Method

Thirty-six children and adolescents aged from seven to 13 years participated in this research. This age group was chosen for its higher probability of ALL development and diagnosis. These participants were subdivided into two groups: G1 and G2. G1 is composed by 18 children and adolescents who: survived ALL, underwent chemotherapy exclusively (systemic and intrathecal), are students in Rio Grande do Norte state (in either public or private schools), and have a mean age of 9,7 years. Out of the 18 children and adolescents, seven are female and eleven are male. G2 is composed by 18 healthy children and adolescents, matched to G1 by sex, age, and school system.

Instruments

Child Behavior Checklist (CBCL): The CBCL questionnaire has been designed for parents in order to investigate behavioral, emotional, and adaptive functioning problems in children aged 6 to 18 years. It consists of 138 items, divided into two scales: Social Competence Scale (20 items) and Behavioral Problems Scale (118 items). The latter is subdivided into two scales: Internalizing Behavior Problems Scale (subdivided in anxiety/depression, withdrawal/depression, somatic complaints) and Externalizing Behavior Problems Scale (subdivided in rule-breaking behavior and aggressive behavior) (^{Achenbach, 1991}).

Deasy-Spinetta Behavioral Questionnaire (DSBQ): This questionnaire is intended for teachers and aims at investigating the process of school reintegration of children and adolescents surviving ALL. It consists of 38 dichotomous questions (affirmative or negative) and investigates three major domains: learning, socialization, and emotional aspects (^{Gomes et al., 2017}).

Data Collection Procedure

The selection of children and adolescents who survived ALL was carried out by consulting the records of the pediatric sectors of two public healthcare facilities that are local references in the treatment of childhood cancer. Initially, 63 individuals who met the criteria were identified, but this number was reduced to 18 due to factors such as death, hospital records not being kept up to date, patients who live far from the capital (location of hospitals) only sporadically visiting it, presence of other diseases, patients or family members not interested in participating in the research, and recurrences during evaluation period.

After the parents signed the Informed Consent Form, they were asked to individually answer the CBCL in approximately 30 minutes. Later, in school visits, teachers were also asked for consent, plus DSBQ completion. The teachers were chosen based on the length of time they had been in contact with the patients, at least one semester being required. It is noteworthy that for the control group only the CBCL was used, since the DSBQ was developed exclusively for teachers of ALL survivors, and therefore consists of specific questions that are not applicable to the typical learning process.

Data Analysis Procedure

CBCL data were computed using the Assessment Data Manager (ADM) software. In the analysis of the results, there was a separation of behavior frequency into three nominal classification ranges: Borderline, Clinical and Normal. However, given that the objective of the present study is the detection of behavioral differences when compared to normative data and to the control group (not the diagnosis of clinical conditions), the Borderline and Clinical categories were grouped, which leaves two rating ranges: Normal and Altered. This methodological approach is supported by the guidelines contained in the questionnaire manual, which states that, since the objective is dichotomizing the performance between normal and altered, the Borderline and Clinical classifications can be grouped (^{Achenbach, 1991}). The results from the DSBQ were obtained from the answers attributed to each of the 38 questions, as well as from the grouping of these into the three domains that make up the instrument: learning, socialization and emotional aspects.

Initially, descriptive statistical analysis provided the CBCL and DSBQ results frequencies, which were classified as normal or altered. Then, in order to compare the results obtained by the two groups at the CBCL, Pearson’s Chi-square test was performed. This is a statistical test applied to categorical data which adopts the value p <0.05 as the rejection parameter of the null hypothesis. In addition, Pearson's chi-square test was also applied to assess whether there was a statistically significant difference between the answers provided by parents and teachers for two of the domains assessed by the instruments: Learning and Behavior.

Results

Child Behavior Checklist (CBCL)

Among the four CBCL scales that measure Social Competence, the highest percentage of normal classification for both groups was identified in the Activity Competence Scale, Social Interaction Scale and School Activity Scale. However, in the Total Competence scale, a higher percentage of changes in G1 was identified, as described inTable 1.

In the Behavioral Problems Scale, G1 had a high prevalence of internalizing problems, as well as G2, making it relevant to emphasize the percentage of changes presented by G2 participants, given that there is a higher frequency of changes regarding these problems, which will be discussed later. However, a pattern contrary to this was identified in the Externalizing Problems Scale, with higher percentages of normality being identified in both groups, although there is a difference between them, with G1 showing a higher percentage of change when compared to G2. In the Total Problems Scale, the performances of the groups differ, with G1 presenting a higher frequency of altered results, as shown in Table 2.

As for the distribution of behavioral syndromes, data is shown in Table 3.

