Family Systems Model of Family Influence
Around the same time that Bruch (1978) was publishing her theories on the development of AN, Minuchin, Rosman, and Baker (1978) developed a family systems model that explained eating disorders as a manifestation of disturbed family relationships. According to this model, all families are made up of subsystems (e.g., spousal, parental, and sibling) involving different roles and responsibilities. Boundaries between subsystems can range from enmeshed to disengaged. Enmeshment refers to poor differentiation of boundaries. For example, children might be invited to take part in transactions that would typically remain between husband and wife (spousal subsystem), as when one parent seeks advice from a child about how to deal with frustration over the other parent’s behavior. Disengagement refers to the existence of boundaries that create isolation among family members. For example, children might feel like an afterthought in the context of their parents’ busy lives. Clarity of boundaries may differ across family members and may change within a family over time. Indeed, family systems need to change in response to the normal processes of development (i.e., as children grow to adulthood) and in response to external events (e.g., the family moves).
Minuchin and his colleagues (1978) posited that families in which AN arose were marked by the following characteristics: enmeshment within the family; overprotection of children, resulting in rigid boundaries separating family members from extrafamilial relationships; conflict avoidance; and concern over bodily functions, including physical symptoms, eating, and diets. According to Minuchin et al.’s model, the lack of boundaries among members of these families triggered conflict as girls approached adolescence. However, conflict in these families had no means for outlet, because distress was suppressed behind a facade of closeness. Girls in enmeshed families were stifled in their attempts to achieve independence and establish relationships outside the family. Because there was no viable option for expressing interpersonal conflict, conflicts were somaticized; that is, they were expressed as physical conditions. Within Minuchin et al.’s (1978) framework, conflict could be revealed in a variety of physical maladies in any member of the family. Applied to AN, this model viewed the daughter’s symptoms as evidence not of intrapsychic conflict, as proposed by Bruch (1978), but of intrafamilial conflict. Thus treatment needed to focus on healing family interactions.
A number of studies have examined the quality of family relationships of individuals with eating disorders. These studies have generally indicated that the families of such individuals show greater dysfunction than do families of healthy controls (Holtom-Viesel & Allan, 2013). In a systematic review of the literature, Holtom-Viesel and Allan (2013) concluded, however, that there was no typical pattern of family dysfunction that characterized or differentiated among eating disorders. For example, studies have variously supported greater conflict, lower conflict, and no significant differences in conflict in families of individuals affected by eating disorders when compared with control families. Notably, the studies reviewed by Holtom-Viesel and Allan (2013) examined concurrent associations between eating pathology and family relationships. Thus the casual implications of any significant associations are unclear. Dysfunctional family interactions might be responses to eating disorders rather than contributors to their development.
Shoebridge and Gowers (2000) used a retrospective, follow-back design to examine whether overprotective parenting precedes the onset of AN. Compared with mothers of matched controls, index mothers were more likely to report providing near-exclusive child care (defined as 95% or more of child care to the exclusion of care by fathers), infant sleep difficulties, maternal distress at first separation, higher maternal trait anxiety, and a later age for the daughter’s first sleepover. In seeking to explain the maternal anxieties, Shoebridge and Gowers found that obstetric records revealed that 25% of the index mothers had experienced a severe obstetric loss (e.g., miscarriage) before the birth of the child in the study, compared with only 7.5% of the control mothers. In 90% of the cases of obstetric loss in index mothers, the daughter with AN was the next-born female child. Unsurprisingly, more index mothers than control mothers reported worrying significantly about miscarriage during the pregnancy that led to the birth of the child in the study. Of interest, no significant differences were found between index and control mothers’ reports of child eating difficulties in the first five years of children’s lives.
Taborelli et al. (2013) extended Shoebridge and Gowers’s (2000) approach by comparing maternal experiences of anxiety in relation to daughters with AN and their healthy sisters. Mothers reported having experienced greater anxiety while pregnant with the daughters who later developed AN and also reported waiting until an older age to leave these daughters in the care of another adult. Taborelli et al. (2013) found no differences in maternal anxiety or overprotectiveness between daughters with BN and their healthy sisters.
Byely, Archibald, Graber, and Brooks-Gunn (2000) used a prospective design to evaluate the relationship between girls’ perceptions of family relationships and changes in dieting and body image disturbance. Over a one-year follow-up period, 10- to 14-year-old girls reported increased problems in family relationships. Negative family relationships predicted dieting at follow-up even after controlling for dieting at baseline. Family relationships, however, did not predict changes in body image. Byely et al. (2000) also evaluated maternal modeling of dieting behaviors as a potential predictor of dieting and body image disturbance in girls and found no association.
Laliberte, Boland, and Leichner (1999) examined the effects of family relationships in terms of two categories of variables: “traditional family process variables (conflict, cohesion, expressiveness)” and “ ‘family climate for eating disorders’ . . . Family Body Satisfaction, Family Social Appearance Orientation, and Family Achievement Emphasis” (p. 1021). Both types of family relationship variables significantly predicted disordered eating attitudes and behaviors in college women, with the family climate for eating disorders demonstrating a stronger relationship. In comparisons of eating-disordered patients, depressed controls, and healthy controls, the family process variables distinguished both patient groups from controls but not from each other. The family climate for eating disorders, however, differentiated the eating-disordered group from both the depressed and healthy controls.
Examination of the role of family factors in the development of eating disorders thus far demonstrates some similarities across theories. First, most theories suggest a role for high parental expectations in the form reduced responsiveness to infant cues, which may contribute to a more critical family environment. Second, most theories acknowledge that familial concerns with food, eating, and weight may shape the content of difficult interactions. The next section examines the extent to which children may learn disordered eating attitudes and behaviors from their families’ disordered eating attitudes and behaviors. This social learning model is consistent with A. Stein et al.’s (2006) finding that duration of exposure to maternal eating disorders predicted severity of disordered eating in children.