 12 to 23 mo old children 0.2 ± 0.7 2.2 ± 1.3 2.0 ± 1.1 2.8 ± 1.6 2.5 ± 1.3 3.0 ± 1.7 2.6 ± 0.9 2.2 ± 1.2 2.0 ± 1.0 2.5 ± 1.2 2.4 ± 0.9 Child feeding indexb 3.2 ± 0.8 4.1 ± 1.2 3.8 ± 1.1 3.6 ± 1.3 3.6 ± 1.3 4.2 ± 1.2 4.2 ± 1.1 4.0 ± 1.2 3.3 ± 1.3 4.0 ± 1.1 4.3 ± 1.1 DRC Democratic Republic of the Congo a Sample includes only households with complete data on children’s height, age, and haemoglobin levels. Values represent mean ± standard deviation or percentage b Child Feeding Index: A child feeding index was created based on age-specific (6 to 8 months, 9 to 23 months) feeding recommendations on breastfeeding, use of feeding bottle, diet diversity, and meal frequency (Kramer, 2012 [22]; Ruel and Menon, 2002 [17]; WHO, 2010 [21, 24]) C hristian et al. BMC Nutrition (2023) 9:9 Page 9 of 13 Fig. 1 Bars show the observed prevalence of co‑occurrence of anemia and stunting (CAS) by child feeding index and country, among children aged 6 to 23 months (n = 33,846) in 22 sub‑Saharan African countries (pooled data, Demographic and Health Survey data sets, 2008 to 2017). Stars above bars indicate significant difference (P < 0.05) in the prevalence of CAS within levels of child feeding scores (i.e., High, Medium, and Low) SSA followed by child feeding index. Result of the influ- socioeconomic status that in-turn affects the proximate ence of child feeding practices on comorbid malnutrition, determinants of health and, either directly or indirectly, corroborates with earlier research conducted in other affect the health and nutritional status of children [37]. global regional blocks [32, 33]. IYCF score was associated This is however influenced by other socio-environmental with the decreased likelihood of stunting and CAS, but factors [38]. An analysis of the effect of parent educa- rather the increased odds of children being anaemic. tion on children nutritional status from 56 developing The Lancet series on child survival identified IYCF countries, concluded that formal education of the parent practice as a major factor in children survival [34, 35]. might have greater impact on nutrition when the educa- This analysis buttresses the importance of infant and tion directly improved nutritional knowledge of parents young child feeding practices in child nutritional status [39]. and growth. There is a need for continued interventions Overall, type of water and sanitation were also impor- such as counselling of caregivers about benefits of appro- tant predictors of CAS. This finding would suggest sup- priate or adequate child feeding practices. In the first two port of interventions that addressed environmental years of life, appropriate breastfeeding and complimen- enteric dysfunction an underlying cause of both stunting tary feeding techniques are channels for optimal nutri- and anaemia, particularly among people living in condi- tion that provide defence against poor growth and boost tions of poor water, sanitation, and hygiene [40]. How- immunity against illnesses. ever, the SHINE study in Zimbabwe yielded no additional That IYCF was associated with anaemia but not in the benefit with respect to reduction in stunting and anae- expected direction. This may be because that there are mia. The association of sanitation and the presence of several other factors such as non-iron deficiency (other fever in children with the CAS reinforces the important micronutrients), infections and blood disorders that role infections play in child growth. Infections causing contributes significantly to anaemia in the sub-Saharan illnesses such as malaria and diarrhoea result in anaemia African region. The causes of childhood anaemia are and growth faltering in children [41, 42]. Child nutrition multifaceted and intricately related to one another. There and infections have a bidirectional