Psychology

Mothers and Eating Disorders
Introduction
It is generally accepted that Western cultural beliefs about what it means to be fat or thin have a greater impact on women  than men in the West ( Wooley   Wooley, 1980). This suggests that increased vulnerability to eating disorders in women may, at least in part, be linked to recent cultural forces acting selectively on women. Relevant forces include female sexual liberation ( Bennett  Gurin, 1982), changes in the role of women, and changes in cultural demands and expectations. These pressures can be linked to specific beliefs, at the cultural or subcultural level, about what it means to be fat or thin. Among other things, it has been suggested that thinness represents a rejection of the feminine stereotype ( Szyrynski, 1973), andor an expression of female sexual liberation ( Bennett  Gurin, 1982), with the attributes or meanings associated with thinness (e.g., athleticism, nonreproductive sexuality) representing a kind of androgynous independence. Others, such as Bruch ( 1978), have emphasized the process rather than the meaning and have suggested that, faced with too many conflicting demands and choices, some women  choose weight and shape as means to control their fives or as realms to be successful in. This is described by Vitousek and Hollon ( 1990) as a New Years resolution style of cognitive process. As well as affecting vulnerability, it also is possible that cultural pressure acting on women  and not men may affect the expression of eating disorders. Western women are, traditionally and unlike men, rewarded for affiliation and dependence. Thus self-esteem, as well as being tied to internal or dispositional factors, also tends to be more strongly tied to interpersonal approval ( Bardwick, 1971). These two processes may be reflected in womens vulnerability to develop weight and shape assumptions tied to both self-acceptance and to acceptance by others.

Eating disorders are more likely to develop in adolescent girls and young women. It has been suggested that adolescence is more difficult for girls than for boys in Western culture ( Hsu, 1990). In particular, it has been suggested that the onset of puberty may make denial or avoidance of the expectations associated with adolescence more difficult for girls and that, in the West, recent cultural changes may magnify the stress.

Negative Self-Beliefs
The concept of schemas ( Neisser, 1967) may be relevant to understanding negative self-beliefs, particularly the concept of self-schemas. A schema may be defined as a cognitive structure that represents knowledge about a concept or type of stimulus, including its attributes and the relations among those attributes ( Fiske  Taylor, 1991). A self-schema is a schema about oneself that is, ones self-concept. Negative self-beliefs represent the content of schema. Such structures possess schematic properties that influence the way we perceive, understand, make sense of, and recall the information presented to us ( Eysenck, 1993). Cultural forces might influence the existence of negatively toned schemas, their structure, and the processes associated with them.

Typically, a self-schema includes objective information or descriptions of physical characteristics (e.g., Im female, Im tall, Im fat) as well as personality descriptors (e.g., Im honest, Im successful, Im caring). To date, it is usually the  more objective or physical descriptors that have been studied ( Markus, Hamill,  Sentis, 1987). While social cognition theory tends to treat these two types of schema as equivalent, this tendency is not apparent in cognitive theories of eating disorders. In particular, personality descriptors appear to exist at a deeper level of meaning than physical descriptors, particularly the belief that one is fat or thin. They can be derived from physicalobjective descriptors by asking, What does being (believing that you are) fat mean or say about you This level of meaning lies at the core of the self-concept in cognitive theory and therapy for eating disorders. Despite this difference in view of the self, the concept of self schema remains useful as a framework for furthering our understanding of cultural influences on eating disorders.

With respect to the existence of negatively toned schemas, preliminary evidence suggests that the self-concept or self-schema may be similar in women with eating disorders and women with depression, at least in Western cultures ( Cooper  Hunt, 1998) that is, in the presence of negative self-beliefs. With respect to cultural differences, Koenig ( 1997) notes that depression may manifest itself in different ways in West and non-Western cultures. In particular, cognitivemotivational symptoms (e.g., guilt, self-deprecation, despair, suicidal ideation) seem to be more common in Western cultures. Koenig links this difference in symptoms to differences in the cultural self way ( Markus, Mullally,  Kitamaya, 1997).

Selfways are characteristic ways of engaging in the social world. Defined as a pattern that establishes or strengthens certain kinds of self-concepts ( Neisser, 1997, p. 5), selfways, like the individual self-concept, have schematic properties. They guide what people notice and think about, what they feel moved to do, what they feel, how they feel, and how they organize, understand and give meaning to their experiences ( Markus, Mullally,  Kitamaya, 1997, p. 15).

