Social determinants of malnutrition in Chilean children aged up to five

Malnutrition in children under five—either under or overnutrition—is a matter that requires serious attention because it increases the risk of illness or death, worsens comorbidities, and negatively impacts physical, intellectual, and psychomotor development [2, 6, 8, 9]. Moreover, children with undernutrition are more likely to live in a context of poverty and inequality in their adulthood, which perpetuates the risk of undernutrition for future generations, creating a harmful cycle that is difficult to eradicate [4]. On the other hand, those with overnutrition have detrimental academic performance and, in addition, mental health problems linked to low self-esteem. Likewise, child malnutrition implies losing a healthy lifetime [10], increasing health costs [9], and negatively affecting future human capital [11].

Malnutrition in children under five can result either from a deficit or excess in energy intake, both within the same population, or at different times in a child’s lifespan. For example, in Latin America and the Caribbean, children’s obesity has increased dramatically, while undernutrition continues to be a challenge [8]. This so-called “double burden” of child malnutrition is related to the nutritional transition process, a public health problem that deserves greater attention from an epidemiological point of view [12]. It should be noted that, although both types of malnutrition are associated with different energy intakes and reveal certain differences in their SDs [5, 8], they should be approached from a comprehensive perspective [13].

To classify the underlying and immediate SDs of malnutrition, Harris and Nisbett [14] identified three overall factors: a) material, human, and social sources; b) structural (society, markets, law, political); and c) ideas and beliefs. Previous studies have found conditions that could encompass this classification. For example, regarding undernutrition of children from zero to five, some key SDs are parents’ educational level—mainly the mother, access to drinking water, and sanitation facilities [4, 6, 15, 16]. In this vein, the mother figure is critical, as a greater vulnerability is found in children of single mothers with a low income, chronic diseases, and lacking health insurance. Combined, these conditions mean it is less likely the mother could stay with the child or provide a substitute caregiver [5]. On the contrary, maternal higher body mass index and higher educational level are associated with a lower likelihood of child undernutrition [6]. Unsurprisingly, underlying those conditions is household income, which is a significant determinant of childhood undernutrition [3, 4, 6, 7, 15]. Similarly, the community’s role is relevant through social support and knowledge sharing concerning access to social protection programs and enabling collective demand for better public services [4, 5]. In addition, Hossain et al. [3] revealed that the odds of undernutrition decrease in children breastfed for periods of longer than six months or with complementary feeding. In terms of gender, there appear to be geographical differences. For example, in Nepal, a lower risk of undernutrition was found in male children [3], while the opposite is true in Indonesia [17]. Additionally, higher probabilities of undernutrition have been observed in some countries’ geographical areas, among ethnic communities, and in less urban locations [7, 15].

On the other hand, overnutrition (i.e., overweight and obese) is related to increasing urbanization [18, 19], new ways of both transportation and food commercialization, and the development and use of technologies that increase physical inactivity and unhealthy food consumption [20‐22]. Considering the transition from greater undernutrition to greater overnutrition, several observed mediating variables are the change from rural to urban, fewer people and children in the household, and access to prenatal care for mothers [18]. Overnutrition is a rising problem worldwide [23‐25], which for under-fives is related to babies younger than 36 months and overweight parents [13, 24], with a higher prevalence in some geographical areas [26]. However, there are other SDs related to overnutrition in under-fives that have less consistency in previous research. For example, higher odds of being overweight have been reported when the mother has a higher education level and when the child belongs to families with higher incomes than those with lower incomes [8, 24]. However, other studies reveal a relationship between overnutrition with lower household income, neighborhood insecurity [27] and lower parental education [28].

Chile has undergone a rapid socio-economic and nutritional transition, leading to a co-existence of the two extremes of child malnutrition, although with lower undernutrition and greater overnutrition (overweight and obese) [29]. Consequently, some researchers have claimed that the “double burden” of child malnutrition would not be present in Chile because of the low rates of undernutrition. They also emphasize that the country’s main problem would be overnutrition [30]. Nevertheless, the authorities recently highlighted the topic of undernutrition because it continues to be a public health problem that requires attention [31]; among Chilean children under five years old, the undernutrition rate is 1.8%, while 9.3% regarding overnutrition [32]. However, due to increases with age [33], it reached a worrying 54% in the case of primary school children in 2020 [34].

