Prevalence and factors associated with adverse early childhood experiences: a population-based study in Ceará, Brazil

Objective: To estimate the prevalence of adverse childhood experiences and identify associated factors. Methods: A population-based cross-sectional study comprised data from a sample of 3,200 households with 3,566 children under 6 years of age, representative of the state of Ceará, Brazil. A multistage sampling approach was used, with stratification among the state capital, Fortaleza, and the 28 countryside municipalities, in which 160 census tracts were randomly selected, each one with a cluster of 20 households. The outcome variable was structured based on adverse childhood experiences as suggested by the Center for Disease Control and Prevention, according to the number of situations to which the child was exposed: 0–2, 3–5, and 6–9. Ordinal logistic regression multivariate model was applied to assess associations. Results: Among the 3,566 children studied, 89.7% (95%CI 88.7−90.7) were exposed to at least one adverse experience, of which the most prevalent were neglect, and emotional/physical abuse. The main factors associated were maternal advanced age and smoking, paternal absence, low education level of the head of the family, food insecurity and lack of a social support network. Conclusion: The study found a high occurrence of adverse early childhood experiences, particularly among preschool children born to mothers of older age, solo, who smoke and in a situation of social and economic vulnerability, including food insecurity, who should be target of control and prevention measures.


INTRODUCTION
Adverse childhood experiences (ACEs) are potentially traumatic events that occur in childhood and can directly influence child development. About 1 billion children and adolescents worldwide are estimated to be exposed to violent behavior, which may be associated with up to 45.0% of childhood disorders and up to 32.0% of those that appear later 1 . In a survey of more than 26,000 adults conducted by the Centers for Disease Control and Prevention (CDC) in the USA, it was observed that 59.4% of the sample had contact with at least one ACE 2 . A meta-analysis study shows a 12.7% prevalence of childhood sexual abuse in the world 3 . The highest prevalence and intensity of ACEs are associated with negative outcomes such as mental health problems, chronic diseases, infectious diseases, and risk behaviors such as smoking and alcoholism 4,5 .
Previous studies in Brazil, especially in the South and Southeast states, reveal the occurrence of at least one ACE, range from 53.6 to 85% [6][7][8] . In the Southern states study, it was reported a 7% of physical abuse, 1.4% of sexual abuse and 20% of emotional neglect 8 .
These experiences result in delays in child development, biological and psychosocial impairment in adult life, with a higher prevalence of chronic diseases, cancer, cardiovascular diseases, sexually transmitted diseases, health risk behaviors, sedentary and inappropriate lifestyle habits, psychological distress, self-perception of health and low quality of life, disruptive behaviors and increased risk of early mortality 1,9,10 .
They are public health problems and are divided into three major categories: types of abuse (physical, sexual, and mental), family dysfunction, and physical or emotional neglect 4 . The prevalence of ACEs varies according to the location (influence of cultural beliefs and values, race and sociodemographic characteristics), the collection instrument (type, quantity and format of the questions) used, and the evaluated sample (specificities of each group, selection of sample and type of collection) 3 .
Although many studies worldwide have estimated the prevalence, very few studies are focused on identifying risk or protective factors associated with ACEs. Thus, the present study aims to both estimate the prevalence and identify factors associated with adverse early childhood experiences.

METHODS
We analyzed data from the Maternal and Child Health Survey Study in Ceará (Pesquisa de Saúde Materno-Infantil no Ceará [PESMIC]), and full details of the methods can be found in another reference 11 . The PESMIC is a population-based cross-sectional study on the health and nutritional status of preschool children up to 72 months of age living in the state of Ceará, in Northeastern Brazil. Ceará is one of the poorest states in Brazil, with a population of approximately 6 million inhabitants living in a semiarid climate, representing about 65.0% of state's total population 12 . Fortaleza (2.3 million inhabitants) is the capital and urban commercial center of Ceará. In the state's rural areas, subsistence farming is dominant.
