Association of Attachment and Reflective Function with Baseline Symptoms in Child and Adolescent Psychodynamic Psychotherapy

ABSTRACT

The present study aimed to verify the associations between symptoms reported at baseline, attachment style and reflective function (RF) in children and adolescents. For this, we conducted a cross-sectional and naturalistic study, including 90 children and adolescents aged between 9 and 17 years old (M = 13.04, SD = 2.72). Instruments were a demographic form, the Child Behavior Checklist, the Friends and Family Interview and the Reflective Function Questionnaire for Youths. From our findings, internalizing symptoms were reported in 74.4% of the cases, and externalizing symptoms in 55.6%. Concerning the attachment styles, 46.7% of the cases were classified as insecure-dismissing, 38.9% as insecure-preoccupied, 10% as secure and 4.4% as disorganized. Participants’ scores for RF were low. We found associations between attachment styles and anxiety, depression and withdrawal symptoms. We found significant differences between the insecure-dismissing style and the insecure-preoccupied and disorganized styles groups regarding anxiety and depression symptoms. The secure attachment style group showed significant differences in withdrawal symptoms when compared to insecure attachment style groups. Further studies exploring associations between attachment styles, RF and psychopathology in childhood and adolescence, could contribute to the evaluation and planning of psychotherapies processes with this population.

Introduction

Children and adolescents’ relationship patterns, or their attachment style (Bowlby, 1988), as well as their ability to identify and understand their own and others’ emotional states, which came to be called mentalizing or reflective functioning (Fonagy et al., 2002) impact on their mental health and psychopathological symptoms. In addition, they also influence how a child or adolescent behaves in psychotherapy, thus affecting their therapy process. Therefore, it is important that these constructs are taken into consideration at the beginning of psychological interventions.

Previous studies have pointed to different connections between attachment and psychotherapy. Firstly, the attachment theory has provided important guidance for some therapeutic practices, such as fostering a healthy therapy relationship and exploring how the patient relates to others (Innerhofer, 2013; Malberg & Mayes, 2013). In addition, attachment can be seen as a key factor embedded in the therapy process, with the way the patient interacts with the therapist providing important clues about their functioning (Levy et al., 2012; Mikulincer et al., 2013). Furthermore, it has been well established in the literature that all psychotherapies involve, in addition to exploring relationships with significant others, a focus on understanding the mental states of the self and others (Adshead & Fonagy, 2012; Fonagy & Adshead, 2012; Goodman et al., 2016).

Bowlby (1988) claimed that the attachment experienced during childhood would underlie one’s subsequent capacity to establish affective connections, as well as a wide range of dysfunctions during adulthood, including marital problems and problems with children, neurotic symptoms, and personality disorders. Therefore, the initial attachment experiences have long-lasting effects that tend to persist throughout life, being among the main determinants of one’s personality formation.

Numerous studies have been discussing the attachment styles’ potential influence on the development of internalizing and externalizing symptoms and behaviors during childhood and adolescence (Colonnesi et al., 2011; Fearon et al., 2010; Groh et al., 2012; Madigan et al., 2013; Spruit et al., 2020). All of them support Bowlby’s theory (Bowlby, 1988), indicating that attachment insecurity is associated with higher levels of internalizing (Colonnesi et al., 2011; Groh et al., 2012; Madigan et al., 2016; Spruit et al., 2020) and externalizing problems (Fearon et al., 2010; Hoeve et al., 2012; Madigan et al., 2016).

In the following paragraphs, we briefly address the constructs of attachment and mentalization (or reflexive function) and then synthesize empirical evidence that shows the relationship of these constructs with psychopathological symptoms in childhood and adolescence.

Attachment and reflective function

Attachments are fundamental in human development, as they establish the foundations of how individuals experience social and emotional relationships throughout their lives. During early childhood, the first experiences with caregivers lead to the development of mental representations about the self, the world, and others. From these interactions, children develop internal working models of significant relationships (Bowlby, 1988).

Internal working models are the representations or expectations that guide one’s own behavior. They serve as a basis for predicting and interpreting behaviors of whom someone is attached. These models are related to the feeling of attachment figures’ overall availability, the likelihood of receiving emotional support from them in stressful moments, and the general interaction dynamics with those figures (Bowlby, 1988).