Deasy-Spinetta Behavioral Questionnaire (DSBQ)

As for the three DSBQ domains aimed at assessing school reintegration of ALL survivors, it was found that the emotional dimension is the one in which the highest percentages of changes are concentrated. Data regarding frequency and percentages obtained in the three domains can be found inTable 4.

Comparison between groups - CBCL

Regarding the comparison of both groups’ CBCL data, statistically significant differences were found in the Behavioral Problems and Social Competence scales. The ALL survivors group (G1) presented greater difficulties in specific domains, as shown inTable 5.

Comparison between CBCL and DSBQ answers of parents and teachers of ALL students

Research into possible differences in parent and teacher positioning regarding learning and behavioral dimensions indicates that there is a high level of agreement on behavioral aspects. This agreement was verified by cross-checking the CBCL Behavior Problems Scale data and the DSBQ Emotional Aspects Domain data - both with content associated with behavioral aspects - and concluding that there is no statistically significant difference in any of the variables that make up these domains.

However, analysis of the intersection of the results from the School Activity dimension of the CBCL Social Competence Scale and the DSBQ Learning Domain shows different positions when it comes to the learning process of the surviving children and adolescents. Such differences regard the presence of school difficulties ((² = 5.513; p = 0.01), difficulties in memorizing or organizing tasks ((² = 7.200; p = 0.007) and reading difficulties ((² = 5.513; p = 0.01). For all the dimensions mentioned, the evaluations performed by teachers show more altered scores that those carried out by parents.

Discussion

Behavioral and learning profile of ALL surviving children and adolescents from a parent perspective

Perceptions of their children's behavior by parents of ALL surviving children and adolescents and parents of healthy individuals were compared. The first dimension analyzed refers to the Social Competence Scale, which identified a statistically significant difference between the groups in all its aspects, in favor of the population of healthy children and adolescents. In general, this scale is associated with social interactions, requiring the mobilization of social cognition skills and emotional regulation, also considering school activities (^{Achenbach, 1991}).

Understanding the commitments reported by parents involves the recognition of childhood and, especially, adolescence as stages in which social life is being experienced. However, this experience usually occurs far from hospital environments, where patients are immersed in a rigid and extensive routine of medical procedures, showing greater propensity to reduce interpersonal relationships, which causes difficulties in creating bonds of friendship (^{Jeon & Kim, 2014}).

In this domain, the findings of the present study are consistent with the research conducted by ^{An and Lee (2019}), which highlights that the majority of ALL survivors present difficulties in school and friendship in and out of school. For the authors, these difficulties are associated with different aspects, such as changes in physical appearance resulting from treatment and concerns about possible return of the disease.

In this sense, school reintegration is characterized by living with friends that the survivors did not have consistent interactions because they were away from school for long periods during treatment. Thus, they may be depressed for reasons that involve difficulties in friendships and frequent absences from school due to long periods of hospitalization (^{You, 2006}).

Changes in appearance of leukemia survivors, such as very short hair, can affect self-esteem and, consequently, psychosocial relationships, which may lead to anxiety, depression, despair, social isolation, and other diseases (^{Mertens et al., 2014}). The study by ^{An and Lee (2019}) points out that ALL surviving children and adolescents felt that their peers' attention was focused on their short hair and other physical signs of illness and treatment, which made them experience heartache and negative feelings.

In addition, there was a fear that teachers' attention could be assessed as a special treatment when they wanted to be recognized as normal, without depending on different treatment. The identification of this possible disparity between reality and the perception of survivors may underlie the psychosocial problems indicated and perceived by the parents who participated in this study.

The second domain analyzed refers to the results of this study that came from the Behavior Problems Scale, notably the dimensions of externalizing behaviors, which evidenced significantly higher occurrence in ALL surviving children and adolescents compared to the control group, with emphasis on the “rule violation behavior” dimension. Worldwide, the prevalence of mental health problems in children and adolescents ranges from 10% to 20%, with behavioral and emotional problems being the most important causes of these changes (^{Stewart-Brown, 2003}). Similarly, with regard to the Brazilian reality, authors warn of the high prevalence of children with behavioral disorders (^{Anselmi et al., 2004}).

The term “Behavior Problems” is controversial as to its definition, although it is generally defined by behaviors or actions considered socially inadequate, representing deficits or behavioral surpluses that undermine the child's interaction with peers. According to the ^{APA (2014)}, behavioral problems are disruptive and contain the following disorders: conduct, challenging opponents, and attention.

^{Borsa et al. (2011}) discuss that studies have shown divergences regarding the prevalence of externalizing and internalizing behavior problems in the juvenile population. While some studies suggest that externalizing behaviors are prevalent over internalizing behaviors (^{Anselmi et al., 2004}), other suggest that internalizing behaviors are predominant (^{Keegstra et al., 2010}). Externalizing problems involve impulsivity, physical or verbal aggression, agitation and teasing. Internalizes can be observed when there is excessive worry, withdrawal, sadness, shyness, insecurity and fears. These are often manifested in disorders such as depression, social isolation and anxiety (^{Achenbach, 1991}).