relationship. Whereas are many non-dietary factors, such as intestinal para- frequent infections can lead to diarrhoea and/or impair sites like soil-transmitted helminths (STH) and Schisto- absorption of needed nutrients thus resulting in mal- soma; malaria; HIV infection; and chronic diseases like nutrition among children, a malnourished child is also sickle cell disease, in addition to dietary factors (such as more susceptible to infections. Generally, this study also iron deficiency, other micronutrient deficiencies such as revealed the disadvantage of poor socio-demographics folate, vitamin B12, and vitamin A), that may be linked (via wealth) on the nutrition outcomes, here CAS. This to IYCF. finding is also consistent with previous studies conducted The role mother’s formal education plays in the over- in other global regional blocks [43, 44]. all well-being and specifically children’s nutrition and In addition to food security and health care services, growth has long been documented [36]. Generally, higher the UNICEF conceptual framework considers the impor- educational status of caregivers reflects an improved tance of immediate determinants such as adequate Christian et al. BMC Nutrition (2023) 9:9 Page 10 of 13 Table 4 Logistic regression results showing the overall relationship between child feeding index and Childhood anaemia and stunting and their Co‑occurrence among children 6 to 23 monthsa Indicator of malnutrition Anaemia Stunted CAS Child feeding indexb 1.08*** 0.83*** 0.86*** (1.049—1.108) (0.805—0.848) (0.836—0.884) High wealth (Yes)c 0.91** 0.76*** 0.75*** (0.833—0.987) (0.696—0.824) (0.690—0.824) Household size (#) 1.04*** 0.97*** 0.98*** (1.034—1.054) (0.960—0.974) (0.978—0.992) Household head (Male) 1 0.92* 0.92 (0.915—1.092) (0.851—1.004) (0.837—1.006) Household head’s education (At least secondary) 0.83*** 0.78*** 0.76*** (0.754—0.910) (0.718—0.849) (0.696—0.836) Improved water source for drinking (Yes)d 0.99 0.99 1.02 (0.921—1.069) (0.930—1.064) (0.949—1.091) Improved sanitation (Yes)e 0.90*** 0.82*** 0.81*** (0.831—0.972) (0.764—0.885) (0.745—0.869) Caregiver’s age (1 = 25 to 49 y; 0 = 15 to 24 y) 0.75*** 0.94* 0.87*** (0.699—0.811) (0.880—1.006) (0.813—0.938) Caregiver’s education (At least secondary) 0.72*** 0.73*** 0.69*** (0.656—0.794) (0.666—0.810) (0.616—0.767) Child had diarrhoea in past 2 wk. (Yes) 1.03 1.13*** 1.13*** (0.945—1.116) (1.047—1.216) (1.040—1.230) Child had fever in past 2 wk. (Yes) 1.34*** 1.01 1.12*** (1.230—1.453) (0.941—1.090) (1.037—1.213) Child given iron tablet in past 6 mo (Yes) 1.07 1.01 1.01 (0.939—1.214) (0.906—1.130) (0.898—1.134) Observations 29,408 29,408 29,408 a CAS: Co-occurrence of anaemia (haemoglobin level < 11.0 mmol/L) and stunting (length/height-for-age Z-score < -2). Values are odds ratios (95% confidence intervals) from logistic regression models. Adjusted models are multiple logistic regression with CAS as the dependent variable controlling for all covariates shown; ***p < 0.001, **p < 0.05 b Child Feeding Index: A child feeding index was created based on age-specific (6 to 8 months, 9 to 23 months) feeding recommendations on breastfeeding, use of feeding bottle, diet diversity, and meal frequency (Kramer, 2012 [22]; Ruel and Menon, 2002 [17]; WHO, 2010 [21, 24]) c Wealth quintiles were calculated from an asset-based wealth index using assigned asset weights from a principal components analysis to create standardized asset scores d Improved water source: piped household water connection, public standpipe, borehole, protected dug well, protected spring and rainwater e Improved sanitation: ventilated improved pit latrines and pit latrines with a slab or covered pit; NB: There is no data on wealth for South Africa caregiving practices, including IYCF, to ensure the opti- improve