If the caregiver is consistently sensible to the child’s proximity seeking behavior, they will be perceived as a secure base and a safe haven from which the environment can be explored (Bowlby, 1988). When the caregivers are inconsistently sensible, display a lack of sensitivity or even scare the child, the latter is at risk of developing insecure attachment behaviors (Ainsworth et al., 1978; Main & Solomon, 1990).

From these dynamics, the child’s attachment styles with their caregiver are formed, which can be classified as secure, insecure-avoidant, insecure-ambivalent and disorganized. Children with secure attachment present a good balance between their need of exploring the environment and their attachment behaviors, such as seeking proximity with the caregiver (Ainsworth et al., 1978). Children with insecure-avoidant attachment style devalue the attachment figures since they experienced rejection and low sensibility from their caregivers. Children with insecure-ambivalent attachment experience inconsistent relationships with their caregivers and, thus, maximize their attachment behaviors, to keep the caregiver’s attention (Ainsworth et al., 1978). Lastly, the disorganized/disoriented attachment style (Main & Solomon, 1990) is characterized by the child perceiving their caregiver as a source of comfort and fear. This fear hinders the development of an organized strategy for reaching out to the attachment figure when in fear and results in a combination of different types of insecure behaviors, combined with the caregiver’s fearful reactions.

Advances in attachment theory have made possible the development of the concepts of mentalizing and reflective functioning. According to these perspectives, the main attachment figure needs to have the capacity to perceive the child as someone who has a separate mind, with their own mental states, such as feelings, beliefs, and desires. As a consequence of those attachment figure’s capacity, the child will also be able to recognize their own mental states (Fonagy et al., 2002). The mentalizing capacity thus developed allows for the prediction of others’ behavior, favoring affective regulation, making it possible to distinguish between perceived and real reality, strengthening attachment security. Therefore, the process related to the making of internal working models enables a ‘reflective functioning” – the operationalisation of the child’s mentalizing capacity- determinant for the organization of the self.

The acquisition of the capacity to mentalize is developed in the context of secure attachment relationships, and it provides the child with the possibility to attain the regulation and control of their own emotions, developing internal security, self-esteem, and autonomy (Fonagy et al., 2002). Secure attachments work as a protective factor for psychopathology, while insecure attachment relationships are considered a risk factor (Colonnesi et al., 2011; Groh et al., 2012; Madigan et al., 2013).

Insecure attachment as a risk factor for psychopathological symptoms

The theoretical assumption that the foundation for the processing of internal experiences is influenced by the children’s attachment to their main caregivers has been one of the pillars in the study of psychopathology and child development in the last half-century (Madigan et al., 2016). When an attachment figure cannot perceive the child’s evident behaviors as a sign of what is going through their emotional world, they are, consequently, not able to reflect this understanding to the child or will reflect it in an insensitive manner. This makes it difficult for the child to recognize and regulate their own affective world, and, as a result, they may present behavioral problems (Ensink et al., 2017).

Behavioral problems characterized by internalization and externalization are related to different childhood experiences. The internalizing problems are related to lack of boundaries, overprotection, insensibility, and lack of parental support in moments they are needed by the child (Shamir-Essakow et al., 2004). These problems are also related to the child having an unfavorable view of themselves (Goodman et al., 2012), as well as negative expectations about themselves, others, and the world in general (Warren et al., 2000). On the other hand, externalizing problems derive from severe parental punishments, hostility, unavailability, lack of communication with the mother or emotional distortion of the experience, maternal intrusiveness, negativity, role reversal, and disorganized parental behaviors (Şentürk, 2018).

Studies that have explored the attachment styles’ potential influence on the development of internalizing and externalizing symptoms and behaviors during childhood and adolescence discussed what might be the most prominent risk factors – through effect sizes-, and what are the variables that moderate – or not – these results. Some studies also try to demonstrate the association between attachment styles and their role in symptom development.

The studies that addressed internalizing behaviors have found associations with attachment styles ranging from weak (Groh et al., 2012) to moderate (Madigan et al., 2013). However, when examined separately, the association between attachment and specific anxiety and depression symptoms seem to be stronger than total internalizing problems (Colonnesi et al., 2011; Madigan et al., 2016; Spruit et al., 2020). Taken altogether, these findings suggest that having an insecure attachment style can be considered a risk factor for depression (Spruit et al., 2020).

