Satsanga.buffalo.edu

Journal of Child Psychology and Psychiatry 45:1 (2004), pp 63–83 Management of child and adolescent eating 1University of Liverpool, UK; 2School of Medicine, University of Southampton, and Great Ormond Street Hospital NHS Although eating disorders in children and adolescents remain a serious cause of morbidity and mor-tality, the evidence base for effective interventions is surprisingly weak. The adult literature is growingsteadily, but this is mainly with regard to psychological therapies for bulimia nervosa and to someextent in the field of pharmacotherapy. This review summarises the recent research literature coveringmanagement in three areas, namely physical management, psychological therapies, and service issues,and identifies prognostic variables. Findings from the adult literature are presented where there is goodreason to believe that these might be applied to younger patients. Evidence-based good practice rec-ommendations from published clinical guidelines are also discussed. Suggestions for future researchare made, focusing on 1) the need for trials of psychological therapies in anorexia nervosa, 2) applica-tions of evidence-based treatments for adult bulimia nervosa to the treatment of adolescents, and 3)clarification of the benefits and costs of different service models.
interval; RCT: randomised controlled trial; RR: relative risk.
This review addresses current knowledge and rec- Children and adolescents may also present with ommendations about the management of eating other types of clinical eating disturbance, including disorders in young people between the ages of 8 and ‘food avoidance emotional disorder’ (Higgs, Goodyer, 18. It does not describe or refer to the literature on & Birch, 1989), ‘selective eating’ (Nicholls, Christie, feeding problems and eating difficulties in younger Randall, & Lask, 2001), and other phobic disorders children, which are common, but present with very with eating difficulties as prominent presenting fea- different symptom patterns (see, e.g., Crist & Napier- tures (Bryant-Waugh, 2000). Management of these Phillips, 2001, and Hutchinson, 1999, for reviews other types of eating disturbance does not form part and data on feeding problems in this younger age- of this review, as they appear to be quite distinct from group). The term eating disorder is used here to the classic eating disorders of the AN and BN type in indicate anorexia nervosa (AN), bulimia nervosa (BN) terms of core psychopathology, the characteristic and associated disorders. Anorexia nervosa can overvaluation of weight and/or shape being absent arise from the age of around 8 years, whilst full (Cooper, Watkins, Bryant-Waugh, & Lask, 2002).
bulimia nervosa appears very rare in those under 12 For the purposes of this review ‘children’ will (Bryant-Waugh, 2000). Clinically significant variants generally refer to those between the ages of 8 and 12, of AN and BN do, however, occur in children and and ‘adolescents’ to those between 13 and 18.
adolescents, probably at higher rates than full syn- However, the literature is very limited in terms of drome disorders. Such presentations usually involve children, whilst those over 16 are often treated as a significant preoccupation with food, weight or adults and included in adult research series.
shape, accompanied by eating disturbance, but do Difficulties in matching clinical presentations seen not meet full criteria for AN or BN. Some of the dif- in childhood and adolescence to existing diagnostic ficulties in applying diagnostic criteria are specific to criteria for eating disorders particularly arise in the this younger age group, whilst others reflect wider case of AN in children (see Nicholls, Chater, & Lask, problems in terms of matching clinical populations 2000). Whilst this may be less of an issue in terms of to the existing classification systems. For example, it general clinical practice, it can pose a significant is known that across all ages, around half of all problem for researchers, who need to be able to eating disorder patients do not meet full criteria for define groups of patients studied with a degree of AN or BN (Turner & Bryant-Waugh, in press; Ricca et al., 2001). Current classification systems attempt In children and adolescents there tend to be two to address this issue within the diagnostic category types of age-related problem: firstly, there are diffi- ‘Eating Disorders Not Otherwise Specified’ (EDNOS) culties inherent in the reliable assessment of psy- (DSM-IV, American Psychiatric Association, 1994) or chopathology in this age group, and secondly, there ‘atypical’ anorexia or bulimia nervosa (ICD 10, World are those related to the strict application of existing diagnostic criteria. Assessment can be difficult, as Ó Association for Child Psychology and Psychiatry, 2004.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA children in particular may be unable to describe their exclusively on children, and adolescents have often thoughts, attitudes and behaviours clearly, or be been included in population-based studies of adults.
unwilling or scared to do so truthfully. This means A recent chapter reviewing epidemiological studies that information from parents or others can be suggests that the average prevalence rate of AN in helpful, but should not form an exclusive alternative, young females is around .3%, with incidence rates and underlying psychopathology should not be highest for females aged 15–19, who represent inferred. It also means that care needs to be taken to approximately 40% of all identified cases and 60% of elicit information using age-appropriate measures female cases (van Hoeken, Seidell, & Hoek, 2003). The and means. Standardised assessment of eating age range in the included studies was from 11 to 35 disorder psychopathology has been difficult in early- years, with various different screening methods and onset cases in part because of the lack of psycho- diagnostic criteria used. The four studies in the review metrically sound measures for use in this age group.
dating from 1993 onwards are all of adolescents (aged Some child adaptations of existing adult measures 11–20), with an average prevalence rate of .5% do exist, for example the child version of the Eating (Rathner & Messner, 1993; Wlodarczyk-Bisaga & Attitudes Test – ChEAT (Maloney, McGuire, & Seli- Dolan, 1996; Steinhausen, Winkler, & Meier, 1997; bowitz, 1988), the child version of the EDI (Eating Nobakht & Dezhkam, 2000). Incidence and pre- Disorder Inventory), known as the KEDS (Kids Eating valence rates of AN in males are more rarely reported, 1 Disorder Survey) (Childress et al., 1993), and the but it has been noted that where they are, the female to child version of the Eating Disorder Examination male ratio is around 11 to 1 (van Hoeken et al., 2003).
(Bryant-Waugh, Cooper, Taylor, & Lask, 1996).
For bulimia nervosa, two-stage surveys of pre- There are also other measures related to DSM-IV valence rates in 11–20-year-olds published since diagnostic categories, primarily designed for use with 1993 suggest an average rate of just under 1% children and adolescents, which have eating disorder (Rathner & Messner, 1993; Wlodarcyzk-Bisaga & modules (e.g., the Diagnostic Interview for Children Dolan, 1996; Santonastaso et al., 1996; Steinhausen and Adolescents (DICA), which comes in both a child et al., 1997; Nobakht & Dezhkam, 2000). Reported and an adolescent version – Welner et al., 1987).
female to male ratios in the incidence of bulimia However, all these measures tend to suffer from a nervosa range from 33:1 to 27:1 (van Hoeken et al., range of psychometric and practical drawbacks.
Difficulties related to the strict application of A recent Office for National Statistics survey (2000) existing criteria include the continuing lack of reported a prevalence rate of eating disorders gen- agreement and consistency regarding the ‘weight erally in UK 11–15-year-olds of 4 per 1,000, whilst a criterion’. Both the ICD-10 and DSM-IV classifica- catchment area total population study in the NW of tion systems require a generally similar level of body England (the TOuCAN trial) found that the median weight. In ICD-10, this is ‘at least 15% below normal number of referrals to a generic Child and Adoles- expected weight for age and height’, and in DSM-IV: cent Mental Health Service is 3 cases per year ‘refusal to maintain weight at or above a minimally normal weight for age and height – (e.g., .body The question of whether rates of eating disorders weight less than 85% of that expected.)’. It is very are increasing in the younger age group is difficult to difficult to apply such a criterion cleanly as the cal- answer, despite repeated assertions in the popular culation of expected weight needs to take into ac- press that this is the case. Some argue that apparent count any stunting of height as a consequence of increases in incidence can be accounted for by a dietary restriction, plus, ideally, access to premorbid number of factors, including general population weight and height charts as well as parental height trends, changes in access to and the use of health and population norms. In adults a diagnostic cut-off care, and improved recognition of eating disorders.
of BMI at or below 17.5 is often used (e.g., Treasure, Perhaps of more immediate relevance to healthcare 1999), but this is not useful in a younger population providers and clinicians is the fairly widespread where BMI norms are age and gender specific. An observation that over the past decade there has been equivalent cut-off of BMI below the 2nd centile has an increase in the numbers of children and adoles- been proposed for children and adolescents (Royal cents presenting for treatment, and that healthcare College of Psychiatrists, 2002), but not universally expenditure in relation to eating disorders has risen accepted. The interpretation of such criteria and how they are applied to young people varies greatly,raising issues about consistency in diagnosis.
It is difficult to be specific about the incidence and prevalence of eating disorders in children and ado- Anorexia nervosa. Case series of anorexia nervosa lescents. Varying rates have been reported and some have existed in the medical literature for well over of this variation is likely to be due to inconsistencies in 100 years (Gowers, 2001) and there is now a detailed the definition and diagnosis of eating disorders, as body of evidence from cohort studies detailing out- well as the method used in the process of case iden- come and prognostic factors. However, research into tification. No epidemiological studies have focused the aetiology and psychology of the condition Management of child and adolescent eating disorders significantly outstrips that on management and the prevalent than AN in young females (van Hoeken evidence base for the efficacy of treatments across all et al., 2003), an older mean age of onset means that age groups is very weak (Treasure & Schmidt, 2002; this, too, is a rare disorder in younger adolescents.
NICE, 2003). Treasure and Schmidt (2003), on the The systematic reviews that have been published basis of their systematic review, concluded that have unanimously found benefits for cognitive there was very little Level I or Level II evidence (that behaviour therapy (CBT) in improving the specific obtained from one or more randomised controlled symptoms and eating behaviours of bulimia nervosa trials (RCTs)) to support specific interventions for and non-specific symptoms such as depression.
anorexia nervosa at any age. Indeed this review found support for only two positive conclusions and therapy (IPT), have yielded more modest findings (see these were tentative. Firstly, there was limited evid- below). Whittal et al. (1999) and Bacaltchuk et al.
ence from one RCT (Dare, Eisler, Russell, Treasure, (2000), in reviews of antidepressant treatments, both & Dodge, 2001) that various psychotherapies, in- found short-term reductions in bulimic symptoms cluding focal therapy, cognitive analytic therapy and a small reduction in depressive symptoms. One (CAT) and family therapy, were more effective than systematic review (Bacaltchuk et al., 2000) found ‘treatment as usual’ (non-specific routine follow-up evidence for advantages of combination therapy by a junior psychiatrist) for adults. Secondly, they (antidepressants plus psychotherapy) in producing found limited evidence from one small RCT (Crisp remission and mood compared with antidepressants et al., 1991) that outpatient treatment was as alone, but not in reducing binge frequency.
effective as inpatient treatment in those adolescentsand adults not so severely ill as to warrant emer-gency medical treatment. A further ten RCTs failed to detect a difference in the efficacy of various psycho- In the absence of a strong body of research in the therapies, or between psychotherapy and dietary child and adolescent eating disorder literature, it is advice, whilst a similar number of controlled drug tempting to draw conclusions from adult findings, trials failed to provide good evidence for their effect- but one should carefully consider the validity and aspects of the disorder. These reviewers noted, however, the small size of many of the trials, whichwere unlikely to have been powered to detect a dif- • Adolescence is a developmental stage which is not ference between treatments had there been any, and defined merely by age. It can be argued that many also a wide variability in the quality and reporting of young adults with eating disorders are still in the studies in terms of the CONSORT standards (Moher, throes of addressing the challenges of adolescence and indeed adolescent developmental difficulties There are no comprehensive systematic treatment have been thought to underlie the aetiology of reviews focusing on the child and adolescent age anorexia nervosa in particular (e.g., Crisp, 1995).
group (Treasure & Schmidt, 2003). In part the ex- • The essential features of anorexia nervosa and planation could be that anorexia nervosa of ICD-10 bulimia nervosa are consistent across the age or DSM-IV diagnostic severity is a disorder of rela- spectrum – in terms of characteristic behaviours tively low incidence, such that non-specialist ser- vices recruit at a low rate. In addition, as already pathology (over-evaluation of the self in terms of discussed, children may receive a diagnosis of atyp- weight and shape) and non-specific features (low ical AN or EDNOS because existing diagnostic cri- teria are not sufficiently developmentally sensitive.
