Family Structure in Eating Disorders

Whitney Aline Spannuth

We are all genetically and socially affected by our families. Families serve as the matrix of our identity. It is through interactions within the family that we develop a sense of who we are and how we fit in (Minuchin, Rosman & Baker, 1978). Parents serve as role models, providing examples for attitudes, coping skills, and eating habits, as well as setting standards for perfection, ambition and acceptance (Hall & Cohn, 1992). Many researchers claim that family dynamics are at the root of eating disorders such as anorexia nervosa. The role of dysfunctional family interactions in the pathogenesis of anorexia nervosa has been given a prominent place in the research field. Evidence for a specific family constellation in this disorder, however, has been conflicting. While the majority of studies argue for a specific family interaction style, further studies must be conducted to identify distinguishing characteristics of anorexic subtypes and to determine whether these characteristics are of a causal or consequential nature (Minuchin, Rosman & Baker, 1978).

Family focused treatments for anorexia nervosa have been developed based on accounts in family therapy literature of the "typical" anorexic or "psychosomatic" family (Weme & Yalom, 1996). Anorexic families may appear to have a perfect or ideal environment on the surface, but upon close observation little expression of affection or warmth is seen. Members of these families seldom take specific stands on issues, and conflict is avoided at all costs. Underlying dissatisfaction and tension is often present within the parental dyad. It has been suggested that parents of anorexic offspring put high expectations on their children to over-compensate for the lack of love in their own marriage (Blinder, Chaitin & Goldstein, 1988). The anorexic is then capable of using the illness to unite his/her parents.

In a review article on anorexia and family issues, Yager describes how anecdotal reports of child-parent interactions and personality styles of parents show a great deal of variability. The relationships between mothers and daughters are reported by some to be rejecting and by others to be ambivalent or overinvolved. Although these mother-child interactions are contradictory, several general themes are present (Blinder, Chaitin & Goldstein, 1988). Anorexic mothers tend to focus all of their attention on the well-being of their children (Minuchin, Rosman & Baker, 1978). They set high expectations and foster ambitions for external achievement. The mothers of anorexics may be involved socially, they usually lack intimate friends. In many cases, the daughter becomes the mother's confidant. This overinvolvement creates separation difficulty later in life (Blinder, Chaitin & Goldstein, 1988). A great amount of variability exists in father-daughter dyads as well. Some anorexic fathers have been described as kind and affectionate, while others have been described as passive and ineffectual. These fathers are often peripheral to the family (Blinder, Chaitin & Goldstein, 1988).

Linear and systems models of anorexia nervosa have been postulated to explain the development and treatment of this disorder. Systems concepts concerning anorexia nervosa have been most clearly outlined by Minuchin et al.(Minuchin, Rosman & Baker, 19178). Munichin et al. have recognized a group of family system characteristics that reflect the family dynamics of patients with anorexia ncrvosa: enmeshment, overprotectiveness, rigidity, and avoidance of conflict and lack of conflict resolution. These four transactional characteristics provide the context for the anorexic child to use his/her illness as a means for communicating avoided messages as well as family and parental conflict (Blinder, Chaitin & Goldstein, 1988).

Enmeshment is a transactional style where family members are highly involved with one another. Their is excessive togetherness, intrusion on other's thoughts, feelings and actions, lack of privacy, and weak family boundaries. Members often speak for one another, and perception of the self and other family members is poorly differentiated (Weiss, Katzman & Wolchik, 1985). A child growing up in this type of family learns that family loyalty is of primary importance. This pattern of interaction hinders separation and individuation later in life (Blinder, Chaitin & Goldstein, 1988).

Another characteristic of anorexic families postulated by Minuchin and colleagues, overprotectiveness, refers to the excessive nurturing and protective responses commonly observed. Pacifying behaviors and somatization are prevalent (Weiss, Katzman & Wolchik, 1985).

The third principal characteristic is rigidity; these families are heavily committed to maintaining the status quo. The need for change is denied, thereby preserving accustomed patterns of interaction and behavioral mechanisms. Rigidity is commonly observed in the family cycle during periods of natural change where accommodation is necessary for proper growth and develop. Such an example would be when an adolescent begins to request more independence (Blinder, Chaitin & Goldstein, 1988).

The fourth main characteristic is avoidance of conflict and lack of conflict resolution. Family members have a low tolerance for overt conflict, and discussions involving differences of opinion are avoided at all costs. Problems are often left unresolved and are prolonged by avoidance maneuvers (Blinder, Chaitin & Goldstein, 1988).

Evidence for the psychosomatic family model in anorexia nervosa has been conflicting. To test for the presence of a specific family constellation, Minuchin conducted a formal controlled study involving forty-five psychosomatic families, eleven of which were "anorexic families." A standardized interactive task was set up for each family, and each session was rated on operationalized constructs of enmeshment, overprotectiveness, conflict avoidance and rigidity. The results showed that the more dysfunctional families were more extreme on the above dimensions compared to normal controls. The anorexic families showed the most extreme patterns compared to the dysfunctional family contrast groups. Because a detailed account was not given of how the transactional constructs were operationalized, the outcome of the results cannot be scientifically objectively judged (Blinder, Chaitin & Goldstein, 1988).