Concerning the association between attachment and externalizing behaviors, only the disorganized attachment style was associated with externalizing problems among children and adolescents in previous meta-analyses (Fearon et al., 2010; Fonagy et al., 2016). Within this same domain, a lack of attachment to the parents also seems to be associated with delinquency in boys and girls (Hoeve et al., 2012).

Despite the important contributions of these studies in highlighting possible relationships between attachment and symptoms, none of these meta-analyses included measures associated with attachment – such as reflective function (RF)- among children and adolescents. In that sense, most research has focused on the parents’ mentalizing capacity associated with the child’s (Şentürk, 2018), rather than the specific study on the RF of children and adolescents.

The few studies that examined RF scores and symptoms in children and adolescents have identified tentative trends and risk factors. RF seems to be lower in clinical samples in comparison to non-clinical samples (Both et al., 2019; Breinholst et al., 2018), indicating that these patients present themselves with lower RF or their symptoms hinder their reflective functioning. Furthermore, RF seems to be negatively correlated with BPD features (Ha et al., 2013) and general internalizing and externalizing symptoms in adolescents (Both et al., 2019).

The studies developed thus far indicate that there are significant associations between attachment and RF with internalizing and externalizing problems. Although there is some research examining those variables with broader symptom domains, less attention has been given to specific symptoms and their relation to attachment styles and RF. Furthermore, to the best of our knowledge, no studies examined if insecure attachment, as assessed by an interview focusing on the representational dimension of attachment and RF, is associated with specific externalizing symptoms, such as oppositional-defiant behavior and conduct problems. It is also not clear the influence of attachment relationships with the mother and the father on the development of internalizing and externalizing behaviors in general, and what attachment styles are associated with internalizing and externalizing problems (Madigan et al., 2016).

Considering this background, this study aimed to examine the association between specific symptoms reported at baseline with the attachment styles and RF of children and adolescents that were referred to psychodynamic psychotherapy in a community clinic in southern Brazil. It also aimed to describe this population’s attachment patterns and RF, to verify if specific attachment styles were associated with specific symptoms, and to analyze if the patients’ representations of “safe haven” and “secure base” concerning their mothers and fathers would be correlated to their symptoms. Examining the associations between attachment and RF with specific symptoms may add valuable knowledge to the field of research and child and adolescent psychotherapy. At the same time, the present study contributes to exploring characteristics of the relationship with mothers and fathers as well and its association with psychopathological symptoms.

Method

This study adopted a naturalistic, cross-sectional, and correlational design.

Participants

90 children and adolescents aged from 9–17 years old (M = 13.04, SD = 2.72) participated in this study, 52.2% of the sample being female (n = 47). Most patients were adolescents (68.9% aged between 12–17 years old, n = 62). These young people, alongside their parents or caregivers, sought psychological treatment in a community clinic. Children and adolescents based in institutional shelters, the ones who did not fill the instruments at baseline and those who did not consent to participate in this research were excluded (n = 7).

Instruments

Demographic form

Completed by the patients’ parents or caregivers, addressing information such as the patients’ and their parents/caregivers’ age and education level, and family income.

Child Behavior Checklist (CBCL/6-18; Achenbach, 1991)

The CBCL assesses symptoms and behaviors in patients aged from 6 to 18 years old. It is comprised of 138 sentences that are answered by the patients’ parents or caregivers addressing social and school competencies, besides investigating symptoms in eight specific domains: withdrawal, anxiety/depression, somatic complaints, social problems, thought problems, attention problems, aggressive behavior, and delinquent behavior. According to the scores measured by these scales, the child or adolescent may be included in a clinical, borderline, or normal range concerning their global functioning and internalizing/externalizing profiles. The CBCL is validated internationally and in the country of this study (Bordin et al., 2013).