In Western cultures, the cultural selfway is reflected in the drive to be distinct and unique, as compared to many non-Western cultures where the drive is to fit in and be a useful member of the group. This difference means that, in the West, negative affect is linked to low self-esteem, which is then reflected in negative self-beliefs. In non-Western cultures, however, negative affect is linked to not fitting in, and does not have the effect of encouraging the development of negative self-beliefs. This may, at a broad level, increase vulnerability to the development of an eating disorder.

Within a specific cultural selfway, and consistent with the suggestion made earlier, Markus et al. ( 1997) also note that females experience may be rather different from that of men, particularly in the West where cultures are more likely to emphasize dependent selves to women that is, maintaining relationships and connection to others. However, in the West, connecting to others is a distinguishing attribute of the self, not fitting in or being part of a relationship. This means that women in the West appear to have both positive, strong, articulated selves and an emphasis on connectivity and relatedness to others. These two themes in the self-concept may help to explain, more clearly and in more detail  than previous researchers have done, why women with eating disorders are concerned not only about weight and shape as a means to enhance their own selfesteem, but also about weight and shape as a means to be accepted by others.

Differences in the type of selfway developed and the emphasis on a dependent self for women in Western cultures may affect vulnerability to eating disorders. Different cultures may also stress different kinds of self-concept. The type found in Western cultures may not be universal and may also relate, at a general level, to vulnerability to eating disorders. For example, the Penobscot Indians have a concept of the self in which each individual is made up of two parts, the body and a vital self dependent on the body but able to have out of body experiences ( Speck, 1920). This structure differs from the coherent, whole, and integrated structure characteristic of the Western self-concept.
Attributional theory may also be extended to negative self-beliefs, specifically to the attribution of responsibility for the belief that one is, for example, a failure or worthless. As with underlying assumptions, responsibility for being worthless or a failure can be located in the self or in external factors. As with its application to underlying assumptions, both the inferential processes and the heuristics used to construct explanations may be open to cultural influence. Again, differences between cultures in these may affect the likelihood of developing eating disorders.

Schema-Driven Processing
Attributional theory is also applicable to schema-driven processing. In relation to eating disorders, the self-serving bias (i.e., the tendency to enhance or protect self-esteem) is relevant to schema compensation in which dieting is seen as a way to overcome or make up for perceived negative qualities. The existence of this bias is open to cultural influence. In the developmental cognitive theory outlined above, schema compensation processes reflect dieting as a means to overcome negative self-beliefs. This process, self-esteem enhancement, seems analogous to the concept of self-serving biases that, within the context of attributional theory, serve to protect or enhance self-esteem or self-image.
While research in Western cultures finds self-serving and self-protective biases in which individuals take credit for their success and attribute it to their own abilities, the opposite pattern is found in other cultures. In Japan, for example, individuals explain success in terms of situational factors ( Kitayama, Takagi,  Matsumoto, 1995). Cultural differences in the existence of the self-serving bias may contribute to individual vulnerability to developing eating disorders.

Culture and the Expression of Eating Disorders
It seems clear that there are cultural differences in the expression of eating disorders, both currently across cultures and historically within cultures. Cases of eating disorders in some less developed countries do not always appear to have  all the typical features of those in Western cultures. For instance, fear of fatness may not be common in non-Western anorexia nervosa ( Khandelwal, Sharan,  Saxena, 1995). Instead, decrease in food intake may be related to fasting for religious purposes or to eccentric nutritional ideas. It has been suggested that degree of Westernization may affect findings such as this ( Fedoroff  McFarlane, 1998), with atypical presentations being more common in less Westernized cultures. Historically, symptoms are not always related to a sense of fatness or to pressure against fatness ( Littlewood, 1995).

Cultural differences can, within a cognitive framework focused on the cultural meaning system ( DAndrade, 1984), explain these differing manifestations. In Western cultures, DSM-IV ( American Psychiatric Association, 1994) provides a set of rules that create eating disorders. These include a focus on dieting or not eating to avoid becoming fat. An examination of subjective experience can highlight many different culturally determined meanings for refusal to eat. In the Middle Ages one got closer to God through fasting. It was this that made fastingdieting good and acceptable. Exploring differences in cultural meaning also may help to explain the rather different meanings attached to weight loss and dieting. For example, the emphasis on control ( Fairburn, Shafran,  Cooper, 1999), as opposed to other outcomes (e.g., self-acceptance), may simply be another instance of a culturally determined meaning attached to dieting andor fasting.

