How adolescents with anorexia nervosa and their parents perceive family functioning?

Introduction

Anorexia nervosa (AN) is a serious health condition, given the association with several negative physical and psychosocial outcomes, including role and social impairment (e.g. difficulties in interpersonal interaction and the preference for activities that do not involve others, such as watching TV and reading), comorbidity with other mental disorders such as mood, anxiety, obsessive-compulsive disorders, and substance abuse, and high rates of mortality and suicide (Bühren et al., 2014; Krug et al., 2013; Preti et al., 2009; Smink et al., 2012; Swanson et al., 2011).

According to the Diagnostic and Statistical Manual of Mental Disorders (5th edn; DSM-5) (American Psychiatric Association (APA), 2013), AN is characterized by three key features: an intense fear of gaining weight that cannot usually be reduced by weight loss, a distorted perception of body weight or shape, and a significantly low body weight compared to normal minimum weight for age or height.

AN typically occurs during adolescence and has an estimated lifetime prevalence of .3 percent in US adolescents aged 13–18 years (Swanson et al., 2011). The course of the disorder is often chronic (Wentz et al., 2009), with medical complications and poor treatment response (Arcelus et al., 2011). Indeed, individuals with AN have a tendency to deny their condition and it takes long before they become aware that they are dealing with AN, as they often have a positive view of the onset of the eating disorder (e.g. they feel they have more control over eating) (Fox and Diab, 2015). Therefore, they do not seek for professional help, and it has been noted that only a small number of adolescents with this disorder receive specific care for their eating problems (Swanson et al., 2011).

In order to improve the effectiveness of intervention programs, several qualitative studies have underlined the importance of considering patients’ own perception and experiences of treatment they receive for AN. For instance, it has been found (Smith et al., 2016) that during treatment, patients with AN perceive an increase in control over their eating behavior, give great importance to peer and staff support, and feel the desire to be more involved in the treatment process, such as in decisions regarding meal protocols. In fact, patients reported difficulties during mealtimes (including physical, cognitive, and emotional problems) and the need for more autonomy and responsibility (Long et al., 2012).

In addition, in primary care, where interventions should be brief and practicable, patients with AN seem to benefit from both emotion-focused interventions—such as Cognitive Remediation and Emotion Skills Training (CREST), which allow them to learn abilities to identify and manage emotions and increase assertiveness—and guided self-help cognitive behavioral therapy, via the use of a self-help book, which provides information about learning strategies concerning, for example, self-monitoring and psychoeducation (Allen and Dalton, 2011; Money et al., 2011).

Based on family systems theory, the Circumplex Model of Marital and Family Systems has offered an important contribution to the study of family functioning as perceived by family members (Laghi et al., 2016, 2015; Olson, 2000; Olson et al., 1983). According to this widely used and recognized model, there are three dimensions believed to be critical for understanding family functioning: family flexibility, cohesion, and communication. Family flexibility has been previously described as the family ability to change leadership roles, relationships, and rules; more recently, it has been given a new definition (Olson, 2011; Olson and Gorall, 2006), that is, the quality and expression of family leadership, organization, roles, rules, and negotiations. A very low level of flexibility results in rigidity and a very high level of flexibility leads to chaotic families. Cohesion is defined as the emotional bonding between family members. A very low level of cohesion results in disengagement, and a very high level of cohesion leads to enmeshment. Finally, communication refers to a facilitating skill that can help family members to change their levels of flexibility and cohesion. The main assumption on which the model is based suggests that balanced levels of cohesion and flexibility are related to more functional families, whereas very low or very high levels of these two dimensions characterize problematic family systems. To the best of our knowledge, no study so far has investigated the different dimensions of family functioning according to the new model proposed by Olson (2011).

Literature focusing on risk factors involved in eating disorders highlighted that family functioning can play an important role in the development and maintenance of these disorders (Laghi et al., 2012a, 2012b). According to family systems theory, families are evaluated as a whole in which members are interconnected, and the family organization can affect the behavior of all family members (Holtom-Viesel and Allan, 2014; Laghi et al., 2012c, 2014a, 2014b). Family systems concepts specifically regarding AN have been most clearly defined by Minuchin et al. (1978), who suggested that specific family patterns contribute to the development and maintenance of AN, and identified a group of family systems characteristics considered as typical in families of patients with AN: excessive enmeshment, which refers to the clarity of boundaries in family interactions; rigidity, or the tendency to maintain the status quo; over-protectiveness, characterized by the high degree of concern for each other’s well-being; and finally, avoidance of conflict. Family systems with these features were named “psychosomatic,” because the symptoms of a family member are considered as the expression of the discomfort of the entire family. Therefore, much attention has been dedicated to analyzing the different aspects of family functioning, including communication, and the presence of clear boundaries between family members (Minuchin et al., 1975).