Friends and Family Interview (FFI; Kriss et al., 2013)

The FFI is an interview protocol theoretically guided by the Adult Attachment Interview (AAI; George et al., 1985), but dimensioned for the developmental capacities of children and adolescents. The FFI allows for the evaluators to attribute an attachment classification (secure, insecure-preoccupied, insecure-dismissing, or disorganized/disoriented) based on the interview’s global coherence and in dimensions such as safe haven and secure base evidence, RF, social competence, self-esteem, peer-relationships, siblings, among others. Safe Haven (SH) and Secure Base (SB) evidence assess the young people’s representations referring to their perception of their mother and father as a safe haven and a secure base. The caregiver’s safe haven function is to provide protection and care, particularly when the child is suffering, distressed, or upset. The secure base function is to encourage the exploration away from the parents, offering support when needed to reinforce the child’s autonomy. The FFI was adapted and validated to the Brazilian context, having met good reliability and fidelity criteria, as well as construct and criterion validity (Schmidt, 2020). Previous studies using the FFI have demonstrated robust agreement between judges and construct validity in community samples (Kriss et al., 2013; Psouni et al., 2020). In this study, the FFI presented high internal consistency (Cronbach’s alpha = .82).

Reflective Function Questionnaire for Youths (RFQY)

The RFQY (Sharp et al., 2009) is an adaptation of the Reflective Function Questionnaire (RFQ; Fonagy et al., 2016) for the adolescent population. The RFQY is a self-report measure that is consisted of 46 items filled on a 6-points Likert scale. Higher scores reflect higher RF, with 12 being the highest possible mark on the total scale. The instrument’s internal consistency in its validation study was .71. The RFQY was adapted and validated in the country of this study (Both et al., 2019). In this study, the RFQY had a Cronbach’s alpha = .81.

Procedures

This study took place in a community clinic. According to the clinic’s usual protocol, the patients and their parents/caregivers scheduled a screening interview. On the day of this first meeting, the parents would complete the demographic form and the CBCL. After the screening interview, the patients who were referred for psychotherapy were invited to participate in this study. The young people and parents who gave consent were included in this research. The child or adolescent was, then, asked to fill in the instruments in a room separate from their parents. The FFI was conducted by a trained member of our research team. All interviews were audio-recorded and, by the end of it, the patients filled in the RFQY. The interviews lasted between 30–60 minutes.

This study followed the guidelines of the National Health Council, Resolution 510/2016, and was approved by the Research Ethics Committee of the university where it was developed (Protocol 2.459.618). All participants received detailed information on the study and signed an assent form, while their parents signed a consent form.

The FFI interviews were recorded and transcribed by a group of trained judges, according to a detailed and standardized Coding Manual (Steele et al., 2009). All interviews were double coded by independent ratters, who formed random pairs and who were blinded concerning the patients’ clinical characteristics. We identified high interrater reliability, with a mean intraclass correlation (ICC) of .87, ranging from .71 to .99. The CBCL was rated through a programme designed by the instrument’s author.

We recorded the results in a database elaborated in the software IBM SPSS v23.0. Descriptive and correlational analyses (Spearman, Chi-square) were used to assess the associations between the patients’ RF, attachment, and baseline symptoms. We considered significant results with p-values<.05.

Results

Socioeconomic and clinical data of the sample

The majority of the participants were enrolled in middle school (n = 66, 73.3%) followed by high school (n = 19, 21.1%), while 5.5% (n = 5) had already finished secondary education. Family income ranged between 2 to 3 minimum wages for 60% of the families (n = 54), followed by 16.7% (n = 15) living on minimum wage, 14.4% (n = 13) ranging between 4 to 6 minimum wages, while 7.8% (n = 7) earned between 7 to 10 minimum wages and one family more than 11 minimum wages. Most patients lived only with their mothers (41.1%, n = 37), 33.3% (n = 30) lived with both parents, 17.8% (n = 16) with their mothers and another partner, 4.4% (n = 4) lived with their grandparents, 2.2% (n = 2) only with their fathers and one participant lived with their father and his partner. The mothers’ mean age was 40.99 years old (SD = 7.11), and the fathers’ 44.11 years old (SD = 8.27). Most of the mothers had finished secondary education (n = 76, 84.3%), and out of those 28 (31%) were enrolled or already finished graduate education. Concerning the fathers, 79.9% (n = 46) had finished secondary education and 29.2% (n = 19) were enrolled or already finished graduate education.

The reasons that made the children and adolescents seek psychological treatment, alongside the CBCL results, are described in Table 1. CBCL results were analyzed considering clinical or non-clinical scores. The borderline category was merged into the clinical group since it is considered that borderline scores already present some risk. The parents or caregivers reported that the symptoms were present on average for 28.08 months (SD = 39.3), which means that 64.4% of the patients (n = 58) presented symptoms for more than two years, on average.