PESMIC surveys were conducted in 1987,1990,1994,2001,2007, and 2017 using the same methodology. For this analysis, we used data from the sixth 2017 survey carried out from August to November 2017. The study's total sample consisted of 3,200 households with children up to 72 months old. A multistage sampling approach was used to select these households with state representativeness. Firstly, there was a stratification among Fortaleza and the 28 municipalities of the interior, with strata samples proportional to their population size. Thus, the capital city sample comprised 30.0% of the study's total sample size (960 households). Secondly, systematic sampling was used to select the countryside municipalities, proportionally to their population size. Finally, a cluster sampling approach was used to select the households to be visited. In this process, in each countryside municipality, four census tracts of the Brazilian Institute of Geography and Statistics (IBGE) were randomly selected, while in Fortaleza were 48. By the end, there were a total of 160 clusters of 20 households, totalizing the sample size of 3,200 households. Once a census tract was selected and its corresponding map obtained, the location of the cluster of 20 houses to be visited was determined. The starting point of the cluster (the first home to be visited) was randomly selected utilizing ArcGIS ® (GIS Inc). Households were visited consecutively in a counterclockwise direction. All children from 0 to 72 months old and women from 10 to 49 years old living in the homes were included. Absent families with children under 6 years old were revisited up to three times within a day in an attempt to obtain data. In each household, information about all children were obtained through the mother or primary caregiver report, and child anthropometric measurements were taken by trained staff.
The outcome variable was structured based on the situations of adverse experiences conceived by the Centers for Disease Control and Prevention (CDC), presented according to nine groups, as follows: 1. Emotional abuse (verbal threats); 2. Physical abuse against child (physical aggression); 3. Child emotional neglect (child who do not live with the father, unwanted and unplanned pregnancy and non-acceptance); 4. Physical neglect with the children (physical integrity put at risk); 5. Domestic violence (peer violence, feeling of unsafety at home, and physical/verbal aggression); 6. Alcohol and drug abuse in the family; 7. Presence of maternal common mental disorder (CMD) (eight or more positive answers to the Self-Reporting Questionnaire [SRQ-20]); https://doi.org/10.1590/1980-549720220035 8. Divorce or dispute over custody of the child, and 9. Family member arrested or who has died 13,14 . In Table 1, the CDC classification is related to the study variables applied. As a whole, the variable ACE was composed of 32 variables, grouped into nine categories, based on the CDC's 10 categories of ACE, with the exception of the 'sexual abuse', which was not investigated in the present study 4 . In Supplementary Table 1, there is a table with the description of the 32 variables mentioned above. The final outcome variable was constructed with three categories based on the sum of ACEs: 0 to 2, 3 to 5, and 6 or more.
The SRQ-20 was used for assessing maternal CMD. This questionnaire has been validated in several countries, including Brazil 15 . Each SRQ-20 affirmative answer scores a value of 1 to compose the final score by summing all 20 questions. The scores obtained are related to the probabil-ity of the presence of nonpsychotic disorder, ranging from 0 (no probability) to 20 (extreme probability). Cases with scores equal to or greater than eight were considered positive 16 . Intimate partner violence was assessed using the Hurt, Insulted, Threaten, Scream (HITS) questionnaire 17 . Adverse experiences were self-reported by the participants' mothers.
The independent variables were divided into two assessment groups, as follows: 1. Maternal factors i.e., education, age at childbirth, marital status, current inadequate conditions (smoking, obesity, and increased abdominal circumference) and risk behaviors during pregnancy (smoking and drinking), and 2. Socio-economic factors determined by poverty (lower social classes) through the Brazilian Association of Companies and Research (ABEP) classification, inadequate sanitary conditions (with/without flushing toilet), low ACEs: adverse childhood experiences; OR: odds ratio; 95%CI: confidence interval 95%; AOR: adjusted odds ratio. All study data were double-entered using Epi Info 2000 and tested for concordance, and analyzed using the Statistical Package for the Social Sciences (SPSS) version 23 (IBM Inc).
In the initial analysis, bivariate analysis was carried out to identify maternal and socio-economic factors associated with ACEs. After that, a multivariate analysis was performed selecting the variables significantly associated with the outcome at the p<0.05 level. The ordinal logistic regression model was used in a backwards approach, considering the groups of maternal/family and socio-economic factors, based on the CDC model for regression analysis of toxic stress 13 , as shown in Supplementary Figure 1.
Written informed consent was obtained from participating women. Written consent for children was given by mothers, and consent for adolescent minors was obtained from their parents or legal guardians. The PESMIC survey was approved by the Research Ethics Committee in Brazil under opinion number 2.255.063.

