Assessment of Early Childhood Mental Health in Immigrant and Local Children in Chile

Little is known about mental health during early childhood and differences in mental health problems between migrant and local preschool children in Chile. This research aimed to study the early mental health of children and disparities in mental health indicators between immigrant and local children in the context of a middle-income country. Parents and teachers assessed immigrant (n = 120) and non-immigrant (n = 383) children between 3 and 4 years old with the Sistema de Evaluación de Niños y Adolescentes (Child and Adolescent Evaluation System, SENA). A multivariate analysis of covariance (MANCOVA) was used to study both main and interaction effects while controlling for the clustering of children on center-based care. Results show that immigrant parents tended to report more internalizing problems than locals’ parents did in terms of depression, anxiety, and somatic complaints. Also, they reported more attention problems, developmental delay, unusual behavior, rigidity, and isolation in immigrant children. However, teachers’ assessments regarding immigrant children were similar to non-immigrant children’s evaluations in many cases. Finally, girls presented fewer emotional and behavioral problems than boys. Our study shows the importance of assessing mental health during early childhood, especially in immigrant children, and the difficulty of carrying out this evaluation through different information sources from parents and early childhood teachers.


Introduction
Children worldwide experience mental and emotional disorders, which frequently remain unrecognized. The prevalence rates are difficult to determine because of a lack of knowledge and difficulty in assessing and diagnosing young children (Lyons-Ruth et al., 2017). Existing information suggests that the rates of mental disorders among young children are comparable to those of older children and adolescents. Some epidemiological studies have revealed that the prevalence of such conditions among children aged 1-5 years is 16-18%, with just over half (8-9%) of these being severely affected (Von Klitzing et al., 2015). It has also been found that the highest prevalence rates can be observed in developing nations (World Health Organization, 2005. The lack of broader recognition of childhood disorders particularly concerns because of the position that childhood occupies as a central pillar in development throughout life (Fox et al., 2010;Knudsen, 2004;National Scientific Council on the Developing Child, 2007). However, the first years of life can be a risky stage because trauma and violence can have an impact on children's development (Black et al., 2017;Champagne 2010;Shonkoff et al., 2012;Walker et al., 2007).
The evidence of the last few decades points to the importance of early diagnosis of symptoms and not waiting until children are of school age (Egger & Angold, 2006). Mental health problems can be identified early during infancy and childhood, and they can affect individuals throughout adolescence, and adulthood (Cote et al., 2009;Rutter et al., 2006). Moreover, there is evidence that severe behavioral problems in the early years of life are associated with a moderate to a strong tendency toward long-term problems (Campbell et al., 2000).
In terms of cognitive development, such disorders as depression in the early stages of life usually include difficulties in attention, concentration, comprehension, and retention, affecting learning processes (Graber, 2013). The prevalence of depression among children and adolescents globally is 6% (Erskine et al., 2017). In addition, anxiety disorders are a significant health problem among children, with a worldwide prevalence of 1-27% and the highest rates in Latin America (Erskine et al., 2017;Wlodarczyk et al., 2016). Behavioral disorders are another category of problems that can occur in childhood (Batstra et al., 2012). Diagnoses of this type of disorder are observed more frequently in males, with more cases of attention deficit hyperactivity disorder, a disorder that significantly affects socialization and learning processes (Batstra et al., 2012).