Since the early work of Minuchin and colleagues, there is more and more evidence that families need to play an active role in the treatment process and facilitate the change in several aspects of family patterns. For this reason, families are often considered as a significant resource and are included in intervention programs (e.g. Le Grange, 2005; Le Grange et al., 2010; Pace et al., 2015). Specifically for eating disorders, intervention studies reported that adolescents with AN respond well to family therapy, in particular, family-based interventions seem to be useful and effective if they are the main form of treatment (e.g. Eisler et al., 2003; Le Grange and Eisler, 2008; Lock et al., 2006)

Family functioning has been measured both as a unitary construct reflecting the general functioning of the family as a whole and as a set of several aspects, such as flexibility, cohesion, and communication. Existing evidence suggests that there is a substantial agreement within family members dealing with AN in perceiving an impairment of the general family functioning and a strong dissatisfaction with family life (Casper and Troiani, 2001; Ciao et al., 2015; Cook-Darzens et al., 2005; Emanuelli et al., 2004; McDermott et al., 2002; Szabo et al., 1999); on the other hand, when considering specific dimensions of family functioning, significant differences are found between girls with AN and their parents, even though research shows variable and sometimes conflicting results (Ciao et al., 2015; Holtom-Viesel and Allan, 2014).

It has been noted that patients’ perception of family functioning is the most impaired; they perceive their families as less communicative, cohesive, flexible, and having difficulties in problem solving, compared to their parents (Ciao et al., 2015; Cook-Darzens et al., 2005; Emanuelli et al., 2004; Vidovic et al., 2005). Furthermore, eating-disordered girls perceive their families as having less facilitative and more constraining family rules, such as the control that family exerts on thoughts, self, and feelings of its members, compared to what their parents perceive (Gillett et al., 2009). In addition, Ciao et al. (2015) showed that eating-disordered girls reported a poorer clarity of rules within the family and a more inappropriate affective involvement—that is, a lack of interest and concern for each other within the family—compared to their mothers. Moreover, girls and their fathers perceived a greater impairment in the family’s ability to express feelings and emotions compared to what mothers perceived.

Unlike eating-disordered girls, their mothers resulted to be less critical of their families, reporting healthy and less chaotic family functioning and a lower presence of boundary problems that can create enmeshment between family members (Dancyger et al., 2005; Rowa et al., 2001). Fewer differences were also found between mothers and fathers concerning aspects of problem solving, cohesion, and affective responsiveness, with fathers having higher mean scores (Cook-Darzens et al., 2005; Dancyger et al., 2005).

A number of studies also found differences in family functioning perception in relation to AN subtypes: families with a daughter with restricting type of AN tend to share a healthy perception of family functioning and report to be satisfied and in harmony with family environment (Casper, 1981; Casper and Troiani, 2001; Garfinkel et al., 1983; Kog and Vandereycken, 1989; North et al., 1995). In particular, parents considered their families as cohesive and poorly conflicting, with satisfied relationships with their daughters (Garner et al., 1985). With regard to clinical girls, they perceived their families as flexible and cohesive, and they reported a good level of communication with their mothers (Vidovic et al., 2005). In addition, it was noted (Casper and Troiani, 2001) that girls with restricting type of AN rated their mothers’ performance within the family better than their mothers did. These findings suggest that parents’ and daughters’ positive view of family interactions may be due to denial and conflict avoidance, influencing their judgments, and leading them to idealize family relationships (Casper and Troiani, 2001; Vidovic et al., 2005).

Several studies have also investigated how girls with AN and controls differ in the perception of family functioning. The results showed that with regard to a general level, clinical groups were less satisfied about family environment and rated their families as having a worse functioning compared to controls (Casper and Troiani, 2001; Emanuelli et al., 2004; McDermott et al., 2002; Waller et al., 1989). When studies considered specific aspects of family functioning, it was found that adolescents with AN considered their families as poorer in cohesion, adaptability (Cook-Darzens et al., 2005; Shisslak et al., 1990; Stern et al., 1989; Vidovic et al., 2005), communication, appropriate expression of feelings and emotions, and clarity of roles and rules (Casper and Troiani, 2001; Emanuelli et al., 2004; Maglio and Molina, 2012; Shisslak et al., 1990; Waller et al., 1989) compared to controls. Furthermore, patients perceived in their families less encouragement to autonomy and independence, in particular by mothers (Karwautz et al., 2003; Shisslak et al., 1990), less emotional involvement with other family members (Waller et al., 1990), as well as more difficulties in family ability to resolve problems or conflicts (Casper and Troiani, 2001; Emanuelli et al., 2004) compared to controls.

