Social Determinants of Health and Adverse Childhood Experiences: How Are These Intertwined to Determine Childhood Overweight and Obesity in Argentina?

Abstract

Argentina is one of the countries in the Latin American region with the highest incidence of childhood obesity. The incidence of childhood overweight and obesity among children under five years of age is approximately 15%, according to the 2022 MICS survey. Obesity is a nutritional disorder that is explained not only by biological factors but also by socioeconomic and family circumstances. The objective of this study is to investigate the factors and risks that determine childhood overweight and obesity in Argentine children under five years of age. To explain this, this article presents an integrative model of two theoretical frameworks, social determinants of health (SDH) and adverse childhood experiences (ACEs), called hereafter “IM-SDH-ACEs”, in order to understand how these determinants are related to influence childhood malnutrition. This study performs a Chi-square test to empirically corroborate that SDH and ACEs influence this type of weight disorder; thus, it is shown that socioeconomic status (SES) and age groups together with ACEs are factors associated with childhood overweight and obesity. Likewise, a multinomial logistic regression model is estimated, and the results show that the highest risks of suffering from overweight and obesity in childhood are the predictors of physical violence (PV) inflicted on the child as well as an interaction of the variables of deprivation in health care and a secondary education level of mothers or carers. It is concluded that a more inclusive social policy is required to reduce childhood obesity in Latin American societies.

1. Introduction

The development of knowledge regarding nutritional problems is relevant in public health, because weight abnormalities determine health conditions in early childhood can generate negative impacts at later ages, such as suffering from diseases in late childhood, adolescence, or adulthood due to eating disorders (Pulgarón 2013). In addition, malnutrition can be a health marker of underlying mental health and functional deficits throughout the life cycle of children (Black et al. 2013). Moreover, overweight and obesity can have consequences for the person when suffering from non-communicable diseases (NCDs) such as cardiovascular diseases, diabetes, cancers, neurological disorders, chronic respiratory diseases, and digestive disorders (GBD 2020).

Malnutrition is defined as deficiencies or excesses in nutrient intake or imbalances of basic nutrients (UNICEF 2023; WHO 2024); while being overweight means excessive fat deposits, obesity is a more complex disease because it is characterized by chronic conditions of fat deposits, and both types of the diseases can have a negative impact on health (WHO 2024). The global prevalence of overweight in children under 5 years of age was 37 million in 2022, and more than 390 million children and adolescents between the ages of 5 and 19 were overweight; of this figure, 160 million were living with obesity (WHO 2014).

Poor and developing countries increasingly face the double burden of malnutrition (GBD 2020), which refers to both unresolved undernutrition and increased overweight and obesity in the general population, as well as among children. These problems coexist even within the same community or household (UNICEF 2023). Additionally, children in low- and middle-income countries are more vulnerable to inadequate nutrition (Black et al. 2013). At the same time, people living in poverty tend to be exposed to high-fat foods, sugars, salt, and energy-dense snack foods poor in micronutrients (Drewnowski and Specter 2004), which tend to be cheaper but also of lower nutritional quality than those from food-at-home, along with low physical activity habits (Mancino et al. 2018; WHO 2024).

Argentina is a Latin American country with the highest prevalence of childhood obesity in children under 5 years of age in Latin America, followed by Paraguay (UNICEF 2021a). The Multiple Income Clusters Survey (MICS) of Argentina carried out in 2019 and 2020 reports a prevalence of 15% of obese children of this age; of these, 10% are overweight and 5% are obese children, according to the Body Mass Index (BMI)1.

It is important to investigate the determinants that influence childhood malnutrition, which include various socioeconomic factors such as poverty (Orden et al. 2019). It is noteworthy that more than half of Argentine children lived in households experiencing income poverty in 2019, 52% (Paz 2024, p. 12). Additionally, it is necessary to investigate the existence of other types of factors that determine malnutrition, such as adverse childhood experiences (ACEs).

There are previous studies that show their analysis from the perspectives of ACEs and the social determinants of health (SDH), with the purpose of understanding how these determinants influence infant mortality, child malnutrition or other health conditions, etc. (Camacho and Henderson 2022; Thumm et al. 2022; Harper et al. 2023; Burton et al. 2024). The following study builds on previous research and evidence on the influence of these determinants on children’s health and well-being; however, this study integrates both theoretical frameworks, SDH and ACEs, and empirically tests them by analyzing the associations and interactions of some of these factors in influencing of childhood malnutrition, particularly overweight and obesity. This theoretical integration will be referred to hereafter as the “Integrative Model of Social Determinants of Health (SDH) and Adverse Childhood Experiences (ACEs)” and abbreviated as “IM-SDH-ACEs”. This study also explains the proximate and intermediate determinants of well-being and health in early childhood to classify SDH and ACE factors based on Mosley and Chen (1984) and Göran and Whitehead (1991).

This study is based on secondary data analysis to empirically corroborate the proposed integrated theoretical model. The empirical analysis is based on the Chi-square test to examine associations between weight disorders and SDH-ACEs as explanatory factors. Additionally, we fit a multinomial logistic regression model using the standardized Body Mass Index (BMI) as a categorical dependent variable for the types of malnutrition (WHO 2008b, 2014, 2024) in order to obtain the relative risks (RRRs) of being an overweight and obese child in Argentina, according to some of the determinants mentioned above. So, the research question is posed as follows: how do SDH and ACE factors influence childhood overweight and obesity in Argentine children under 5 years of age?

1.1. Social Determinants of Health and Adverse Childhood Experiences as Predictors of Health

We conceive child well-being based on Greeley and Dubowitz (2014) as a healthy and social environment that allows children to have stable family and social relationships, safe environments, and adequate nutrition, among other aspects that include the most transcendental factors for child development. The ecological framework proposed by Bronfenbrenner and Morris (1998) also allows for the inclusion of socioeconomic determinants that influence children’s environment and their development. From this perspective, Garner et al. (2015) point out that health is the result of a complex interaction between the child’s genetic endowment, the ecological environment in which he or she lives, and the neurobiological inscription of lived experiences.

It is also acknowledged by Arruabarrena and De Paúl (2012) that childhood is a vulnerable stage because it is there where physical, cognitive, emotional, and social resources are acquired to achieve well-being. However, it is also a stage of vulnerability to adversity, since at this age there are still no tools to satisfactorily regulate stress. From these perspectives, this study proposes an association or determination of SDH-ACEs in the well-being and health of children through several factors that have important impacts throughout their entire life cycle.

The World Health Organization (WHO) proposed the SDH framework to study the lifestyle of the population and how this is affected by broad social, economic, and political factors, which consequently influence the quality of health or well-being (OPS 2020).

The analytical approach of the SDH incorporates a broad set of indicators. The basic components of this conceptual framework include structural determinants, factors of the political and socioeconomic context, and intermediate determinants. The first mentioned include gender, income, education, occupation, social class, and ethnicity. The second type of factors refers to the set of state interventions in terms of implemented health policies for people accessing health goods and services, and the intermediate determinants include environmental, cultural, psychosocial, and behavioral factors seen as community-level factors. The former generate social stratification and together shape the health opportunities of social groups (Göran and Whitehead 1991; Borrell and Malmusi 2010).