RESULTS
From the proposed sample size of 3,200 households, 3,180 were visited, and data were obtained from 3,566 children below 72 months of age and their respective mothers. In 20 (0.6%) eligible selected households there was a refusal of participation. Among the nine categories of the ACEs analyzed, the most prevalent were physical (57.2%) and emotional (55.1%) child abuse, intimate partner violence (33.8%), emotional neglect (29.3%), prison or death of a household member (28.4%) and maternal CMD (20.9%) (Supplementary Table 2).
The main maternal and household characteristics of the sample are presented in Table 1. Almost half of the mothers were in their 20's, and one quarter lived without a partner. More than half belonged to the poorer socio-economic classes D/E, one out of four families had someone who lost OR: odds ratio; 95%CI: confidence interval 95%; LCL: lower control limit; UCL: upper control limit; ACEs: adverse childhood experiences.  Table 3).
The outcome variable that ranged from 0 to 9 was summarized into three categories according to the number of child exposures to ACEs: 0-2, 3-5, and 6-9 (Supplementary Table 4). While there was no statistical difference in the level of exposure to ACEs between male and female children (p=0.284), in relation to age, children two years old or above presented a prevalence of high exposure to ACEs (6-9) three times higher than children below two years of age (p<0.001).
Children of younger mothers tended to show a lower ACE score compared to children of older mothers (Table 1). Among teenage mothers, the chance of a higher ACE score was 46% lower as compared to mothers aged 40-49 years (AOR: 0.54; 95%CI 0.5-0.7). The chance of having a higher ACE score among children who do not live with a father was 4.4 times greater than among those who have a father figure present at home (AOR: 4.37; 95%CI 3.6-5.2). Children of smoking mothers were 71% more likely to have higher ACE scores than children of non-smoking mothers (AOR: 1.71; 95%CI 1.2-2.3). Maternal education and alcoholism were not associated with ACEs in the adjusted analysis. Families with at least one member who lost his/her job had 40% more chance of their children presenting ACEs, as compared to families that did not (AOR: 1.40; 95%CI 1.1-1.6). The same is true for families whose mothers reported the lack of social support from friends and/or family members, with a 2.8 more chance of a higher score of ACEs, as compared to mothers who did not report (AOR: 2.77; CI95% 1.8-4.3). The level of education of the household's head was also an important factor associated with ACEs, with the presence of a risk gradient that increased with reduced educational attainment ( Table 2). The chance of achieving ACEs higher score was 62% more likely for those with no formal education, than those with university level (AOR: 1.62; CI95% 1.0-2.7). Moderate to severe food insecure households also showed 77% more chance of scoring higher on ACEs, as compared to food secure or mildly insecure households (AOR: 1.77; CI95% 1.4-2.1). Social classes, family income, and availability of water and sanitation were not associated with ACEs in the adjusted analysis.
In the final regression model shown in Figure 1, younger mothers were less likely to provide an environment with the presence of ACE in childhood. Meanwhile, factors like the absence of a partner, lack of social support network, mother's smoking habit, food insecurity, and unemployment increased the chance of child exposure to ACEs.

DISCUSSION
Our results show that adverse experiences with the highest prevalence were neglect and emotional/physical abuse, affecting more than half of the children, followed by events that make the home environment unwelcoming, such as intimate partner violence, the arrest or death of a family member, and maternal depression, affecting 20.0 to 30.0% of the families. The main factor associated with ACEs found in the present study were mother's older age and smoking habit, paternal absence, low education of the head of the family, food insecurity, and lack of social support network.
Recent studies carried out in the US found ACEs of a similar nature to those of the present study but with lower levels of prevalence, such as emotional abuse (33.5%), intimate partner violence (17.8%), physical abuse (17.5%), mental illness (16.2%) and incarcerated family member (8.1%) 18,19 .
The presence of an ACE is related to the prediction of other types of ACEs. Thus, the cumulative effect of ACEs acts on both behavior and cognition of the individual, resulting in lower academic performance, behavioral problems and long-term consequences on health and social aspects, increasing the chances of having more health problems by up to 22.0% 9,20 .
In the present study, about half of the children were exposed to a considerable number of ACEs (3 or more), with increased prevalence proportional to increased age (children over two years old). In agreement, other studies that followed children also observed a rise in ACEs concomitant to the aging of children 21 , resulting in older children resorting to more coping strategies with ACEs 22 .