Immigrant Children and Mental Health Problems
The scarcity of studies on mental health among preschool children is even more evident in the migrant population, although we know that the experience of migration can negatively affect the psychological wellbeing of migrant children (DuPlessis & Cora-Bramble, 2005;Stevens & Vollebergh, 2008).
Immigrant children are defined as those who have at least one foreign-born parent. They include both the first generation born outside the host country and secondgeneration immigrants born within the host country (Suárez-Orozco et al., 2015).
Unlike their non-immigrant peers, immigrant children face significant acculturative challenges, including learning the characteristics of the new culture, while at the same time learning and maintaining their family's culture of origin (Oppedal & Toppelberg, 2016). Although many immigrant children do well and even thrive in many areas, others encounter difficulties (Masten et al., 2012).
Immigrant children can experience distress and problems with psychological adjustment (Motti-Stefanidi et al., 2012). The most common mental health problems in the population of immigrant children are internalized difficulties, such as anxiety, depression, or somatic complaints (Kim et al., 2018;Rousseau & Frounfelker, 2019). There is also a higher prevalence of externalized problems, such as disruptive or aggressive behavior, and other problems, such as substance abuse (Fernández, 2018;Kim et al., 2018). In a systematic review that only included studies conducted in Europe and another systematic review of studies from the United States and Canada, Kouider et al., (2014Kouider et al., ( , 2015 failed to establish conclusively whether migratory status is a risk factor for mental health in children. In Germany, children with a migration background showed lower levels of wellbeing than local children did. However, those in the former group who attended initial education benefited from it (Kaiser & Bauer, 2019). In the same country, a study carried out with migrant children of preschool age found that the children showed more significant externalizing problems than the native group (Schreyer & Petermann, 2010), while another study showed differences in internalizing problems (Kuschel et al., 2008). Finally, Daglar et al., (2011) compared the behavior of Turkish preschoolers who migrated to the United Kingdom, natives living in Turkey, and migrants living in Turkey. The two migrant groups showed greater externalizing problems than the non-immigrant group did.
Despite the above findings and how complex the migratory process may be for children; some positive consequences have been observed in this process. Thus, in some contexts, immigrant children have mental health problems similar to the children in the receiving culture. They may perform as well as or better than non-immigrant children in psychosocial adaptation, despite the inherent adverse experiences in the migration process (socioeconomic difficulties). This phenomenon is known as the "migration paradox," which has given rise to claims that immigrant children are more resilient (Marks et al., 2014;Masaud et al., 2010).

Chilean Context
An increase in migration has been observed in Chile, mainly from Latin American countries. In 2002, 187,008 migrants resided in our country, constituting 1.3% of the total, while the 2017 Census showed 746,465 migrants, corresponding to 4.4% of the total population (Instituto Nacional de Estadísticas, 2019).
In Chile, although the migratory flow has been increasing, research on mental health in the immigrant population has been sparse (Abarca & Carreño, 2014;Urzúa & Cabieses, 2018). However, the literature that has been performed has mainly focused on the adult population, reporting that high levels of depressive and anxiety symptoms and social adjustment problems are associated with migration (Urzúa et al., 2017). Recently, Caqueo-Urízar et al. (2020) found no significant differences in mental health indicators between immigrant and non-immigrant children at both elementary and secondary levels.
It is particularly relevant to study this phenomenon in Chile because there are substantial inequality gaps between immigrant children and their Chilean counterparts. The former have higher biopsychosocial risk (Cabieses et al., 2017), difficulties accessing health services, and a higher prevalence of common mental disorders, such as mood and anxiety disorders (Rojas et al., 2011).
Assessing and diagnosing very young children's mental health is a difficult task. Most studies on preschool mental health and experiences of child migration have been conducted in developed countries . For this reason, the objective of this study is to contribute to knowledge by describing the mental health indicators of preschool children aged 3-4 years, a population that has generally been studied little and establishing differences between immigrant and non-immigrant children in the context of a middle-income country.

Participants
The study utilized data from 520 children attending the higher secondary level of publicly run kindergartens administered by the National Board of Kindergartens (Junta Nacional de Jardines Infantiles, JUNJI) Arica y Parinacota and Metropolitan regions of Chile. We invited 21 center-based care in the Metropolitan region with more migrant children (between 10 and 30%) to participate in our study. In Arica and Parinacota Region, we invited all center-based care in the Arica district. The study used a non-probabilistic cluster sample. Twenty-eight classrooms were selected: nine at center-based care in the Arica y Parinacota region and 19 in the Metropolitan region.
Of the 520 cases, seven were removed from the analyses because they showed scores equal to or greater than 1.6 on the inconsistency scale, as recommended as a cutoff by the TEA publishing house (Fernández-Pinto et al., 2015). There were also 10 cases where information on the child's immigration status could not be obtained. Therefore, for analysis of the children's mental health, the final sample consisted of 503 valid cases where both the parents and the early childhood teachers completed their questionnaire versions. Table 1 shows the characteristics of the sample. In the sample analyzed, 47.5% (n = 239) were boys and 52.5% (n = 264) were girls. Moreover, 48.9% (n = 246) were 3 years of age, and 51.1% (n = 257) were 4 years old. A total of 37.8% (n = 193) were from the city of Arica, and the rest were from 11 districts of Santiago.
Of the children, 23.9% were immigrants (n = 120) and 76.1% were non-immigrants (n = 383). In Table 1, when comparing both subgroups (immigrant versus nonimmigrant), differences can be observed in the mothers' educational level (t = 2.47, df = 156.2, p = 0.01), with the mothers of non-immigrants obtaining a higher level. There are also significant differences regarding the number of people in the household (t = 2.01, df = 202.04, p = 0.05), indicating that the households of immigrant children are smaller than those of non-immigrant children. Immigrant children tend to be entering kindergarten later compared with non-immigrant children (t = − 2.67, df = 193.78, p < 0.01).