With regard to specific variables and their relationship with behavioral changes, studies indicate that, in general, boys are the ones with the most behavioral problems, especially externalizing ones. In turn, girls with fewer problems in general and complaints refer more intensively to internalizing behaviors (^{Anselmi et al., 2004}). The gender variable needs to be considered in the cultural context, especially those in which the stereotype is identified that boys are more active, aggressive, independent when compared to girls. In turn, girls are more passive, less aggressive and have better social interactions (^{Borsa et al., 2011}).

In the present study, the results regarding the behavioral profiles of children and adolescents surviving ALL, from the perspective of parents, reveal a high prevalence of internalizing behavior problems. These results corroborate previous studies identifying significant symptoms of anxiety and depression in these groups (^{Sharkey et al., 2019}).

^{Liu et al. (2018}) suggest that the experience of hospitalization and painful procedures resulting from cancer treatment may indeed contribute to the development of emotional and behavioral problems. Thus, these difficulties are not considered to be due to the chemotherapy treatment itself, since the treatment of leukemia is long lasting, interfering with the school, family and social routine of children and adolescents. However, it is important to highlight that studies suggest that the immature brain is in the process of development, being strongly susceptible to the impact of stress, causing anatomical and functional changes directly associated with the emergence of behavioral problems, notably depression and anxiety (^{Kosir et al., 2019}).

Finally, despite the absence of significant difference between the prevalence of internalizing problems in the two groups investigated in the present study, similar to what was identified in the above studies regarding cancer survivors, one aspect that can be highlighted is the equally high prevalence of internalizing problems in the control group, since it reaches a level above 25%, while in the other dimensions investigated the average percentage of changes is 5%. Such results are compatible with several recent Brazilian studies (^{Borges & Pacheco, 2018}). ^{Thiengo et al. (2014}) conducted a literature review about the prevalence of depression among young people in the country, identifying a variation of 0.6 to 30% of occurrence of depressive disorder. In this review, more than half of the studies indicated prevalence in the range between 5.9% and 12.5%.

Regarding the Behavioral Syndromes Scale, notably the dimensions of Social Contact Problems, Thinking Problems and Rule Violation, the first two dimensions contemplated above seem to reflect difficulties arising from changes in psychosocial relationships, as well as symptoms of depression and anxiety. However, the dimension of difficulties in meeting social norms requires further discussion.

The study by ^{Liu et al. (2018}) demonstrates the presence of a persistent pattern of anger and irritable mood among ALL survivors, as well as argumentative and challenging behavior toward authority figures such as parents and teachers. Such profile may negatively affect the social, academic and / or educational functioning of this group and seems to be associated with identified executive dysfunctions as well as conflicts between parents and children.

Corroborating the above results regarding the executive dimension, the study by ^{Fellah et al. (2019}) compared the performance of ALL survivors in relation to population norms, demonstrating the presence of significant changes in attention and executive functions in the first group. Surviving individuals underwent functional magnetic resonance imaging during the performance of the tasks, revealing hypoactivation in the parietal, temporal and right hippocampal cortices, especially in patients diagnosed earlier. Activation was also reduced in the frontal, bilateral temporal and right parietal lobes, due to the greater exposure to methotrexate, which is the basis of antileukemic chemotherapy treatment. These findings provide evidence of brain alterations, which may aid in understanding the neural substrates underlying attentional deficits and the opposing challenging pattern commonly reported among ALL survivors.

Behavioral and Learning Profile of ALL surviving children and adolescents from the teachers' perspective

Regarding the perception of the teachers participating in the present study, three specific domains of change within the clinical group were predominantly mentioned: learning, emotional and social. As to the first domain of learning, it was found that teachers highlighted difficulties in memorization, reading, mathematics and reasoning activities. These results are consistent with those pointed by the study by Plas et al. (2017), which investigated the executive functioning, intelligence, motor skills and mathematical skills of 130 ALL survivors and 158 control subjects aged 8 to 18 years. The survivors group had significant deficits compared to healthy subjects in the domains of working memory, processing speed, addition and subtraction, and numerical operations.

Difficulties in mathematics, commonly identified among ALL survivors, are directly associated with problems in the field of visual attention and changes in visomotor skills resulting from methotrexate administration (^{Richard et al., 2018}). Still in this direction, the study by Gordon (2016) with 16 surviving individuals of ALL and their twin brothers reveals that the clinical group performed worse in the comprehension tasks of reading and reading, as well as in the activities of logical mathematical reasoning.