The maternal factors associated with ACEs found in the present study were older age and smoking. Adolescent mothers were associated with a lower exposure of their children to ACEs, with a gradient of greater probability concurrent with increased maternal age. This may be on account of the fact that, considering the local cultural context, family members usually support very young mothers, often the grandmothers, both in economic aspects and in child care. Older mothers, on the other hand, often do not even have the support of the child's own father. A study in the state of Rio de Janeiro showed that in cases where the mother had her child in the limit age groups (below 20 and above 35), physical neglect and abuse were more common 7 . Additionally, another study in a hospital in the state of Amazon, in turn, observed that younger mothers tend to exercise their parenting more continuously, showing greater availability in their daily lives, perhaps due to the inexistence of a formal job 23 .
The absence of a father figure makes the child four times more likely to have high score of ACE, appearing in the present study as the factor most strongly associated with this outcome. It is worth noting that separations of couples usually involve periods of intense and prolonged stress, with emotional and financial repercussions on the family and the child itself. This finding is corroborated by a study that showed a trend towards an increase in the number of ACEs in environments where children did not live with both parents, favoring the experience of emotional neglect and physical abuse, in addition to compromising child development by reducing child protection, family income and emotional security 7 .
Although it showed a low prevalence among the investigated mothers (6%), smoking was presented as a factor associated with the child's exposure to a greater number of ACEs. Smoking has often been associated with maternal depression 24 and the study's findings may be an extension of common maternal mental disorders.
The study identified the low education level of the head of the family as a factor strongly associated with a greater number of ACEs in the child. Families headed by people with little or no education showed an increased chance of 30.0% and 60.0%, respectively, compared to those with a higher education level. It is worth noting that in this study, 26.0% of the families were headed by their mothers. The influence of social class and socio-economic environment on ACEs were also identified in a study in the USA, in which, a sample of white people with high level of education and income showed that about 42.0% of children had no contact with ACEs 18 .
Schooling as a determining factor for the occurrence of ACEs is related to parental care, which, due to poor argumentative ability, has difficulty resolving family conflicts and may even resort to violent acts 7 .
In the present study, food insecurity was also an essential associated factor with an increased number of ACEs, in children from families with moderate or severe food insecurity, presenting a 77.0% greater probability of being exposed to an excessive number of ACEs, in comparison to children from food-secure or mildly insecure households. Also, moderate/severe food insecurity was associated with lower income and social class, children under two years of age at home, and the presence of family dysfunctions. In addition, studies show the influence of food insecurity on children's health, growth and development, with lifelong consequences on cognition, schooling and economic productivity 25 .
Unemployment presented as adversity that significantly favored the child's exposure to a more significant number of ACES in our study. One in four families reported that at least one of their members lost his/her job in the previous 12 months. Other studies have also found unemployment as a driver of a high number of ACEs in the family. The instability resulting from the loss of a job places the family in a context of greater vulnerability due to the imminence of financial problems, changes in the family routine, the anxiety of the heads of the family, and high exposure to situations of social, emotional and affective instability 26,27 .
About 4% of the mothers reported that they felt a lack of support from other family members and friends. This lack of support was identified as the second factor most strongly associated with adverse experiences, increasing by almost three times the probability of the occurrence of such outcome. Social support is characterized by formal or informal interpersonal relationships that are important for the family to support the child's care. Its importance is recognized in strengthening families in the face of adverse experiences, and in preventing health and mortality problems, including mental suffering 28,29 .
This study has some limitations. First, as much as we have used validated tools, they are not diagnostic of maternal depression (SRQ-20) and intimate partner violence (HITS). Second, the cross-sectional design of the study does not allow to evaluate direct causal relation to the ACEs. Finally, the results obtained are representative of the study population, and may not be generalized in other contexts.
Through the results presented, it is possible to observe that the occurrence of ACEs arises amid the action of unfavorable socio-environmental factors, given the characteristics found of low education, inadequate health behavior (poor diet, smoking, alcoholism), low socio-economic status, insecure environment and absence of family support ranging from the absence of a partner to the lack of a social/ family support network. In this way, it is emphasized that in a context in which knowledge about the consequences of ACEs in the short, medium and long term is increasingly consolidated, knowing their associated factors is relevant for the establishment of effective preventive programs and interventional strategies. AUTHORS' CONTRIBUTIONS: MISM: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. HALR: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. AJML: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. SGMOR: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -original draft, writing -review & editing. DABSA: supervision, validation, visualization, writing -review & editing. MMTM: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. JSC: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. EGMS: supervision, validation, visualization, writing -review & editing. ACS: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing. LLC: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing -review & editing.