Instruments
The Sistema de Evaluación de Niños y Adolescentes (Child and Adolescent Evaluation System, SENA) was developed by specialists in psychopathology and psychological assessment. SENA's aim is to help identify a wide range of emotional and behavioral problems in individuals ranging from 3 to 18 years old. It should be noted that it has been fully developed and validated in Spanish (Sánchez-Sánchez et al., 2016). The constructs this instrument assesses are as follows: • Internalized problems: depression, anxiety, social anxiety, and somatic complaints.
• Externalized problems: attention problems, hyperactivity/impulsivity, anger control, aggression, and defiant behavior. • Other problems: delay in development and unusual behavior-behaviors associated with autism spectrum disorder. • Vulnerabilities: problems in emotional regulation (difficulties regulating [inhibiting-facilitating] or controlling certain emotional and behavioral reactions), rigidity (lack of adaptability to changes and lack of flexibility), and isolation.
SENA also assesses the presence of certain personal resources that individuals can use to address the personal difficulties they may encounter. These protective or support factors are social integration and competence, as well as emotional intelligence, defined as a set of skills to perceive and appropriately manage both own emotions and those of others. Finally, the instrument includes scales of control, such as positive impression (informs the presence of a markedly favorable response pattern of the individual assessed), negative impression (informs the presence of a markedly negative response pattern of the individual assessed), and inconsistency. The last component evaluates whether the individual has answered the test items without paying attention to what was appropriate or random (Fernández-Pinto et al., 2015).
SENA uses a Likert scale between 1 and 5 in the response options for each item, ranging from "never or almost never" (1) to "always or almost always" (5). The total of each dimension is the average of the total of the responses that constitute it, which ranges from 1 to 5. The reliability of subscales is greater than 0.7 in Spain (Sánchez-Sánchez et al., 2016). In Chile, the scale shows adequate psychometric properties with reliability ranging from 0.67 to 0.80 (Caqueo-Urízar et al., 2021).
This study used the children's version of SENA, which covers children aged 3-6 years, with scales for each year of age. The two available versions, Family and School, were used, for which reports were made by the family members and teachers. These instruments contain 131 and 122 items, respectively. The response rates from the School SENA was 96.6% and 90.1% from the Family SENA.

Sociodemographic Questionnaire
A sociodemographic questionnaire was administered to parents to collect such variables as the culture of origin of the children-expressed in terms of the country of birth of the child, the mother, and the father, and whether they belonged to a certain native people-age, sex, and level of entry to preschool education. Finally, the variable on the educational level of the parents was also included, which was defined as the highest level of education completed, distinguishing between complete and include levels of secondary education. This study defines immigrant child as any child born outside Chile or any child with a parent who was born abroad (Portes & Zhou, 1993).

Procedure and Ethical Considerations
Approval for this study was received from the Ethics Committee of the Pontificia Universidad Católica de Chile. Nine educational establishments in the Arica y Parinacota Region were invited to take part, all of which accepted. Consent was requested from the parents of all the participating children after explaining the purpose and scope of the study. The informed consent of the participating educators was obtained in the same way.
The assessments of the families were carried out in three stages. In the first stage, meetings were held in which the instructions and individualized support were provided when needed. These meetings lasted approximately 45 min. Subsequently, the families that did not attend were contacted in person or by phone to schedule an appointment. Finally, if they were not available, the protocols were sent to their home, along with specific instructions.
In the case of the classroom team, they were accompanied by interviewers who were trained in the instrument. Many of the instruments were employed together to ensure the necessary feedback. The entire field process was carried out between September 2019 and January 2020.