Table 1. Patients’ symptoms at baseline according to the CBCL and their parents’ report on the reason for referral.

	N	%
Parents and caregivers’ reports on the reason for referral
Anxiety/depression	36	40
Withdrawal/depression	18	20
Learning problems	12	13.3
Relationships problems	8	8.9
Attention problems	6	6.7
Aggressive behavior	5	5.6
Oppositional/defiant behavior	3	3.3
Thought problems	1	1.1
Somatic complaints	1	1.1
CBCL – Clinical results
Activity scale	35	38.9
Social competence	38	42.2
School competence	35	38.9
Total social competence	61	67.8
Anxiety and depression	55	61.1
Withdrawal	49	54.4
Somatic complaints	31	34.4
Relationship problems	40	44.4
Thought problems	37	41.1
Attention problems	37	41.1
Oppositional/defiant behavior problems	22	24.4
Aggressive behavior problems	33	36.7
Internalizing problems	67	74.4
Externalizing problems	50	55.6
Total problems (internalizing and externalizing)	68	75.6

Concerning the attachment assessment, 46.7% (n = 42) of the patients presented insecure-dismissing attachment style, 38.9% (n = 35) insecure-preoccupied, only 10% (n = 9) secure attachment, and 4.4% (n = 4) disorganized-disoriented attachment. The M mother SB was 1.65 (SD = .69), and mother SH 1.69 (SD = .69). Concerning the relationship with the father, the SB M score was 1.35 (SD = .61), and the SH M was 1.30 (SD = .60).

According to the RFQY, this sample’s mean RF score was 7.97 (SD = .86). We identified, through this instrument, that only 13.5% of the patients (n = 12) had good RF (scores higher than 9), while 86.5% (n = 77) presented low RF (scores equal to or lower than 8).

Correlation between attachment, RF, and clinical symptoms

Initially, we ran Chi-square tests to verify possible associations between attachment styles and the broad domains of internalizing and externalizing problems assessed by the CBCL. The CBCL symptoms were analyzed as dichotomous variables (clinical and non-clinical). No associations were found between attachment styles with internalizing symptoms (χ² = 2.203, df = 3, p = .53), externalizing symptoms (χ² = .82, df = 3, p = .85), or total problems (χ² = 1.80, df = 3, p = .62).

Afterward, we ran Chi-square tests between attachment styles and the CBCL specific symptom scales. We found significant associations between attachment styles and anxiety and depression (Fisher’s exact test = 9.14, p = .02) and withdrawal (Fisher’s exact test = 8.01, p = .03). In Table 2 we present the differences found between the clinical and non-clinical groups concerning their anxiety and depression and withdrawal symptoms with attachment styles. Referring to anxiety and depression, the insecure-avoidant attachment style was significantly different from the insecure-preoccupied and disorganized groups. As for the withdrawal symptoms, the secure attachment differed significantly from the insecure ones.

Table 2. Distribution of attachment styles in the symptom domains of anxiety, depression, withdrawal and somatic complaints.

CBCL			Attachment				Total
			Secure	Insecure-avoidant	Insecure-preoccupied	Disorganized
Anxiety and Depression	Non-clinical	N %	3a,b 33.3	23b 54.8	9a 25.7	0a 0	35 38.9
	Clinical	N %	6a,b 66.7	19b 45.2	26a 74.3	4a 100	55 61.1
Withdrawal	Non-clinical	N %	8a 88.9	18b 42.9	13b 37.1	2a,b 50	43 47.8
	Clinical	N %	1a 11.1	24b 57.1	22b 62.9	2a,b 50	47 52.2

Note. a, b: attachment subsets. The subsets differ significantly considering p < .05.

To examine possible group differences in terms of gender and age we ran a series of Chi-square tests focusing on possible associations between attachment styles and those variables. Concerning gender, we found significant differences (Fisher’s exact test = 10.73, p = .08), with the secure attachment group composed only of female participants, while the other group, including the insecure-avoidant, insecure-preoccupied, and disorganized attachment styles, were composed of girls and boys with no significant differences. In terms of age groups, we divided the sample between children (up to 11 years and 11 months old) and adolescents (12 to 17 years old), but no significant differences were found (χ² = .79, df = 3, p = .85). No significant differences were also found between age and gender in terms of anxiety and depression (age: χ² = .03, df = 1, p = .96/ gender: χ² = .31, df = 1, p = .58) and withdrawal (age: χ² = 1.05, df = 1, p = .31/ gender: χ² = 3.78, df = 1, p = .52).