Other obstacles to treatment research are reviewed • Much of the literature reports combined adoles- cent/adult case series without separate analysis.
• Finally, some of the treatments which have been Bulimia nervosa. This is a condition which has only found to be effective in adult eating disorders are appeared in the literature for a quarter of a century effective in the treatment of adolescents with other (Russell, 1979); however, treatments were beginning conditions; that is to say, it is not developmentally to be developed and tested within two years of its inappropriate to use them in this age group.
description (Fairburn, 1981). Indeed the body of Examples include the use of cognitive behaviour treatment research for bulimia nervosa is now much therapy (CBT) (Harrington, Whittaker, Shoebridge, greater than that for anorexia nervosa and generally of & Campbell, 1998) and antidepressants (Alderman, better quality (Treasure & Schmidt, 2003). A number Wolkow, Chung, & Johnston, 1998) in adolescent of systematic reviews have been published (Whittal, depression and obsessive-compulsive disorder.
Agras, & Gould, 1999; Bacaltchuk, Hay, & Mari, 2000; Hay & Bacaltchuk, 2002), but no controlledadolescent treatment trials have as yet been reported.
• In younger patients, eating disorders less com- Although BN is said to be around three times more monly fall neatly into the ICD-10 or DSM-IV categories; that is to say, atypical forms (EDNOS) prior to this date, reference to earlier work is occur more commonly. As treatments for EDNOS included. It covers management in the areas of are poorly developed in adults, there might be little physical management, psychological therapies and service provision, and identifies prognostic features.
• The treatment aims, particularly in AN, are often different in adolescence, because of the differentphysical issues involved, i.e., where the onset is before growth and development are complete, The three key clinical areas included in this review treatment needs to address the completion of (physical management, psychological therapies, and puberty and growth in psychological as well as service issues) were identified in association with physical terms (Nicholls & Bryant-Waugh, 2003).
members of the UK National Institute for Clinical Whereas in the treatment of adults with AN re- Excellence (NICE) Guideline Development Group covery usually involves returning to a pre-morbid (NICE, 2003). Following electronic searches for sys- healthy physical condition, in younger patients it tematic reviews and high-quality randomised con- may be more a case of discovering and adjusting to trolled trials addressing treatment efficacy in the a new state. In terms of weight targets this requires general field of eating disorders, specific attention constantly revising upwards as healthy weight is was paid in this review to evidence relating to chil- recalculated with the attainment of greater height dren and adolescents. A search was also undertaken (Gowers, 2001). All this might indicate a need for a of published and unpublished clinical guidelines.
longer duration of treatment for younger cases, Electronic searches were made of the major elec- whilst a shorter duration of illness before treat- tronic databases (MEDLINE, EMBASE, PsychINFO, ment, in some, might argue for the opposite.
CINAHL), the Cochrane Database of Systematic • When considering the literature on pharmacother- Reviews, the NHS R & D Health Technology Assess- apy, one should be aware of the different phar- ment database and Evidence Based Mental Health & macodynamics and pharmacokinetics in children.
In general the latter means that children andadolescents require higher doses of drugs per kgbody weight to attain similar blood levels and therapeutic effect, owing to the child’s more rapidliver metabolism and more efficient clearance by The accepted management of child and adolescent the kidney (Cawthron, 2001). A number of psycho- eating disorders is based mainly on expert clinical tropic drugs are not licensed for use in children, opinion and cohort studies rather than research possibly limiting pharmacology trials.
trials. A number of academic bodies (The American • Irrespective of any consideration of aetiological Psychiatric Association (APA), 2000; The Royal Col- variables, parents will usually need to be involved lege of Psychiatrists, 2002; the National Institute for in the management of younger patients (Lock, Le Clinical Excellence (NICE), 2003; the Finnish Med- Grange, Agras, & Fairburn, 2001). This is espe- ical Society – Ebeling et al., 2003; The Society for cially so if they are at risk and parental involve- Adolescent Medicine – Kreipe et al., 1995) have ment is believed likely to reduce the risk. The published consensus guidelines, the last two spe- treatment of both AN and BN includes aspects of cifically in relation to the management of children behavioural management and parents will need to and adolescents. There is much greater emphasis in be involved if handling these is to be effective; at a these on the physical management of AN than of BN.
practical level, parents usually have a role in In the absence of RCT findings, the key issues in shopping for food, meal planning and mealtime these guidelines with respect to physical manage- management. The involvement of siblings is gen- erally regarded as beneficial, for the sibling if notfor the patient, as this provides an opportunity for them to express fears or guilt and to dispel anyfalse ideas about the nature of the condition, its Kreipe et al. (1995), in the Society for Adolescent likely causes and prognosis (Lock et al., 2001).
Medicine’s position paper on eating disorders in • Finally additional attention will need to be given to adolescents, refer to the potentially irreversible ef- the different social and educational needs of this fects on physical growth and development and argue age group in treatment, particularly when treated that the threshold for medical intervention in ado- in hospital (Nicholls & Bryant-Waugh, 2003).
lescents should be lower than in adults. Of particu-lar importance, they say, is the potential forpermanent growth retardation if the disorder occurs before fusion of the epiphyses, and impaired bone This review provides a comprehensive summary of calcification and mass during the second decade of research published since 1993. In a small number of life, predisposing to osteoporosis and increased areas, where practice is influenced by key research fracture risk later on. They say that these features Management of child and adolescent eating disorders emphasise the importance of immediate medical consider a weight at (Ebeling et al., 2003) or close to management and ongoing monitoring by physicians (Lask, 1993) 100% weight for height to be desirable, who understand normal adolescent growth and based on findings from ovarian ultrasonography.
Medical complications can occur in younger sub- jects before evidence of significant weight loss (Kre-ipe et al., 1995). In treating the malnourished There is little in the guidelines to direct the physical patient, care should be taken to avoid the re-feeding management of BN. A key objective in planning syndrome, by regular monitoring of heart rate, dietary programmes is to break the vicious cycle be- tween dieting and binge eating (Ebeling et al., 2003).
including phosphorus, glucose, magnesium and Lethal medical complications are rare in BN (NICE, potassium (Royal College of Psychiatrists, 2002).
2003), but trauma to the gastro-intestinal tract, fluid This review, however, has drawn attention to the and electrolyte imbalance and renal dysfunction can limitations of serum electrolyte levels in assessment occur. As in anorexia nervosa, attention to the adverse of total body electrolytes, which may be depleted dental effects of vomiting and specific preventative with normal serum levels. Also, it notes that guidance on oral hygiene is recommended (NICE, re-feeding syndrome is more common with paren- 2003; Ebeling et al., 2003). Neither Kreipe et al.
teral than enteral feeding and regular serum elec- (1995) or Rome et al. (2003) make significant refer- ence to aspects of the physical management in BN.
frequently in those eating food in hospital.
nutritional management, Kreipe et al. (1995) state In the absence of research trials in the area of that adolescents have specific nutritional require- physical management, the field is very open. Some ments, taking into account their pubertal status and areas do not lend themselves easily to RCT design.
activity level. The Royal College of Psychiatrists There are considerable gaps in knowledge around (2002) recommends an energy intake in excess of the long-term consequences of physical aspects of 3000 kcal/day, while the American Psychiatric eating disorders and their treatment. For example, Association (APA, 2000) suggest 70–100 kcal/kg what are the very long-term consequences of mal- body weight/day during weight gain and 40–60 nutrition on bone density, fertility and growth and to kcal/kg/day during the weight maintenance phase.
what extent are these reduced by energetic inter- Rome et al. (2003) suggest that food intake should be vention to achieve 100% expected body weight as expected to achieve a weight gain of .3–.4 lb (130–180 g) per day during a life-threatening phase and 1–2 lb(450–900 g) per week if treated as an outpatient.
Untoward effects of re-feeding caused by a suddenincrease in metabolic load can be reduced by start- Drug studies. The use of psychotropic medication is not considered a first-line treatment of choice in increasing slowly (Royal College of Psychiatrists, eating disorders. However, the appropriate use of 2002), and the Finnish guideline (Ebeling et al., medication can have a place in management as part 2003) sets a more modest target in the early stages of of a more comprehensive treatment package. This section reviews drug studies published over the pastdecade, and summarises current recommendations Management of medical complications. The Royal and practice with regard to the use of medication in College of Psychiatrists (2002) has issued detailed guidance on the management of electrolyte replace-ment in the event of specific deficiencies. They Drug studies in AN. There are very few randomised recommend that intravenous replacement should controlled trials of the use of medication in the only be considered under the supervision of a phy- treatment of AN, and only one systematic review 3 (Treasure & Schmidt, 2002). Recent research (RCTs The treatment of osteopenia and vitamin defi- published 1993 or later only) is summarised below ciency is reviewed below. NICE (2003) and Ebeling et al. (2003) draw attention to the adverse dentaleffects of vomiting and recommend specific preven- studies investigating the use of fluoxetine in AN, oneas an adjunct to an inpatient regime, and the other Target weights. Ebeling et al. (2003) argue that when administered post-discharge after weight gain defining a target weight is essential, with the mini- in hospital: Attia, Haiman, Walsh, and Flater (1998) mum objective being a weight which enables found that fluoxetine made no significant difference resumption of a normal menstrual cycle. Most to weight gain, eating symptoms or depressive symptoms compared to placebo when added to an resulted in some improvement in bone turnover and inpatient regime. However, Kaye et al. (2001) found osteocalcin levels, but that bone mineral density that the administration of fluoxetine after discharge (BMD) and body composition did not show signific- from inpatient treatment (involving weight gain) did ant improvements. The second study (Gordon et al., have a significant beneficial effect in terms of pre- 2002), which ran for a year, compared the use of oral venting relapse. Ten out of 16 in the fluoxetine group DHEA and HRT in postmenarcheal young women remained well one year post-discharge compared to with AN. This showed no significant change with the only 3 out of 19 in the placebo group. Both these use of either drug in terms of lumbar BMD, but did studies involved relatively small numbers of adult find significant improvement in hip BMD in both women with a DSM IV diagnosis of AN.