Statistical Analyses
Descriptive statistics were employed for the dimensions assessed in the Family and School versions of the SENA instrument. A first approximation of the differences was determined using a direct comparison via a t-test. Then, multivariate analysis of covariance (MANCOVA) was used considering sociodemographic covariates. MANCOVA goes beyond a corroboration for two reasons: first, it allows us to control for the nested structure of the data by including fixed effects to control for the clustering of children in preschools, and second, by testing not only of the main effect of children immigration status and who did the assessment (parents or teachers), but also the effect of their interaction. Stata version 14 was used to analyze the data.

Children's Mental Health Problems and Personal Resources by Teachers' and Parents' Reports
The results shown in Table 2 illustrate that the problems most frequently reported by both parents and teachers were associated with externalizing problems, such as attention problems and hyperactivity/impulsivity. At the same time, in the area of internalizing problems, parents and teachers reported problems of social anxiety. Meanwhile, parents reported a higher frequency of problems related to externalization, such as problems with anger management, defiant behavior, and emotional regulation problems, compared with the assessment of teachers. Finally, both the parents and teachers report good integration and social competence of the children, as well as good levels of emotional intelligence (Table 3, 4).

Children's Mental Health Problems and Personal Resources by Migrant Status
To determine how the assessment of immigrant versus non-immigrant children varies when assessed in different contexts, namely the school and the family, we performed multivariate analysis of variance (MANCOVA). The mental health reports were modeled as follows: (i) a function of the children's situation (immigrant/non-immigrant), (ii) a function of who made the evaluation (parents/teachers), and (iii) the interaction of these two variables. In addition, the model controlled several covariates, such as children's sex and children's age, geographical region, and the kindergarten they attended to control for clustering effects. Table 4 presents the results of the MANCOVA.
The omnibus MANCOVA test was statistically significant (p < 0.001) according to the four statistics that are traditionally considered, namely Wilks' lambda, the Lawley-Hotelling trace, Pillai's trace, and Roy's largest root, indicating consistently statistically significant differences of mental health scores as a function of the predictors considered in the model. In terms of the specific predictors, the MANCOVA results indicated a statistically significant effect across the 16 dependent variables for the immigrant status of children, the source of the mental health assessment, and the interaction of these two factors, as well as the significant effect of gender and the factor associated with the kindergarten to which they attended, included to control for correlation among the children due to this clustering. The MANOVA model did not find statistically significant effects for two of the control covariates included in the model-the age of the children and the geographical region in which they reside. Details of the overall MANCOVA results are presented in Table 3. Table 4 presents the results for each of the 16 regression models considered in the MANCOVA. In terms of the control covariates included in the model, the only one that was statistically significant overall was gender, indicating that girls were systematically assessed as having fewer problems than boys, in particular with lower levels of depression, lower externalizing problems (in all dimensions), lower vulnerability, and lower contextual problems. In addition, girls were consistently assessed as having more personal resources compared with male children.
In relation to one of the two main predictors of interest, Table 4 shows the results of the contrast between the reports that parents make of non-immigrant children versus the reports of immigrant parents. It can seem that immigrant parents tend to report greater degrees of problems than parents of non-immigrant children. In particular, they reported higher levels of depression, anxiety and somatic complaints (internalizing problems), and attention problems. Similarly, immigrant parents report more developmental delay, unusual behavior, rigidity, and isolation in immigrant children. Finally, immigrant parents also reported lower levels of emotional intelligence. Table 4 also presents the results associated with the second predictor of interest, contrasting the mental health reports made by parents versus teachers of non-immi-  grant children. In this case, we can see that teachers tend to report better outcomes for non-immigrant children compared with the assessments of their parents, reporting lower levels of internalizing problems, specifically anxiety and somatic complaints, and lower levels of externalizing problems (except for depression). In addition, teachers report fewer vulnerabilities in non-immigrant children and fewer problems with emotional regulation. In contrast, teachers report higher levels of depression and rigidity than non-migrant parents do. Finally, teachers report significantly fewer personal resources compared with the resources perceived by the parents of non-migrants.
As the MANCOVA analysis indicated, there was a statistically significant interaction on mental health results when teachers assessed immigrant children. These interactions consistently show a negative effect pointing toward lower levels of reporting of depression, anxiety, somatic complaints, attention problems, rigidity, and isolation in the assessments that teachers make of immigrant children. When considering the overall results, including both main effects and interaction effects, the model indicates that teachers' assessment regarding immigrant children is often similar to the evaluations of non-immigrant children, although immigrant parents tend to report higher levels. A representative pattern of results for the variables where the interaction was statistically significant is presented in Fig. 1, where we can see how the reported   levels of anxiety vary for immigrants and non-immigrants children when assessed by their parents versus teachers. We can see that parents of immigrant children reported higher levels of anxiety compared with parents of non-immigrant children; however, no statistically significant difference is evident when comparing the teachers' reports for immigrant and non-immigrant children.