In Chile, measures in children under six indicate that males are more likely to be severely obese, while females are more likely to be overweight and of normal weight [34]. Similarly, male children have a slightly higher percentage when undernutrition categories are grouped (“undernourished” and “underweight”). Overnutrition arises mainly between three and five years of age [10, 35]. Likewise, previous works have evidenced that native ethnicity becomes a significant SD of undernutrition in children under five in Latin America [36] and Chile [37]. In this vein, research on indigenous Chilean children between six and eight years old has observed a higher prevalence of obesity [38].

Moreoever, Pavez-Soto et al. [39] highlighted the challenge of studying migrant children’s nutritional conditions, stating that the migrant population in Chile comes mainly from Latin American countries with similar sociocultural conditions. Thus, it would be expected to find high levels of overnutrition, as it is a characteristic seen in this region [8]. Similarly, recent works have discovered that migrant children in Chile present a higher biopsychosocial risk, lower access to social programs, and a higher proportion of multidimensional poverty [40], all factors associated with a higher probability of undernutrition [3, 4, 6, 7, 15]. Thus, all the variables mentioned above will be considered in this study.

In Chile, it is worth noting that there is a National Supplementary Feeding Program (PNAC) aimed mainly at families with public health insurance. The PNAC has a basic, single, and identical scheme for all children under six years old (only differentiating by age) but considers a reinforcement for undernourished risk [41].

Therefore, bearing in mind the relevance of malnutrition in Chile, which requires awareness of its multidimensional causes, the present study aims to identify and describe the SDs related to malnutrition in children up to five. For such purposes, the data provided by the Casen 2017 survey will be considered [42]. A multinomial logistic regression will be performed. This probabilistic model is used when the predicted outcome variable is nominal and has more than two categories—without a given rank or order. The model can be used with any number of independent variables, either categorical or continuous. Such methodology has been used previously in studies regarding the nutritional status of children and adults [43, 44].

The data used in this study are from the National Socioeconomic Characterization Survey, CASEN 2017 [42], conducted by the Surveys and Longitudinal Studies Center of the Catholic University of Chile (PUC) for the Ministry of Social Development and Family (MINDES). The sample units of the CASEN 2017 survey are dwellings selected in a probabilistic, stratified, and multistage manner. Inside dwellings, all family units are identified (households). Interviewers applied a paper-and-pencil survey in a face-to-face procedure, interviewing the head of the family unit or an adult member in their absence. Data consisted of a sample of 26,249 households with children under five years old, which, using the expansion factor (EXPR), reached 1,270,485 children. The EXPR allows us to obtain the results of persons and family units expanded at the national and regional levels, as well as by area (urban and rural), being representative of the country’s overall population [42]. All interviewees participated voluntarily, and anonymity is ensured under Law 17.374. CASEN 2017 is used as a secondary data source [45].

CASEN 2017 asks interviewees what the child’s nutritional status is. Based on medical check-up reports, response options are 1. Undernourished or at risk of undernourishment; 2. Normal; 3. Overweight; 4. Obese. In this study, responses were grouped as undernutrition (1), normal (2), and overnutrition (3 and 4).

Independent variables were extracted from the previous findings as long as they were present in CASEN 2017. CASEN has seven modules (registration-residents, education, work, income, health, identities-networks-participation, and housing-environment), from which three types of dimensions grouped the variables of interest in this research. First is the demographic dimension (including gender, age, nationality, ethnicity, urban or rural area, and geographic location). Second is the social dimension (comprising household income, educational level, type of health insurance, and child participation in the supplementary feeding program). Finally, the third dimension grouped the child’s personal features (health problems, health checkup, staying at home, and kindergarten attendance). All independent variables are dichotomous: 1 (belongs to the category) and 0 (does not) (see Table 1). It is worth noting that the nationality, belonging to a native ethnic group, household income, and household educational level variables were considered based on the survey respondents’ information.