Fassino et al. (2002) conducted a trial of citalo- The effects of recombinant human insulin-like pram versus placebo in women aged between 16 and growth factor 1 (rhIGF-1) on bone turnover and bone 35 with AN treated on an outpatient basis and found density have been investigated by Grinspoon and no statistically significant difference in weight gain colleagues. In one study the authors concluded that short-term administration of rhIGF-1 increases boneturnover in a dose-dependent manner in women with Anti-psychotics. Despite the apparent increasing DSM-IV AN (Grinspoon et al., 1996). A subsequent use of some of the newer antipsychotics in the study (Grinspoon, Thomas, Miller, Herzog, & Kli- management of AN, there are no published RCTs to banski, 2002) additionally explored the effects of oral support this practice. The only study of anti-psych- contraceptive administration on bone density, and otics in the past decade is that of Ruggerio et al.
concluded that the administration of rhIGF-1 but not (2001), who compared the use of fluoxetine, amis- oral contraceptives resulted in significant change in ulpiride and clomipramine in adult inpatients with a spinal bone density, and that rhIGF-1 also improved DSM-IV diagnosis of AN. They found no significant differences in weight gain across the three groups, Finally, an RCT by Klibanski and colleagues and no significant differences on other variables, investigating the effects of oestrogen on trabecular including weight phobia, body image disturbance, bone loss in young women with AN concluded that amenorrhoea, or binge/purge frequency.
oestrogen supplementation did not confer signific-antly beneficial effects in terms of bone health (Kli- Cisapride. There is one RCT investigating the use of banski, Biller, Schoenfeld, Herzog, & Saxe, 1995).
cisapride in the inpatient management of adults withAN (Szmukler, Young, Miller, Lichtenstein, & Binns, Existing practice and recommendations around the 1995). Results showed no significant difference in use of drugs in the management of AN in gen- weight gain compared to placebo over an 8-week trial eral. The above studies have led to the widely held period. However, cisapride is not recommended in the view that the regular use of drugs is not justified in treatment of AN, due to increased risk of cardiac the management of primary anorexia nervosa, and irregularities, and has been withdrawn in many should be reserved for cases complicated by comor- countries because of this (Treasure & Schmidt, 2001).
bid diagnoses. With regard to depression, opinion isdivided. Some hold that the depression that is com- Zinc. Low levels of zinc in patients with AN have monly associated with low weight AN tends to lift with been thought by some to contribute significantly to restoration of physical health, and should be man- reduced dietary intake, resulting in the practice of aged through psychotherapy accompanying weight zinc supplementation. There is one RCT in the past gain. Others favour the use of selective serotonin re- decade which compared rates of weight gain in older uptake inhibitors (SSRIs) even at low weight, al- adolescent and adult inpatients receiving zinc gluc- though there is little evidence to support this. Clearly, onate versus placebo (Birmingham, Goldner, & in presentations complicated by a worsening of Bakan, 1994). No difference in average daily weight depressive symptoms, severe anxiety or obsessive- compulsive disorder, the use of medication can beappropriately considered. Tranquillisers or antihis- The management of osteoporosis. Anorexia ner- tamines are also often used symptomatically to re- vosa, and the endocrine disturbance that accom- duce the high levels of anxiety present with AN.
panies it, is known to have a negative effect on bone Although there are no controlled studies, low doses of density. Consequently, patients may receive medi- the atypical antipsychotics are being used to alleviate cation to manage or prevent the development of os- anxiety during re-feeding (Bruna & Fogteloo, 2003).
colleagues have published two RCTs on the use of Existing recommendations around the use of drugs oral dehydroepiandrosterone (DHEA) in young wo- in the management of AN in children and adoles- men with anorexia nervosa. The first of these (Gor- cents. A recent article from the US aiming to build don et al., 1999) found that the administration of on existing background and position papers on the DHEA in varying doses over a three-month period management of children and adolescents with eating Management of child and adolescent eating disorders disorders (Rome et al., 2003) provides some guid- antidepressants reduced bulimic symptoms. Bac- ance on the use of drugs in this age group, stating altchuk and colleagues further concluded that there that ‘supplementary multivitamins, calcium, zinc, was no significant difference in effect between dif- iron, or folate’ might be prescribed for young people ferent classes of antidepressants, but also that there with eating disorders ‘as needed’. It further suggests had been too few trials to exclude a clinically that ‘if delayed gastric emptying is delaying refeed- important difference (Hay & Bacaltchuk, 2002).
ing, cisapride or metoclopramide can be prescribed’, Fluoxetine has been shown to be effective in the adding that extreme caution should be used in the reduction of bulimic behaviours in the short term, at event that the patient is bradycardic, has prolonged three times the dose recommended for depressive QT interval, is extremely malnourished, or is on disorders (60 mg in BN). It has more recently been SSRIs – which together arguably include most shown to be of potential value in preventing relapse.
patients with acute AN. Rome et al. (2003) also Romano, Halmi, Sarkar, Koke, and Lee (2002) suggest that where purging or reflux has resulted in showed that continued fluoxetine treatment was oesophagitis, histamine-2 blockers and/or proton- associated with a significantly longer time to relapse, pump inhibitors can be used in adolescents. Finally, although this study had an extremely high attrition they state that SSRIs may be appropriately pre- rate (131 out of 150 participants left the study early).
scribed in near normal weight adolescents with eat-ing disorders, or weight restored AN patients.
The optimum treatment of osteopenia and vitamin been investigated in terms of its effects on controlling deficiency is controversial, but Kreipe et al. (1995) bingeing and purging behaviours in BN. Faris et al.
recommend calcium 1300–1500 mg/day and Vita- 4 (2000) found that the mean number of binge and min D (400 IU/day). They consider sex hormone purge episodes was halved in BN patients following a replacement therapy to be unhelpful as it can cause 4-week administration period. However, this drug is growth arrest and the illusion of a healthy repro- not currently recommended for routine prescription in the absence of sufficient trials, plus knowledge Another practice guideline for the treatment of about physiological mechanisms in BN (Bruna & children and adolescents with eating disorders, produced by a multidisciplinary group from Finland(Ebeling et al., 2003), contains more cautious rec- Existing recommendations around the use of drugs ommendations. Here the use of fluoxetine as sup- in the management of BN in general. The above portive medication in weight restored patients is put studies can be taken to demonstrate that the forward as possibly being of benefit, and short-acting appropriate use of medication can be of clear benefit benzodiazepine administered before meals sugges- to people with BN. Antidepressants have been shown ted as a means of reducing eating-related anxiety. In to reduce bulimic symptoms in the short term, but view of the known risk of adverse medication-related evidence supporting their use in maintenance effects in severely malnourished patients, these treatment is lacking (Hay & Bacaltchuk, 2001).
authors suggest that the use of medication should Many experts in the field of eating disorders believe really only be justified in weight restored patients – that psychotherapy (CBT or IPT) remains the treat- i.e., that medication is not normally justified in the ment of choice (see below). In some cases, patients management of acute primary AN. It is clear that may have to wait to access such treatment, or they existing guidelines regarding pharmacological treat- may not have access to therapists trained and ments in children and adolescents differ greatly, experienced in the use of the evidence-based psy- perhaps more related to local and national practice, chotherapies in eating disorders. Under such cir- than on the basis of research evidence.
cumstances, the use of medication, which cancontribute to a reduction in bulimic behaviours, can Drug studies in BN. Over the past decade, more be considered. There is also some evidence that the RCTs have been carried out exploring the effects of use of medication can add modestly to the benefits of drugs in the management of BN than of AN. Hay and psychological treatment in BN (Walsh et al., 1997).
Bacaltchuk (2001) have conducted a systematic re- In summary, it appears that the use of medication view of recent research in this area. They conclude alone will rarely be sufficient for full and lasting that although antidepressants of various types have been shown to reduce bulimic behaviours in the shortterm (by achieving reduction or cessation of bingeing Existing recommendations around the use of drugs and/or purging behaviours), there is inconclusive in the management of BN in adolescents. Ebeling evidence about the persistence of these effects.
et al. (2003) briefly review the drug studies in BN andconclude that ‘there is no evidence justifying the use Antidepressants. There are two recent systematic of medication as the only or primary treatment for reviews of the use of antidepressants in the treat- bulimia in children and adolescents’. However, tak- ment of bulimia nervosa (Whittal et al., 1999; Bac- ing into account the general reservations above, altchuk et al., 2000). These reviews both found that cautious extrapolation of research findings to older adolescents justifies the use of antidepressants as therapy being studied. Others temporarily hos- pitalise those whose weight falls below a certainlevel, without much consideration of the impact ofthis on the intervention being studied.
In anorexia nervosa, drug research is hampered by concerns about unwanted effects on the physicallycompromised patient. Much interest to date has fo- cused on the potential of drugs to enhance weight importance of the therapeutic relationship in treat- gain, rather than to influence the psychological as- ing adolescents with AN. In particular, they stress pects of the disorder and thereby, longer-term out- the desirability of a relationship which can be comes. Trials of post weight-restoration treatment maintained over time and an empathic engagement (e.g., Ebeling et al., 2003; NICE, 2003). Many eating In bulimia nervosa further research is required to disordered young people find it hard to acknowledge ascertain the impact of antidepressants on binge that they have a problem and are ambivalent about eating and mood in the younger age group and in change, in part because of the positive value placed particular the persistence of any beneficial effect by those with anorexia nervosa on their behaviour.
In BN, the young person may fear that the therapistwill share their feelings of guilt and shame aroundbingeing and vomiting; dispelling these beliefs is an early therapeutic goal. Kreipe et al. (1995), in theSociety for Adolescent Medicine’s position statement, Although there are a considerable number of studies recommend that psychological interventions should of psychological therapies in the recent eating dis- address not only the characteristic eating psycho- orders literature, a number of methodological issues pathology but also mastery of the developmental make for difficulties in combining results in meta- tasks of adolescence and the psychosocial issues analysis and reaching firm conclusions about the central to this age group. They consider family therapy should also be central to the treatment.
• Heterogeneity within therapies of the same name.