Lastly, we tested the Spearman correlation between the FFI three domains: SB, SH, and RF with baseline symptoms according to the CBCL and also the RFQY results with the CBCL scores. We employed Spearman correlation since not all results obtained from the FFI and CBCL presented a normal distribution. All CBCL scales values were included as numeric variables in this analysis, as well as the RFQY scores and the participants’ mean scores on their parental SB and SH. The perception of the parents as an SB and an SH is a constitutional element in the attachment representations (Verschueren, 2020), and, as they are dimensions assessed by the FFI, we also examined them specifically, aiming to identify possible associations between these variables and symptoms assessed by the CBCL. In Table 3 we present the correlations between these variables.

Table 3. Correlation between clinical symptoms, safe haven, secure base and RF.

	Mother SH	Father SH	Mother SB	Father SB	RFQY
Activity Scale	.09	.30**	.13	.20	−.02
Social Competence	.23*	.33**	.24*	.25*	.19
School Competence	.23*	.27*	.30**	.16	.26*
Total Social Competence	.10	.27**	.12	.18	.22*
Anxiety/Depression	−.10	−.17	.08	−.08	−.06
Withdrawal	−.19	−.14	−.12	−.02	−.15
Somatic complaints	.03	−.10	.14	−.08	−.02
Relationship problems	−.27**	−.28**	−.09	−.15	−.29**
Thought problems	−.15	−.09	−.13	.01	−.10
Thought problems	−.08	−.09	−.09	−.02	−.41**
Defiant/oppositional/delinquent behavior	−.09	−.06	−.10	−.12	−.29**
Aggressive behavior	−.19	−.16	−.17	−.11	−.32**
Internalizing problems	−.17	−.21*	−.01	−.08	−.19
Externalizing problems	−.22*	−.12	−.18	−.09	−.33**
Externalizing and internalizing problems	−.19	−.19	−.08	−.06	−.33**

Note. SH – Safe Haven, SB – Secure Base, RF – Reflective function. **p < .01 (bilateral) *p < .05 (bilateral).

Discussion

This study aimed to analyze the association between attachment and RF with baseline symptoms of children and adolescents who sought psychotherapy. Firstly, it is important to discuss this population’s characteristics. By being a naturalistic study, there was no type of control concerning the number of cases and psychopathologic conditions, for instance. We included all young people who sought psychotherapy within a determined time window.

Most patients came from low-income families. This may be explained by the institution being a training center that is open to the public and offers lower consultation fees when compared to private practice. We noted that most children and adolescents did not live in the same household as their fathers (64%). This factor can have a socioeconomic impact on those patients, as the literature points out that socioeconomic risk factors are associated with attachment problems as well as internalizing and externalizing symptoms (Madigan et al., 2016).

The reasons for referral as stated by the parents were predominantly anxiety, and depression and withdrawal, which is in line with previous studies (Fernandes et al., 2017). According to the CBCL, which was completed by the same individuals, internalizing, externalizing, and global problems were considered clinical for most of the patients. The most prominent categories were, likewise, withdrawal, and anxiety and depression.

By analyzing the associations between attachment styles and the domains of internalizing and externalizing problems as assessed by the CBCL, we did not find significant results. However, when looking into specific symptom categories within each dimension, we identified some significant correlations, which is in accordance with previous studies (Madigan et al., 2016; Spruit et al., 2020).

The symptoms that presented significant associations with attachment styles were withdrawal, and anxiety and depression, all of them belonging to the internalizing symptoms dimension. In the same direction, associations between attachment styles and anxiety, depression, and withdrawal/isolation have been discussed in the literature (Colonnesi et al., 2011; Madigan et al., 2016; Spruit et al., 2020).

When examining each attachment style, the preoccupied and disorganized ones were associated with anxiety and depressive symptoms, as already demonstrated in previous studies (Colonnesi et al., 2011), while the insecure-preoccupied and insecure-avoidant styles were associated with withdrawal. On the other hand, we identified that the perception of the mother and the father as an SH and SB was associated with higher social and school competencies as assessed by the CBCL. Bowlby (1988) had already pointed out that secure attachments were a protective factor for psychopathology.

