Maternal Depression, Behavioral Profile and School Performance in School-Age Children

The aim of this study was to compare behavioral profile and school performance of school-age children living with a mother who presents clinical history of recurrent depression, diagnosed according to CID-10 criteria in order to verify the influences of such adversity. Thirty-eight mother-child dyads were evaluated using tests, interviews and questionnaires. Approximately two-thirds of the children presented behavioral and school performance difficulties with predominance of emotional and relationship problems, and impairment in the three areas of school performance which were assessed (writing, arithmetic and reading). Such difficulties may be associated with the negative impact of maternal depression. One-third of the children did not present difficulties, which suggests the use of protective mechanisms. The study highlights the importance of considering differences in children’s profiles for the planning of mental health

The theoretical approach of Development Psychopathology stresses the conditions that favor either competence or malfunction along the different stages of human development (Linhares, Bordin, & Carvalho, 2004;Masten & Gewirtz, 2006).More recently, studies in this area have given emphasis to the conditions related to risk and resilience (Yunes, 2006), which have become more and more visible due to its applications on preventive psychology and in the proposal of public policies (Masten & Gewirtz, 2006;Sameroff & Fiese, 2005).
The concepts of resilience and risk are associated, with the last one being identified as the individual or his/her family faces stressful conditions which threaten satisfactory development or adaptation.Wright and Masten (2006) point out that an indicator for the assessment of resilience conditions is the competence to perform expected tasks for each developmental stage.
In the present study, specific developmental tasks for school-age children (6-12 years old) are highlighted, which, according to Marturano and Loureiro (2003) are related to school performance, which is expressed by children's adjustment to school activities and their ability to have relationships with their peers and adults.In addition to personal dispositions to accomplish such developmental tasks, the quality of the support offered by school, family and community must also be stressed as it is considered a condition that can influence and orientate the developmental course (Marturano & Loureiro, 2003).
Regarding family environment, parents' recurrent psychiatric disorders have been considered as adverse con-ditions for their children's full-potential development (Luoma et al., 2001;Najman et al., 2001).Maternal psychopathology must be specially stressed out, considering the relevance of maternal role in the context of child development (Elgar, Mills, Mcgrath, Waschbusch, & Brownridges, 2007).
Among several psychopathologies, depression is one of the major causes of impairment, malfunction and suffering for people's everyday lives, according to the World Health Organization (Organização Mundial da Saúde, 2009).Such effects are related to depression's duration, recurrence and its own symptomatology, characterized by sadness, apathy and low disposition to perform everyday tasks.As such disorder affects young women, its implications for the exercise of maternity influence family routine and child development.
Epidemiological data show an incidence of 2:1 when comparing depression rates along men and women, with increased incidence in populations exposed to economical, educational and social disadvantages (Hammen, Bistricky, & Ingram, 2010).
The gravity and chronicity of depressive symptomatology were highlighted by Garber and Little (1999) as possible potentiators of depression's effect over child development.Such findings are also reported by Hammen (2003) and Leiferman (2002), who stressed that the gravity of maternal depressive symptomatology is a moderating variable over school-age children's behavior.According to Walsh (2005), stressing events display a higher probability of affecting the individual's routine adversely when they are unexpected, grave or persistent, pointing out to depression's recurrence as an important variable to be studied when focusing on child development.
As indicated in previous studies, children of mothers with depression end up being exposed to other stressors, which are commonly associated to peculiarities of depressive symptomatology, such as family discord and occupational problems (Mian, Tango, Lopes, & Loureiro, 2009), a lower level of support and family cohesion (Langrock, Compas, Keller, Merchant, & Copeland, 2002), family and conjugal conflicts (Burt et al., 2005;Whiffen, Matthew, & Kallos-Lilly, 2005) and communication difficulties (Elgar, Mcgrath, Waschbusch, Stewart, & Curtis, 2004).Therefore it is clear that such stressors interfere in the quality of interaction and in the organization of family environment, exposing the child to chronic family stress (Essex et al., 2006).
The development of children who cope with maternal depression may be affected in several domains, especially in emotional manifestations, involving difficulties in social abilities and impulse control (S.H. Goodman & Gotlib, 1999) and emotional regulation (Silk, Shaw, Forbes, Tonya, & Kovacs, 2006).Mendes, Loureiro and Crippa (2008), based on a extensive literature review on the impact of maternal depression over school-age children, indicate the negative impact of such psychopatho-logy over children's school performance, intellectual level, behavior and mental health, based on data from empirical studies.
When regarding school-related difficulties, recent empirical studies presented evidences on the negative impact of maternal depression over children's school performance (Essex et al., 2006;Murray et al., 2006).Such relation can be aggravated by the negative impact of maternal depression over children's cognitive functioning, potentiating school-related difficulties (Joormann, Talbot, & Gotlib, 2007).
Based on a study with a community sample Marturano and Ferreira (2004) reported that children with low school performance commonly present associated social-emotional problems, which was also identified in a study with children exposed to maternal depression, in a clinical, convenience sample (Mian et al., 2009).
In all studies herein mentioned, the predominance of the negative impact of maternal depression over several domains of child development was observed.However, the authors stress that it is not an exclusive outcome, emphasizing the importance of studies that comparatively assess children exposed to maternal depression who do not present difficulties in specific developmental domains.The present study is included within this context.
The aim was to compare behavior profiles and school performance of children coping with recurring maternal depression in an attempt to verify the influence of such adverse condition of the family environment on typical, developmental tasks of the school period.