Two examples are the range of different models offamily therapy (e.g., ‘Behavioural Family Systems NICE (2003) concluded that there is limited evidence Therapy’, Robin et al., 1999; ‘Emotionally Fo- that a range of specific psychological treatments for cussed Family Therapy’, Johnson, Maddeaux, & AN with more therapeutic contact is superior to Blouin 1998) and ‘generic’ CBT as opposed to CBT ‘treatment as usual’ (with a lower rate of contact) in for eating disorders (Fairburn et al., 1991).
terms of mean weight gain and the proportion of • A range of outcome measures. In anorexia nervosa patients recovered (based on a meta-analysis of three these vary from measures of weight gain to multi- dimensional composite measures of physical and There is insufficient evidence from 6 small RCTs to psychosocial wellbeing (e.g., the Morgan–Russell suggest that any particular specialist psychotherapy Outcome Assessment Scale – Morgan & Hayward, (Cognitive Analytic Therapy (CBT), Interpersonal 1988). In bulimia nervosa, some studies report the Therapy, family therapy, or focal psychodynamic number of subjects achieving abstinence in bin- therapy) is superior to others in the treatment of geing or purging while others merely report re- adult patients with AN either by the end of treatment or at follow-up (NICE, 2003, based on 6 studies, • Timing of follow-up. This varies in different studies n ¼ 297). These trials also provided insufficient evid- from end of treatment to later follow-up of variable ence to conclude that any one specific psychother- timing. In anorexia nervosa outcome is sometimes apy was more acceptable to patients than others.
measured at discharge from inpatient treatment There are no controlled treatment trials of adoles- and in other studies considerably later.
cents with BN. NICE (2003) concludes that subject to • Entry criteria. Treatment is commenced in some adaptation for age and level of development, ado- studies at presentation to the service, i.e., as a first- lescents with BN should receive the same type of line treatment, whilst in other reports (e.g., Russell treatment as adults with the disorder, though con- et al., 1987; Eisler et al., 1997) it follows weight sideration should be given to involvement of the restoration in hospitalisation – this is sometimes referred to as a ‘relapse prevention’ paradigm.
• Other concurrent therapy. In the treatment of anorexia nervosa in particular, many studies arecarried out on inpatients who will be receiving a In anorexia nervosa, a handful of studies have range of other treatments alongside the specific examined the efficacy of CBT (Channon, de Silva, Management of child and adolescent eating disorders Hemsley, & Perkins 1989; Serfaty et al., 2002; Pike, In anorexia nervosa, one adult trial is under way Walsh, Vitousek, Wilson, & Bauer, in press). These (McIntosh, Bulik, McKenzie, Luty, & Jordan, 2000) studies suggest that individual CBT may be moder- comparing CBT, IPT and Focal Supportive Psy- ately effective in this condition, but possibly no more so than other focal therapies. It may be more effec-tive, however, at the symptomatic level, for example Behaviour therapy and exposure with response in reducing body image disturbance (Norris, 1984).
In bulimia nervosa, by contrast, there have been more than 30 RCTs exploring the efficacy of CBT, Agras (1989) found that ERP with a supportive pro- which have led to the conclusion that a specific form gramme of behaviour therapy was effective in achie- of CBT that focuses on modifying abnormal eating ving abstinence from purging (10/22 compared with behaviours and weight- and shape-related cognitions 1/19 controls, p ¼ .006) in bulimia nervosa, but is currently the most effective treatment (Fairburn & researchers have described the behavioural com- Harrison, 2003). The optimum treatment protocol ponent (requiring binge eating with prevention of involves about 20 weekly treatment sessions, with vomiting) as an unpleasant, aversive treatment to most studies achieving complete remission in about administer (NICE, 2003). Bulik, Sullivan, Joyce, 40% of cases (Wilson & Fairburn, 2002).
Carter, and McIntosh (1998) concluded that ERP adds nothing to the benefits of CBT administered alone.
(Jacobi et al., 1997; Whittal et al., 1999; Hay &Bacaltchuk, 2001, 2002) have demonstrated the advantages of CBT over placebo or waiting list con-trol in terms of numbers of patients achieving ab- Although the earliest models of psychological ther- stinence from bingeing (Griffiths, 1994) and purging apy for anorexia nervosa utilised psychodynamic (Agras, 1989), and clinically significant reductions in ideas, these have not generally been studied sys- bingeing (NICE, 2003; Griffiths, Hadzi-Pavlovic, & tematically. Herzog and Hartmann (1997) have pro- Channon-Little 1994; Treasure et al., 1994) and vided a review. Dare et al. (2001) compared focal purging (NICE, 2003; Griffiths et al., 1994; Agras, psychoanalytic psychotherapy (n ¼ 21) with family Schneider, Arnow, Raeburn, & Telch 1989; Freeman, therapy (n ¼ 22), cognitive analytic therapy (CAT) 1988). As well as having an effect on bulimia, CBT (n ¼ 22) and ‘routine treatment’ (n ¼ 19) in an RCT also causes a significant reduction in depression for adults with anorexia nervosa not so ill as to re- scores (Agras, 1989). A systematic review of 10 RCTs quire urgent admission. Focal psychoanalytic ther- (Hay & Bacaltchuk, 2001) also demonstrated no apy was significantly better than routine treatment difference in weight change between those receiving in producing weight gain (f ¼ 5.4, p ¼ .02) and in CBT and controls. In comparison with Interpersonal terms of overall progress at one year (recovered and Therapy (IPT), CBT has been found to be effective significantly improved vs. improved and poor out- more quickly, achieving remission by end of treat- come, RR ¼ .70, 95% CI .51–.97), but there was little ment more often, though by 8-month follow-up there to distinguish between the different specific therap- appears to be no difference between treatments ies. There may also have been a ‘dose effect’ in that ((Agras, 2000). There have been no RCTs on child or the routine treatment often involved fewer sessions and, in addition, those providing the specific ther-apies were more senior and experienced.
This is a specific form of focal psychotherapy whichaims to help patients identify and address interper- The psychosomatic conceptual model of Minuchin sonal difficulties associated with the onset or main- tenance of the eating disorder. Originally developed interest in the use of family interventions in anorexia as a treatment for major depression (Klerman, nervosa, particularly in adolescents. Initially the Weissman, Rounsaville, & Chevron, 1984), it has rationale was based on the notion of the ‘anorexo- been successfully developed for BN and Binge Eating genic family’, but empirical study has failed to sup- port the aetiological role of family dysfunction and Although CBT produces more rapid remission and the model fuels concern about blaming parents.
reduction in symptoms in BN, several studies (Agras Family interventions have thus developed as treat- et al., 2000; Fairburn et al., 1991, 1995) have con- ments which mobilise family resources, whether sistently shown that IPT is of equal efficacy in the delivered as ‘conjoint’ family therapy, separated FT longer term (with follow-up at 8–12 months). The (in which parents and the child or adolescent patient Agras et al. (Agras, Walsh, Fairburn, Wilson, & Kra- are seen separately) or ‘parental counselling’. There 9 mer, 2000) study suggested that BN patients found have been a number of RCTs. Russell et al. (1987), in IPT more theoretically ‘appropriate’ for their diffi- a trial of adolescents and adults whose weight had culties than CBT and expected more success with it.
been restored in a specialist inpatient service prior to randomisation, found that for a small group (n ¼ 21) asymptomatic, a poor outcome reflecting a high of adolescents with short duration of illness, family number with binge-purging AN, rather than BN.
therapy was superior to individual therapy. The There were no differences between the groups.
findings in relation to those who had been ill for more Dodge, Hodes, Eisler, and Dare (1995) reported a than three years were inconclusive and the outcomes small series of 8 who received outpatient FT and had were generally poor. At five-year follow-up (Eisler significant improvements in bulimic behaviours.
et al., 1997), the adolescent short duration subgroup Good or intermediate outcomes were achieved by six continued to do well, with 90% of those who had (using the composite Morgan–Russell scale).
received FT having a good outcome, compared with36% receiving individual therapy.
Four studies have compared different forms of family intervention in adolescent AN. Geist, Heine- The apparent effectiveness of family interventions 11 man, Stephens, Davis, and Katzman (2000) compared with children and adolescents with AN and the need family therapy with family group psycho-education to develop more intensive family-based interventions (n ¼ 25, mean age 15). There was no difference in for those who require it led to the development of this weight gain between the two interventions, or signi- treatment approach. The therapy aims to help family ficant difference in self-reported psychological out- members learn by identifying with members of other comes. All patients were concurrently hospitalised.
families with the same condition, by analogy (Asen, Robin et al. (1999) compared the effect of Beha- 13 in press). It is generally delivered within a day hos- vioural Family Systems Therapy (BFST) with Ego- pital programme, in which up to 10 families with an Oriented Individual therapy (EOIT) in 37 adolescents adolescent with AN attend a mixture of whole family with AN. Parents in the EOIT group received separate group discussions, parallel meetings of parents and parental counselling. There was no significant dif- adolescents and creative activities. Preparation of ference on end-point weight, or on psychological lunch and communal eating is a central part of the measures; however, the BSFT group had a greater programme. There is generally a four/five-day block change in BMI over time (F ¼ 12.6, p < .001), re- of therapy followed by a limited number of day flecting different baseline values. By 1-year follow-up attendances at approximately monthly intervals 94% of the BFST group had resumed menstruation (Scholz & Asen, 2001; Dare & Eisler, 2000). This compared with 64% of the EOIT group (p < .03).
treatment is at an early stage of evaluation but pre- Forty-three per cent of this sample had been hospi- liminary findings suggest a high degree of accept- talised when their weight fell below 75%.
ability and promising outcomes, particularly in Le Grange, Eisler, Dare, and Russell (1992) and terms of a reduced need for hospitalisation (Scholz & Eisler et al. (2000) compared conjoint family therapy with separated family therapy (SFT) in which patientswere seen on their own and parents seen separately by the same therapist. Both treatments were deliv-ered as outpatients, though 4/40 in the Eisler study There is insufficient evidence to determine the effi- required hospitalisation during treatment. The over- cacy of nutritional counselling given alone, though all results were similar in the two trials. The Le many services offer it as an adjunct to other specific Grange trial (n ¼ 18, mean age 15) found a non-sig- therapies. In one remarkable RCT (Serfaty, 1999), 35 nificant trend for the separated FT group to do patients with AN (mean age 21, youngest ¼ 16) were slightly better in terms of weight gain and on the randomly allocated to CBT (n ¼ 25) or nutritional composite Morgan–Russell Outcome Measure than counselling (n ¼ 10). All patients receiving nutri- the Conjoint FT group, though there were baseline tional counselling had dropped out by 3 months, trends in this direction. The Eisler et al. study resulting in a lack of follow-up data for this group. At (n ¼ 40, mean age 16), found a trend favouring SFT follow-up, 16 /23 of the CBT group no longer met in terms of Morgan–Russell Outcomes at one year based on comparison between good vs. intermediateand poor outcomes (n ¼ 40, RR ¼ 1.41, 95% CI .86– 2.29). A small subgroup with high maternal ex-pressed emotion did markedly better with SFT.