One group in Western countries which is frequently exposed to involuntary under-nourishment and malnutrition is the homeless. The numbers of people who become homeless is rising and mothers with dependent children make up 65 per cent of homeless people and are the most rapidly expanding group in the homeless population most frequently they become homeless following domestic violence (Vostanis et al., 1996). Food deprivation is reported particularly by mothers with children (DiBlasio and Belcher, 1995). For those living in hostels for the homeless, providing food for the family may be less of a concern, since food is often provided. However, the need for hostels to provide food on a limited budget results in a diet that is high in fat (Killion, 1995). Pregnancy rates for homeless women are twice the normal rate and pregnant homeless women have a number of experiences that are associated with complications of pregnancy, including malnutrition (Killion, 1995). Contributing factors to malnutrition among the homeless are poor access to transportation, which means the homeless obtain food from local stores that are often more expensive and stock little fresh food. The lack of usual cooking, storage and refrigeration facilities encourages them to eat more filling convenience foods that do not meet their nutritional requirements. Furthermore, food obtained is sometimes stolen from them. The most extreme under-nourishment and malnutrition is associated with high substance use (Killion, 1995).

Though there is a continuum from good to poor eating practices, poor practices are common. At the extreme they become labeled eating disorders. There are a range of proposed eating disorders, but agreement about their definition is sometimes poor. Cases where there is wide agreement include anorexia nervosa and bulimia. However, probably the most common disorder is compulsive eating. There is variability in the estimates of the prevalence of eating disorders, but it is thought that in Western society about 1 per cent of women will experience anorexia at some time in their lives and 5 per cent will experience bulimia. The prevalence appears to have increased over the last thirty years.
Anorexia nervosa is characterized by a severe loss of weight and a dream of becoming fat and frequently an associated sense of guilt when consuming food. It is predominantly experienced by girls and young women. Anorexia can have a permanent impact on health. An inadequate diet leads to an inadequate supply of vitamins and minerals including low levels of electrolytes such as sodium and potassium. This can result in extreme low blood pressure, heart damage, or cardiac arrhythmia. There has been speculation that anorexic women have a distorted body image, believing themselves to be fat when they are not. Despite the diagnostic criteria of the American Psychiatric Association (Table 1.1), evidence suggests that anorexic women feel fat but are as accurate as other women in making judgments about actual body size and shape nevertheless, there is some distortion in the body image of both eating disordered and normal eaters.

1 Refusal to maintain minimal normal weight Body weight is less than 85 of that expected for age and height 2 Fearful of becoming fat or gaining weight 3 Disturbance in body weight and shape perception. Body image linked to self-esteem 4 Absence of at least three normal non-drug-induced consecutive menstrual cycles Source American Psychiatric Association (1994) TABLE 1.1 Diagnostic criteria for anorexia nervosa

A recent study examined a number of dimensions relating to body image, including fear of fatness, preference for thinness, body size distortion, body dissatisfaction and actual body size, and looked at their relationship to restrictive eating. Fear of fatness, preference for thinness and body size distortion had an effect on body dissatisfaction over and above the effects of actual body size. Fear of fatness was the best predictor of restrictive eating (Gleaves et al., 1995). Bulimia nervosa is associated with binge eating and a feeling that eating is out of the individuals control (see Table 1.2). Binge eating is accompanied by depression and self-deprecation. Self-induced vomiting occurs in the majority of cases. Bulimia can cause a wide range of medical problems including inflammation of the digestive tract and, as with anorexia, it is associated with cardiac problems resulting from low levels of electrolytes.

1 Recurrent episodes of binge eating accompanied by a sense of loss of control 2 Recurrent compensatory behavior, e.g. self-induced vomiting, use of laxatives or enema 3 Binges and compensating behaviors occurring about twice a week for three months 4 Body shape and weight are critical in self-evaluation 5 This experience does not only occur during a period of anorexia nervosa Source American Psychiatric Association (1994) TABLE 1.2 Diagnostic criteria for bulimia nervosa
Weight loss programs that involve the parents are more effective than those that do not. This is partly because obesity runs in families and so weight loss is easiest to achieve when the whole family changes its eating behavior as has been noted above, parents influence their childrens eating and activity levels. One important feature of programs that involve parents is that fewer children drop out of the program. Programs that involve parents are most effective if they provide skills training, such as training parents in behavior modification techniques and problem-solving strategies (Brezinka, 1992). Involving parents in behavior modification programs results in greater weight loss in children than when the family is not involved. Furthermore, involved parents also lose weight (Brownell and Cohen, 1995). The inclusion of physical exercise in weight loss programs assists weight loss among adults, but has less impact on childrens and adolescents weight. Children achieve similar levels of weight loss and maintain weight loss equally on programs that do and do not include exercise. The inclusion of exercise improves fitness but is not a necessary component of weight loss programs for children and adolescents (Brezinka, 1992).