On the other side, studies that discriminate between AN subgroups showed different and opposing findings: with regard to restrictive type of AN, no significant difference was found between clinical and control groups for specific aspects such as cohesion, flexibility, communication with family members (in particular with mothers), family’s abilities to express emotions in an appropriate way, concern and interest for each other, ability to solve problems, and the presence of clear rules within the family (Casper and Troiani, 2001; Vidovic et al., 2005). In addition, girls with restrictive type of AN reported a higher level of appropriateness in the differentiation of tasks and roles within the family, and adherence to values, rules, and standards that families establish and to which family members have to comply (Casper and Troiani, 2001).

This study aims at thoroughly analyzing the perception of the different aspects of family functioning, such as cohesion, flexibility, and communication within the family in adolescent females diagnosed with AN, from the perspective of multiple family members. To our knowledge, this is the first study exploring different perspectives on dimensions of family functioning according to the new Circumplex Model of Marital and Family Systems (Olson, 2011). A further improvement brought on by the Balanced and Unbalanced ratio score is that it offers a method to assess curvilinearity of Cohesion and Flexibility dimensions. The following hypotheses were generated: First, when compared with healthy controls (HCs), girls with AN would endorse reduced family cohesion and flexibility. Second, according to previous studies (Casper and Troiani, 2001; Emanuelli et al., 2004; Maglio and Molina, 2012; McDermott et al., 2002; Shisslak et al., 1990; Waller et al., 1989), girls with AN would also endorse reduced family communication and satisfaction.

Methods

Results

Family functioning similarity between parents and adolescents

On average, there were no significant differences, for the clinical sample, between mothers and fathers for all dimensions, as reported in Table 1. On average, only mothers’ self-report of Rigid scale of FACES-IV was significantly higher compared to their daughters. The effect size for this significant difference can be classified as medium (Cohen’s d = −.45). Additionally, paired-sample t-tests did not reveal significant differences between fathers and daughters for all dimensions.

OPEN IN VIEWER

Table 1.

Means, standard deviations, and paired t-tests for FACES-IV dimensions, Family Communication, and Family Satisfaction.

	Fathers (N = 36)		Mothers (N = 36)		Daughters (N = 36)		Fathers/mothers	Fathers/daughters	Mothers/daughters
	M	SD	M	SD	M	SD	t (d)	t (d)	t (d)
FACES-IV dimensions
Cohesion	27.59	4.92	27.78	4.12	27.36	4.01	.19	.10	−.04
Flexibility	26.06	3.17	25.59	3.54	24.33	4.06	−.65	1.59	1.09
Disengaged	15.47	4.69	16.03	4.40	17.00	4.28	.62	−.99	−.91
Enmeshed	16.84	4.30	16.09	4.15	15.64	4.62	−.70	1.38	.52
Rigid	18.63	3.23	19.59	3.06	17.81	4.21	1.28	1.50	2.37* (−.45)
Chaotic	16.75	4.29	16.53	4.05	17.44	4.92	−.21	−.52	−1.04
Family Communication	36.38	6.28	36.03	7.71	35.50	6.71	−.22	.50	.04
Family Satisfaction	32.53	6.30	31.13	6.93	33.42	7.45	−1.00	−.68	−1.58

FACES-IV: Family Adaptability and Cohesion Evaluation Scales; SD: standard deviation.

Cohen’s d are reported in parentheses (effect sizes of .20 are small, .50 are medium, and .80 are large).

*

Paired t-test p < .05.

Family functioning in the AN and control groups

Preliminary analyses revealed no differences between participants in the AN group and those in the HC group for any of the demographic and clinical variables considered, except for BMI, F(1, 70) = 86.81, p < .001. The differences between AN and HCs were investigated by conducting a MANOVA on the FACES-IV dimensions. The MANOVA revealed a main effect for group, λ = .38, F(8, 63) = 12.70, p < .001, partial η² = .62. The results of the univariate tests revealed differences between the groups on the subscales of Cohesion, F(1, 70) = 83.67, p < .001, partial η² = .54; Flexibility, F(1, 70) = 36.75, p < .001, partial η² = .34; Family Communication, F(1, 70) = 33.70, p < .001, partial η² = .32; and Satisfaction, F(1, 70) = 26.50, p < .001, partial η² = .27, where the AN group showed lower mean scores than the HCs. The AN group exhibited a higher mean score on Disengaged scale than HCs, F(1, 70) = 21.55, p < .001, partial η² = .23 (see Table 2).

OPEN IN VIEWER

Table 2.

Differences between clinical and control groups on the FACES-IV dimensions, Family Communication, and Family Satisfaction.