This is a type of behaviour therapy that views emo- Studies of FT in bulimia nervosa have been more tional dysregulation as the core problem in BN, with limited and there is only one published RCT, though bingeing and purging viewed as attempts to control two are under way. Russell et al.’s (1987) series painful emotional states. DBT was found to be more included a subgroup of 23 adults with bulimia (some effective than a waiting list control in achieving at low weight), randomly allocated to individual abstinence from bingeing and purging (4/16 com- therapy or FT. At one year the outcomes were gen- pared with 0/15) in one small adult study (Safer, erally poor, with no significant difference between Telch, & Agras, 2001). A further small uncontrolled the groups; 19 were followed up at 5 years (Eisler 15 trial (Palmer et al., 2003) of seven adult patients with an eating disorder and comorbid borderline Management of child and adolescent eating disorders personality disorder (BPD) found all patients stayed Given that most young people with AN are treated in therapy and were improved (though not in remis- without admission to hospital (Gowers, Weetman, Shore, Hossain, & Elvins, 2000), further evaluationof the efficacy of psychological therapies designed tobe delivered on an outpatient basis is required. This should explore content and who the therapies are This therapy is rooted in attempts to combine cog- nitive elements into psychoanalytic methods, deliv- Further research is needed to identify what the ered in a brief focal therapy. There are only two small most effective intervention is to challenge the prim- adult studies of CAT in the eating disorders litera- ary cognitive distortion, in which the young person ture. In a small pilot study of those with AN (Tre- over-evaluates themselves in terms of their weight asure et al., 1995), CAT was compared with educational behaviour therapy (EBT). CAT resulted We also need to understand further how beha- in greater subjective improvement at one year and in vioural management can best be effected on an objective outcomes on the composite Morgan–Rus- sell scales, though these were not statistically sig- Given the importance of parental involvement in nificant in view of the very small sample size. The managing young people with AN, further studies are Dare et al. (2001) study of 4 therapies, including CAT needed to investigate the relative merits of indi- (see above), found no benefit of any one specific vidual, family or combination treatments. The relat- ive benefits of separated vs. conjoint family therapyapproaches for young person, parents and siblingsrequire further study, as do the range of outcomes by Motivational therapy and therapeutic engagement which the success of such therapies is measured.
Recently there has been considerable interest in the Despite there being good evidence-based treat- importance of motivational interventions in the ments for bulimia nervosa in adults (particularly engagement and treatment of people with AN (Geller, CBT and IPT), these have been insufficiently ex- Cockell, & Drab, 2001; Treasure & Ward, 1997; plored in terms of application to adolescent-onset Vitousek, Watson, & Wilson, 1998), based on the trans-theoretical model of change of DiClimente and including a degree of parental involvement and age- Prochaska (1998). Motivational interviewing is a related motivational issues, require further study. As potentially useful technique which aims to move a with adults, it is unlikely that CBT will be successful person to a position where they are more prepared to and/or acceptable as the primary treatment for all adolescents with BN, suggesting a need for further therapy (MET) compared with CBT (4 sessions each) development and evaluation of a range of other out- was found in one small study of BN to lead to no differences in short-term outcome (Treasure et al.,1999). More comprehensive RCTs in this area are asyet lacking.
Most young people with anorexia nervosa, bulimia Comparisons between psychological therapies nervosa and related eating disorders can be man- aged on an outpatient basis, with inpatient care Five trials have compared antidepressants with CBT usually only being required for a minority with in BN. In meta-analysis (n ¼ 270) they provided anorexia nervosa, where there are serious com- limited evidence that CBT is superior in terms of plications related to comorbid diagnoses, or where remission from bingeing and purging by end of there is high physical and/or psychiatric risk treatment, but little evidence is available about dif- (Nicholls & Bryant-Waugh, 2003). When admission ferences in frequencies of these behaviours or at is deemed necessary this may be to a paediatric ward, a general child or adolescent psychiatric unit, There is insufficient evidence to conclude on the or a specialist eating disorder service. In UK practice, relative efficacy of antidepressants and other psy- the latter includes specialist adult units, and both independent and public sector services. There arerelatively few dedicated NHS beds for the manage-ment of children and adolescents with eating dis- orders, and existing services have been unevenly Evaluation of interventions to improve motivation distributed. A survey by the Royal College of Psy- and adherence to treatment are particularly required chiatrists carried out in 1997/1998 found that 4 in the younger population, as many children and regions, representing 25% of the UK population, had adolescents are brought to treatment by others ra- no specialist provision for young people with eating ther than actively seek treatment themselves.
disorders, and that 69% of clinics who identified themselves as providing treatment for children and 1994). The majority of people with AN are treated on younger adolescents with eating disorders were in an outpatient basis (Palmer, Gatward, Black, & the South East of England (Royal College of Psychi- Park, 2000), although such treatment tends to be atrists, 2000). This situation is slowly being rectified, poorly described and documented and presumably with increased activity in the development and varies considerably between services.
commissioning of eating disorder services for young Specialised day-patient treatment for AN has been people throughout the UK over the past few years described in the UK and abroad (Gerlinghof, Back- (Great Ormond Street National Map Project – per- mund, & Franzen, 1998; Birchall, Palmer, Waine, Gadsby, & Gatward, 2002; Zipfel et al., 2002; Perhaps related to the fact that there are so few Robinson, 2003). These studies report short-term dedicated inpatient beds is the finding that children positive outcomes in older adolescents and adults.
and adolescents with eating disorders occupy a sig- However, there are no RCTs and it is not always clear nificant percentage of all available generic inpatient whether, in the absence of the day care offered, the beds. A recent one-day census of bed occupancy by patients included in the study would have been diagnosis in child and adolescent units in the UK treated as inpatients or outpatients. Although the revealed that more beds were occupied by young addition of a day programme to a comprehensive people with eating disorders than any other dia- service has been found to reduce the use of inpatient beds in an adult service (Birchall et al., 2002) it A summary of current research in the area of ser- seems unlikely that inpatient treatment will cease vice provision is set out below. This covers studies to be needed. The relative effectiveness and cost- over the past ten years that have attempted to effectiveness of the two forms of more intensive investigate the relative merits of inpatient, day- treatment have yet to be adequately studied.
patient and outpatient treatment delivery, and the It is widely believed that there may be benefits in relative effectiveness of treatment by a specialist the treatment of severe AN within a specialised ter- eating disorder service vs. a more general setting.
tiary eating disorders service compared with lessspecialised secondary services. Both competenceand confidence tend to develop in settings where such treatment is a regular and ongoing activity.
Research in the area of service provision is limited.
This is regarded as a particular problem in the case There is one systematic review summarising what is of very young onset AN, which is relatively rare.
known about the issue of the relative effectiveness of However, there is a lack of studies that might provide inpatient and outpatient care in the management of anorexia nervosa (Meads, Gold, & Burls, 2001).
However, the review is based on only one small RCT with a five-year follow-up, often referred to as the StGeorges study (Crisp, Norton, Gowers, 1991; Gow- In the UK, very few people with BN are treated on an 17 ers, Norton, Halek, & Crisp, 1994) plus a number inpatient basis. Admission tends to occur only in of very varied case series making meaningful con- those with severe physical complications or with clusions difficult. The main conclusions of the sys- comorbid presentations. It is generally recommen- tematic review are that outpatient treatment for AN ded that the great majority of adolescents and adults at a specialist tertiary referral eating disorder service with BN should be treated on an outpatient basis was as effective as inpatient treatment in those not (NICE, 2003). The idea of ‘stepped care’ has been put so severely ill as to warrant emergency intervention, forward in the context of managing BN (Fairburn & and that outpatient care is in general cheaper than Peveler, 1990; Dalle Grave, Ricca, & Todesco, 2001), with patients being offered simpler and less expen- Gowers et al. (2000) carried out a non-randomised, sive interventions first with more complex and naturalistic comparison of outcome (at 2–7 years) in expensive interventions reserved for those who have adolescents with anorexia nervosa treated as inpa- tients and outpatients. They found those treated as A range of different types of intervention for BN has inpatients did less well, with admission status being been studied (see above) but there are no systematic the main predictive variable. Their findings suggest comparisons of outcome with different service levels.
caution in assessing the benefits of inpatient treat- All of the current evidence-based therapies for BN ment, but care should be taken about conclusions are designed to be delivered in an outpatient setting.
drawn from this study in the absence of a random- The place of inpatient treatment for BN is not clearly supported by research evidence. Special inpatient Whilst the St Georges study lacked power and had and day-patient treatment regimes have been de- other difficulties, it did clearly demonstrate that scribed (Zipfel et al., 2002), in relation to extreme many older adolescent and adult patients with AN severity, comorbidity or suicidal risk. There are some were able to make progress with fairly modest out- reports on special treatment programmes for severe patient treatment (Gowers, Norton, Halek, & Crisp, BN complicated by self-harm, substance abuse and Management of child and adolescent eating disorders similar behaviours in patients who often fulfil and recommendations around consent to treat- criteria for borderline personality disorder (Lacey & ment, the assessment of the young person’s Evans, 1986). There are no specific studies in- capacity to make treatment-related decisions, and vestigating these issues in adolescents.
the legal framework within which young peoplemay be treated against their stated wishes in thosecases where treated is deemed essential. A number of helpful documents and papers can be recom-mended in this respect (e.g., Manley, Smye, & Given the very limited amount of research indicating Srikameswaran, 2001; Honig & Bentovim, 1996).
which service configurations are most effective in the • In the case of older adolescents with ongoing management of young people, current provision treatment needs, transition to adult services from tends to be guided by recommendations found in child and adolescent services should be planned national and professional guidelines. The recommen- dations below are representative of current inter-national thinking about services for children andadolescents with eating disorders, and have been drawn from a number of published guidelines (Eating There is only a limited amount of information on the Disorders Association, 1994; Kreipe et al., 1995; experience and views of young people with eating Royal College of Psychiatrists, 2000; NICE, 2003): disorders and their families about the treatment they • Services for children and adolescents should be set receive. Information of this type can be considered up and run in a way that involves parents or prim- an important variable in the assessment of the ary carers, plus other significant family members.
relative merits of different service configurations.