child feeding practices among caregivers con- mal growth of children. This study’s findings demon- tinue, as a policy recommendation there is an increased strated the impact of IYCF and its importance in a child’s need for early childhood educators to be educated on development. The effects of some underlining determina- appropriate child feeding practices. This is necessary as tions in the framework, such as health (which includes we see an increasing number of young children attend- diarrheal status), education, and sanitation, on children’s ing day care centres and prekindergarten programs. The nutritional status are also buttressed in the current study WHO and others rightly acknowledge the role these early [45]. childhood educators play in optimal infant and child With respect to policy, countries are urged by the health [47, 48]. Global Strategy on Infant and Young Child Nutrition to A major strength of this work is that the current analy- create and put into effect comprehensive IYCF guide- sis was drawn from twenty-two nationally representa- lines that safeguard, encourage, and promote breastfeed- tive large samples. Additionally, to our knowledge, this is ing and complementary feeding [46]. While efforts to the first analysis using representative data across SSA to C hristian et al. BMC Nutrition (2023) 9:9 Page 11 of 13 Fig. 2 A summary of the decision tree analysis showing the three most important correlates of co‑occurrence of anaemia and stunting (CAS) 1Improved water source: piped household water connection, public standpipe, borehole, protected dug well, protected spring and rainwater investigate the effect of the WHO’s recommended infant associated with the endogenous variable being predicted and young children feeding practices on the co-occur- in the first stage equation but not associated with CAS. rence of stunting and anaemia at the individual level. This This may ensure better estimates. Unfortunately, after notwithstanding, the study is not without some limita- much exploration we found no such variable or variables tions. First, it uses cross-sectional data occurring at dif- in the DHS dataset. ferent time periods in different countries and that the nutrition outcomes of interest, that is childhood anae- Conclusions mia and stunting could be influenced by prevailing some Although there is no “silver bullet” that could solve the country-specific microeconomic variables which are not CAS, these results strengthen the argument for the captured in the DHS, thus not included in our current improvement of IYCF and maternal formal education models. Information on the use of micronutrients sup- may serve as a pathway to the improvement of childhood plements and/or other drugs by children that could influ- nutrition. Interventions aimed at addressing the burden ence their anaemia status in children were not accounted of stunting and anaemia and their co-occurrence are for in the DHS thus could not be controlled for in our in synchrony with the Sustainable Development Goals models. aimed at reducing hunger and improving the wellbeing of Another probable limitation of our estimation tech- the populace, particularly women and children. nique, particularly in the logistic regression may be that the child feeding index (score) variable may be endog- Supplementary Information enous to our model, in that it may be influenced by The online version contains supplementary material available at https:// doi. some covariates that also influenced the co-occurrence org/ 10.1 186/ s40795‑ 022‑ 00667‑9. of stunting and anaemia. The existence of such an issue (endogeneity) could lead to having biased estimates Additional file 1: Supplementary Table 1. Logistic regression results showing the relationship between child feeding index and CAS among [49]. This