The results indicated that being female would be associated with secure attachment. Numerous previous studies have discussed how gender is related to the factors here examined. Spruit et al. (2020), for instance, analyzed the association between attachment and depression among children and adolescents and reported that female samples presented effect sizes significantly higher than male ones. Other studies revealed that insecure attachments are, in general, more associated with externalizing (Fearon et al., 2010) and internalizing (Madigan et al., 2013) problems in samples of boys versus girls, respectively. Nevertheless, another study that assessed the association between general insecure and disorganized attachment styles with internalizing symptoms did not find significant differences between groups (Colonnesi et al., 2011; Groh et al., 2012). We note, however, that our findings should be interpreted with caution, given our sample size and that only nine patients presented secure attachment.

Although the present study did not aim to investigate causal relations, our results reinforce the theoretical assumption of insecure attachments as a risk factor for the development of psychopathology (Groh et al., 2012). However, clinical symptoms may occur mainly when insecure attachment styles are combined with other risk factors (Kerns & Brumariu, 2014) that were not assessed here.

Based on the Spearman correlations, we observed positive associations between experiencing the mother and the father as an SB and an SH and social and school competencies, according to the CBCL. The security in those bonds seems to effectively provide a base for the young people who participated in this study to be more resourceful in their social and school contexts, in accordance with what has been theoretically described (Bowlby, 1988). On an inverse association, the young people who described their mothers and fathers as a safe haven, according to the FFI, presented fewer relationship problems.

Furthermore, the patients who did not perceive their relationships with their fathers as a safe haven presented more internalizing problems, while the ones who did not perceive their relationships with their mothers as a safe haven had more externalizing problems. That is, there was an impact of those attachment elements in the internalizing and externalizing symptoms domains, but we did not find corresponding associations in the Chi-square tests when focusing on specific symptoms such as withdrawal and depression. A possible hypothesis to explain these results is that when running the Chi-square test the attachment styles were taken as categorical variables based on the FFI as a whole. Likewise, the CBCL symptoms were also considered categorical variables (clinical or non-clinical). For the Spearman correlation, we analyzed only the mean scores obtained in the dimensions of SH and SB as assessed through the FFI, and the symptoms and their domains were considered according to their raw scores. Perhaps a larger sample would be necessary to confirm or refute these findings. Considering the variables as a continuum could also contribute to clarifying this issue. Some studies suggest that using continuous variables for the assessment of attachment can favor the correlations’ effect size (Spruit et al., 2020). Future studies can address these questions.

Although the strength of the correlations ranged from weak to moderate, it was clear that the bonds with the parents have an impact on their children’s social lives during childhood and adolescence. This finding follows the premise that the parental attachment quality shapes the representational models the children have about themselves (Bureau et al., 2020). We identified that the father figure exerted influence in the three areas assessed by the CBCL (the young person’s activities, social and school competencies), while the mother was associated with the social and school competencies. There might be a difference in the way mothers and fathers talk and interact with their children and, therefore, in the way that they influence them, perhaps for them assuming different functions (Bögels & Phares, 2008; Verschueren, 2020). According to Verschueren (2020), fathers seemed to exert an important role as an SB, stimulating the child through plays and challenges, leading to support and encouragement for the exploration of the social environment.

We also identified that higher RF, as assessed by a self-report questionnaire, was positively associated with higher school competence and lower relationships and attention problems, lower aggressive, defiant, oppositional, and delinquent behaviors. These findings highlight the importance of the participants’ RF, emphasizing their role, just like the attachment styles, in the development – or not – of psychopathological symptoms (Ha et al., 2011).

The literature discusses what are the most appropriate forms to assess children and adolescents’ symptoms, as well as their attachment and RF (Colonnesi et al., 2011; Groh et al., 2012; Madigan et al., 2016). Some authors defend that Likert scales, answered by the patients themselves, are associated with stronger correlations (Madigan et al., 2016). Yet other studies indicate that the type of measure, either representational or questionnaire, is not a factor that moderates the results (Madigan et al., 2013). Nevertheless, it seems to be a consensus that having the child or adolescent as the informant favors larger effect sizes (Madigan et al., 2013).