Clear expectations around communication be- Newton (2001) reports that although various surveys tween all individuals and agencies involved should have identified strengths and weaknesses in existing be established and implemented. This would nor- service provision, this information seems to have had mally include the child, the parents, the general little impact on service planning. Assessment of user practitioner, the child’s school, and the treating and carer satisfaction specifically in relation to ser- team in relation to the eating disorder. Other vice setting is rarely carried out. Similarly, patient individuals or agencies, such as social services, adherence and drop-out, specifically in relation to other medical practitioners including paediatri- service setting, is not usually investigated (Mahon, cians, etc. may be also involved. Care needs to be taken to respect the young person’s right to confi- The major focus of existing studies has been on dentiality, and to adhere to existing local and improving the acceptability of services, which may professional guidelines around this.
have benefits in terms of improved attendance rates, • Services should be delivered in an age-appropriate but also increased involvement with, and effective- manner and setting, taking account of develop- ness of, programmes and treatments prescribed mental, social and educational needs. Wherever (Matoff & Matoff, 2001; Swain-Campbell, Surgenor, possible, children and adolescents should be & Snell, 2001). Taking account of user and carer treated locally. Assessment and ongoing manage- perceptions when designing and delivering services ment should be multidisciplinary, and provided by may also facilitate help seeking over a prolonged healthcare providers who have experience in the period in people with recurrent mental health prob- management of young people with eating disorders lems (Buston, 2002). This in turn may contribute to and who have knowledge about normal physical Individuals with eating disorders, and AN in par- • When inpatient care is required, young people ticular, are often described as being ambivalent should by preference always be admitted to units about seeking treatment. Unlike most other psychi- with regular and continuing experience in the atric conditions, core features of eating disorders can management of eating disorders in their age group, be highly valued by the patient. In addition, the making a distinction between children and ado- hospital environment can contribute to a sense of passivity and vulnerability, which can be linked to should be flexible depending on their level of an increased sense of loss of control, one of the maturity and locally available services. Adoles- central characteristics of an eating disorder (Eivors, cents should be admitted to the most suitable Button, Warner, & Turner, 2003). The acceptability service with experience of eating disorders. Written of inpatient treatment for AN in adolescence has guidelines should be drawn up for monitoring the been rated as low. They often report feeling pres- physical progress of all young people treated for sured and watched, with authoritarian and restrict- • Services involved in the management of young ambivalence (Brinch, Isager, & Tolstrup, 1988).
people with eating disorders will need to ensure Such factors can contribute to a degree of reluctance that all staff members are familiar with guidance to engage fully in interventions, resulting in relatively high levels of treatment refusal and premature drop- Much of the research on prognostic factors suffers out, with related implications for long-term recovery from methodological limitation. Firstly, much of it and healthcare costs (Kahn & Pike, 2001; Swain- has been based on clinical samples attending spe- Campbell et al., 2001). People receiving inpatient cialist clinics, which may result in selection biases.
treatment for AN have been found to be twice as The general outcomes of patients with AN, in par- likely to drop out of treatment compared to general ticular, treated in specialist services seems poor psychiatric inpatients (Kahn & Pike, 2001). Reasons (e.g., Russell et al., 1987), probably reflecting the for drop-out are likely to be varied and complex.
severity of disorder being treated there, though it is Such findings suggested a complicated relationship between service setting, clinical outcome and patient As with the treatment research in general, wide experience that is difficult to tease out.
variations in outcome measures and timing have It is common to find that individuals remain am- been reported. Prognostic factors studied have bivalent about treatment received, particularly those comprised a mixture of pre-treatment variables, with AN (Carnell, 1998), even when followed up after variables relating to adherence and response to many years. Those who have AN in adolescence ap- treatment and end of treatment predictors of out- pear most likely to recall their treatment (whether come. There is little in the literature matching prog- inpatient or outpatient) in negative terms. This atti- nostic features to a particular treatment (NICE, tude tends to persist and does not appear to be related 2003). Different factors may influence speed of re- to treatment duration or intensity (Buston, 2002).
sponse to treatment, outcome at end of treatment or Parents of adolescents have identified a lack of, and need for, support, involvement and education Although a number of potential predictors of out- come have been measured, these are chiefly ones Schrader, Maren, Rey, Touyz, & Beaumont, 1993).
which are easily measured at presentation. Some Such parents have also reported feeling blamed for factors, however, which are assessed more rarely, their child’s eating disorder by clinicians providing such as motivation for change or over concern with treatment (Sharkey–Orgnero, 1999). Lengthy waiting body weight and shape, may be at least as crucial in times for outpatient treatment have been identified determining outcome. Indeed most studies have as a major reason for being dissatisfied with health found that the contribution of any risk or mainten- care, leading to unacceptable stress and anxiety ance factor to outcome is small, implying either that multivariate models are necessary to predict out-come or that the most important factors have notbeen measured. Finally, few studies have included the person with an eating disorder’s own perspective.
A significant minority of young people with AN are Two recent systematic reviews of prospective and currently treated on an inpatient basis; however, experimental studies have considered the evidence the benefits and risks of different service settings for maintaining and prognostic factors (Stice, 2002; remain poorly understood. Further research is NICE, 2003), while Steinhausen (2002), in a more needed on the advantages and disadvantages of inclusive review of outcome predictors from 119 different treatment settings (including inpatient, studies, has also included publications based on outpatient and day-patient) on all aspects of func- tioning, including physical, psychological and social An important question facing those treating pa- functioning. Long-term comparisons of outcome in tients with eating disorders is how to predict which relation to these different treatment settings are re- young people will respond to treatment. This know- quired, but also treatment delivered in ‘specialist’ ledge might enable more intensive treatments to be eating disorder units vs. more generic units. Patient given to those likely to be more resistant. Intensive and parent perspectives on treatment experience treatments for AN, such as inpatient management, and satisfaction should be sought in an attempt to are expensive, less popular and scarce and therefore contribute towards improved service delivery.
Service issues. There is no good evidence on the The aetiology of eating disorders, in common with outcomes of those who do not access formal medical most other psychiatric disorders, is generally con- care (Treasure & Schmidt, 2002), though Crisp sidered to be multifactorial (Cooper & Steere, 1995).
et al.’s (1991) RCT (adults and older adolescents Following the establishment of an eating disorder, combined) found a significant advantage of special- a similar combination of risk and protective factors ised inpatient and outpatient therapies over assess- is thought to maintain the condition, determine ment only, at one-year follow-up. In adolescents, one cohort study found that only 3/21 of those treated as inpatients had a good outcome at 4 years compared Management of child and adolescent eating disorders to 31/51 of those who had never been admitted Physical features. Vomiting, bulimia and profound (Gowers et al., 2000). This paper raises the contro- weight loss are associated with a poor outcome versial issue of the potential adverse consequences (Treasure & Schmidt, 2002). BMI centile alone may of admission, which, given the inevitably more se- therefore not be that helpful in predicting outcome verely ill nature of those selected for it, is difficult to given that binge-purgers do not generally achieve the address without an RCT design. The TOuCAN trial of very low weights seen in restricting AN and indeed inpatient vs. outpatient management (Gowers et al., Rome et al. (2002) conclude that in young patients it in preparation) may help to rectify this. A similar is the restricters rather than purgers who have the difficulty bedevils the evaluation of compulsory treatment and treatment predictors of mortality.
High serum creatinine levels (>1.5mg/100ml) are Those compulsorily treated have a poorer outcome associated with poor outcome in children (Rome (Ramsay, Ward, Treasure, & Russell, 1999), but most treatment guidelines (e.g., NICE, 2003) con-clude that there are considerable benefits, including Dropout from treatment is often cited as a poor Keel and Mitchell (1997), in a narrative review of prognostic indicator (NICE, 2003), though those with predictors of outcome for bulimia nervosa based on other unfavourable features may disengage more 60 studies, and Hay and Bacaltchuk (2002), in their systematic review, concluded that there were fewconsistent predictors of outcome. NICE (2003) re- Predictors of mortality. Neilsen et al. (1998) reviewed viewed 60 studies of sample size >50 and follow-up the mortality rate in AN, based on published out- of greater than 1 year and concluded that meta- come studies across the age range. They concluded analysis was not possible owing to the variety of that the Standard Mortality Ratio (SMR) in anorexia methods employed. NICE (2003) give no prognostic nervosa was raised in those with a lower presenting indicators specifically for adolescents.
BMI and those presenting in adulthood (aged 20–29), Good prognosis has been associated with the fol- that is to say the adolescent-onset condition con- lowing pre-treatment variables: shorter duration of ferred a better prognosis in terms of lethal outcome.
illness, higher social class, younger onset and family Andersen (1992) has concluded, on the basis of a 20 history of alcoholism (Collings & King, 1994). Bell review of a number of large series, that the outcome is (2002), meanwhile, concluded that low self-esteem no better or worse for males than females.
Comorbidity. A number of papers have reviewed the Body mass. In bulimia nervosa higher body mass impact of comorbid conditions in adolescence. Pre- does not appear to act as a maintenance factor for treatment depression was found not to influence bulimic symptoms (Stice & Agras, 1998; Fairburn, outcome at 1 year (North & Gowers, 1999), while 21 Cooper, Doll, Norman, & O’Connor, 2000).
obsessive-compulsive disorder was associated with a Perceived pressure to be thin. In an adolescent poorer outcome (Higgs et al., 1989). Residual OCD sample, perceived pressure to be thin was found to symptoms at end of treatment are also a negative predict maintenance of bulimic symptoms in an adolescent sample followed up for 9 months (Stice &Agras, 1998). The same study also found that Age of onset. An early age of onset has been con- maintenance of bulimic symptoms was also related sistently reported as conferring a good prognosis to higher rates of presenting body dissatisfaction.
(Treasure & Schmidt, 2002), along with a short dur-ation of illness before treatment, which may be a Dieting. The dietary restraint model argues that confounding variable. Gowers, Crisp, Joughin, and calorie restriction contributes to the maintenance of Bhat (1991) suggested, however, that there may be a binge eating. This proposal was supported in the sub-set of very early onset (pre-menarcheal) cases, adolescent study of Stice and Agras (1998), but not with especially poor physical, social and personality in adults (Fairburn et al., in press).
development, who might have a poor outcome. (Bry-ant-Waugh, Knibbs, Fosson, Kaminski, and Lask Negative affect. This has been found to be a non- (1988) found a poor outcome in those developing AN significant predictor of bulimic symptom mainten- under the age of 11 and Rome et al. (2003) suggest ance in general (NICE, 2003) and in adolescents that asociality in childhood predicts poor outcome.
(Stice & Agras, 1998). Leon, Fulkerson, Perry, Keel,Klump (1999), meanwhile, found that it predicted an Life events. In a prospective adolescent series, increase in general eating pathology in a large sam- North, Gowers, and Byram (1997) found that those ple of adolescents followed up for three years.
with a severe negative life event precipitant (i.e., acuteonset) had a good prognosis, probably reflecting the Perfectionism. Santonastaso, Friederici, and Favaro healthier premorbid adjustment of this subgroup.
(1999) followed up 72 adolescents and found that high initial perfectionism predicted maintenance of relevant sections above. Finally, further clarification is required to guide treatments for those with ‘atyp-ical’ or ‘not otherwise specified’ eating disorders.