A multinomial logistic model was estimated to establish the relationship between children up to five’s nutritional status and the independent variables. For each of the interactions, average marginal effects, likelihood, and relative risk were calculated. A statistical analysis was performed with Stata14 software. A set of coefficients, namely β¹, β², and β³, representing normal nutrition, overnutrition, and undernutrition, respectively, were estimated, corresponding to:Relative likelihood of y = 2 to base result isThe relative risk ratio for a one-unit increase/decrease in X_i is:

The estimation of the model is based on 1,270,485 observations. The log-likelihood value is −706,371.45, χ² (40) = 36,668.99, p < 0.0001, thus rejecting the null hypothesis that all parameters included in the model are equal to zero. Likewise, goodness of fit, Count R2 = 0.81, indicates the model has good explanatory power. The independent variables explain the outcome variable between 2.8% (Cox-Snell) and 4.1% (Nagelkerke) (see Table 2).

The model, by default, set normal nutrition as the reference group and estimated a model for overnutrition and a model for undernutrition, both of which were relative to normal nutrition. According to the confidence intervals for the relative risk, with 95% confidence, the relative risk ratio is proper.

Table 3 illustrates the increasing or decreasing likelihood of a child under five having normal nutrition, overnutrition, and undernutrition due to the independent variables. The marginal effects (likelihood) were slightly more significant for overnutrition compared to either undernutrition or normal nutrition.

The variables that increased the likelihood of normal nutrition are household educational level (4.4%), at home most of the time (3.7%), and Great North Zone (3.0%), while variables that decreased this are: quintile II (−9.3%), nationality (−7.5%), and health problems (−4.5%) (see Table 3).

Likewise, variables that increase the likelihood of overnutrition are nationality (9.1%), quintile II (6.2%), health problems (3.6%), participations in the state supplementary feeding program (3.4%), and type of health insurance (3.1%). On the contrary, variables that decreased this likelihood are Great North Zone (−4.0%), at home most of the time (−3.0%), and age ≤ 2 years (3.0%) (see Table 3).

Regarding undernutrition, variables that increased the probabilities are participation in the state supplementary feeding program (5.5%), quintile II (3.1%), and age ≤ 2 years (1.1%), while those that decreased it are: household educational level (−6.0%) and nationality (−1.6%) (see Table 3).

From the first analysis level based on relative risk, it is possible to infer the likelihood that a Chilean child up to five years of age shows over or undernutrition regarding a normal nutritional state (see Table 4).

There is a greater probability of over and undernutrition in male children, those who belong to native ethnic groups, those who attend kindergarten, those with health problems, and those who attend well-child checkups through the public health system (Table 4). In addition, the medical checkup and receiving complementary food variables increase the probability of over and undernutrition, compared to a normal nutritional status. Likewise, if the child’s home is in an urban area, the likelihood of over or undernutrition decreases by the same amount; 0.84 times less.

Likewise, there is a 1.04, 2.73, and 1.17 times greater probability of overnutrition if the child is male, has a Chilean background, and belongs to a native ethnic group, respectively. In parallel, if the child attends kindergarten and has health problems, they are more likely to be overnourished by 1.27 and 1.32, respectively (Table 4). In addition, if the child’s home is in the South and Austral Zone of the country, the probability of overnutrition increases.

On the other hand, there is a 1.21, 1.41, and 1.09 times greater probability of undernutrition in male children, those under two years old, and those that belong to native ethnic groups, respectively. If the child attends kindergarten and has health problems, the probability of undernutrition increases by 1.10 and 1.38, respectively. Likewise, if the child’s home is in the north zone of the country (Great North), the probability of undernutrition increases by 1.27 (see Table 4).

Moreover, as income decreases, i.e., moving down from Quintile IV to Quintile I, the probability of both under- and overnutrition among children increases compared to children in the highest income Quintile (V). However, this trend differs regarding Quintile I, which has a lower probability of over and undernutrition than Quintile II (see Table 4).