The SHD framework focuses on social inequalities as determinants of people’s health. In this sense, Solar and Irwin (2010) state that socioeconomic conditions, such as socioeconomic strata (SES) and household material conditions, directly affect people’s health. This framework allows us to understand that there are specific population groups with greater vulnerability to suffering from diseases, given their conditions of inequality or poverty (Orden et al. 2019).

Mosley and Chen’s (1984) child survival framework precedes the SDH model and further classifies risk factors that determine both child mortality and anthropometric measures, which are used as dependent variables. The researchers posit that there are proximate determinants that directly influence children’s risks of morbidity, malnutrition, or death. For example, they identify factors related to the mother, such as her age, parity, birth interval, breastfeeding, etc. However, the authors classify the environmental determinant not only as proximate factors but also as underlying factors (called intermediate factors in terms outlined by Göran and Whitehead 1991). Proximate environmental determinants include ecological risk factors that could influence children’s health, such as air pollution, food, water pollution, and other vectors.

In the second categorization concerning the factors that determine child survival, Mosley and Chen consider the underlying determinants, such as some family factors (knowledge, time, father’s work capacity, income, assets, services, savings, resources, and family composition), as well as cultural factors; institutional factors (economy, politics, infrastructure, social institutions, bureaucracy, and health programs). From the perspective of researchers, the underlying factors influence the health of children through proximate factors. In addition, environmental factors are also included in this classification, such as climate, land, water, topography, etc. (Mosley and Chen 1984, p. 27; Mosley 1988, p. 319). From these perspectives, this study advocates for classifying determining factors into proximate and intermediate, and the additional contribution of our approach is the integration of ACE determinants to determine their impact on development and general health.

Felitti et al. (1998) described different types of ACEs, such as physical or emotional abuse, mental illness of someone in the home, sexual abuse, mother or stepmother treated violently, and substance abuse by a family member. Cronholm et al. (2015) and Finkelhor et al. (2015) also incorporate public insecurity and violence in the community.

There are several consequences of the ACEs on people’s lives. ACEs act as joint risks to the health of the infant because children are not able to regulate their impact on their own and generate stress known as toxic, which, in the short or long term, has a cumulative effect that affects the development of functions involved in the physiological regulation of stress (Kelly-Irving et al. 2013; Soares et al. 2016).

For example, Kaminer et al. (2023) demonstrated the association of ACEs with mental health in a study of seven countries. One of the countries is Argentina, and they found that there is a high association of ACEs with mental health in Argentine students, who showed high incidence rates of experiencing ACEs in terms of different aspects, for example, regarding community violence (80.5%), violence against household members (77.4%), emotional abuse (75.9%), physical abuse (64.5%), and frequent violence against household members (34%), among others.

Furthermore, Roberts et al. (2013) have shown evidence that ACEs affect preschool children in the relationship between their quality of life and their health. There is also evidence regarding this study on the relationships between ACEs and childhood obesity (Burke et al. 2011), and some researchers revealed from a large sample of US adolescents that ACEs are related to an increased risk of obesity, but not underweight (Davis et al. 2019).

The problem of ACEs and their influence on children’s health has been highlighted in many countries around the world, although there is not enough information in developing countries. Argentina is one of these and is not the only country that shows childhood obesity due to the influence of ACEs. This country shows a high incidence of obesity, and it is necessary to investigate the risks of this eating disorder due to verbal or physical abuse in childhood.

Furthermore, Hanć et al. (2022) pointed out the importance of analyzing factors and their joint effects on children’s well-being and health, as they found that the effect of ACEs and SDH on weight abnormalities was significant; their results indicate that interactions between childhood abuse, socioeconomic status, and family factors (including parental education) are possible modulators of the association between ACEs and BMI. The authors conducted their study on Polish children.

1.2. Integrative Theoretical Model of SDH and ACEs for the Explanation of Child Obesity

This study acknowledges previous studies that address child health from the perspectives of SDH and ACEs. However, these visions are not integrated into a comprehensive model in which an interrelationship between the different factors that make up both frameworks can be conceptualized and formalized. Some of these previous studies were conducted by Camacho and Henderson (2022), who found that ACEs are risk factors in rural and urban settings and shape access to resources differently due to their intersection with inequitable environments and the presence of marginalization. The scholars developed the Intersectional Nature of ACEs Framework, which also builds on previous views on ACEs; however, their framework focuses particularly on the ACEs that children experience in various place-based contexts.

Our main focus is the SDH framework outlined by Göran and Whitehead (1991) and Solar and Irwin (2010); however, our study proposes that SDHs and ACEs are factors that are dynamically interrelated. The reason is that ACEs affect children’s health and imply disease risks in childhood or later life, impairing their well-being.

Therefore, we integrated the SDH and ACEs frameworks in this study and called it the “Integrative Model of Social Determinants of Health and adverse childhood experiences” (IM-SDH-ACE). This model is composed of explanatory factors that determine people’s health and well-being from childhood, so we took the example of factors that can trigger diseases in childhood, such as obesity.

Figure 1 shows that the core of the IM-SDH-ACEs model is the child. From the center emerges his or her relationship with maternal and family factors, as well as with the social and natural environment. Children are born with a genetic load (Duggal and Petri Jr 2018); however, their environment shapes their personality based on biological, environmental, and ecological factors (Kovas et al. 2007; Wells 2018), along with family, social, and cultural factors, through which their health and well-being are determined (Mosley and Chen 1984; Göran and Whitehead 1991; Bronfenbrenner and Morris 1998; WHO 2008a; Solar and Irwin 2010; Hughes et al. 2017; OPS 2020).

IM-SDH-ACEs uses the previously described classification of factors into proximate and intermediate. The first ones mentioned are the individual, maternal, family, and household determinants, which are at the center of the child in Figure 1. The ACEs also interact with these factors:

(A)

The individual determinants are shaped by the genetics and biology of the children and directly determine their health, as well as the risk possibilities of suffering from specific diseases (Gentner and O’Connor 2019). Also, we find the psychosocial determinants at the individual level that come from the mother, father, caregivers, or family members who interact in the stages of the development of the child (Gaspar et al. 2022). This stage also involves the acquisition of cognitive and linguistic skills, particularly from zero to five or six years (Fernandez 2014). (Araujo et al. 2015) state that children’s language development begins before they utter their first word and is a complex process related to other processes, such as attention, memory, language, and motivation.

Subsequently, as the child grows, school attendance and education, including daycare, become part of their training, as well as recreation, sports, and culture (Tuñón et al. 2015). Therefore, we classified these determinants as individual factors, because they have a close association with child well-being, the way children and adolescents take care of their own health by exercising and gaining knowledge, probably from their parents or school, that proper nutrition can prevent obesity.

Another determinant is school non-attendance, which is a negative factor for the well-being of the child (Sasaki et al. 2024), and child labor is also a negative condition, which can be considered as ACEs in most cases, because it generally arises from the existence of social deprivation in the home and is related to poor growth, malnutrition, infections, and emotional disorders, among others (Ibrahim et al. 2019).

(B)

Maternal factors are proximate to childhood survival, as defined by Mosley and Chen (1984). In this case, infants are directly exposed to their closest environment, which determines their well-being and health. Maternal factors include breastfeeding, the interpregnancy interval, dietary intake, and ablactation, among other factors (Bongaarts 1978; Mosley and Chen 1984; Mosley 1988).