Method
A cross-sectional study with group comparison was performed with participants chosen from a clinical sample of women/mothers with history of recurring depression.

Ethical Aspects
The present study is part of a wider project, which was appraised and approved by the Research Ethics Committee of the Faculty of Philosophy, Sciences and Letters of Ribeirão Preto -University of São Paulo, process CEP-FFCLRP nº 267/2006.1.1179.59.1, and authorized by the mental health services in which data was collected.Mothers were invited to participate when they sought the mental health services, and were then informed of the objectives of the study, of the voluntary aspects of their participation and that there was no risk of damage or harm from their participation.Mothers who accepted to participate signed the Consent Form, the objectives of the study were also explained to the children who agreed to partici-pate.Feedback sessions were offered to the mothers and when difficulties were detected they received orientation and were referred to professional care.

Participants
Thirty-eight mother/child pairs were assessed in which mothers were initially chosen from a careful analysis of their medical records provided by the services were they received treatment in a medium-sized municipality in São Paulo State.Mothers presented clinical history of recurring depressive disorder with at least one moderate or severe episode in the past two years and their diagnostics were confirmed by the structured clinical interview for DSM-IV (Del-Ben et al., 2001), which was performed in data collection.
Women who presented with moderate or severe symptoms in the six months previous to data collection were excluded in order to prevent the depressive symptomatology from interfering in the information given by patients.Women who had a diagnostic of light depression or comorbidity with other mental disorders were also excluded to avoid the interference of these clinical conditions.
Children were identified through their mothers.Children of both sexes participated in the study, the criteria for participation included children living with their biological mothers, in school age (from seven to twelve years old) and having at least the inferior average percentile in Raven's Coloured Progressive Matrices Test.Children with history of chronic organic diseases or apparent sensory disabilities and mental disabilities were excluded in order to avoid possible confusing variables which could interfere in the child's school performance and behavior.
Behavior and school performance indicators which were systematically assessed by specific instruments were used to identify aspects of socialization and performance in development tasks expected for the age.

Procedures and Instruments
Data collection occurred during individual face-to-face sessions at the mental health service where most mothers were identified.Four other mothers preferred to be interviewed at their homes, preserving their privacy and with no external interference.
The assessment of mothers was conducted by two psychologists with clinical experience and training, the assessment of the children had the support of two scientific initiation scholarship students and a technician, who were properly trained to apply the instruments.
Mothers' Assessment.After introducing the study's objectives and collecting the signature on the Consent Form there was a first session in which the Sociodemographic Form was filled to gather information on children's sex, family composition and economic level, age and education level of parents and children.Then the Structured Clinical Interview for DSM-IV (SCID) was applied, including the general review, modules A, B, C and D of SCID I and SCID II, translated and adapted by Del-Ben et al. (2001), to confirm the recurring depression which is essential for the inclusion in this study.On the second session mothers answered the Strengths and Diffi-culties Questionnaire (SDQ) proposed by R. Goodman (1997), translated and validated for Brazil by Fleitlich, Cortázar and R. Goodman (2000).The questionnaire has 25 items which address children's behavior in the past six months, 10 items are related to strengths and 15 to difficulties.The items are distributed evenly in 5 subscales, the first four are Emotional Symptoms, Conduct Problems, Hyperactivity, Peer Problems, which are related to difficulties, and the Pro-social Behavior subscale is related to resources.Each of the 25 items can be answered as "not true", "somewhat true" or "certainly true".
Children are classified into three categories (normal, borderline or abnormal) for each of the scales and according to the total score given by the sum of the results of the first four scales.Rank values vary according to the scales.The cutoff for behavior problems suggested by the Brazilian study of Fleitlich et al. (2000) referred to a total score above 16.The instrument's version used was the version for parents, available at www.sdqinfo.org.
Children's Assessment.It was held after the assessment with their mothers in a single session, initiated by a brief rapport and followed by the application of Raven's Colored Progressive Matrices Test -Special Scale, adapted to assess the intellectual level of Brazilian children (Angelini, Alves, Custódio, Duarte, & Duarte, 1999) comprising 36 items, equally divided into three series, and the School Achievement Test, standardized by Stein (1994), which aims to assess the performance of school children, divided into three parts: Writing, Arithmetic and Reading.For the application of both tests specific technical recommendations were followed.
Specific technical recommendations were also adopted for coding the tests and questionnaire.Data from the Socio-Demographic Sheet were treated by frequency and proportion.