It should be noted that in this study we opted to employ the CBCL completed by the parents for the symptom assessment since this instrument was already adopted by the clinic where this investigation was performed, being already part of the institution’s screening process. We prioritized not to interfere in how the treatments would normally take place, even if that meant a possible divergence between the symptom reports (Spruit et al., 2020). On the other hand, we used a measure of attachment that was based on young people’s own representations.

The present study confirms the existing evidence concerning the association between attachment style and symptoms like withdrawal, anxiety, and depression (Colonnesi et al., 2011; Madigan et al., 2016; Spruit et al., 2020). In addition, when exploring the perception of the mother and the father as SH and SB we found that they were associated with higher scores of social and school competencies as well as with fewer relationship problems. As far as we know, this is a new finding on the role of these aspects of attachment representations and their association with those competencies. Furthermore, we found that internalizing and externalizing symptoms were differently associated with perceiving the mother and the father as a safe haven. When the father was not perceived as a safe haven, the youth presented with more internalizing problems, while in relation to the mother the result was the opposite, and the youth would present more externalizing problems. This is a preliminary finding, and further studies are needed to confirm or refute such associations.

In summary, we found associations between attachment and symptoms of anxiety, depression, and withdrawal. Significant differences were identified between the insecure-avoidant group and the insecure-preoccupied and disorganized groups regarding symptoms of anxiety and depression. As for the withdrawal symptom, the secure attachment group differed significantly from the insecure attachment groups. Another important finding was the identification of the attachment style and the reflexive functioning of children and adolescents who seek psychotherapy, demonstrating that they are young people with high insecurity in relationships, along with a difficulty in being able to understanding their internal mental states, both in a cognitive and in an affective way. In this sense, identifying the patterns of attachment and reflexive function of young people at the beginning of the therapeutic process is extremely important, because they relate to trust in relationships in general, and the therapeutic relationship in particular. It is possible to hypothesize that the characteristics of attachment and reflective functioning of young people can impact the therapeutic alliance and the course of the therapeutic process, as well as being related to the symptoms presented and their intensity.

Conclusions

This study aimed to verify the associations between attachment and RF with symptoms of children and adolescents who sought psychotherapy. We did not find associations between attachment styles and global symptom scores. Nevertheless, when the symptoms were assessed separately, we found significant associations between depressive and anxiety symptoms with insecure-preoccupied and disorganized attachment styles, and between withdrawal and insecure attachment. To the best of our knowledge, this is the first Brazilian study that examined the association between attachment styles with broader symptom domains as well as specific symptom categories in children and adolescents.

An important finding for psychotherapy initial assessment and planning is that most children and adolescents who sought treatment presented difficulties concerning their attachment bonds and their capacity to mentalize and being able to understand their own mental states as well as others’. On the other hand, better relationships of SB and SH with the parents reflected better social performance in the young people. Curiously, we also noted that perceiving the mother as a SH was inversely correlated with externalizing problems while perceiving the father as a SH correlated negatively with internalizing problems.

The study of attachment is important for its role in complex processes such as mental representations, RF and consequently in social relationships and emotional regulation. Although attachment styles may not be direct causes of certain outcomes, they might be one of the most important aspects to be assessed during childhood in psychological and emotional terms.

Thus, assessing the RF among children and adolescents who seek psychological treatment is of extreme relevance. Their attachment style indicates how the child relates, trusts, and invests in their relationships in general. Therefore, they will be expressed in their interaction with their therapist, their parents, and the way they deal with the world around them. All this information could help in treatment planning. Furthermore, RF informs about higher or lower aptitude for mental processes that may be fundamental in the psychotherapy process.

This study has some limitations. Being naturalistic, it does not count with a control group and, although the sample size is reasonable, it included diverse symptoms, which can have interfered with the correlations or lowered their strength. Another limitation is that the symptoms analyzed were based on a single source – the CBCL completed by the patients’ parents. Future studies can include self-report measures for the symptoms’ assessment, as well as a bigger range of instruments and sources.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Financial support was provided by the National Council for Scientific and Technological Development (309544/2017-3), and Coordination for the Improvement of Higher Education Personnel (88887.150777/2017-00 and 88881.170035/2018-01), Brazil.