(C)

Family factors are of various types, and we classify them in this study into behavioral, time, and knowledge determinants. All three have a proximate influence on the health and well-being of the child and are associated with the treatment and nurturing of children provided by parents or caregivers. Behavioral factors refer to the treatment that children receive from the moment they are born and throughout their childhood and adolescence, whether they are raised with love and respect or whether they are mistreated (Chamberlain et al. 2019; Negriff 2020). These types of factors include ACEs that are triggered in the family or are exerted directly on the child, whether with verbal, psychological, or physical violence by a family member, caregiver, or their parents (Felitti et al. 1998; Cronholm et al. 2015; Kaminer et al. 2023).

Several scholars state that child abuse prevents the full potential of brain development because its effect is negative on the child’s cognitive ability and educational achievement. The effects also include anxiety, depression, and social isolation (Belsky and de Haan 2011). So, domestic violence is an important indicator for evaluating a child’s well-being and health (García and Ritterbusch 2015).

The second type of factor refers to the time that parents dedicate to interaction, recreation, and play with the child, in addition to their education. Tuñón et al. (2015) studied multidimensional child poverty in Argentina and considered the lack of early stimulation as a child deprivation. The scholars include indicators such as reading stories, playing with children, and attending educational centers, among others. Furthermore, Bundy et al. (2009) have considered indicators related to artifacts such as toys, balls, and dolls as an important part of play and recreation activities, which give evidence of child development in its different stages.

The factor referring to knowledge indicates the actions taken by the mother, father, or caregivers to provide preventive measures, health, or curative services to children, such as medical consultations, medications, vaccines, oral serum, etc. (Mosley and Chen 1984). In this case, the education of the mother, caregiver, or head of the household also acts directly and includes the provision of adequate nutrition (Bongaarts 1978; Mosley 1988). All these determinants can be analyzed at any age of childhood and adolescence.

(D)

The determinants classified as household are diverse and are identified as part of their standard of living. These include household goods and services that also determine the health and well-being of children and adolescents. In this sense, the household may lack or be deprived due to a lack of a basic food and non-food basket; a lack of sanitation, a lack of a stove to cook food, overcrowding, a lack of drinking water, a lack of proper drainage, a lack of access to health and social security services, etc. (UNICEF and CONEVAL 2019). A lack of these basic services can lead to diarrhea, malnutrition, and other diseases.

Furthermore, low socioeconomic status (LSES) is an important factor that is analyzed at the household level. Ecological studies have shown that socioeconomic status (SES) as a measure of household wealth is one of the most relevant factors associated with obesity. SES has been measured through household income, parental education, or type of employment and other well-being indices (Orden et al. 2019). LSES has an impact on child well-being because this aspect is part of multidimensional child poverty, which in turn affects children’s health and development (Guillén-Fernández and Vargas-Chanes 2021; Guillén-Fernández 2024).

There are also intermediate determinants at the household level—for example, access to information, such as the internet, which is conceived as a right in different countries around the world and should be included in measurements of child well-being and poverty (Guillén-Fernández 2023a). Information and communication technologies play a crucial role in children’s education and health (Baños et al. 2013; DeMartini et al. 2013; Park et al. 2016). Information and communication technologies (ICTs) can also play a crucial role in medical treatment and can be accessed through state public services or health and education institutions.

Other intermediate factors are the education and employment of the head of household, as well as indicators of SES, which are associated with positive behaviors and access to social rights such as access to health institutions (Mirowsky and Ross 2003; Hoffmann et al. 2019). In addition, the type of locality is an indicator associated with socioeconomic status. There is evidence that some populations in rural areas do not have access to health services, such as infrastructure, doctors, and medicines, and localities where there is no access to clean water are a negative factor for health. Under these conditions, some diseases prevail (Ekbrand and Halleröd 2018). All these aspects of the standard of living are characteristics of LSES. Therefore, persistent poverty negatively affects childhood by hindering children’s capabilities and achievements (Duncan and Brooks-Gunn 2003).

(E)

Natural environmental factors may include the presence of food crises, either due to economic crises or pandemic crises such as COVID-19 (Álvarez-Gálvez et al. 2019; Barker and Russell 2020). These crises directly determine the type of nutrition for the child and her family (Mosley and Chen 1984) and consequently affect the child’s growth and development (Bronfenbrenner and Morris 1998). Additionally, there is an influence of the natural environment on the child’s survival in the presence of pollution, adverse environmental cycles, natural disasters, epidemics, or pandemics (Mosley and Chen 1984; Mosley 1988; Gentner and O’Connor 2019).

Additionally, ACEs can also originate from natural disasters and have consequences in illness or psychological functioning for children or later ages (Inoue et al. 2019; Dutta et al. 2022). These natural determinants could influence in an intermediary way when, for example, a drought prevents the production of one of the essential childhood foods in a given place; however, there may be substitutes for the food.

(F)

The social environment can also act as a proximate factor or as an intermediary. The influence is proximate, for instance, in cases of street children, who are exposed to a risky environment (Siersbaek et al. 2021). The older a child is and therefore more involved in the social sphere, the more the social environment will directly determine their well-being and health. For example, we refer to the problem of bullying, which also has negative effects on health in childhood and adolescence (Gentner and O’Connor 2019; GBD 2020). The proposed IM-SDH-ACEs distinguish ACEs that come from the family from those that come from the social sphere (Narayan et al. 2021; Pierce et al. 2022). The latter can also have a proximate influence, as in the case of delinquency, sexual abuse in childhood, violence in the community (Zhao et al. 2023), or wars (Cabrera et al. 2007; Gentner and O’Connor 2019).

On the other hand, the economic context, such as unemployment rates or economic crises, has an intermediate effect on the well-being and health of the child. Another intermediate element may be the existence of social action groups to raise awareness about obesity through social networks.

(G)

Intermediate factors are all those that incorporate society itself and the state and are of the meso-social, macroeconomic, or institutional sphere, such as the design and implementation of public policies for the creation of infrastructure (schools, hospitals, sports areas in localities, parks, cinemas, and theaters), cultural promotion, access to adequate and quality levels of schooling, the establishment and promotion of democracy, community services, social service such as the construction of rehabilitation centers, community participation, and social inclusion, among others. All these factors are relevant since the lack of access to parks or green areas can influence the increase in risks of stress, teenage pregnancy, and drug use in adolescents (Fredricks and Eccles 2008; García and Ritterbusch 2015), and their provision or lack thereof influences in an intermediate way the health and well-being of children (Mosley and Chen 1984; Göran and Whitehead 1991).

The IM-SDH-ACEs further show in Figure 1 that there is a dynamic interrelationship between all factors. This integrative model is proposed in accordance with the concept of “child well-being” which involves health care and is defined as follows: “states of quality of life that imply the realization of all the rights of children, so that they can achieve their child development (physical, emotional, cognitive), to enjoy health, develop their capacities and interact with the environment in which they develop, at all stages of their childhood life, and considering the different aspects of their life that influence their living conditions: health, food and nutrition, child development, care, natural environment and family environment, education at all levels; be free from child labor, access to information, social security, equality and other rights” (Guillén-Fernández 2023b, p. 248). To this conceptualization, we add being free of violence, living in a healthy natural and social environment, and access to sports and culture.

