Complementary feeding practices and associated factors among Mongolian children 6–23 months of age

Abstract Little is known about factors influencing children's dietary intake in Mongolia, a country undergoing rapid nutrition transition. Using nationally representative data from the 2017 Mongolia National Nutrition Survey, we assessed the nutritional status of children aged <2 years and examined household, maternal, and child factors associated with feeding practices among children aged 6–23 months (n = 938). Multivariable logistic regression models were used to identify predictors of minimum meal frequency (MMF), minimum dietary diversity (MDD), and minimum acceptable diet (MAD). The prevalence of child stunting (length/height‐for‐age Z‐score < −2 SD) was 6.3%, and the prevalence of overweight (weight‐for‐height Z‐score > +2 SD) was 16.8%. The prevalence of anaemia and iron deficiency was 39.0% and 32.2%, respectively, and 73.5% and 85.5% of children had inadequate vitamin A and vitamin D status, respectively. Of children aged 6‐23 months, 92.1% (n = 864) had MMF, 49.6% (n = 465) had MDD, and 43.8% (n = 411) achieved MAD. Increased household wealth was positively associated with all three indicators, whereas severe food insecurity was not associated with MMF, MDD, or MAD. Older child age (odds ratio, 95% CI: 1.09 [1.06, 1.12]; p < .001) and maternal dietary diversity (odds ratio, 95% CI: 2.36 [1.67, 3.34]; p < .001) were positively associated with child MDD. Nutrition‐specific and nutrition‐sensitive efforts are needed to improve the dietary quality of infants and young children in Mongolia and reduce the high burdens of child micronutrient deficiency and overweight in the country.

malnutrition, characterized by caloric sufficiency, but underconsumption of nutrient-rich foods, is increasingly recognized at the global, national, subnational, household, and individual levels (WHO, 2017). Recent evidence indicates a growing prevalence of child stunting and adult overweight coexisting within the same households, as well as the cooccurrence of micronutrient malnutrition and overweight at the individual level (Dang & Meenakshi, 2017;Gubert, Spaniol, Segall-Corrêa, & Pérez-Escamilla, 2017).
In Mongolia, rapid socio-economic development in recent years has led to extensive migration from rural to urban areas, with~70% of the population currently residing in urban settings and the majority living  95% CI), and a two-sided significance level was used (α = .05).
Analyses were conducted using IBM SPSS Version 23.0 (IBM Corporation, Armonk, NY).

| RESULTS
A total of 938 children 6-23 months were included in the analysis.
When analysing the nutritional status of children as a function of household food security, the percentages of stunted children, anaemic children, and children having iron and vitamin D deficiency were similar across food security categories (Table 2). However, the proportion of overweight children was significantly higher in food secure/mildly food insecure households, as compared with moderately and severely food insecure households (p < .001), and the prevalence of child vitamin A deficiency was highest in severely food insecure households (p = .033).