concern is often addressed using instrumental children 6 to 23 months by study countries. variables and a two-stage least squares methods. Thus, future studies could explore identifying a variable that is Acknowledgements The authors are grateful to IPUMS and DHS for granting access to the data. Christian et al. BMC Nutrition (2023) 9:9 Page 12 of 13 Authors’ contributions 11. Allali S, Brousse V, Sacri AS, Chalumeau M, de Montalembert M. Anemia A.K.C. and E.A.D. conceptualized, extracted dataset, and conducted the analy‑ in children: prevalence, causes, diagnostic work‑up, and long‑term conse‑ sis for the study. G.S.M. provided oversight with the conception of study and quences. Expert Rev Hematol. 2017;10(11):1023–8. interpretation or results. All authors reviewed the manuscript. The author(s) 12. Gosdin L, Martorell R, Bartolini RM, Mehta R, Srikantiah S, Young MF. The read and approved the final manuscript. co‑occurrence of anaemia and stunting in young children. Matern Child Nutr. 2018;14(3):e12597. Funding 13. Mohammed SH, Larijani B, Esmaillzadeh A. Concurrent anemia and stunt‑ This research received no specific grant from any funding agency, commercial ing in young children: prevalence, dietary and non‑dietary associated or not‑for‑profit sectors. factors. Nutr J. 2019;18(1):10. 14. Piwoz EG, Huffman SL, Quinn VJ. Promotion and advocacy for improved Availability of data and materials complementary feeding: Can we apply the lessons learned from breast‑ All data generated or analysed during this study are included in this published feeding? Food Nutr Bull. 2003;24(1):29–44. article [and its supplementary information files]. 15. Daelmans B, Mangasaryan N, Martines J, Saadeh R, Casanovas C, Arabi M. Strengthening actions to improve feeding of infants and young children 6 to 23 months of age: Summary of a recent World Health Organization/ Declarations UN ICEF technical meeting, Geneva, 6–9 October 2008. Food Nutr Bull. 2009;30(2). Available from: www.t hela ncet.c om/ journ als/ lancet/ articl e/ Ethics approval and consent to participate PIIS01 40‑.[cited 2022 Sep 30]. The study was carried out according to the guidelines and regulations laid 16. WHO. World health statistics 2016: monitoring health for the SDGs sus‑ down in the declaration of Helsinki. The DHS Program uses protocols reviewed tainable development goals. 2016. and approved by the ICF Institutional Review Board (IRB). In addition, IRBs 17. Ruel MT, Menon P. Child feeding practices are associated with child in study countries review country‑specific DHS survey protocols to make nutritional status in Latin America: innovative uses of the demographic sure they comply with local laws on the inclusion of respondents, including and health surveys. J Nutr. 2002;132(6):1180–7. minors and adults, in its surveys. In accordance with the ethics of conducting 18. Boyle EH, King ML, Garcia S, Culver C, Bourdeaux J. Contextual data in research with human subjects, voluntary consent to participate in the survey IPUMS DHS: physical and social environment variables linked to the was sought from respondents after a clear description of the study, procedure, Demographic and Health Surveys. Popul Environ. 2020;41(4):529–49. duration, possible risks and benefits of the study was explained for each study 19. Vollmer S, Wójcik J. The Long‑Term Consequences of the Global 1918 respondent.. Written informed consent was obtained from all study partici‑ Influenza Pandemic: A Systematic Analysis of 117 IPUMS International pants. IPUMS DHS granted access to the data. Census Data Sets. 2017. Available at SSRN: https:// papers. ssrn. com/ sol3/ papers. cfm?a bstr act_ id= 308358 4 or https:// doi.o rg/1 0.2 139/ ssrn. 30835 Consent for publication 84. Not applicable. 20. DHS. The DHS Program ‑ Demographic and Health Survey (DHS). 