Severity. Higher rates of bingeing and vomiting are Currently these terms are used to include a hetero- associated with poor outcome (NICE, 2003).
geneous mix of clinical presentations, for whomthere are no evidence-based treatments. In those not Comorbidity. Substance misuse confers a poor quite meeting criteria for AN and BN, studies are prognosis (Keel, Mitchell, Miller, Davis, & Crow, required to investigate the acceptability and efficacy 1999; NICE, 2003). Premorbid obesity in childhood of treatments for AN and BN. Predictors of outcome has also been cited (Fairburn et al., 1995; Bulik, within this group also need further study, as part of Sullivan, Joyce, Carter, & McIntosh, 1998; NICE, a process of improving the ability to match inter- 2003). Personality disorder or disturbance is con- ventions to individual presentations.
sistently associated with poor outcome, particularlyCluster B (Rossiter, Agras, Telch, & Schneider, 1993;NICE, 2003) and impulsivity (Keel et al., 1999).
Grateful thanks are due to the members of the NICE Eating Disorders Guideline Development Group whocontributed many of the clinical questions referred to Those with poor motivation for change do poorly in this review and to the searches and analysis.
(NICE, 2003). Continuing bulimic behaviours at theend of treatment are associated with poor outcome,thus treatment should aim for complete abstinence rather than reduction in these behaviours (NICE,2003). The association of continuing abnormal atti- Simon Gowers, Professor of Adolescent Psychiatry, tudes, body dissatisfaction, drive for thinness and Academic Unit, 79 Liverpool Road, Chester CH2 low mood, at end of treatment with poor outcome 1AW, UK; Email: [email protected] (NICE, 2003), indicates that both cognitive andbehavioural change are vital to long-term recovery.
Agras, W.S., Schneider, J.A., Arnow, B., Raeburn, S.D., & Telch, C.F. (1989). Cognitive behavioural andresponse prevention treatments for bulimia nervosa.
Despite the very long history of treating eating dis- Journal of Consulting and Clinical Psychology, 57, orders, good quality research is lacking. For a con- dition which commonly arises in adolescence, the Agras, W.S., Walsh, B.T., Fairburn, C.G., Wilson, C.T., number of adolescent treatment trials is very small.
& Kramer, H.C. (2000). A multicenter comparison of We must conclude that the barriers to research are cognitive-behavioral therapy and interpersonal psy- considerable. The need for intensive medical man- chotherapy for bulimia nervosa. Archives of General agement, often delivered on an inpatient basis, makes research into AN problematical; many studies Alderman, J., Wolkow, R., Chung, M., & Johnston, H.F.
list severe physical compromise as an exclusion (1998). Sertraline treatment of children and adoles- cents with obsessive compulsive disorder or depres-sion. Journal of the American Academy of Child and physical management may draw attention away from Adolescent Psychiatry, 37, 386–394.
psychological aspects of the disorders. Problems American Psychiatric Association. (1994). Diagnostic with engagement and obtaining informed consent and statistical manual of mental disorders (4th edn).
are ubiquitous clinical issues and make for addi- Washington, DC: American Psychiatric Association.
tional difficulty in recruitment into treatment trials.
American Psychiatric Association. (2000). Practice A number of aspects of management may not lend guideline for the treatment of patients with eating themselves to RCT design; nevertheless, well-con- disorders (revision). American Psychiatric Association ducted pragmatic trials of adequate power may well Work Group on Eating Disorders. American Journal of elucidate the more effective and acceptable treat- ments. Although clinical guidelines have identified Andersen, A.E., (1992). Follow-up of males with eating consensus good clinical practice, research is needed disorders. In W. Herzog, H.-C. Deter, & W. Vande-reycken (Eds.), The course of eating disorders. Berlin: to confirm or challenge unsubstantiated belief.
The gaps in the knowledge base regarding effective Asen, K.E. (in press). Developments in multiple family treatments in child and adolescent eating disorders suggest that almost any adequately powered well- Attia, E., Haiman, C., Walsh, B.T., & Flater, S.R. (1998).
conducted trials would add to knowledge. There are Does fluoxetine augment the inpatient treatment of some areas, however, which should be considered anorexia nervosa? American Journal of Psychiatry, priorities and these have been highlighted in the Management of child and adolescent eating disorders Bacaltchuk, J., Hay, P., & Mari, J.J. (2000). Antide- Cooper, P., Watkins, B., Bryant-Waugh, R., & Lask, B.
pressants versus placebo for the treatment of bulimia (2002). The nosological status of early onset anorexia nervosa: A systematic review. Australian and New nervosa. Psychological Medicine, 32, 873–880.
Zealand Journal of Psychiatry, 34, 310–317.
Cooper, P.J., & Steere, J. (1995). A comparison of two Bell, L. (2002). Does concurrent psychopathology at psychological treatments for bulimia nervosa: Impli- presentation influence response to treatment for cations for models of maintenance. Behaviour Re- bulimia nervosa? Eating and Weight Disorders, 7, Crisp, A.H. (1995). Anorexia nervosa; Let me be. Lon- Birchall, H., Palmer, R., Waine, J., Gadsby, K., & Gatward, N. (2002). Intensive day programme treat- Crisp, A.H., Norton, K.W.R., Gowers, S.G., Halek, C., ment for severe anorexia nervosa – the Leicester Levett, G., Yeldham, D., Bowyer, C., & Bhat, A.
experience. Psychiatric Bulletin, 26, 334–336.
(1991). A controlled study of the effect of therapies Birmingham, C.L., Goldner, E.M., & Bakan, R. (1994).
aimed at adolescent and family psychopathology in Controlled trial of zinc supplementation in anorexia Anorexia Nervosa. British Journal of Psychiatry, 159, nervosa. International Journal of Eating Disorders, 15, Crist, W., & Napier-Phillips, A. (2001). Mealtime Brinch, M., Isager, T., & Tolstrup, K. (1988). Patients’ behaviours of young children: A comparison of evaluation of their former treatment for anorexia normative and clinical data. Journal of Developmental nervosa. Nordisk Psykiatrisk Tidsskrift, 42, 445– and Behavioral Pediatrics, 22, 279–286.
Dalle Grave, R., Ricca, V., & Todesco, T. (2001). The Bruna, T., & Fogteloo, J. (2003). Drug treatments. In J.
stepped-care approach in anorexia nervosa and Treasure, U. Schmidt, & E. Furth van (Eds.), Hand- bulimia nervosa: Progress and problems. Eating and book of eating disorders (2nd edn, pp. 311–323).
Dare, C., & Eisler, I. (2000). A multi-family group day Bryant-Waugh, R. (2000). Overview of the eating disor- treatment programme for adolescent eating disorder.
ders. In B. Lask & R. Bryant-Waugh (Eds.), Anorexia European Eating Disorders Review, 8, 4–18.
nervosa and related eating disorders in childhood and Dare, C., Eisler, I., Russell, G., Treasure, J., & Dodge, L.
adolescence (pp. 27–40). Hove: Psychology Press.
(2001). Psychological therapies for adults with anor- Bryant-Waugh, R., Cooper, P., Taylor, C., & Lask, B.
exia nervosa: Randomised controlled trial of out- (1996). The use of the eating disorder examination patient treatments. British Journal of Psychiatry, 178, with children: A pilot study. International Journal of DiClemente, C. & Prochaska, J.O. (1998). Towards a Bryant-Waugh, R., Knibbs, J., Fosson, A., Kaminski, Z., comprehensive, transtheoretical model of change. In & Lask, B. (1988). Long term follow up of patients W. Miller & N. Heather (Eds.), Treating addictive with early onset anorexia nervosa. Archives of Dis- behaviours (pp. 3–24). New York: Plenum Press.
Dodge, E., Hodes, M., Eisler, I., & Dare, C. (1995).
Bulik, C.M., Sullivan, P.F., Carter, F.A., McIntosh, V.V., Family therapy for bulimia nervosa in adolescents: An & Joyce, P.R. (1998). The role of exposure with exploratory study. Journal of Family Therapy, 17, 59– response prevention in the cognitive-behavioural therapy for bulimia nervosa. Psychological Medicine, Eating Disorders Association. (1994). Eating disorders: A guide to purchasing and providing services. Nor- Bulik, C.M., Sullivan, P.F., Joyce, P.R., Carter, F.A., & McIntosh, V.V. (1998). Predictors of 1-year treatment Ebeling, H., Tapanainen, P., Joutsenoja, A., Koskinen, outcome in bulimia nervosa. Comprehensive Psychi- M., Morin-Papunen, L., Jarvi, L., Hassinen, R., Keski- Rahkonen, A., Rissanen, A., & Wahlbeck, K. (2003).
Buston, K. (2002). Adolescents with mental health Practice guideline for treatment of eating disorders in problems: What do they say about health services? children and adolescents. Helsinki: Finnish Medical Journal of Adolescence, 25, 231–242.
Carnell, K. (1998). Eating disorders: The patient’s Eisler, I., Dare, C., Hodes, M., Russell, G.F.M., Dodge, perspective. Australian Prescriber, 21, 103–105.
E., & Le Grange, D. (2000). Family therapy for Cawthron, P. (2001). Adolescent psychopharmacology.
adolescent anorexia nervosa: The results of a con- In S.G. Gowers (Ed.), Adolescent psychiatry in clinical trolled comparison of two family interventions. (In practice (pp. 425–448). London: Arnold.
Process Citation). Journal of Child Psychology and Channon, S., de Silva, P., Hemsley, D., & Perkins, R.
(1989) A controlled trial of cognitive-behavioural and Eisler, I., Dare, C., Russell, G.F.M., Szmukler, G., Le behavioural treatment of anorexia nervosa. Behavior- Grange, D., & Dodge, E. (1997). Family and indi- al Research and Therapy, 27, 529–535.
vidual therapy in anorexia nervosa. A 5-year follow- Childress, A., Brewerton, T., Hodges, E., & Jarrell, M.
up. Archives of General Psychiatry, 54, 1025–1030.
(1993). The kids’ eating disorders survey (KEDS): Eivors, A., Button, E., Warner, S., & Turner, K. (2003).
A study of middle school children. Journal of the Understanding the experience of drop-out from treat- Academy of Child and Adolescent Psychiatry, 32, ment for anorexia nervosa. European Eating Disor- Collings, S., & King, M. (1994). Ten year follow-up of 50 Fairburn, C., & Peveler, R. (1990). Bulimia nervosa and patients with bulimia nervosa. British Journal of a stepped care approach to management. Gut, 31, Fairburn, C.G. (1981). A cognitive behavioural ap- Gowers, S.G. (2001). Eating disorders. In S.G. Gowers proach to the treatment of bulimia. Psychological (Ed.), Adolescent psychiatry in clinical practice (pp.
Fairburn, C.G, Cooper, Z., Doll, H.A., Norman, P., & Gowers, S.G., Crisp, A.H., Joughin, N., & Bhat, A.
O’Connor, M. (2000). The natural course of bulimia (1991). Premenarcheal anorexia nervosa. Journal of nervosa and binge eating disorder in young women.
Child Psychology and Psychiatry, 32, 515–524.
Archives of General Psychiatry, 57, 659–665.