An Italian psychiatrist, Selvini-Palazzoli, worked with anorexic families at the Milan Center of Family Studies. She observed family system characteristics similar to the ones described by Minuchin. She reported faulty communication patterns; family members often discounted the messages sent by others. The parents appeared unable to take responsibility or a leadership role, typically blaming one another for bad decisions. According to Selvini-Palazzoli, the parents of anorexic children feel victimized and view their position in the family as a personal sacrifice (Blinder, Chaitin & Goldstein, 1988).

In a study by Sabovich et al., psychological measures were used to systematically examine the dependency, insecurity, boundary difficulties and maladaptive management of aggressive impulses in the parents of anorexic children. Sabovich matched the sample of anorexic parents with other parents of emotionally disturbed inpatient and outpatient girls. Results showed the parental dyad to exhibit more insecurity, dependency, and difficulty with impulse management when compared to the other contrast groups (Blinder, Chaitin & Goldstein, 1988).

Kramer (1983) studied the family systems characteristics of anorexic restricters, anorexic-bulimics, and nortnal-weight bulimics and compared them to a normal-weight control group. The Family Environment Scale and Structural Family Interaction Scale were used in the study. Kramer found that all the eating disorder groups were more similar than different on family systems characteristics. Eating disorder groups also showed more dysfunctional family patterns. The anorexic-bulimic group exhibited more dysfunctional patterns than the other contrast groups. Anorexic-bulimics perceived more mother overprotection, father overprotection, less flexibility, and less mother-child conflict resolution compared with the other groups and controls. Anorexic restricters perceived less parent management and greater triangulation maneuvers (each parent trying to get their child to be in coalition against the other parent). Overall, the findings by Kramer were consistent with Minuchin's psychosomatic paradigm. His findings also extended the family systems model to include anorexic subtypes (Blinder, Chaitin & Goldstein, 1988).

An empirical study by Sonne et al. (1981) investigated family systems transactions in families with anorexic offspring. The study focused mainly on patterns of enmeshment and conflict avoidance within these families. The study included I I anorexic female inpatients, five of other inpatient emotionally disturbed adolescents, and 27 outpatient emotionally disturbed teenagers. The interactions focused on mother-daughter and father-daughter dyads. The results of this study were opposite of those hypothesized. Anorexic families showed less enmeshed behaviors compared with other contrast groups. This finding was due to low parental intrusion. Anorexic adolescent girls were, however, seen as equal or more controlling toward parents when compared to other adolescent groups. Anorexics also made more inferences about the feelings of their parents without checking to see if their perceptions were valid. Anorexic families did show a much higher conflict avoidant pattern in mother-daughter interaction compared to contrast groups, but characteristics of conflict avoidance between father-daughter dyads were not significantly different from the contrast groups (Blinder, Chaitin & Goldstein, 1988).

The findings by Sonne of less enmeshment in anorexic adolescents contradicts Minuchin's findings. She explains that the operationalized behavior she used may not have truly matched the construct of enmeshment. In addition, the control groups used by Sonne differed from Minuchin's; they were comprised of other emotionally disturbed adolescents in comparison with Minuchin's use of other psychosomatic physically ill and "normal" groups. The emphasis Sonne placed on conflict in experimental interactions could have been another confounding variable (Blinder, Chaitin & Goldstein, 1988).

Two family measures were administered to the parents of anorexic and anorexic-bulimic patients in a study conducted by Strober (1981). Each set of parents completed the Moos Family Environment Scale (FES). On the FES, anorexic families showed greater cohesiveness (mutual support and concern among members) and organization (clarity of structure and rules). On the marital adjustment scale, disharmony was found in both groups. The findings by Strober contradict the majority of observations of anorexic families by system theorists. Anorexic families perceived themselves as significantly higher on cohesiveness than anorexic-bulimics. Anorexic-bulimics, however, scored lower on the cohesiveness scale when compared with FES normal family groups.

According to Minuchin's theory, dysfunctional families should be seen on the extremes of the cohesiveness dimension. Extremely high cohesion is labeled as enmeshment and extremely low cohesion as disengagement. Functional normal families are referred to as those families having a balance on the enmeshment-disengagement continuum. One can therefore conclude that Strober's results showed the anorexic-bulimic families to be more disengaged. This finding contradicts Minuchin's theory, as anorexic families are viewed as all extremely cohesive or enmeshed. Minuchin's theory also maintains that all anorexic families should be more extreme on conflict avoidance than functional-normal families. Strober found that both anorexic and anorexic-bulimic groups scored higher on the FES conflict scale than normals on this dimension (Blinder, Chaitin & Goldstein, 1988).

Several confounding variables in Strober's study have been cited. Family measures were analyzed from only the perspective of the parents; the patient's perception of the family was never taken into consideration. In addition, parents were asked to fill out the FES retrospectively, before their child's symptoms ever arose; perception of the family environment could have been distorted due to the time lag between mean onset of symptoms and hospitalization. The question as to whether the FES is applicable to clinical populations must also be taken into account (Blinder, Chaitin & Goldstein, 1988).