	AN (N = 36)		HC (N = 36)		F(1, 70)
	M	SD	M	SD
FACES-IV dimensions
Cohesion	27.36	4.01	33.58	.73	83.67*
Flexibility	24.33	4.06	29.53	3.16	36.75*
Disengaged	17.00	4.28	12.67	3.62	21.55*
Enmeshed	15.64	4.62	15.53	3.81	.012
Rigid	17.81	4.21	18.31	5.13	.204
Chaotic	17.44	4.92	16.50	3.87	.819
Family Communication	35.50	6.71	43.28	4.43	33.70*
Family Satisfaction	33.42	7.45	41.69	6.14	26.50*

FACES-IV: Family Adaptability and Cohesion Evaluation Scales; AN: anorexia nervosa; HC: healthy controls; SD: standard deviation.

*

p < .001.

Discussion

The first aim of this study was to examine whether adolescents with AN and their parents differ in the perception of the different aspects of family functioning. Our findings highlighted a general agreement among family members, except that mothers and daughters had a different view of the dimension of rigidity; in particular, AN girls rated their families as characterized by less rigidity compared to mothers. Rigidity refers to family systems having a strict discipline and defined roles, where negotiations are limited and decisions are imposed by an authoritarian leadership (Olson, 2000). For these reasons, these families have difficulties in evolving in response to circumstances, either originated from internal factors to the family or from the external environment. These results confirmed earlier findings showing a unified view between daughters and fathers (Dancyger et al., 2005) and between mothers and fathers (Ma, 2011). Moreover, this study results were in accordance with earlier studies that found differences in mothers’ and daughters’ perception of family functioning (Ciao et al., 2015; Cook-Darzens et al., 2005; Dancyger et al., 2005; Rowa et al., 2001; Vidovic et al., 2005), as well as with Casper and Troiani (2001), who found that girls with restrictive type of AN tended to have a more positive view of family functioning than their mothers. These findings support the theory that judgment of patients with AN could be influenced by the tendency to idealize their families, avoid conflict, and deny potential problems with family members. In fact, according to “psychosomatic” families (Minuchin et al., 1978), a low tolerance for conflict and the difficulty in identifying and resolving issues are characteristics of a common pattern in families of patients with AN.

When compared to HCs, findings confirmed that girls with AN had a worse perception of family functioning and were less satisfied about their families, as it was hypothesized. In particular, in line with previous studies (Cook-Darzens et al., 2005; Shisslak et al., 1990; Stern et al., 1989; Vidovic et al., 2005), adolescents with AN rated family environment as poorly communicative, characterized by a low level of emotional bonding between family members and by a low ability to change when circumstances require, compared to controls.

Additionally, this study results suggested that compared to controls, clinical group perceived interactions among their families as having less emotional closeness and more difficulty in focusing on experiences, activities, and interests of family members. Moreover, AN group found it more difficult to confide in one another within the family and seek emotional support when needed. This last result is in accordance with Waller et al. (1989), who found that girls with AN considered their families as characterized by less affective involvement and lower levels of interest and concern for family members, compared to controls.

The findings of this research seem to suggest that adolescent girls with AN tend to have an overall worse perception of their family environment, compared to HCs. Nevertheless, when compared to their parents’ perception, mothers seem to be more dissatisfied of their families; in particular, they have the feeling that family functioning is characterized by a low level of flexibility resulting in difficulties with shifting or evolving in response to changes due to life events.

This study provides useful information on aspects of family functioning as perceived by family members dealing with AN. When dealing with eating disorders, family members are often considered as an important resource rather than an obstacle (e.g. Le Grange, 2005; Le Grange et al., 2010) and are therefore included in the treatment process. For this reason, it is extremely important to obtain a better knowledge on family environment, in order to best involve parents in family-based interventions. Furthermore, to our knowledge, this is the first study aimed to analyze the different dimensions of family functioning according to the new Circumplex Model of Marital and Family Systems (Olson, 2011).

Despite the strengths of this study, some limitations need to be taken into consideration. First, the HC group only included healthy adolescents matched by education and age, but did not include their parents to compare perception of family functioning in families with a daughter with AN and families of HCs. In fact, it remains unknown whether the difference found between mothers and daughters concerning the dimension of rigidity is specific to the clinical group or would also be found in controls. Second, all the information was obtained through self-report measures. Third, the cross-sectional design did not allow to investigate direction of causal relationships between the considered variables. In light of the above limitations, further research should be addressed to both replicate these findings with the inclusion of control families and examine the results in a longitudinal approach in order to investigate the presence of important developmental changes in the relationship between family functioning and AN across adolescence.

Finally, these findings have further important implications. This study showed that families with a daughter with AN are perceived as having major weaknesses in several aspects of family functioning, including poor communication and low levels of cohesion and flexibility. These dimensions could represent the aspects on which treatment should be focused, in order to enable family members to improve their abilities and learn how to act with a better use of functional patterns. Moreover, a deep examination of the relationship between the different aspects of family functioning and AN may contribute to a better understanding of how to support families during the treatment process. In addition, the discrepancy found between daughters and mothers stresses the importance of taking into consideration the different views of family members, as such information may represent an essential resource in healthcare settings.