2. Materials and Methods

The objective of the empirical analysis in this research is to empirically corroborate SDH and ACEs as determinants of childhood obesity and overweight based on secondary data analysis. However, for this study, we focused on the analysis of individual, family, and household factors.

This study is based on data provided by the MICS (Multiple Income Cluster Indicators) carried out by UNICEF for Argentina. This microdata survey belongs to round six of MICS surveys, which is the most recent MICS and was conducted between 2019 and 2020. It is a specialized survey on child well-being that includes not only anthropometric measures and health indicators but also socioeconomic and family indicators. The MICS are based on data collected at the community, family (household), and individual levels and are representative at both the national and subnational levels (UNICEF 2023). So, estimates obtained in this study are generalizable to the Argentine context, as the survey is representative, and the results are valid for children under 5 years of age and their households when these are significant. For this study, the Argentine MICS survey was used because it provides information on malnutrition, SDH variables, and ACEs; MICS from other countries do not provide information on ACEs.

The MICS survey was conducted on a sample of multi-stage households identified and selected based on the Argentine Census framework. These households are composed of children aged 0 to 17 years. Personal interviews were conducted with mothers or caregivers to collect data on information about the children and their households (UNICEF 2021b).

This study bases its analysis on two types of statistical methodologies to evaluate the hypothesis that SDH and ACEs factors have an influence on overweight and obesity in children at an early age. First, we carried out a Chi-square test to examine the independence of the categorical variables; the null hypothesis (H₀) states that variables are independent, and the alternative hypothesis (H_A) states that there is a dependency between them. This study also uses contingency tables to show associations between two variables and multidimensional contingency tables to display associations between three variables (Mendenhall and Beaver 2003). Then, prevalences are calculated and are shown in contingency tables; the prevalence of a particular disease, in this case overweight or obesity in 2020 for Argentinian children, refers to the number of subjects with the disease at a time point between the total number of subjects in the population (Noordzij et al. 2010, p. c18).

In addition, a Spearman-based correlation test was performed to determine BMI and the weight-for-height (WH) indicator to test their association and assess whether BMI is an adequate measure. Spearman was used because it is a non-parametric technique used to measure the strength of a relationship—in this case, between two variables, because the data are in a hierarchical form (Stevenson 1981), so we tested the null hypothesis of independence of a pair of random variables.

2.1. The Multinomial Logistic Regression Model to Estimate RRR of Child Overweight and Obesity

Secondly, this research article estimates a multinomial logistic regression model with the purpose of accounting for the dependent variable (BMI), composed of four categories: underweight, normal weight, overweight, and obesity, as well as its relationship with SDH and ACEs. As well as exploring whether there are significant interactions between determining factors that also show risks for presenting overweight and obesity in childhood.

As indicated by Hosmer and Lemeshow (2000, p. 262), Formula (1) is the function of the multinomial logistic regression model:

$$\mathrm{P}\left(Y=j|x\right)={\displaystyle \frac{e^{g_j\left(x\right)}}{\sum _{k=0}^2{e}^{g_{i\left(X\right)}}}}$$

(1)

where β₀ = 0 and g₀(X) = 0.

Therefore, in this case, the model will indicate the probability that a child aged 0 to 5 years has some type of malnutrition (overweight or obesity).

Based on previous evidence, such as that uncovered by Murer et al. (2016), who state that childhood obesity is multifactorial and the influential determinants must be taken into account to design effective treatment and prevention programs for childhood obesity, we estimate the relative risk ratios (RRR) from the multinomial logistic regression model, which are obtained from the model in order to compare the risks of being overweight or obese in a comparative way based on a reference category according to the exposure or risk factor. I.e., a relative risk is the ratio of the probability of specific event occurring in the exposed group versus the probability of the event that occurs in the non-exposed group (Tenny and Hoffman 2023), for example: the relative risk of developing overweight due to the experience of ACEs (physical violence) vs. not experiencing it. The logarithm of the likelihood ratio of outcome m to k, or log[Prob(Y_i = m)/Prob(Y_i = k)]) is calculated by taking its exponential. RRR are estimated based on Borooah (2001) from Formula (2):

$${\displaystyle \frac{\mathrm{Pr}\left(Yi=m\right)}{\mathrm{Pr}\left(Yi=1\right)}}=exp{\sum }_{r=1}^R{\beta }_{mr}{x}_{ir}$$

(2)

Furthermore, this research advocates interaction-based analysis to also obtain the RRR with combined factors (e.g., ACEs with SDH) in order to assess their influence on children’s weight disorders, particularly on overweight and obesity. Significant interaction effects are explained in the results section of the model.

On the other hand, the validity of the regression model will be assessed based on the χ² (chi2) statistic test. This is obtained from the fact that L is the log likelihood function evaluated in the unrestricted estimator. This statistic is distributed χ² with k degrees of freedom, where k is the number of coefficients, with the following hypotheses:

Ho: B = 0

Hi: B = 0

(Hosmer and Lemeshow 2000; Greene 1999; Borooah 2001).

Thus, the multinomial logistic regression model estimated in this study shows a χ² value of 37.61 with a significance range of 95% (0.038); therefore, we rejected the null hypothesis.

2.2. Measures

The analysis of child overweight and obesity implies the estimation of the dependent variable of the model and is a type of anthropometric measure. Also, independent variables were identified in the MICS as the factors that determine child malnutrition, according to our theoretical model of IM-SDH-ACEs. It should be mentioned that this study is focused on the individual, family, and household levels, so only this type of indicators is considered for the analysis. The calculation of these variables is explained as follows:

2.2.1. The Dependent Variable

The Body Mass Index (BMI) is the dependent variable used for estimating child malnutrition in Argentina. Overweight and obesity are measured by calculating the BMI by dividing people’s weight and their squared height: (kg)/(m²) (WHO 2024), and this value can be analyzed and compared based on the tables provided by the WHO with standardized measurements (De Waard 1978; WHO 2008b).

UNICEF (2021b) calculated the BMI from the weight and height of the children surveyed based on UNICEF’s (2018) own anthropometric recommendations and calculated the standardized BMI variable in the MICS, taking into account the age data in months based on the birth dates collected during the interview (UNICEF 2021b). We use the already standardized BMI variable, so the WHO tables are used to estimate overweight. This is measured through a BMI indicator greater than 2 standard deviations (SDs) above the median established in the WHO child growth standards, and obesity is estimated at 3 standard deviations above the median (WHO 2014, 2024).

BMI is one of the indicators to measure obesity in childhood; others are weight for age, weight for height, and DXA (Diagnosis Indexes of Abdominal Obesity), which includes waist circumference, etc. In this study, we measured the ratio of a person’s weight relative to their height, rather than adiposity, which indicates the proportion of fat relative to total body mass. Regarding BMI, it has been discussed that the use of inappropriate cut-off points may explain the inadequate reporting of many obese children not being reported as overweight (Freedman et al. 2005). In this case, BMI is the appropriate indicator to measure overweight and obesity in children under 5 years of age. To verify this, the correlation of BMI with weight for height referred to in the information document of the Argentine MICS survey (UNICEF 2021b) was evaluated. In this case, Spearman’s rho was significant, with a correlation coefficient of 0.92 and Prob > /t/ = 0.0, so we reject the null hypothesis that BMI and WH are independent. Thus, BMI is valid to assess overweight and obesity in Argentine children under 5 years of age from the MICS survey.