Discussion
The study's objective was to describe preschool mental health indicators and establish differences according to children's migrant/non-migrant status. Considering all children, the highest scores, although not high enough to be clinically significant, were observed in attention problems, hyperactivity/impulsivity, anger control problems, defiant behavior, and problems of emotional regulation. Parents reported higher scores for all these problems compared with the assessments made by teachers. Both reports generally indicate good social integration and competence on the children's part.
For the comparison between immigrant and non-immigrant status, from the perspective of parents, there are statistically significant differences in various symptoms associated with internalizing problems, such as depression, anxiety, and somatic complaints. The presence of internalizing problems in the preschool immigrant population is consistent with international findings (Kim et al., 2018;Reijneveld et al., 2005;Rousseau & Frounfelker, 2019). It is also consistent with previous findings for immigrant children in primary education in Chile, where the immigrant school population also obtained higher scores on internalizing problems and lower scores on self-esteem subscales, as well as having greater difficulties in terms of their integration and social competence than the non-immigrant population did (Caqueo-Urízar et al., 2019).
Between immigrant and non-immigrant children, there are also differences in externalizing problems, such as attention problems, and contextual problems, such as emotional regulation, rigidity, and isolation. These results are also consistent with international evidence that shows a higher prevalence of externalized problems in the immigrant population (Daglar et al., 2011;Fernández, 2018;Kim et al., 2018).
Certain vulnerabilities are expressed to a greater extent in the immigrant population, such as developmental delay and unusual behavior. All the aforementioned symptoms are greater in immigrants according to the perspective of parents, while, from the perspective of the schools-that is, based on the assessment of the educators-no significant differences were found in any of these indicators. Some authors also found no differences depending on the culture in the broad indicators according to the teachers' views when comparing the means (Stevenset al., 2003).
The differences between parents and teachers reports about children's mental health and resources show the difficulty involved in assessing and diagnosing young children. The detection of these children's problems is affected by who is being judged and where the judgment is carried out. Clinical assessments often include reports from multiple informants who observe children's behavior in different contexts (De los Reyes et al., 2015). For this reason, to better determine how the assess-ment of immigrant versus non-immigrant children varies when assessed in different contexts (school or families), we performed a MANCOVA.
First, it can seem that immigrant parents tend to report greater degrees of problems and lower levels of personal resources than families of non-immigrant children. Second, when comparing the mental health reports made by parents versus the teachers of non-immigrant children, it can seem that the teachers tend to report fewer problems, except depression and rigidity. Finally, teachers' assessment regarding immigrant children in many cases is similar to the evaluations of non-immigrant children, even though immigrant parents tend to report higher levels of emotional and behavioral problems. This has already been reported in previous literature, both for emotional and behavioral problems (Liu et al., 2011) and for executive function problems (O'Meagher et al., 2020).
There are several potential explanations for indicators where the parents' assessment of children's mental health problems is higher than the teachers' evaluation of children's mental health problems. First, parents may overfocus on their child, making problems seem more salient. Second, children's behavior in school and at home may differ. Third, given the teachers' more extensive experience with different children, they are likely to have more expertise in differentiating different types of problem behaviors than parents, who have only a few children at home. Fourth, since teachers rate immigrant and non-immigrant children, similarly, may there also be an effect of the own parents' level of stress. A hypothesis could be that immigrant parents potentially interpret children's behaviors more "negatively" given their own levels of stress. Future studies that use objective and independent direct observation other than rating scales from different informants (parents and teachers) may shed light on the best way to measure preschoolers' mental health.
When looking in more detail at possible differences by gender, we observe that boys may have a greater risk of suffering from depression and externalizing problems (attention problems, hyperactivity/impulsivity, problems with anger control, aggression, and defiant behavior). This is consistent with the results that found higher reporting of externalized problems in males (Liu et al., 2011). This trend has been reported in various cultures, where boys show more aggression problems than girls do (Chen, 2010;LaFreniere et al., 2002).