In turn, in the present study, the difficulties pointed out by teachers in the emotional and social domains were previously discussed. In the emotional dimension, the difficulties highlighted by the surviving children and adolescents to trust others, the presence of shyness and the need for approval were highlighted. Such characteristics can be understood in the context of psychosocial difficulties, notably in the fear of lack of approval and of being considered as requiring special attention. These aspects are directly related to the weaknesses highlighted by teachers in the social domain, namely, the lack of sharing about their activities and the low initiative in seeking colleagues to perform different activities (^{An & Lee, 2019}).

Approximations and distances between the conceptions of parents and teachers

No significant differences were identified in the perceptions of parents and teachers in the domain of behavior. Both groups refer to difficulties in the domain of interpersonal relationships, with the presence of behaviors that denote insecurity, as well as signs of depression and anxiety.

However, this study identified a significant discrepancy between the perceptions of parents and teachers in the learning domain. For teachers, more than half of survivors have difficulties at school, especially when memory, attention, and the specific fields of math and reading are required. However, for parents, only 33% reveal weaknesses in school learning processes.

In order to understand the above data obtained in this investigation, we highlight the results of the study by ^{Sharkey et al. (2019}) in which the associations between the executive functions of children surviving ALL and their parents' perception of their children's executive functioning were examined. The results suggest that parents respond to the difficulties presented by their overprotected children, which in their view may be adaptive behavior, but it does not contribute to the development of the child's independence.

In this sense, the study by ^{Huang et al. (2018}) investigated the influence of variables related to family profile and functioning of 213 ALL survivors, concluding that family and parental factors contribute to the health of this group. Among the relationships pointed out by the research, one highlights that indicates that mothers with low education tend to have overprotective behavior and their children more symptoms of behavioral changes. Thus, when it is considered that the families that participated in this study are totally from the most disadvantaged socioeconomic levels, one can explain the discrepancy between parents and teachers regarding the presence of school difficulties in the group investigated.

It is important to conclude that learning disabilities in this population may also be associated with the fact that treatment results in prolonged absence from school or school absenteeism. The child survivor of ALL finds several barriers to the realization of their experience in the school context, among which stand out: the restrictions of the antineoplastic treatment itself and its side effects; the unpreparedness of the school staff to welcome the student with cancer; families' fear of keeping the child in school; and the difficulty in communication between the health team and school professionals (^{Gomes et al., 2017}).

In view of this, it is concluded that the diagnosis of cancer, especially acute lymphoid leukemia, associated with long-term treatment and the repercussions they bring to the lives of children and adolescents, is often a traumatic experience. Advances in medicine have elevated cure rates, but the presence of psychosocial difficulties experienced before, during and after treatment is becoming increasingly evident.

The results of this study highlighted how cancer presents itself as a disruptive factor in the developmental trajectory of this group, since it interferes with social activities and imposes new demands on survivors and their families, besides experiencing symptoms of anxiety, depression, school difficulties and somatic complaints. The etiology of these changes is multifactorial, involving cultural and historical aspects that are co-authors of social stereotypes, patterns of family relationships, which in turn are risk factors or protective factors.

Limitations of this study are, in particular, the low sample number. In this sense, there is the need to conduct research to increase the sample size, making it possible to establish prevalence rates of behavioral, emotional and learning disorders among ALL survivors, comparing them with those identified in the population of children and healthy adolescents. Additionally, future studies may contribute significantly to the discussion by investigating possible protective factors such as sports and family support and risk factors. Certainly, the broader understanding of the behavioral, emotional and reintegration difficulties experienced by ALL survivors may support the development and implementation of therapeutic programs that contribute to minimizing the impacts of the disease process.

Acknowledgments

To collaborators from LAPEN for aiding in collecting, interpreting and analyzing data. To the teams of Hospital Infantil Varela Santiago and Liga Norte-Riograndense Contra o Câncer. Authors also want to thank the reviewers and the editor for helpful comments on how to improve the manuscript. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001.

All authors share the responsibility for the final version of the work submitted and published. The first author, Mrs. Raquel Campos, confirm that she had full access to all the data in the study and takes responsibility for the integrity of the data in the study and the accuracy of the data analysis. The authors have declared that they have no competing or potential conflicts of interest.

Authors contribution

Raquel Campos: Concetualização, Investigação, Metodologia, Redação do rascunho original.

Danielle Garcia: Curadoria dos dados.

Débora Leite: Curadoria dos dados.

Ediana Gomes: Curadoria dos dados.

Hedyanne Pereira: Redação - revisão e edição.

Laura Aragão: Redação - revisão e edição.

Mariane Dantas: Curadoria dos dados.

Izabel Hazin: Concetualização, Análise formal, Supervisão, Redação do rascunho original.