| DISCUSSION
This study revealed a high prevalence of food insecurity among Mongolian households. One in five households experienced severe food deprivation, with a larger proportion of food insecure households   in urban areas. The association between urbanization and low food security has been observed in low-income countries (Dangura & Gebremedhin, 2017;Matuschke & Kohler, 2014;Ruel, Garrett, Yosef, & Olivier, 2017;Szabo, 2016). Although more than 70% of the poorest households experienced some level of food insecurity, approximately 40% of households in the wealthiest quintile also experienced moderate or severe food insecurity, suggesting that factors other than food availability and affordability are likely influencing children's diets in this context.
Our study findings indicate that recommended complementary feeding practices are not being achieved in Mongolia. The fact that almost all children had appropriate meal frequency, whereas only half received a minimally diverse diet (provided foods from ≥4 food groups), suggests that children's inability to achieve an adequate diet is primarily due to limited dietary diversity and low consumption of nutrient-rich foods, rather than insufficient caloric intake. Children consuming foods from at least four food groups are more likely to consume at least one animal-source food and one fruit or vegetable, in addition to a starchy staple food. As expected, a significantly lower However, as the prevalence of child stunting, anaemia, iron, and vitamin D deficiency was comparable across food secure and food insecure households, our findings suggest that children's vulnerability to dietary inadequacy exists across socioeconomic strata. The lack of effect of household size on children's feeding practices also points to factors other than food/resource availability that may be underlying nutritional inadequacy among Mongolian children. It would be useful to examine IYCF practices across a variety of food security experiences to understand influencers of food choices and feeding behaviours such as parental knowledge, cultural norms, and intrahousehold food allocation practices in the Mongolian context.
In our study, the likelihood of achieving adequate dietary diversity increased with child age, with infants aged 6-11 months at higher risk for a poor quality diet compared with children 12-23 months of age. The transition from exclusive breastfeeding to complementary feeding is challenging for many mothers and is commonly associated with delayed introduction of first foods and provision of inadequate quantity and/or quality of food for optimal child growth and development (Abeshu, Lelisa, & Geleta, 2016;Dewey, 2013;Issaka et al., 2015a(Issaka et al., 2015b; Mitchodigni et al., 2017). The fact that maternal dietary diversity was positively associated with children's dietary diversity suggests that mothers who consume items from a variety of food groups are more likely to ensure their children do so as well.
The low prevalence of child underweight, wasting, and stunting reflects the substantial improvements in child growth achieved by the government of Mongolia in recent years (Joshi, Bolorhon, Narula, Zhu, & Manaseki-Hollan, 2017;Mongolia Ministry of Health, 2017).
These are laudable successes and provide compelling evidence that a sizable reduction in child stunting is within reach in low-income countries. However, Mongolia is now faced with a rapidly increasing burden of adult and child overweight and obesity. The~10% overweight prevalence among children in moderately and severely food insecure households and the fact that less than half of overweight children had an acceptable diet in our study were unexpected.
However, emerging data from LMICs indicate an association between food insecurity and consumption of energy-dense foods, resulting in higher risks for overweight, obesity, and micronutrient malnutrition (Jomaa, Naja, Cheaib, & Hwalla, 2017;Schlüssel, Silva, Pérez-Escamilla, & Kac, 2013;Vuong, Gallegos, & Ramsey, 2015). Low-resource households tend to allocate a larger share of their income to food purchases, resulting in predominantly higher consumption of cereals/tubers and increasingly affordable highly processed poor quality foods at the expense of a diverse and nutritious diet World Bank, 2018). The fact that overweight is being established at a young age in Mongolia has serious implications for the onset of diabetes, hypertension, and other chronic illnesses in early life (WHO & FAO, 2003).  In addition to these food-based, supplementation, and fortification strategies, investing in the prevention of childhood overweight and obesity is important for reducing the risk of developing these conditions and comorbidities into adolescence and adulthood. This necessitates a broad approach addressing the socioeconomic and environmental influences governing food choices. Key to preventing excess child weight gain are parental messages promoting healthy and nutritious foods, encouraging good feeding habits, and discouraging the purchase of high-fat and high-sugar "junk" foods for their children. Mandatory nutrition labelling on prepackaged foods should be considered to help parents make healthier food choices. In addition, market-based strategies such as taxes on sugar-sweetened beverages and restricted advertising to children to reduce obesogenic influences, as well as family focused physical activity programmes, should be considered to prevent and reduce overweight in the Mongolian population (WHO, 2012).
Our study included a large nationally representative sample of children and mothers in Mongolia. As this was a cross-sectional survey, we were not able to infer causation from observed associations between dependent and outcome-influencing variables. Survey data were collected during a 2-month period at the end of autumn/beginning of winter and, consequently, may not represent children's dietary patterns and risk of nutritional inadequacy during other seasons in Mongolia.
Finally, as the dietary quality assessment was based on foods children consumed during a 24-hr period, it may not reflect the usual feeding habits of children who were ill or experiencing other atypical events.
However, as consistent associations were observed across the three diet-related indicators measured, this unlikely affected our results.
Nutritionally inadequate diets, consisting of limited variety and higher consumption of energy-dense nutrient-poor foods, are contributing to the high dual burden of malnutrition among Mongolian children. Multisectoral equity-focused efforts are necessary to improve children's dietary intake amidst shifting food consumption patterns associated with Mongolia's nutrition transition.

ACKNOWLEDGMENTS
We acknowledge the Ministry of Health and National Center for Public Health in Mongolia for leading and conducting the survey from which data were utilized for this study.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

SOURCE OF FUNDING
None declared.

CONTRIBUTIONS
AJ and JLB designed the study; AJ and JLB conducted the data analysis; AJ, ML, NB, BE, JB, and JLB interpreted the data; AJ and JLB drafted the manuscript; and ML, NB, BE, and JB contributed to the preparation of the manuscript. All authors approved the final manuscript. AJ and JLB had primary responsibility for the final content.