The target population of this study is 5649 children aged 0 to 5 years from the MICS Argentina 2019–2020, which reports anthropometric measurements; of these children, 82.7% have normal weight, 2.4% are underweight, and 14.9% suffer from childhood obesity (9.8% are overweight and 5.1% are obese children) according to the Body Mass Index (BMI) (

Table 1. The dependent variable.

Variable Name	Label	Codes	Percentage and Number of Cases (n) of the Dependent Variable	Variables in MICS and Questionnaire
BMI	Body Mass Index (BMI) standardized to the WHO measures	0: Underweight	2.42% (137)	ZBMI—Body Mass Index z-score WHO
		1: Normal weight	82.69% (4671)
		2: Overweight	9.77% (552)
		3: Obesity	5.12% (289)
		Total number of cases (N):	100% (5649)

Source: own calculation based on the 2019–2020 Argentina MICS (UNICEF 2024) and tables of standard deviations (SDs) provided by WHO (2024). Notes for abbreviations: BMI: Body Mass Index; WHO: World Health Organization; ZBMI: Standardized Body Mass Index.

).

2.2.2. The Explanatory Variables

The explanatory variables of this study are the determining factors that we classified as SDH or ACES in the integrated theoretical framework. Those that were used in this research are classified in

Table 2. The explanatory variables (determinant factors).

Variable Names and Classification by SDH and ACE Factors	Label	Codes	Percentage and Number of Cases in the Survey	Variables in the MICS and Labels in Its Questionnaire
SDH
sex	The sex of the child	0: boy	51.6% (2916)	HL4—sex
		1: girl	48.4% (2733)
age	The age of the child	0: age < 2 years	34.9% (1972)	CAGE_6—age
		1: ≤2 age < 4	42.4% (2395)
		2: ≤4 age < 5	22.6% (1278)
SES	Socioeconomic strata *	0: High socioeconomic stratum (HSES)	14.8% (841)	windex5—wealth income quintiles
		1: Medium socioeconomic stratum (MSES)	36.5% (2060)
		2: Low socioeconomic stratum (LSES)	48.7% (2748)
melevel	Mother’s educational attainment	0: Low educational attainment (LEA) (below secondary education)	43.5% (2459)	melevel—the highest educational attainment of the mother
		1: Secondary level completed, medium educational attainment (MEA)	40.8% (2304)
		2: High school completed and University level, high educational attainment (HEA)	15.6% (883)
HC	Healthcare access	0: Household’s members are entitled (not deprived)	54.5% (3077)	UB9—right to health insurance fmaportante—mother or father contributing to social security
		1: Household’s members are not entitled (deprived)	45.3% (2560)
Ind	Indigenous ancestry of the head of the household	0: Do not have indigenous ancestry	92.1% (5203)	ethnicity—Indigenous belonging of the head of household
		1: Do have an indigenous ancestry	5.6% (318)
Variable Names and Classification by SDH and ACE Factors	Label	Codes	Percentage and Number of Cases in the Survey	Variables in MICS and Questionnaire
ACEs
PV	Physical violence	0: The child does not experience PV at home.	50.9% (2877)	UCD2C, UCD2F, UCD2G, UCD2I, UCD2J, UCD2K—variables in which respondents report any type of PV, such as hitting the child with the hand, an object, or a belt
		1: The child experiences PV at home.	31.9% (1804)
VV	Verbal violence	0: The child does not experience VV at home.	45.3% (2559)	UCD2D, UCD2H—variables indicating yelling at their child
		1: The child experiences VV at home.	38.5% (2174)

Source: own elaboration based on variables and data from the 2019–2020 Argentina MICS. Notes: SDH refers to the social determinants of health; ACEs are adverse childhood experiences; MICS is the Multiple Income Clusters Survey. * The first and the second quintiles sum up the low socioeconomic stratum or poor income; third and fourth quintiles sum up the medium socioeconomic stratum, and the fifth quintile is the high socioeconomic stratum.

. These determining factors were identified and measured with data used in the MICS of Argentina for early childhood (0 to 5 years).

Table 2 shows the specific indicators of childhood, which are the sex of the infant and their age2. Household and family characteristics are the household’s socioeconomic level (SES), measured as income poverty stratum; the mother’s educational level (melevel); access to health services (HC); Indigenous ancestry (Ind); and variables of ACEs.

Table 2 shows the percentages for each of the children’s characteristics. A total of 48.4% of the sample are girls, and 51.6% are boys. The Argentine MICS sample shows that children aged 0 to 2 years comprise 34.9%, those aged 2 to 4 years comprise 42.4% of the sample, and those aged 4 to 5 years comprise 22.6%.

Children under five years of age living in low socioeconomic stratum (LSES) and poor-income households comprise 48.7%. Children who experience deprivation in access to health care comprise 45.3%, children whose mother has a secondary education level comprise 40.8%, and those with education below the mandatory level of the mother comprise 43.5%. Children with Indigenous ancestry comprise 5.6% of this demographic group.

The family factors classified as characteristic of ACEs are physical violence (PV) and verbal violence (VV). As explained in the approach proposed by IM-SDH-ACE, the well-being and health of the child are negatively affected by this type of trauma. It is worth mentioning that the MICS questionnaire asks about these risk behaviors in the past; however, the answers are related to the parents’ attitudes regarding their children’s obedience in their current life, i.e., it is not a retrospective view. Therefore, the risks are explained only with respect to the association with the explanatory variables at a point in time3.

Among the population subgroup that suffers from ACEs, around 31.9% of children aged 0 to 5 years suffer from PV, and 38.5% suffer VV4.

It should be noted in Table 2 that the categories for all these variables were coded as 1 for the population most vulnerable to overweight or obesity based on the literature review. Other categories were coded as 0 when there was less vulnerability.

3. Results

In this section, we apply the methodologies described for two purposes; the first is to evaluate the associations between the prevalence of overweight and obesity with respect to SDH and ACE determinants based on the Chi-square test. Secondly, we apply a multinomial logistic regression model to evaluate the relationships and their interrelation with the dependent variable of BMI.

3.1. Prevalence of Overweight and Obesity Grouped by Different Factors Among Argentinian Children Under 5 Years

In this section, we estimate the prevalence rates among children under 5 years of age presenting with weight disorders (overweight and obesity) and their association with several SDH and ACE factors. For this purpose, we use the Chi-square test to assess whether the associations between variables are significant; in this case, the alternative hypothesis is accepted only when there is dependence between variables. This statistical analysis is based on the study by Murer et al. (2016, p. 624) showing the prevalence of childhood overweight and obesity in its association with several sociodemographic characteristics or risk factors among Swiss primary school children. Therefore, the Chi-square statistical test shows the aggregated information regarding the statistical validity between the association of the variables, in addition to this,

Table 3. Prevalence of overweight and obesity in children under five years of age in Argentina according to SDH and presence of ACEs (% = percentages, n = cases per cell, and N = cases per row).