There may be multiple interpretations about the gender differences in externalizing problems. One explanation could be that preschool girls, relative to their boy counterparts, are more likely to develop earlier and better language skills. Better language skills enable them to regulate their own emotions and behavior better and communicate their needs more effectively (Keenan & Shaw, 1997;Petersen & LeBeau, 2021). Another possible explanation could be from a cultural perspective, where gender variations in externalizing problems may reflect divergent socialization patterns among parents (Endendijk et al., 2017;Nikapota, 2009).
In Chile, measurements of mental health in preschool children are particularly significant, given the high rates of mental disorders (27.8%) previously observed in children between 4 and 11 years of age (Vicente et al., 2012). In addition, the increase in the migrant population in the school system (Ministerio de Educación de Chile, 2018), means that it is especially important to study the preschool migrant population. It is necessary to diagnose mental health problems early because they can affect normal functioning throughout life (Rutter et al., 2006). Migration is a complex social process that subjects' children to physical and social stresses, and it can provoke mental health problems. To protect the mental wellbeing of migrants, it is important to conduct an early diagnosis of mental health vulnerabilities in this population, as well as the risk factors associated with the socioeconomic circumstances in which they live (Cabieses et al., 2017).
Some interventions in developed countries suggest that the school could be a suitable place to assess and address mental health needs in a safe and comfortable environment, with the family involved (Ellis et al., 2010(Ellis et al., , 2011. Some studies have suggested that a strong alliance between parents and teachers in the face of the challenges of education from an early age would create stable conditions for children's learning and personal growth. Specifically, it has been observed that a closer relationship between teachers and parents is associated with better indicators of emotional regulation (Acar et al., 2019). Similarly, when there is effective coordination and communication between the family and the school, it is possible to prevent or identify and treat emotional or behavioral conflicts successfully and appropriately, given that systematic and heterogeneous observation of repetitive patterns or changing dispositions in the behavior of children in different interactive settings can warn of the presence of cognitive, emotional, or socio-emotional disorders (Sheridan et al., 2012). Definitely, school-based mental health programs that connect children, families, schools, and communities can contribute to promote children mental health (DiGirolamo et al., 2021). More research is needed in this area, and in particular, on effective interventions that can influence mental health outcomes in immigrant and non-immigrant children in early childhood education (Chan et al., 2009).
The limitations of this study include that it is not possible to generalize for all of Chile, and finally, it is necessary to consider that, during the data collection period there was a social uprising in Chile, which caused transportation difficulties for the families (Delgado, 2019), numerous episodes of violent confrontations between police and civilians (Bermúdez, 2019), and more than a month of interruption in the provision of initial education in the establishments that participated in the study.
Future studies should cover a larger population, although this study did manage to address two regions that have cultural differences. It would also be useful to conduct longitudinal studies that allow the evolution of mental health to be observed. Finally, it could be interesting to compare the mental health of children who have migrated from different countries, considering the huge variability that exists within the group of migrants. While some come from middle-income Latin American countries, such as Venezuela or Peru, there are others who migrated from such countries as Bolivia and Haiti, historically categorized by low income (Fernandez, 2018World Bank, 202073]. In other studies, differences have been observed between groups of migrants (Stevens et al., 2003), particularly when they are studied according to the income level of the nations of origin, which is a predictor of the prevalence of mental health problems (World Health Organization, 2005;.

Conclusion
The findings of this study show that parents report significant differences in internalizing problems and behavioral problems, with the scores being higher among immigrant children; however, teachers' assessments do not reflect these differences. It is possible that immigrant children manifest greater symptoms in family contexts where there is a high likelihood of situations of poverty, or parental stress. These social determinants of mental health should be approached from a broader perspective. Therefore, the design of public strategies for early diagnosis and mental health interventions among children should be considered a priority for public health (Copeland et al., 2013), particularly for a country with high levels of mental problems in its population.