Data Analyses
Based on the responses of mothers to the Strengths and Difficulties Questionnaire (SDQ) on the behavior of their children, and children's performance in the School Achievement Test (SAT), two groups were organized.Group 1 (G1) was composed of children who had indicators of behavioral and/or school performance difficulties, these characterstics were identified, respectively, by the total score greater than 16 in SDQ and/or lower than expected classification in SAT for age and/or school grade, considering the highest score obtained.Group 2 (G2) was composed of children who had no indicators of behavioral difficulties in school performance, expressed by total score lower than or equal to 16 on SDQ and appropriate scores to their education and/or age in the SAT.
The cutoff point suggested by Fleitlich et al. (2000) was used for coding the SDQ, grouping "borderline" and "normal" classifications in the category without behavioral difficulties and the "abnormal" classification in the category with behavioral difficulties.
Based on the non-normal distribution (Kolmogorov-Smirnov test) of the behavioral and school performance variables, the choice was made for the application of nonparametric statistical procedures.Then, groups the data on total and specific scores and the items of the SDQ were compared using Mann-Whitney U, Chi-square and Fisher's Exact non-parametric tests, the significance level p<.05 was adopted.There was a correlation of variables by means of Spearman's Product Moment Correlation, highlighting for each group the behavioral (SDQ) and school performance (SAT) variables.