(A) Contingency Tables with Two Variables Associated					(B) Contingency Tables with Three Variables Associated
SDH		BMI			SDH and ACEs	BMI			SDH and ACEs	BMI
		Overweight	Obesity	N		Overweight	Obesity	N		Overweight	Obesity	N
Sex		Chi-square = 1.779 p-value = 0.6199			Sex and PV	Chi-square = 5.354 p-value = 0.148			Sex and VV	Chi-square = 2.657 p-value = 0.448
boy	% (n)	9.5 (276)	5.2 (151)	2916	boy with PV	8.4 (82)	4.6 (45)	979	boy with VV	9.1 (103)	4.7 (53)	1136
girl	% (n)	10.1 (276)	5.0 (138)	2733	girl with PV	10.4 (86)	5.3 (44)	825	girl with VV	10.9 (113)	3.9 (41)	1038
Age		Chi-square = 14.845 p-value = 0.021 *			Age and PV	Chi-square = 6.343 p-value = 0.386			Age and VV	Chi-square = 12.225 p-value = 0.050 *
0 < 2	% (n)	10.6 (209)	5.2 (103)	1972	0 < 2 and PV	11.5 (32)	6.1 (17)	278	0 < 2 and VV	12.6 (41)	6.2 (20)	325
2 < 4	% (n)	9.6 (231)	4.5 (108)	2395	2 < 4 and PV	9.0 (85)	3.9 (37)	948	2 < 4 and VV	10.0 (114)	3.3 (38)	1137
4 < 5	% (n)	8.8 (112)	6.0 (77)	1278	4 < 5 and PV	8.8 (51)	5.9 (34)	577	4 < 5 and VV	8.6 (61)	4.9 (35)	710
SES		Chi-square = 12.917 p-value = 0.044 *			SES and PV	Chi-square = 1.161 p-value = 0.979			SES and VV	Chi-square =4.759 p-value = 0.575
HSES	% (n)	9.5 (80)	4.8 (40)	841	HSES and PV	8.3 (16)	4.7 (9)	193	HSES and PV	9.0 (27)	3.7 (11)	301
MSES	% (n)	10.7 (220)	5.8 (120)	2060	MSES and PV	9.8 (65)	5.4 (36)	665	MSES and PV	10.3 (82)	5.1 (41)	799
LSES	% (n)	9.2 (252)	4.7 (129)	2748	LSES and PV	9.2 (87)	4.7 (44)	946	LSES and PV	10.0 (107)	3.9 (42)	1064
melevel		Chi-square = 6.500 p-value = 0.370			melevel and PV	Chi-square = 4.845 p-value = 0.564			melevel and VV	Chi-square = 3.389 p-value = 0.759
HEA	% (n)	9.9 (87)	5.7 (50)	883	HEA and PV	10.6 (23)	6.5 (14)	216	HEA and VV	10.0 (33)	4.9 (16)	324
MEA	% (n)	10.4 (239)	5.6 (128)	2304	MEA and PV	9.2 (67)	5.5 (40)	731	MEA and VV	9.7 (84)	5.1 (44)	867
LEA	% (n)	9.2 (226)	4.5 (111)	2459	LEA and PV	9.1 (78)	4.1 (35)	857	LEA and VV	10.1 (99)	3.5 (34)	983
HC		Chi-square = 3.124 p-value = 0.373			HC and PV	Chi-square = 6.311 p-value = 0.097			HC and VV	Chi-square = 3.269 p-value = 0.352
Not deprived	% (n)	9.9 (305)	5.6 (171)	3077	Not deprived and PV	8.1 (73)	5.9 (53)	903	Not deprived and VV	9.3 (108)	4.9 (57)	1158
Deprived	% (n)	9.6 (245)	4.6 (117)	2560	Deprived and PV	10.4 (93)	3.9 (35)	896	Deprived and VV	10.6 (107)	3.6 (36)	1012
IND		Chi-square = 7.212 p-value = 0.302			IND and PV	Chi-square = 5.580 p-value = 0.472			IND and VV	Chi-square = 3.595 p-value = 0.731
No	% (n)	9.9 (514)	5.1 (267)	5203	No and PV	9.2 (153)	5.1 (85)	1662	No and VV	10.0 (200)	4.2 (85)	2005
Yes	% (n)	8.8 (28)	4.1 (13)	318	Yes and PV	11.7 (13)	1.8 (2)	111	Yes and VV	9.5 (12)	5.6 (7)	126
ACEs
PV		Chi-square = 10.819 p-value = 0.013 *
No	% (n)	10.0 (288)	5.4 (156)	2877
Yes	% (n)	9.3 (168)	4.9 (89)	1804
VV		Chi-square = 11.757 p-value = 0.008 *
No	% (n)	9.7 (249)	5.9 (152)	2559
Yes	% (n)	9.9 (216)	4.3 (94)	2174

Source: own elaboration based on variables and data from the 2019–2020 Argentina MICS. Notes for abbreviations: %: Percentage; Sig: significant; n: number of cases in each cell; N: total cases by row; SDH: social determinants of health; ACEs: adverse childhood experiences; PV: physical violence; VV: verbal violence; SES: socioeconomic strata; HSES: high socioeconomic stratum; MSES: medium socioeconomic stratum; LSES: low socioeconomic stratum; melevel: mother’s educational attainment; HEA: high educational attainment; MEA: medium educational attainment; LEA: low educational attainment; HC: healthcare access; IND: Indigenous ancestry. * Significant at the level of 0.05.

shows the prevalences in percentages (%), frequencies (n), and total number of cases by rows (N), so prevalences are obtained by dividing n by N. In addition, also shown is the sum of the categories of the explanatory variables (SDH); for example, “boy” (2916) and “girl” (2733) are equal to 5649 cases5, which are the total number of cases that we are evaluating in this study for the dependent variable of malnutrition, “BMI”6.

Additionally, Table 3 presents two sections. Section A presents contingency tables between two variables, where malnutrition (overweight and obesity) is associated with different types of determinants that are classified in Table 2 as SDH. At the end of the list of variables in Section A, Table 3 displays the association of BMI with ACEs with physical (PV) or verbal (VV) violence.

On the other hand, section B of the table shows associations with three variables, taking into account the relationship of dependence of BMI with SDH and some type of ACE (PV or VV).

Therefore, the Chi-square test statistic, denoted as χ², shows the associations of the variables that are significant as follows:

-

In section A (Table 3), the SDH variables that are associated with the BMI variable are shown below:

BMI with age, χ² = 14.8, p-value = 0.021; BMI with SES, χ² = 12.9, p-value = 0.044. The ACEs variables associated with BMI are mentioned below:

BMI with PV, χ² = 10.8, p-value = 0.013; BMI with VV, χ² = 11.8, p-value = 0.008.

-

In section B (Table 3), the BMI variable is significantly associated with the following SDH-ACE variables:

BMI with age and VV, χ² = 12.2, p-value = 0.05.

There are relatively high prevalences among the rates of significance; for example, children under two years of age are overweight at a rate of 10.6%. Likewise, children at this age who experience verbal violence (VV) present the highest rate among all the circumstances shown in Table 3; the prevalence is 12.6%. This rate is even above the average rate of overweight among Argentine children between 0 and 5 years of age. Similarly, around 10% of children aged 2 to 4 years who suffer from VV are overweight, and children aged 4 to 5 years who suffer from VV also show relatively high rates of overweight (8.6%).