Gowers, S.G., Weetman, J., Shore, A., Hossain, F., & Fairburn, C.G., & Harrison, P.J. (2003). Eating dis- Elvins, R. (2000). The impact of hospitalisation on the outcome of adolescent anorexia nervosa. British Fairburn, C.G., Jones, R., Peveler, R.C., Carr, S.J., Journal of Psychiatry, 176, 138–141.
Solomon, R.A., O’Connor, M.E., Burton, J., & Hope, Griffiths, R.A., Hadzi-Pavlovic, D., & Channon-Little, L.
R.A. (1991). Three psychological treatments for (1994). A controlled evaluation of hypnobehavioural bulimia nervosa: A comparative trial. Archives of treatment for bulimia nervosa: Immediate pre-post treatment effects. European Eating Disorders Review, Fairburn, C.G., Norman, P.A., Welch, S.L., O’Connor, M.E., Doll, H.A., & Peveler, R.C. (1995). A prospective Grinspoon, S., Baum, H., Lee, K., Anderson, E., Herzog, study of outcome in bulimia nervosa and the long D., & Klibanski, A. (1996). Effects of short term recombinant human insulin-like growth factor I Archives of General Psychiatry, 52, 304–312.
Faris, P., Won Kim, S., Meller, W., Goodale, R., Oak- women with anorexia nervosa. Journal of Clinical man, S., Hofbauer, R., Marshall, A., Daughters, R., Endocrinology and Metabolism, 81, 3864–3870.
Banerjee-Stevens, D., Eckert, E., & Hartman, B.
Grinspoon, S., Thomas, L., Miller, K., Herzog, D., & (2000). Effect of decreasing afferent vagal activity Klibanski, A. (2002). Effects of recombinant human with ondansteron on symptoms of bulimia nervosa: A IGF-I and oral contraceptive administration on bone randomised, double-blind trial. The Lancet, 355, density in anorexia nervosa. Journal of Clinical Endocrinology and Metabolism, 87, 2883–2891.
Fassino, S., Leombruni, P., Daga, G., Brustolin, A., Harrington, R.C., Whittaker, J., Shoebridge, P., & Migliaretti, G., Cavallo, F., & Rovera, G. (2002).
Campbell, F. (1998). Systematic review of the efficacy Efficacy of citalopram in anorexia nervosa: A pilot of cognitive behaviour therapies in child and adoles- cent depressive disorder. British Medical Journal, Geist, R., Heineman, M., Stephens, D., Davis, R., & Hay, P.J., & Bacaltchuk, J. (2001). Psychotherapy for Katzman, D.K. (2000). Comparison of family therapy bulimia nervosa and bingeing (Cochrane Review). The and family group psychoeducation in adolescents with anorexia nervosa. Canadian Journal of Psychi- Hay, P., & Bacaltchuk, J. (2002). Bulimia nervosa.
Geller, J., Cockell, S.J., & Drab, D.L. (2001). Assessing Hay, P., & Balcaltchuk, J. (2002). Bulimia nervosa.
readiness for change in the eating disorders: The psychometric properties of the readiness and motiva- Herzog, T., & Hartmann, A. (1997). Psychoanalytically tion interview. Psychological Assessment, 13, 189– oriented treatment of anorexia nervosa. Methodology- related critical review of the literature using meta- Gerlinghoff, M., Backmund, H., & Franzen, U. (1998).
Evaluation of a day treatment programme for eating Midizinische Psychologie, 47, 299–315.
disorders. European Eating Disorders Review, 6, 96– Higgs, J., Goodyer, I.M., & Birch, J. (1989). Anorexia nervosa and food avoidance emotional disorder.
Gordon, C., Grace, E., Emans, S., Feldman, H., Good- Archives of Disease in Childhood, 64, 346–351.
man, E., Becker, K., Rosen, C., Gundberg, C., & Honig, P., & Bentovim, M. (1996). Treating children Leboff, M. (2002). Effects of oral dehydroepiandros- with eating disorders – ethical and legal issues.
terone on bone density in young women with anorexia Clinical Child Psychology and Psychiatry, 1, 287– nervosa: A randomised trial. Journal of Clinical Endocrinology and Metabolism, 87, 4935–4941.
Hutchinson, H. (1999). Feeding problems in young Gordon, C., Grace, E., Emans, J., Goodman, E., Craw- children: Report of three cases and review of the ford, M., & Leboff, M. (1999). Changes in bone literature. Journal of Human Nutrition and Dietetics, turnover markers and menstrual function after short term oral DHEA in young women with anorexia Jacobi, C., Dahme, B., & Rustenbach, S. (1997).
nervosa. Journal of Bone and Mineral Research, 14, Comparison of controller psycho and pharmacother- apy studies in bulimia anorexia nervosa [in German].
Gowers, S., Norton, K., Halek, C., & Crisp, A.H. (1994).
Psychotherapy Psychosomatik Midizinische Psycholo- Outcome of outpatient psychotherapy in a random allocation treatment study of anorexia nervosa.
Jaffa, A. (2001). National In-Patient Child and Adoles- International Journal of Eating Disorders, 15, 165– cent Psychiatric Survey (NICAPS). Presented at the Royal College of Psychiatrists Annual Conference, Gowers, S.G. (in preparation). A randomised controlled treatment trial of adolescent anorexia nervosa – the Johnson, S.M., Maddeaux, C., & Blouin, J. (1998).
TOuCAN trial. Details on request from the author.
Emotionally focussed family therapy for bulimia: Management of child and adolescent eating disorders Changing attachment patterns. Psychotherapy, 35, children and adolescents with eating disorders: Application of a framework for ethical decision mak- Kahn, C., & Pike, K.M. (2001). In search of predictors of ing. European Eating Disorders Review, 9, 144–166.
dropout from inpatient treatment for anorexia nerv- Matoff, M.L., & Matoff, S.A. (2001). Eating disorder osa. International Journal of Eating Disorders, 30, recovery: Learning from the client’s healing journey.
Kaye, W.H., Nagata, T., Weltzin, T.E., Hsu, G., Sokol, McIntosh, V.V., Bulik, C.M., McKenzie, J.M., Luty, S.E., M.S., Conaha, C.M., Plotnicov, K.H., Weise, J., & & Jordan, J. (2000). Interpersonal psychotherapy for Deep, D. (2001). Double-blind placebo controlled anorexia nervosa. International Journal of Eating administration of fluoxeting in restricting- and re- stricting-purging-type anorexia nervosa. Biological Meads, C., Gold, L., & Burls, A. (2001). How effective is outpatient care compared to inpatient care for treat- Keel, P.K., & Mitchell, J.E. (1997). Outcome in bulimia ment of anorexia nervosa? A systematic review.
nervosa. American Journal of Psychiatry, 154, 313– European Eating Disorders Review, 9, 229–241.
Minuchin, S., Baker, L., Rosman, B.L., Liebman, R., Keel, P.K., Mitchell, J.E., Miller, K.B., Davis, T.L., & Millman, L., & Todd, T.C. (1975). A conceptual model Crow, S.J. (1999). Long-term outcome of bulimia of psychosomatic illness in childhood. Archives of nervosa. Archives of General Psychiatry, 56, 63–69.
General Psychiatry, 32, 1031–1038.
Klerman, G.L., Weissman, M.M., Rounsaville, B.J., & Moher, D., Schulz, K.F., & Altman, D.G. (2001). The Chevron, E.S. (1984). Interpersonal psychotherapy of CONSORT statement: Revised recommendations for improving the quality of reports of parallel group Klibanski, A., Biller, B., Schoenfeld, D., Herzog, D., & randomised trials. Medical Research Methodology, 1, Saxe, V. (1995). The effects of estrogen administration on trabecular bone loss in young women with Morgan, H.G., & Hayward, A.E. (1988). Clinical assess- anorexia nervosa. Journal of Clinical Endocrinology ment of anorexia nervosa. The Morgan–Russell Assessment Schedule. British Journal of Psychiatry, Kopec-Schrader, E.M., Maren, K., Rey, J.M, Touyz, S.W., & Beaumont, P.J. (1993). Parental evaluation of Neilsen, S., Moller-Madsen, S., Isager, T. et al. (1998).
treatment outcome and satisfaction with an inpatient Standardised mortality in eating disorders – quanti- program for eating disorders. Australian and New tative summary of previously published and new Zealand Journal of Psychiatry, 27, 264–269.
evidence. Journal of Psychosomatic Research, 44, Kreipe, R.E., Golden, N.H., Katzman, D.K., Fisher, M., Rees, J., Tonkin, R.S., Silber, T.J., Sigman, G., Newton, T. (2001). Consumer involvement in the ap- Schebendach, J., Ammerman, S.D., et al. (1995).
praisal of treatments for people with eating disorders: Eating disorders in adolescents. A position paper of A neglected area of research. European Eating Disor- the Society for Adolescent Medicine. Journal of Ado- NICE. (2003). Eating disorders: Core interventions in the Lacey, J.H., & Evans, C. (1986). The impulsivist: A treatment and management of anorexia nervosa, multi-impulsive personality disorder. British Journal bulimia nervosa and related eating disorders; a national clinical practice guideline. London: National Lask, B. (1993). Treatment overview. In B. Lask & R.
Bryant-Waugh (Eds.), Childhood onset anorexia nerv- Nicholls, D., & Bryant-Waugh, R. (2003). Children and osa and related eating disorders. Hove: Lawrence young adolescents. In J. Treasure, U. Schmidt, & E.
Van Furth (Eds.), Handbook of eating disorders (2nd Le Grange, D., Eisler, I., Dare, C., & Russell, G.F.M.
edn, pp. 415–434). Chichester: Wiley.
(1992). Evaluation of family therapy in anorexia Nicholls, D., Chater, R., & Lask, B. (2000). Children into nervosa: A pilot suede. International Journal of Eating DSM don’t go: A comparison of classification systems for eating disorders in childhood and early adoles- Leon, G.R., Fulkerson, J.A., Perry, C.L., Keel, P.K., & cence. International Journal of Eating Disorders, 28, Klump, K.L. (1999). Three to four year prospective evaluation of personality and behavioural risk factors Nicholls, D., Christie, D., Randall, L., & Lask, B. (2001).
for later disordered eating in adolescent girls and Selective eating: Symptom, disorder or normal variant.
boys. Journal of Youth and Adolescence, 28, 181–196.
Clinical Child Psychology and Psychiatry, 6, 257–270.
Lock, J., Le Grange, D., Agras, W., & Fairburn, C.
Nobakht, M., & Dezhkam, M. (2000). An epidemiolo- (2001). Treatment manual for anorexia nervosa; A gical study of eating disorders in Iran. International family based approach. New York: Guilford.
Journal of Eating Disorders, 28, 265–271.
Mahon, J. (2000). Dropping out from psychological Norris, D.L. (1984). The effects of mirror confrontation

Source: http://satsanga.buffalo.edu/docs/Guidelines_Psychology_and_Psychiatry.pdf