Results
Data on the distribution of participants in groups, demographic profile of the sample and the assessment of children's intelligence, which was adopted as a criterion for inclusion in the study, are presented descriptively.The assessment data on school performance and behavioral aspects will be presented in the form of tables followed by their descriptions.
Regarding the total of children assessed (N = 38), 23 children (60.5%) had indicators of behavioral and/or school performance difficulties (G1) and 15 children (39.5%) had no indicators of behavioral and school performance difficulties (G2).G1children presented predominance of exclusively behavioral difficulties (39.1%), followed by behavioral and school performance difficulties presented by the same child (34.8%) and school difficulties (26.1%) exclusively.
Concerning the demographic profile of the groups with regard to sex, age and school grade of children, there were no statistically significant differences, making them com-parable to the other variables.Regarding general aspects of the sample, there was predominance of girls (57.9%).
Regarding the education level of mothers, although the difference between groups was not statistically significant, there was a prevalence of mothers who attended between four and ten years of schooling in G1 (91.3%) compared to 66.7% of mothers with the same schooling in G2.As for the socioeconomic status of families, there was a predominance of socioeconomic class C in the sample, followed by D and B. It is noteworthy that 91% of G1families and 80% of G2 families were classified in classes C and D, identifying families of low socioeconomic status with no significant differences between groups.
Regarding family composition, there was predominance of a family structure characterized by the presence of father/stepfather, mother and brothers, for both groups (G1 = 65.2% and G2 = 73.3%).The second most frequent type of family composition was characterized by the presence of mother and brothers and absence of the father, this structure was more common in G1 (21.7%) than in G2 (13.3%).In both groups, the composition of mother with no father and brothers was the least frequent (G1=13.0%and G2= 3.3%).
Aspects concerning the intellectual level of children did not show statistically significant differences between groups.However, considering the Raven classification, most G1 children were classified at the average intellectual level (74%) and while 47% of G2 children were classified at the same level, 53% of children in this group were classified above average or higher.
Table 1 shows the frequency and percentage of behavioral indicators of the children assessed.Regarding the Prosocial Behavior subscale, it was observed that 100% of the children in both groups were classified in the category with no behavioral difficulties, showing skills expected in this area of development.On the other hand, in the Emotional Symptoms subscale, there was a high percentage of children with emotional problems in both groups, predominantly of children in G1.In the Conduct Problems subscale, while none of the children of G2 were classified as having difficulty in this area, 43.5% of G1 children were classified as having diffi-culty, showing a peculiar aspect of the children in G1 and an important differentiator in the profile of the groups.Hyperactivity and PeerRelationship subscales showed that the percentage of G1 children with difficulties in these areas was approximately two times higher than G2.Some of G2 children also had difficulties in some specific areas, especially in Emotional Symptoms and Peer Problems, while a high percentage of children from G1 presented difficulties on several subscales, simultaneously.
Table 2 shows the comparison between group behaviors.There was a statistically significant difference between groups in total score on the Strengths and Difficulties Questionnaire (p <.001) and in the following subscales: Emotional Symptoms (p = .03),Conduct Problems (p <.001) and Hyperactivity (p <.001), being the G1 score significantly higher than that of G2.Thus, there are more indicators of difficulties for G1 for the higher the score, the more difficulties.
In the analysis of items included in the subscales, there were significant differences when comparing the groups in eight items of subscales: Emotional Symptoms (two items), Conduct Problems (three items) and Hyperactivity (three items).In all comparisons, the children in G1 reported more behavioral difficulties than G2 children in the items: complaints of body pain and insecurity (Emotional Symptoms), less obedience, more fights and lies (Conduct Problems) and more agitation and difficulty concentrating and reduced performance in tasks that require more concentration (Hyperactivity).
Table 3 presents data on children's school performance.
There was a statistically significant difference between groups regarding the total score on the School Achieve-ment Test and also in all its subscales.Children in G1 when compared to G2 had more difficulties in writing (p = .008),arithmetic (p = .006),reading (p = .010)and overall performance (p <.0001).
The statistically significant correlations between scores for G1 and G2 children in the SAT and the SDQ are presented in Table 4.
In G1, there were significant negative correlations between writing performance and the Emotional Problems subscale and total SDQ score, showing the association of lower performance in writing to the presence of behavioral difficulties, particularly the emotional symptoms.In the same group there were positive correlations between the Hyperactivity and Conduct Problems subscales and negative correlations between Hyperactivity and Prosocial subscales, which shows, for G1, the association of two areas of difficulty to the presence of minor socialization resources.In G2 there was a significant positive correlation between the subscales Emotional Symptoms and Peer Problems, characterizing the association of less emotional difficulties to less relationship problems.