It is widely accepted that parent child relationships play a central role in childrens psychological development and so we will focus largely on these. The quality and form of these relationships are thought to predict later interpersonal relationships and have a profound inuence on personality development and related psychological functioning, such as in the areas of self-esteem and social condense. Surprisingly, there has been until relatively recently, only a small body of good empirical evidence to support the importance of these relationships in development. It is largely within the area of attachment that a lucid and comprehensive theory of early relationships has evolved. The grounds for attachment theory  were laid by Bowlby from the late 1950s (Bowlby, 1958, 1969, 1973, 1980). However, it only formally emerged as a scientic discipline in the 1980s with the development of appropriate research instruments. A means of assessing individual differences in attachment behavior was later developed through the work of Mary Ainsworth (Ainsworth et al., 1978). This has been enormously helpful in enabling intensive, ongoing attempts to assess the psychological effects of early relationships on development. Parental anxiety may be a constitutional trait or arise from a specic child-related antecedent such as a problematic past obstetric history. The index child may be born after a period of infertility or the loss of an earlier child.

Parental ill-health can have an effect on the quality of attachment through a number of routes. The parent may be unavailable either physically or emotionally and the relationship may suffer frequent disruptions. Where disruptions occur without warning, for example as a result of emergency hospitalization, they are likely to be particularly bewildering, while childrens ability to understand the implications of ill-health will be governed by their developmental maturity. Serious ill-health in one parent may be compensated for by a good quality relationship with the other, or a substitute caregiver, but healthy parents are likely in turn to be adversely affected in terms of their physical availability and their own psychological adjustment to their spouses illness.

Mental ill-health in a parent poses particular challenges for the relationship with the child and his or her psychological development. A parent who is anxious or fearful of the world may transmit such attitudes, while depressive mood will have an impact on emotional responsiveness and availability and may present a gloomy outlook on life. Psychotic disorders may confront a child with both disturbances of behavior and belief. For an older child a negative effect on peer relationships may ensue, with the child being reluctant to bring friends home, he or she may suffer bullying.

Eating disorders provide a good example of the effects of the interplay between parental attitudes and behaviors (in themselves and directed to the child). Mothers who diet or have weight concerns themselves are more likely to bottle feed (Crisp, 1969). The mothers attitude to feeding is likely to be a more important inuence than the direct effect of bottle feeding on infant growth.

Childrens perceptions of their parents health in turn impact on their feelings of security and their view of their own health and resilience. The development of an external locus of control or a feeling of personal ineffectiveness may be particularly potent as risk factors for eating disorders.

Parental Attachment Status
There are good grounds for supposing that parents own experiences of being parented and their attachment relationships in their families of origin will predict the quality of their own childrens attachments. Until recently there were few ways of linking intergenerational attachment representations.
The Adult Attachment Interview (AAI George et al., 1984) is a semi-structured interview based measure for adults which enable evaluation of the quality of past attachments in childhood. It is designed to assess the adults state of mind with respect to attachment, by enquiring about relationships in childhood and evaluating the coherence of their accounts.  It is concerned not so much with what happened as what the subject feels about what happened and whether what the subject says is backed up by evidence. The general quality of child caregiver relationships is probed, together with experiences of early separation, illnesses, losses, rejection, and maltreatment the interviewer probes for specic memories to illustrate general statements. There are three main categories in this classication system Free-autonomous (F), Dismissive (D), and Preoccupied (E). Secure adults categorized (F) are said to value intimate relationships, and acknowledge their effects. In addition, some interviews are characterized by an apparent failure to resolve mourning over loss or abuse, and are separately classied as Unresolved (U). Subjects who simultaneously possess E and D qualities are described as Cannot Classify (CC). All interviews are rated on a number of scales concerning Probable Experience (of attachment gures) and current State of Mind of the interviewee these scales contribute to the overall classication (DEFU).

The AAI has been shown to have predictive validity for the quality of infant attachment in the next generation (van IJzendoorn, 1995), in as much as two-thirds of infant attachments on the ABCD classication match their parent s attachment category on the AAI.