High prevalence rates are also present for children who are obese at the age of 4 to 5 years (the percentage is 6%). In addition, children of this same age who suffer from VV have a prevalence of 5%. Likewise, children in early childhood (0 < 2 years) that present VV have a prevalence of 6.2% for obesity, above the average of childhood obesity among Argentine children (5.12% shown in Table 1). Children from 4 to 5 years who suffer from PV and who are also overweight or obese have considerable prevalence rates between 6% and 9%; however, those mentioned later were not significant.

The contingency table showing the relationship between BMI and household socioeconomic status (SES) shows a prevalence rate of 9.2% for poor children (LSES) who are overweight; the middle class (MSES) also shows a high prevalence of overweight, 10.7%7. Above all, the prevalence of these poor children in the presence of VV stands out at a rate of 10% for overweight, and in the presence of physical violence (PV), it is 9.2%. The prevalence of obese children is lower than that of overweight children; 4.3% of poor children who experience PV and VV, on average, are obese.

Since the association between SES (low income) and BMI was significant, it leads us to evaluate other types of deprivations that influence malnutrition. For instance, health care deprivation (HC) is associated with 9.6% of overweight children and 4.6% of obese children in Argentina. Those children deprived of HC and experiencing PV or VV also have rates of 10.3% on average. Regarding the overweight problem, these later-mentioned rates are higher than children who are not deprived due to healthcare; however, the association of HCs and ACEs and their influence on BMI was not significant.

Furthermore, one of the purposes of this study is to show the association of ACEs only with childhood overweight and obesity. The figures in Table 3, section A, show the prevalence of overweight children in the presence of ACEs: 9.3% is the prevalence rate for those with PV, and it is 9.9% for those with VV. The prevalences are significantly lower for obesity, between 4.9% and 4.3% when children experience either PV or VV.

On the other hand, when ACEs are related to gender, there are no significative differences; however, overweight girls present a higher prevalence than boys for both PV and VV, around 10.4% to 10.9% compared to 8.4% and 9.1%.

In addition, high prevalence rates are also found among children with overweight and obesity in the presence of ACEs, such as VV, when the mothers show a low and medium educational attainment (LEA and MEA); in these cases, the prevalence varies from 9.1% to 10.1%. Similarly, another notable fact is that 11.7% of children who declare having Indigenous ancestry suffer from PV in the presence of overweight. However, more studies should be carried out in this regard since these last-mentioned figures were not significant.

The IM-SDH-ACE theoretical perspective, which states that socioeconomic characteristics and domestic violence problems are related to children’s health and well-being, has been corroborated in this study by showing that the above-mentioned categorical variables are related to BMI. Therefore, we accept H_A, which establishes that there is a dependence between the variables when these associations are significant.

3.2. Model Results

The multinomial logistic regression model proposed in Section 2.1 is estimated below, and its results are shown in

Table 4. Results on the multinomial logistic regression model for childhood obesity in Argentina: normal weight and undernutrition (base outcome) vs. overweight and obesity.

Dependent Variable: Malnutrition	RRR	Std. Err.	P > z
Underweight (Base Outcome)
Normal Weight
ACEs (PV) (r.c. = no PV)
Yes	1.994 *	0.478	0.004
HC (r.c. = Yes)
No	0.373	0.223	0.099
melevel (r.c. = HEA)
MEA	0.596	0.214	0.150
LEA	0.967	0.455	0.943
Age (r.c. = less than 2 years)
≥2 age < 4	1.195	0.288	0.460
≥4 age < 5	1.401	0.409	0.248
Interactions
melevel and HC (r.c. = HEA and Yes)
MEA and No	3.629	2.444	0.056
LEA and No	2.186	1.574	0.278
Constant	34.639 *	12.067	0.000
Overweight
ACEs (PV) (r.c. = no PV)
Yes	1.821 *	0.468	0.020
HC (r.c. = Yes)
No	0.481	0.320	0.272
melevel (r.c. = HEA)
MEA	0.568	0.221	0.148
LEA	1.096	0.550	0.854
Age (r.c. = less than 2 years)
≥2 age < 4	0.997	0.263	0.993
≥4 age < 5	1.094	0.347	0.776
Interactions
melevel and HC (r.c. = HEA and Yes)
MEA and No	3.705 *	2.762	0.019
LEA and No	1.255	0.994	0.774
Constant	4.893 *	1.839	0.000
Obesity
ACEs (PV) (r.c. = no PV)
Yes	1.757 *	0.481	0.039
HC (r.c. = Yes)
No	0.271	0.213	0.098
melevel (r.c. = HEA)
MEA	0.596	0.245	0.210
LEA	0.927	0.492	0.887
Age (r.c. = less than 2 years)
≤2 age < 4	0.913	0.262	0.752
≤4 age < 5	1.437	0.483	0.281
Interactions
melevel and HC (r.c. = HEA and Yes)
MEA and No	4.834	4.191	0.069
LEA and No	2.449	2.232	0.326
Constant	2.820 *	1.123	0.009
Number of obs. = 4663 LR chi2(24) = 37.61 Prob > chi2 = 0.038 ** Pseudo R2 = 0.065

Source: own estimates based on the multinomial logistic regression fitted for the dependent variable of malnutrition and explanatory variables. Dataset used was the 2019–2020 MICS for Argentina. Note for abbreviations: LR: log likelihood ratio; r.c. reference category; ACEs: adverse childhood experiences; PV: physical violence against child; no PV. no physical violence; HC: healthcare access; melevel: mother’s educational attainment; MEA: medium educational attainment; HEA: high educational attainment; LEA: low educational attainment. * Regressors are statistically significant at 95%. ** X² probability is significant at 95%.

, which shows the relative risk ratios (RRRs) for normal weight, overweight, and obesity. The reference category is childhood underweight in the multinomial logistic regression to be able to compare the RRR of obesity and overweight with other types of diseases. The reason is because this study suggests that weight disorders are a function of the SDH and ACE determinants and the relationship between these determinants.

The following analysis considers the possibility of weighting the risks based on how many times the probability of being overweight or obese increases under the influence of some determinants compared to undernutrition, so our view is relative, based on a comparison of weight disorders; otherwise, estimating the RRR from the reference category of normal weight would imply minimizing the effect of childhood overweight and obesity, given that 83% of the population of children aged 0 to 5 years has normal weight.

The validity of the model was also assessed based on the statistical test of X², with a value of 37.61 and a significance of 95% (Prob > chi2 = 0.03); therefore, we reject the null hypothesis and accept the alternative hypothesis that all regressors are different from zero (Table 4).

3.3. SDH-ACE Factors That Influence Childhood Overweight and Obesity in Argentina

Multinomial logistic regression empirically corroborates the study hypothesis from a more integrative view once it has been shown with this model that SDH and ACEs are risk factors that influence the experience of overweight and obesity in Argentine children aged 0 to 5 years. Subsequently, we explain the results shown in Table 4 by focusing on the explanatory variables that were significant in the model with a p value of < 0.05.