Discussion
Approximately two thirds of the school-age children studied who coped with maternal depression were identified with difficulties in achieving the development tasks characteristic of their age, highlighting the possible adverse impact of maternal psychopathology to the behavior and school performance of children.
The presence of different behavioral outcomes, despite the coexistence of both groups with maternal depression, highlights the possible association of maternal psychopathology to other variables of the family environment which may promote or hinder the developmental tasks of children.As for family environment, the socioeconomic status of families in both groups were similar.However, mothers of children in G1 had fewer years of schooling and more families in this group did not have a paternal presence in everyday life, although no statistically significant differences were detected between groups in these two variables.
Socioeconomic level and schooling can be considered variables that affect parental involvement in the school life of children (Marturano & Ferreira, 2004).In the context of living with a depressive disorder, the absence of a father in the family life and low education can favor the lack of support and assistance to children, as highlighted by S. H. Goodman and Gotlib (1999).Such data can be connected to the peculiarities of the studied maternal psychopathology which, for its recurrence, facilitates losses in the quality of family interactions and organization (Essex et al., 2006).
Regarding the intellectual level of children, which was used as a criterion for inclusion in the study, there was no statistically significant difference between groups that  included children without cognitive disabilities.It is noteworthy, however, that a greater number of children in G2, when compared to G1, performed better in the intellectual level (classified as above average or higher).Good cognitive ability was reported by Masten and Gewirtz (2006) as a personal resource of children in coping with maternal depression.Based on a longitudinal study, these authors found that the intelligence and mental health of children were more positive for low-risk groups, thus, considering the lack of accumulated adversities.The family atmosphere indicators already highlighted may, therefore, indicate that the children from G1 might be exposed to different stressors other than maternal depression.
Regarding the distribution of children in groups, when compared to G2 the children of G1presented as predominant profile the presence of: poor school performance and more behavioral difficulties.This profile indicates the presence of accumulation of difficulties for the children of G1, which is consistent with the studies of Weissman and Feder (2004) and Najman et al. (2001) for children coping with maternal depression.
Analyzing the SDQ items, the main difficulties of the children in G1 were expressed by: being less considerate, offering less help, being less obedient, quarreling and lying more, being more agitated and having more difficulty concentrating, even in tasks that require more attention.Such indicators characterize the presence of externalizing behavioral problems expressed by greater agitation and defiant behavior.On the other hand, as indicators of internalizing behavior problems, there was the presence of a high percentage of children with difficulties related to emotional symptoms in both groups as the children of G1, when compared to G2, were facing more problems related to somatization, for the presence of body pain and various complaints.The children in G1 were more insecure when facing new tasks and fought more with their peers which indicated less effective strategies for emotional regulation, as reported in a previous study by Silk et al. (2006) with a similar sample.
Regarding school performance, Elias and Marturano (2004) pointed out that problems in this area can promote personal and social adjustment difficulties, even in later stages of development.Under this perspective the low school performance of children in G1 may increase the chance of future problems.
Two thirds of the G1 children show multiple difficulties regarding development tasks typical of their school grade, with predominance of behavioral problems which can be exclusive or associated with poor school performance to 73.9% of children.Such data highlights the importance of observing this aspect of development for school-age children living with maternal depression as an element that can better guide interventions.
In contrast, approximately one third of children had no indicators of behavioral and school performance problems (G2) pointing to an adaptive pattern, suggesting the presence of resilience resources in the face of adversity, which can be associated with environmental variables, such as better maternal education and the presence of the father in family life.In this study, these aspects were not addressed in a systematic way, suggesting the inclusion of these variables in future studies.
The presence of positive correlation between low scores of emotional problems and relationship problems with peers also suggests that healthy relationships established by these children may be protecting their development, as described by Eckenrode and Gore (1996), which highlighted social relationships as support resources offered by the community to child development.Thus, it can be hypothesized that children with no indicators of behavioral problems and school performance (G2) may be benefiting from the action of protective mechanisms, possibly related to the contextual aspects of the family environment and/or social network.
Characterizing and comparing the socialization profile of a sample of children living with recurrent maternal depression has contributed to identificaty the presence of the negative impact of maternal depression for approximately two thirds of the children, which suggests the need of psychosocial support in when such adverse conditions are present.On the other hand, the identification of children with no indicators of behavioral and school performance difficulties, who were exposed to the same condition, pointed to the positive action of protective mechanisms favoring resilience resources.
Concerning the limitations of this study it is relevant to highlight the adoption of a cross-sectional design, with a small, convenience sample which had only the mothers as informants of children's behavior.Further studies which assess the child's personal variables, contextual variables on the organization of family environment and social support with different design may expand the understanding of conditions which may or may not favor the adaptation of children coping with recurring maternal depression.
It was found that exposure to recurring maternal depression is an adverse condition.Considering that the present study was developed with a clinical sample treated in outpatient mental health services, associated to the National Health System, it is important to highlight these services contributions to scientific knowledge and clinical practice.The identification of the differential impact of maternal depression on children's behavior and school performance draws attention to the need for specific mental health planning practices for families exposed to such adversity.

Table 1
Frequency and Percentages of Children's Classification in SDQ subscales and Total Scores, According to Presence (G1) or Absence (G2) of Behavior and School Performance Difficulties Note.G1= children group with indicators of behavior and/or school performance difficulties; G2= children group without indicators of behavior and school performance difficulties; N= frequency; %= percentage.

Table 2
Means and Standard-deviations of Children's Scores on the SDQ and its Subscales According to Presence (G1) or Absence (G2) of Behavior and School Performance Difficulties G1= children group with indicators of behavior and/or school performance difficulties; G2= children group without indicators of behavior and school performance difficulties; SD= standard-deviation; U= Mann-Whitney non-parametric test; p= significance level; >= higher difficulty; NS= non-significant difference.

Table 4
Significant Correlations Regarding Associations bet-ween School Performance and Behavior Capacities and Difficulties for Children According to Presence (G1) or Absence (G2) of Behavior and School Performance Difficulties

Table 3
Frequency and Percentages of Children's Scores on the School Achievement Test (TDE) and its Subscales According to Presence (G1) or Absence (G2) of Behavior and School Performance Difficulties (N=38) G1= children group with indicators of behavior and/or school performance difficulties; G2= children group without indicators of behavior and school performance difficulties; N= frequency; %= percentage; χ²= Chi-square test; p= significance level; Dif.= difficulty; No Dif.=no difficulty.