In a paper by Le Grange and Rutherford (1994), data from a comparative trial of two forms of family intervention for the management of eating disorders in adolescents was discussed. In the beginning of treatment, family processes were measured using the Expressed Emotion (EE) and Family Adaptability and Cohesion Evaluation Scales (FACES). EE in the families of anorexic and bulimic patients was found to be at low levels; the low levels of parental critical comments represent the conflict avoiding characteristic of families of psychosomatic patients. Surprisingly, these families showed low levels of Emotional Overinvolvement on EE, contradicting clinical descriptions. FACES scores revealed that patients perceived their families to be distant and highly structured. This pattern superficially contradicts accepted clinical descriptions. Parents, however, perceived their family structure to be more flexible and cohesive than did the patients. These findings are more similar to the accepted clinical descriptions. Le Grange and Rutherford found the FACES ideals for family organization scored by patients and parents equate more nearly with the clinical descriptions of enmeshment and lack of boundary structure (Le Grange & Rutherford, 1994).

A review by Broberg (1993) was conducted in order to further explore the concept of "the anorexic family" as used by Minuchin and co-workers. According to Broberg, describing a subgroup of families with an anorexic family member does not provide a valid description of the group as a whole. He maintains that a constellation of familial factors characteristic of the families of all anorexia patients does not exist. Certain familial factors have, however, been over-represented: presence of eating disorders, affective illness and alcoholism among close relatives. "Serious life events", such as the loss of a close relative, have also been shown to be over-represented among teenage girls with this disorder. Broberg warns that treating anorexia nervosa as a uniform entity where the same set of aetiological factors are valid for all cases would be a big mistake (Broberg, 1993).

There are contradictory findings between anorexic subtypes and styles of family interaction. There are many factors that contribute to the variability of findings across eating disorder groups. Many studies involve small sample numbers, calling into question the generalizability of the results. Many studies also recruit subjects that volunteer; these subjects may not provide a true representation of an anorexies. Subjects across different studies also show a varying degree of severity of symptomatology. Another methodological problem involves the use of self-report measures that involve a family member's perceptions of family functioning. An individual's perception of family interaction could be very different from the observational data collected by the researchers themselves (Blinder, Chaitin & Goldstein, 1988).

While eating-disorder families appear to have more disturbed interactions than normal families, most researchers would agree that it is erroneous to overgeneralize about anorexics and their families. Further study must be conducted to clarify whether a characteristic pattern of interaction exists, whether the pattern is a necessary precondition of the pathogenesis of the disorder or occurs as a consequence of the disease, and whether the dysfunctional interactions prolong anorexic behaviors (Weiss, Katzman & Wolchik, 1985). Despite conflicting evidence, the importance of comprehending the relationship between the eating disorder and family interaction should not be overlooked. Whether or not a particular constellation of family characteristics is specific to the condition of anorexia nervosa, an understanding of the child's perceptions of her parents' relationship, as well as her place in it, gives insight into two important aspects of this disorder. its tendency to appear in adolescence and its notably greater incidence in girls than in boys (Gordon, Beresin & Herzog, 1989). To explain these findings does not require that one view the family as the source or origin of the problem. Regardless of how the illness arose, the symptoms of the eating disorder become intertwined with family relationships, causing the entire family to experience suffering and distress (Weme & Yalom, 1996).


Bemporad, J.R., & Herzog, D.B. (1989). Psychoanalysis and Eatiniz Disorders. Guilford Press: New York.

Blinder, B.J., Chaitin, B.F., & Goldstein, R.S. (1988) The Eating Disorders. PMA Publishing: New York.

Broberg, A. (1993). The anorectic family--an old-fashioned concept. Lakartidningen, 5@O 4550-4553.

Hall, L., & Cohn, L. (1992). Bulimia, A Guide To Recovory Gurze Books: CA.

Le Grange, D.C., & Rutherford, J. (1994). Redefining the psychosomatic family; family processes of 26 eating disorder families. International Journal of Eating Disorders. 3, 211-226.

Minuchin, S., Rosman, B.L., & Baker, L. (1978). PYchosomatic Families. Harvard University Press: Mass.

Weiss, L., Katzman, M., & Wolchik, S. (1985). Treating Bulimia. A Psychoeducational Approach. Pergamon Press: New York.

Weme, J., & Yalom, J.D. (1996). Treating Eating Disorders. Jossey-Bass Publishers: San Francisco.


Psychology Department

The Health Psychology Home Page is produced and maintained by David Schlundt, PhD.

Vanderbilt Homepage | Introduction to Vanderbilt | Admissions | Colleges & Schools | Research Centers | News & Media Information | People at Vanderbilt | Libraries | Administrative Departments | Medical 

  Return to the Health Psychology Home Page
  Send E-mail comments or questions to Dr. Schlundt


Search: Vanderbilt University
the Internet

  Help  Advanced

Tip: You can refine your last query by searching only the results by clicking on the tab above the search box

Having Trouble Reading this Page?  Download Microsoft Internet Explorer.