Other Inuences on Attachment
Genetics
There is a growing body of evidence from twin studies to suggest a signicant genetic contribution to attachment patterns. It may be that genetic contributions to the temperamental component of distress proneness is greater than that for securityinsecurity (Goldberg, 2000).

Siblings
Brothers and sisters can facilitate or impair attachment formation in a number of ways. They may display jealousy towards the new child or else their behavior or temperament may inuence parental expectations of subsequent infants.

Living Conditions
Good quality attachment is likely to occur when the family is not pressed by nancial hardship or overcrowding. Good quality family relationships can, however, overcome severe material deprivation.

Effects of Attachment beyond Infancy
Bowlby considered attachment to be a feature of signicant relationships throughout the lifespan, as early experiences are coded as internal working models, which are then carried forward to inuence later personality and behavior. These internal models contain both affective and cognitive information. Emotional expressions become more complex and subtle as children get older and they learn implicit rules about displaying affect, including the masking of negative emotions (Lewis  Michalson, 1983).

Attachment theory has been concerned with two aspects of emotional development the way in which attachment  gures respond to affect and the ways in which attachment relationships vary with individual differences in emotional expression and regulation. Attachment relationships are also thought to inuence information processing through their effects on selective attention and memory. By the age of about 6, children develop a theory of mind, i.e. a notion that they and others have thoughts about the world, which may not be the same and which are independent of external objects. Internal working models of attachment come prominently to include attributions of key relationships, e.g. Mummy likes to bake cakes for me.

Attachment and Health
The notion that internal working models play a key role in linking early attachment experiences to later social and psychological consequences is appealing, but not widely tested. A simple hypothesis might be that insecure attachment increases vulnerability to behavioral problems or psychological disorder. The second part of this chapter will address the evidence for associations between attachment style and the development of eating disorders, followed by a review of non-attachment developmental inuences.

Attachment And Eating Disorders
Historical Perspective
Hilde Bruch, writing in the 1970s, linked emergent attachment ideas to her clinical observations. In her seminal work, Eating Disorders Anorexia Nervosa, Obesity and the Person Within (1974), she offers an unusual insight into Mary Ainsworths thinking, as the precursor of the Strange Situation appears associated with early mother infant feeding interactions

When rated at 12 months, the infants in whom the feeding interaction had been most appropriate to their needs, permitting them active participation, showed the strongest attachment to their mothers, with a clear-cut tendency to seek her proximity, and to express distress at her absence. They made active efforts to gain and maintain contact with her. In contrast babies with inappropriate feeding experiences, showed little or no tendency to seek proximity, interaction, or contact with the mother and little or no tendency to cling when picked up or to resist being released. They tended either to      ignore the mother on her return, or to turn away or go away from her. There was a third group which included the one of pseudo-demand with overfeeding (i.e. mother impatient with the baby and staves them off with food), in which children were distressed by separation, but showed less ability to use the mother as a secure base from which they could enjoy exploring the strange environment. They generally displayed more maladaptive behavior in relation to new and strange situations.

(Ainsworth  Bell, 1969 quoted in Bruch, 1974)
Developing these ideas to describe her own patient population, Bruchs formulation was of a mother child interaction in which mother does not respond appropriately to her infants needs, instead superimposing her own needs such that the infant does not learn to discriminate self. The infant adapts, such that the situation may pass unnoticed throughout a well-behaved childhood, only becoming evident as the adolescent striving for autonomyidentity can no longer be ignored. Although this is clearly an oversimplication of Bruchs rich theory, it serves to highlight the interweaving of attachment ideas with some of the earlier clinical literature, providing a basis from which to explore subsequent developments.

Background Research
Since 1996 there has been increased attention to the topic and the use of instruments has tended to become more rened. In particular, the Adult Attachment Interview (AAI), currently regarded as the gold standard in attachment research, has been applied to eating disordered populations. Thus one might anticipate that greater clarity would emerge from a later review. Questions have also been asked about the association between attachment style and eating disorder diagnosis. Following Bowlby, insecure attachment is usually classied as anxiousresistantavoidantdismissive, or angrypreoccupiedenmeshed. Although no association between attachment status and eating disorder subtype emerged from OKearneys (1996) work, there was a belief that such an association might emerge in a larger review with more rened instruments. Clinically, anorexic women appear avoidant, not only of food but of life in the raw, whereas bulimic women are more often angry and chaotic.