2020. Available from: https://d hspr ogram.c om/ what‑ we‑ do/s urvey‑ Competing interests Types/ dHs. cfm. [cited 2020 May 22]. The authors declare that they have no competing interests. 21. WHO. Anthro for personal computers, version 3.1, 2010: software for assessing growth and development of the world’s children. Geneva: WHO; 2010. p. 2010. Received: 1 December 2021 Accepted: 28 December 2022 22 Kramer M. Optimal duration of exclusive breastfeeding. Cochrane Data‑ base Syst Rev. 2012;2012(8):CD003517. 23. WHO J, Organization WH, UNICEF undefined. Joint WHO/UNICEF Meeting on Infant and Young Child Feeding, Geneva 9–12 October 1979: state‑ ment, recommendations, list of participants. 1981. References 24. WHO. Indicators for assessing infant and young child feeding practices: 1. Subramanian SV, Balarajan Y, Ramakrishnan U, Özaltin E, Shankar AH. part 2: measurement. 2010. Anaemia in low‑income and middle‑income countries. The Lancet. 25. Kim SS, Rawat R, Mwangi EM, Tesfaye R, Abebe Y, Baker J, et al. Exposure 2011;378:2123–35. to Large‑Scale Social and Behavior Change Communication Interven‑ 2. Kassebaum NJ, Jasrasaria R, Naghavi M, Wulf SK, Johns N, Lozano R, et al. tions Is Associated with Improvements in Infant and Young Child Feeding A systematic analysis of global anemia burden from 1990 to 2010. Blood. Practices in Ethiopia. PLoS One. 2016;11(10):e0164800. Available from: 2014;123(5):615–24. https:// journ als. plos.o rg/p loso ne/a rtic le? id= 10.1 371/ journ al. pone. 01648 3. WHO. Haemoglobin concentrations for the diagnosis of anaemia and 00. [cited 2022 Sep 9]. assessment of severity. 2011. 26. UNICEF. Strategy for Improved Nutrition of Children and Women in 4. UNICEF. Levels and trends in child malnutrition UNICEF‑WHO‑World Bank Developing Countries. A UNICEF Policy Review. 1990. Group joint child malnutrition estimates: Key findings of the. New York, 27. Dangour AD, Watson L, Cumming O, Boisson S, Che Y, Velleman Y, et al. NY: UNICEF, WHO & World Bank Group; 2019. Interventions to improve water quality and supply, sanitation and 5. Oruamabo RS. Child malnutrition and the Millennium Development hygiene practices, and their effects on the nutritional status of children. Goals: Much haste but less speed? Arch Dis Child. 2015;100(Suppl Cochrane Database of Systematic Reviews. 2013;2013(8). Available from: 1):S19‑22. https://w ww. cochr aneli brary. com/ cdsr/ doi/1 0.1 002/ 14651 858. CD009 6. United Nations. About the Sustainable Development Goals ‑ United 382.p ub2/ full. [cited 2022 Dec 20]. Nations Sustainable Development. 2015. Available from: https:// www. 28. Brown KH. Diarrhea and Malnutrition. J Nutr. 2003;133(1):328S‑332S un. org/s ustai nable deve lopmen t/ susta inable‑ devel opment‑ goals/. [cited https://a cade mic. oup. com/j n/ artic le/1 33/1/ 328S/ 468756 5. [cited 2022 2019 Dec 8]. Dec 20]. 7. Walter T. 4 Effect of iron‑deficiency anaemia on cognitive skills in infancy 29. Song Y, Psychiatry LYS archives of, 2015 U. Decision tree methods: and childhood. Baillieres Clin Haematol. 1994;7(4):815–27. applications for classification and prediction. Shanghai Arch Psychiatry. 8. Horton S. Chapter 2.1 The Economics of Poor Nutrition: Patterns, conse‑ 2015;27(2):130–5. quences and costs. In: Good Nutrition: Perspectives for the 21st Century. 30. Tong W, Hong H, Fang H, Xie Q, Perkins R. Decision forest: Combining the 2016. p. 83–91. predictions of multiple independent decision tree models. J Chem Inf 9. Horton S, Ross J. The economics of iron deficiency. Food Policy. Comput Sci. 2003;43(2):525–31. 2003;28(1):51–75. 31. WHO. Global Database on Child Growth and Malnutrition. 2012 [cited 10. Dewey KG, Begum K. Long‑term consequences of stunting in early life. 2019 Oct 20]. Available from: https://w ww.w ho.i nt/n utgro wthdb/a bout/ Matern Child Nutr. 2011;7(Suppl. 