The model indicates that ACEs, particularly physical violence (PV) inflicted upon children in their homes compared to no violence, nearly double the risks of being overweight in childhood, with approximately 1.821 RRR compared to underweight children. This statement is significant in the model and is also confirmed for children who experience obesity, with approximately the same RRR, of 1.757, influenced by PV at home.8

This study also explores the effects of the interaction between the mother’s educational level (melevel) and access to healthcare (HC), as we raise in this research the importance of showing the relationships between the factors and their influence on the health of children, as indicated in the “IM-SDH-ACEs”, and these results are also significant in the model. Thus, we found that the interactions between medium educational attainment (MEA) and a lack of access to healthcare (HC) increase the risks of presenting childhood overweight by almost four times (3.705 RRR). This likelihood is also significant in the model. However, this interaction was not significant in the case of childhood obesity (4.834 RRR). The interactions between LEA and a lack of access to HC increase the likelihood of being overweight and obese; however, these are not significant (1.255 and 2.449 RRR).

It is important to mention that the estimated interaction in this model is an effect modifier of the mother’s educational level (melevel), which interacts with one of the main risk factors, health care (HC); that is, there is an inverse effect when one of the main risk factors is considered alone in the model. In this case, there is an RRR of 0.481 of being overweight when there is deprivation due to access to health (HC). The effect is the opposite expected as it reduces the risk to less than one. On the contrary, the interaction between melevel and HC factors, in particular, the medium educational level of mothers (MEA) and health deprivation (HC), are significant and increase the risk of being overweight in children. “A covariate is an effect modifier only when the interaction term added to the model is both clinically meaningful and statistically significant” (Hosmer and Lemeshow 2000, p. 74); other interactions were tested and are not significant in the model.9

Factors relating to low educational attainment (LEA) as well as children under 4 and 5 years of age increase the risk of being overweight by more than one-fold RRR (1.096 and 1.094) compared to the baseline outcome. The 4–5-year-olds are also vulnerable to obesity as their RRRs are 1.437 times more likely than children in the baseline outcome. However, these factors are not significant. However, these variables remain in the model as they do not reduce the overall significance of the model, and it is important to keep them within the model as well as the right to access to health (HC) and the mother’s educational level (melevel) because they are theoretically relevant according to the IM-SDH-ACEs framework.

It is worth mentioning that the model presented in Table 4 is the best-fitted model.

4. Discussion

The research question of the study was empirically corroborated by demonstrating that SDH and ACE are determining factors of overweight and obesity in children under 5 years of age in the Argentine case. The theoretical model “IM-SHD-ACE” has proposed that both types of factors influence child malnutrition.

In general, it can be deduced that a lack of household income, as an indicator of the socioeconomic strata (SES), is associated with the presence of childhood overweight or obesity in the home based on the Chi-square statistical test. In addition, we can see in this study that there are significant dependencies between two types of physical abuse (PV) and verbal violence (VV) in their association with malnutrition (IMC), particularly with childhood overweight and obesity.

The discussion lies in recognizing that low-income households affect access to health care and generate a vicious circle of negative effects on health conditions due to the lack of basic services in the home (Aragonés and Uberto 2023, p. 180), as well as the lack of adequate nutrition (Mancino et al. 2018), particularly during childhood. Consequently, if child abuse at home is added to these conditions of poverty and deprivation, then it is consequently associated with the experience of eating disorders such as overweight and obesity in childhood (Felitti et al. 1998; Tuñón et al. 2015; Orden et al. 2019).

Estimates of the multinomial logistic regression model found that the influence of the domestic violence variable, such as physical violence in children (PV) on malnutrition (overweight and obesity), is empirically corroborated. Thus, ACEs are confirmed as a type of proximal factor in children’s health that increases the risk of presenting these weight disorders. For example, domestic violence is classified within the family factors of the child’s closest environment in Figure 1 and affects children’s health. Previous studies also justify that ACEs are a proximate factor because the stress generated in early childhood can trigger this type of disease related to overweight and obesity (Kelly-Irving et al. 2013; Soares et al. 2016).

Sánchez-Vargas et al. (2023) state that family factors are relevant to children’s lives, as they provide them with opportunities to accumulate human capital, such as the educational level of parents and other social and economic resources, such as family income. These aspects promote children’s education and provide them with knowledge, development, and empowerment. For this study, two joint factors have proven to be relevant in preventing children from being at risk of developing weight disorders, which are the mother’s educational level and access to health services at home.

Furthermore, the theoretical explanations presented in this study suggest that a lack of healthcare could have a negative influence on children’s health and well-being. However, the result of the multinomial logistic model regarding the influence of deprivation of access to healthcare on children’s overweight risks was significant only with the interaction of a mother’s secondary education level. Therefore, the evidence from previous studies, such as the research by Hanć et al. (2022), who show that this interaction is crucial for Polish children, is also corroborated for Argentine children.

Additionally, there is a need for more information from surveys that consider retrospective questions about whether the child suffered verbal or physical abuse at an earlier age since we have evidence in this research that ACEs are associated with overweight and obesity in early childhood. This study shows that the highest prevalence is 13% of children aged 0 to 2 years who suffer verbal abuse are at risk of becoming overweight. In fact, there is evidence that maternal experiences have influences on children’s health, such as toxic stress, which in turn could be associated with their BMI (Kelly-Irving et al. 2013; Condon et al. 2019). In this way, more conclusive results can be achieved on the set of determining factors that influence children’s health.

The IM-SDH-ACEs is a multidimensional and dynamic process in which all factors are interrelated. For instance, the lack of opportunities as well as social exclusion permeates this dynamic process, given that socially excluded groups are not only poor but also vulnerable populations, living in disadvantage, discrimination, marginalization, and deprivation, among other aspects, implying a lack of access to markets and services, such as health care, as well as a disadvantaged socioeconomic position that also leads to the limitation of people to exercise their rights, (Cuesta et al. 2024), including social security, etc.

The importance of this research article also lies in knowing more about the problem of childhood obesity to generate appropriate public and health policies, since in the life cycle of people, experiencing ACEs in early childhood can lead to different types of long-term illnesses. Even childhood obesity can lead to diabetes in the medium and long term, including the stages of late childhood and adolescence.

Limitations of the Study

This study also states that there is a limitation of microdata surveys with information regarding ACEs and malnutrition. The MICS contains information on child maltreatment and abuse as well as types of domestic violence and reports various anthropometric measurements for children from 0 to 5 years old in the MICS of Argentina; however, it does not contain anthropometric variables for children over five years. The MICS also requires a retrospective approach, because it would be important to have a long period of analysis to study this problem of childhood malnutrition in order to reach more solid conclusions about the influence of children’s exposure to ACEs and the determination of these risk factors in childhood obesity.

Furthermore, future research directions on this topic can be tested based on structural equation modeling to identify relationships between proximate or intermediate determinants of the exposed approach, “IM-SDH-ACEs”, for childhood and verify it for other social contexts, since Argentina is not unique in terms of the influence of ACEs and SDH on overweight and obesity in childhood. This research represents an empirical advance in corroborating the integrated model of SDH and ACEs presented in this research.

The IM-SDH-ACEs model proposed here could also be analyzed in future studies through multilevel analysis to incorporate the association of determining factors at the mesosocial and macrosocial levels in childhood obesity, since social, cultural, and natural environments, as well as public policies and social services available in communities, influence the health and well-being of children. Therefore, more appropriate public policy interventions could be designed so that children achieve well-being and health.