3):5–18. intro ducti on/ en/ index5. html. C hristian et al. BMC Nutrition (2023) 9:9 Page 13 of 13 32. Hashmi AH, Nyein PB, Pilaseng K, Paw MK, Darakamon MC, Min AM, et al. Feeding practices and risk factors for chronic infant undernutrition among refugees and migrants along the Thailand‑Myanmar border: a mixed‑methods study. BMC Public Health. 2019;19(1):1586. 33. Hipgrave DB, Fu X, Zhou H, Jin Y, Wang X, Chang S, et al. Poor comple‑ mentary feeding practices and high anaemia prevalence among infants and young children in rural central and western China. Eur J Clin Nutr. 2014;68:916. 34. Bryce J, El Arifeen S, Pariyo G, Lanata CF, Gwatkin D, Habicht JP. Reducing child mortality: Can public health deliver? Lancet. 2003;362(9378):159–64 Elsevier Limited. 35 Black R, Allen L, Bhutta Z, Caulfield L, Lancet MDOT, 2008 U. Maternal and child undernutrition: global and regional exposures and health conse‑ quences. The Lancet. 2008;371(9608):243–60. 36. Caldwell JC. Education as a factor in mortality decline an examination of Nigerian data. Population studies. 1979;33(3):395‑413. 37. Makoka D. The impact of maternal education on child nutrition : evi‑ dence from Malawi,Tanzania and Zimbabwe. DHS Working Papers; 2013. 38 Reed BA, Habicht JP, Niameogo C. The Effects of Maternal Education on Child Nutritional Status Depend on Socio‑Environmental Conditions. Int J Epidemiol. 1996;25(3):585–92. 39. Alderman H, Headey DD. How important is parental education for child nutrition? World Dev. 2017;1(94):448–64. 40. Humphrey JH, Jones AD, Manges A, Maluccio JA, Mbuya MN, Moulton LH, et al. The sanitation hygiene infant nutrition efficacy (SHINE) trial: rationale, design, and methods. Clin Infect Dis. 2015;61(7):S685‑702. 41. Kateera F, Ingabire CM, Hakizimana E, Kalinda P, Mens PF, Grobusch MP, et al. Malaria, anaemia and under‑nutrition: Three frequently co‑ existing conditions among preschool children in rural Rwanda. Malar J. 2015;14(1):440. 42. Danaei G, Andrews KG, Sudfeld CR, Fink G, McCoy DC, Peet E, et al. Risk Factors for Childhood Stunting in 137 Developing Countries: A Compara‑ tive Risk Assessment Analysis at Global, Regional, and Country Levels. PLoS Med. 2016;13(11):e1002164. 43. Khan JR, Awan N, Misu F. Determinants of anemia among 6–59 months aged children in Bangladesh: Evidence from nationally representative data. BMC Pediatr. 2016. 44. Singh MB, Fotedar R, Lakshminarayana J, Anand PK. Studies on the nutri‑ tional status of children aged 0–5 years in a drought‑affected desert area of western Rajasthan. India Public Health Nutr. 2006;9(8):961–7. 45. UNICEF Conceptual Framework on Maternal and Child Nutrition ‑ Google Scholar. [cited 2022 Dec 20]. Available from: https://s chol ar. google. com/ schol ar? hl=e n&a s_ sdt=0% 2C5&q= UNICEF+ Conce ptual+ Frame work+ on+M ater nal+a nd+ Child+ Nutrit ion& btnG=. 46. WHO G. Global strategy for infant and young child feeding. 2003 [cited 2022 Oct 14]; Available from: https:// agris.f ao. org/ agris‑ search/s earch.d o? record ID=X F201 50380 78 47. World Health Organization. Report of the commission on ending child‑ hood obesity. Geneva: World Health Organization; 2016. 48. Organization WH. Global strategy for infant and young child feeding. 2003. Available from: https:// apps.w ho.i nt/ iris/ bitstr eam/ handle/ 10665/ 42590/ 92415 62218. pdf. Cited 1 Oct 2022. 49. Ruel M, Haddad L. The natural history of growth failure: importance of intrauterine and postnatal periods. 2012. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Ready to submit your research ? Choose BMC and benefit from: lished maps and institutional affiliations. • fast, convenient online submission • thorough peer review by experienced rese archers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions