Clinical Psychology Review 34 (2014) 29–43

Contents lists available at ScienceDirect

Clinical Psychology Review

A systematic review of the literature on family functioning across all eating disorder diagnoses in comparison to control families Anita Holtom-Viesel a,1, Steven Allan b,⁎ a b

Coventry Eating Disorders Service, 2 Dover Street, Coventry CV1 3DB, United Kingdom University of Leicester, Clinical Psychology Unit, 104 Regent Road, Leicester LE1 7LT, United Kingdom

H I G H L I G H T S • • • •

Worse family functioning in eating disorder families compared to controls. The notion of a typical pattern of family dysfunction was not supported. Eating disorder patients reported more family dysfunction than parents. Patients with positive perceptions of family functioning had better outcomes.

a r t i c l e

i n f o

Article history: Received 20 February 2013 Received in revised form 26 September 2013 Accepted 27 October 2013 Available online 5 November 2013 Keywords: Eating disorder Family functioning Anorexia nervosa Bulimia nervosa

a b s t r a c t The objectives of this review were to systematically identify and evaluate quantitative research comparing family functioning (a) in eating disorder families with control families, (b) in families with different eating disorder diagnoses (c) perceptions of different family members and (d) the relationship between family functioning and recovery. This adds to the findings of previous reviews of family functioning by including data from control families, the range of diagnoses, and focusing on recovery. Findings were considered in relation to models of family functioning. Using specific search criteria, 17 research papers were identified and evaluated. Findings indicated that eating disorder families reported worse family functioning than control families but there was little evidence for a typical pattern of family dysfunction. A consistent pattern of family dysfunction for different diagnoses was not suggested but patients consistently rated their family as more dysfunctional than one or both of their parents. With respect to outcome and recovery, those with more positive perceptions of family functioning generally had more positive outcomes, irrespective of severity of eating disorder. Conclusions were limited by inconsistent findings and methodological issues. Further research is needed into the relationship between family functioning and outcome and the assessment of family functioning beyond self-report. © 2013 Elsevier Ltd. All rights reserved.

Contents 1.

2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Family functioning and eating disorders . . . . . . . . . . 1.2. Models of family functioning . . . . . . . . . . . . . . . 1.2.1. Family systems theory and the psychosomatic family 1.2.2. The McMaster model of family functioning. . . . . 1.2.3. The process model of family functioning . . . . . . 1.3. Aetiology vs. maintenance . . . . . . . . . . . . . . . . 1.4. Previous literature reviews . . . . . . . . . . . . . . . . 1.5. Rationale and aims of the current review . . . . . . . . . . Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Inclusion criteria . . . . . . . . . . . . . . . . . . . . . 2.2. Quality assessment. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

⁎ Corresponding author. Tel.: +44 116 2232648; fax: +44 116 2231650. E-mail addresses: [email protected] (A. Holtom-Viesel), [email protected] (S. Allan). 1 Tel.: +44 2476 232940; fax: +44 2476500959. 0272-7358/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cpr.2013.10.005

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

30 30 30 30 30 30 31 31 31 31 31 31

30

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

2.3. Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Data extraction and synthesis . . . . . . . . . . . . . . . . . . . . . . . 3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. General description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Comparisons of family functioning between clinical and control samples . . . . 3.2.1. General family functioning . . . . . . . . . . . . . . . . . . . . 3.2.2. Specific elements of family functioning. . . . . . . . . . . . . . . 3.3. Comparisons of family functioning between eating disorder subgroups . . . . 3.3.1. General family functioning . . . . . . . . . . . . . . . . . . . . 3.3.2. Specific elements of family functioning. . . . . . . . . . . . . . . 3.4. Comparisons of family functioning between family members . . . . . . . . . 3.4.1. ED participants' views compared with parents' . . . . . . . . . . . 3.4.2. Mothers' views compare with fathers' . . . . . . . . . . . . . . . 3.4.3. ED participants' views compared with siblings' . . . . . . . . . . . 3.5. Relationship between family functioning and outcome/recovery . . . . . . . 3.6. Study quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1. Samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.2. Measurements. . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.3. Study design . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Differences in family functioning between eating disorder families and controls 4.1.1. Findings in relation to models of family functioning . . . . . . . . . 4.2. Differences in family functioning between eating disorder subgroups . . . . . 4.2.1. General family functioning . . . . . . . . . . . . . . . . . . . . 4.2.2. Specific elements of family functioning. . . . . . . . . . . . . . . 4.3. Differences in the perspectives of family members . . . . . . . . . . . . . . 4.4. Summary of family functioning in eating disorder families . . . . . . . . . . 4.5. Family functioning and outcome/recovery . . . . . . . . . . . . . . . . . 4.6. Quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7. Further investigation and clinical implications . . . . . . . . . . . . . . . . 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix A. Definitions of elements of family functioning . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31 32 32 32 32 33 33 33 33 33 33 33 33 33 33 39 39 39 40 40 40 40 40 40 40 40 41 41 41 41 41 42 42 42

1. Introduction 1.2. Models of family functioning The role of the family in the development and maintenance of eating disorders has long been a subject of interest and research. Earlier research focussed on the causal influence of the family, whilst more recently, the impact that eating disorders can have on family functioning, and the role family functioning may have in the maintenance of the disorder have been explored.

1.1. Family functioning and eating disorders Family functioning has been defined as “the interactions of family members that involve physical, emotional and psychological activities” (Commonwealth of Kentucky, 2001) and “the process by which the family operates as a whole, including communication and manipulation of the environment for problem solving” (Mosby's Medical Dictionary, 8th edition, 2009). Research has measured family functioning as either a unitary concept termed ‘General Functioning’, which is considered to be the overall health or pathology of the family (McDermott, Batik, Roberts, & Gibbon, 2002), or as a composite of several elements such as, cohesion, adaptability and communication. A list of definitions for the different components of family functioning discussed in the current review is presented in the Appendix. The different components of family functioning considered relevant to ED families2are best described in relation to the following models.

2 Throughout the article the term ED family will be used to denote a family where a member has a diagnosed eating disorder.

1.2.1. Family systems theory and the psychosomatic family A core principle of family systems theory is to consider family systems as a whole rather than looking at family members individually. A central assumption is that a family's structure and organisation influence the behaviour of the family members. Minuchin, Rosman, and Baker (1978) highlighted a group of family system characteristics they believed were representative of the families of patients with a diagnosis of anorexia nervosa (AN). These characteristics were: enmeshment, over-protectiveness, rigidity, avoidance of conflict and lack of conflict resolution. Families with these characteristics were labelled ‘psychosomatic’. 1.2.2. The McMaster model of family functioning The McMaster model (Epstein, Bishop, & Levin, 1978) is also based on family systems theory. This model does not profess to cover all areas of family functioning but identifies six dimensions relevant to clinical families: problem-solving, communication, roles, affective responsiveness, affective involvement and behaviour control. These are the dimensions assessed using the Family Assessment Device (Epstein, Baldwin, & Bishop, 1983). 1.2.3. The process model of family functioning The process model (Steinhauer, Santa-Barbara, & Skinner, 1984) differs from the McMaster model in its emphasis on the interaction between the dimensions of family functioning. It is a model of family process rather than family structure. It describes seven key dimensions of family functioning: task accomplishment, role performance, communication, affective expression, affective involvement, control, values and norms. These dimensions are assessed using the Family Assessment Measure (Skinner, Santa-Barbara, & Steinhauer, 1983).

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

1.3. Aetiology vs. maintenance Models of family functioning were initially used to try and establish the role families might play in the development of eating disorders. Research into the causes of eating disorders was reviewed by Polivy and Herman (2002) who considered the contribution of a number of factors, including familial influences, on the development of eating disorders. They reported that the family characteristics found to be significant could be viewed as secondary to the presence of an ill family member rather than causative. Their conclusions suggested that familial factors were not causative in themselves and for the development of an eating disorder additional vulnerability factors, either biological or experiential, were required. Research in this area has been criticised for making unfounded assumptions about cause and effect and presuming that the dysfunction observed in families was a cause rather than a response to the eating disorder (Jack, 2001; Treasure et al., 2008). In addition, this view was seen as being unnecessarily blaming of families. When considering this and the evidence that refutes the idea that families are the primary factor that underlies risk for an eating disorder, the Academy of Eating Disorders reported that it is “firmly against any aetiological model of eating disorders in which family influences are seen as the primary cause of anorexia nervosa or bulimia nervosa” (Le Grange, Lock, Loeb, & Nicholls, 2010). More recently the focus has moved onto how family functioning may act to maintain an eating disorder. This research is significant for considering the role of the family in the recovery process and highlights the importance of examining the role families can play in maintaining eating disorders. As the research with an aetiological focus has been reviewed by Polivy and Herman (2002) and for the reasons outlined above, any empirical papers with the aim to research the causal role of family functioning in eating disorders were not included in the current review.

31

limitations in the studies reviewed, such as unsystematic data collection methods and that few studies employed adequate control groups. This review concluded the need for systematic studies of family functioning between ED families and controls. It also recommended the inclusion of participants across the range of eating disorder diagnoses, the participation of family members and the use of both observational and selfreport measures. 1.5. Rationale and aims of the current review The current review aimed to provide a more up-to-date synthesis of the evidence concerning family function and eating disorders. This review also aimed to address the gaps and limitations of previous reviews by including more recent research that has explored differences between ED families and control families and including research with participants with diagnoses of AN, BN and eating disorder not otherwise specified (EDNOS). As recommended by Kog and Vandereycken (1985), the current review included studies where different family members participated and where their different perceptions of family functioning were assessed. Unlike previous reviews, the current review also included studies that focussed on the impact of family functioning on recovery from an eating disorder. More specifically, the aims of the current review were to systematically review the evidence on family functioning in ED families to answer the following four questions. 1) Are there differences in family functioning between ED families and control families and are these differences in line with any particular models of family functioning? 2) Are there significant differences in family functioning by ED diagnosis? 3) Are there consistent differences in the perceptions of family functioning between family members? 4) Is there any relationship between family functioning and recovery from an eating disorder? 2. Method

1.4. Previous literature reviews

2.1. Inclusion criteria

A literature review in the area of family functioning and eating disorders was conducted by Eisler (2005). One of its aims was to review studies of family functioning in AN families. This narrative review suggested worse family functioning in clinical samples for communication and affective responsiveness. Eisler (2005) concluded that there was no consistent pattern of family functioning in AN families, that there was a lack of support for the psychosomatic family model, and that understanding how families reorganise themselves around a problem is more important for treatment than knowing how the problem developed. Whitney and Eisler (2005) reviewed the literature on the experience of caring for someone with an eating disorder, how the family reorganises itself and inter-personal maintaining factors. This narrative review did not report on the methods of selection or evaluation of the research and consisted mainly of qualitative research papers. The main conclusion of this review was that families could become stuck in unhelpful interactions and lose sight of their strengths and resources. A systematic review by Kog and Vandereycken (1985) explored family relationships in ED families. It excluded studies which focussed on the therapy process or treatment outcome. They found that compared to controls, mothers and daughters from ED families reported more difficulties with task accomplishment, role performance, communication and affective expression. They also found that bulimia nervosa (BN) families had higher levels of conflict and negativity, whereas AN families had higher levels of cohesion, organisation and structure, dependency, interpersonal boundary problems and cross generational blurring. The results across studies were variable when looking at enmeshment, over-protectiveness, rigidity and conflict avoidance both within diagnostic groups as well as across them. These results were not in line with the idea of the ‘psychosomatic family’ as described by Minuchin et al. (1978). Kog and Vandereycken (1985) highlighted a number of

The inclusion criteria for this review were (a) used quantitative methodology, (b) focussed on family functioning and eating disorders and addressed one of the four questions of this review, (c) was not focussing on family functioning as an aetiological factor and (d) included participants with a diagnosed eating disorder using DSM or ICD criteria or their family members. Papers were restricted to those with quantitative methodology as the comparative nature of the review questions lends itself to this. Studies were not restricted to any eating disorder diagnoses or by the family members that were included. 2.2. Quality assessment Commentaries and case studies were excluded as they are low on the hierarchy of evidence (Greenhalgh, 1997) and to enhance generalisability of the findings. With the exception of case studies, no studies were excluded due to quality criteria. However, the quality of the studies included was variable, for example, 13 of the 17 studies used control groups ranging from unmatched controls, clinical controls, to sisters as the controls, and whilst all of the studies used self-report measures only two supplemented this with observational methods and one using interviews. The impact of this is discussed later in the review. 2.3. Search strategy A computerised search was undertaken on 7th January 2013 on Psychinfo, Web of Science and Scopus using the search terms (Eating Disorder AND Family Funct*) OR (Eating Disorder AND Family Maint*) NOT Obesity. The source type was limited to ‘all journals’. The results from the three databases were collated and duplicates removed. A

32

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

visual search of the references from previous reviews revealed no further studies. The remaining articles were reduced to those with full text available to access in English. 2.4. Study selection The results of the search and selection process are represented in Fig. 1. It can be seen in Fig. 1 that the initial search revealed 1556 articles, which were reduced using the limiters described in Section 2.3 to 323 articles. The titles and abstracts of the articles were screened and those where the focus was not on family functioning and eating disorders were excluded. Commentaries, literature reviews, case studies and those with a qualitative methodology were also excluded. The full text was retrieved for the remaining 78 articles. These were screened and included if they met the four inclusion criteria. Duplicate publications were excluded or combined so as not to treat them as separate studies. The most recent paper was cited though findings from earlier papers using the same sample were reported. This resulted in 17 articles being included in the review. 2.5. Data extraction and synthesis Data from the 17 articles was extracted and entered onto a spreadsheet. The data extraction categories were informed by the NHS CRD

(2008) guidelines for data extraction and quality assessment. Information was extracted regarding the study characteristics, participant characteristics, results and information regarding the quality of the study. A narrative description of the data was produced that covered a summary of the study design, sample characteristics, key findings and study quality. A meta-analysis was not conducted given the heterogeneity of the assessment measures, diagnoses, family members and control groups. 3. Results 3.1. General description Of the 17 studies included in the current review, 15 used a crosssectional design, 13 with a comparison group. The remaining two studies used a longitudinal design with no comparison group. Thirteen of the studies reported data from at least one family member; two included parent scores, eight included mothers, six included fathers and two included siblings. Fourteen of the studies used self-report measures only, two used self-report measures and observations of families and one used self-report measures and interviews. A summary of the study characteristics is shown in Table 1. It can be seen that the majority of participants in the studies were female, within the adolescent age range, across the range of eating disorder diagnoses. The sample characteristics and methodology of the studies are presented in Table 2 and a summary of the methodological controls and

Psychinfo, Web of Science and Scopus (including duplicates) n=1556

Excluded: - Focus on obesity: 372 - Not in a journal: 303 - Duplicates: 463 - Full text not available in English: 95

Titles and abstracts screened n=323

Excluded: - Focus not on family functioning or eating disorders: 163 - Commentaries, case studies or literature reviews: 82

Full text screened. n=78

Full papers excluded: - Qualitative methodology: 6 - Family functioning as an aetiological factor: 32 - Participants without eating disorder diagnosis: 6 - Does not relate to the four questions of this review: 15 - Duplicate publications: 2

Studies included in the review. n=17 Fig. 1. Flow of studies for inclusion in the review.

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43 Table 1 Summary of study characteristics. Sample size range

ED participants ED family members Control participants

Total number of ED participants Total number of participants Mean age range (ED participants) Female % (ED participants)a Diagnoses %b Anorexia nervosa (restricting and purging type) Bulimia nervosa EDNOS Cross sectional study design Solely self-report measures used

17–126 0–118 0–1462 693 3562 14.5–24.4 years 97.8% 62.0% 27.4% 10.6% 88% 82%

a Study 8 did not specify the gender of the participants but for this calculation it was assumed they were all female. b Studies 5 and 8 did not detail the number of participants for each diagnosis so they were not included in this calculation.

results is presented in Table 3. This information was used to assess the quality of the studies and to identify the main findings. What follows is a summary of the data in line with the specified aims of this review. Throughout the remainder of this review studies will be referred to by the number assigned to them in Tables 2 and 3. 3.2. Comparisons of family functioning between clinical and control samples 3.2.1. General family functioning With respect to general family functioning, the four studies investigating this found that ED families rated themselves as having significantly worse family functioning than controls (1,4,10,14) and two studies reported that ED families were within the clinical range (9,10). In Study 10, 12% of the community samples were rated within the clinical range for family functioning compared with 62.1% of ED families. The remaining studies comparing clinical and control samples considered elements of family functioning separately. 3.2.2. Specific elements of family functioning Where studies considered specific elements of family functioning separately the results were variable and at times conflicting (1,2,4,8,11–16). The results are presented in Table 4, which highlights the variation of the findings for different components of family functioning. It can be seen that for specific elements of family functioning, for example conflict, there were studies that found higher levels of conflict in elements of the clinical sample (11–13), no significant difference compared to controls (11) and lower levels of conflict in the clinical sample (8). There are elements where clinical samples had either worse levels or no significant difference to controls (affective involvement, communication, task accomplishment and problem solving), which may indicate the elements that could be addressed in treatment to enable families to increase their skills in these areas. Table 4 also shows the differences in the findings across the different eating disorder diagnoses. 3.3. Comparisons of family functioning between eating disorder subgroups 3.3.1. General family functioning Nine studies investigated differences in family functioning between eating disorder subgroups. Two studies found no significant differences between subgroups in general family functioning (3,10); however one study found that AN-B families rated general family functioning significantly worse than AN-R families (1). 3.3.2. Specific elements of family functioning Two studies found no significant differences between ED subgroups on any subscale of the measures used (5,14). Six studies found one or

33

more significant difference between diagnoses on specific subscales. These findings are presented in Table 5. As can be seen from Table 5, there are elements such as cohesion and achievement orientation where those with AN rate significantly differently to those with BN. However, other elements such as planning activities, confiding in each other task accomplishment, communication and effective expression differed between those with and without purging behaviours. Overall, when considering the comparisons of family functioning between different eating disorder subgroups the findings are again mixed, with a number of studies finding significant differences between diagnoses yet others not. 3.4. Comparisons of family functioning between family members In the majority of studies, family members did not have a unified view of their functioning and these differences are detailed below. 3.4.1. ED participants' views compared with parents' In six studies (2–6,13) there were significant differences found between the views of ED participants and one or both their parents. In a study where the parents views were combined, parents rated their families as significantly higher on cohesion, affective expressiveness, intellectual–cultural orientation and moral–religious emphasis and ED participants rated their families higher for conflict (13). Where mothers and fathers views were rated separately, studies found that AN participants rated their family as more dysfunctional on general functioning, problem solving and communication than both their mothers and fathers (4), AN participants rated more dysfunctional than their mothers on every subscale of the FAD (6), mothers and ED participants differed significantly on all FAD subscales except roles (3), scores were significantly different for all family members for adaptability (2) and ED participants rated their families as having fewer facilitative and more constraining family rules than did other family members (5). In contrast, Study 3 found no significant differences in the perceptions of ED participants and their fathers. 3.4.2. Mothers' views compare with fathers' When comparing the views of mothers and fathers, one study found that in AN families, mothers and siblings were significantly more dissatisfied with the family's functioning than fathers (2) and a further study found significant differences in the views of mothers and fathers for problem solving and affective responsiveness (3). In contrast, Study 9 did not find any significant differences between mothers and fathers on any of the subscales of the FAD. One of the studies which found differences between mothers and fathers did not use a control group (3). Therefore is not known whether these differences are specific to the clinical families or would have been present in control families. One study that did find differences in the control families found that control fathers rated problem solving more dysfunctional than control mothers but the opposite was true for general functioning (4). 3.4.3. ED participants' views compared with siblings' When comparing patients with an AN diagnosis with their sisters, there were no significant differences in their perceptions of emotional connectedness towards their parents; however there were significant differences in their perceptions of individual autonomy. Patients perceived they were non-autonomous towards their mothers compared with their sisters. This difference was corroborated by the mothers (5). In summary, the majority of studies that included family members found differences in their perceptions of family functioning. When differences were found, the participant with the eating disorder had a worse perception of family functioning than the other members of their family. In contrast to this, Study 10 found no significant differences in the perception of family functioning between any family members.

34

Table 2 Sample characteristics and methodology.a. Title

Aims related to this review

Design

Sample selection

Groups (N)

Age — mean (range)

F/M

Ethnicity

Measures of family functioning

1.Casper and Troiani (2001) USA

Family functioning in anorexia nervosa differs by subtype.

Cross-sectional

Consecutive referrals to ED unit over 16 months.

AN (22) C (45) P-AN (17) P-C (34)

AN-16.7 C-15.8

F

Caucasian

Family Assessment Measure (FAM)

2. Cook-Darzens, Doyen, Falissard, and Mouren (2005) France

Self-perceived family functioning in 40 French families of anorexic adolescents: implications for therapy.

Cross-sectional

Consecutive referrals to ED unit over 15 months.

37 F 3M

Not detailed

Family Adaptation and Cohesion Evaluation Scale (FACES III)

Do daughters with eating disorders agree with their parents' perception of family functioning?

Cross-sectional

Consecutive referrals to outpatient program over 5 years.

ED-18.3 (13–34)

F

Caucasian

The McMaster Family Assessment Device (FAD).

1. To compare the FF between Fam-AN and Fam-C. 2. To compare FF between family members.

Cross Sectional

Consecutive patients to an ED unit over a 16 month period.

F AN-15.7 (12–17.8) C-14.5 (12–17)

Not detailed

The FAD

Cross sectional

Consecutive patients seeking treatment at ED unit.

AN (40) M-AN (40) F-AN (40) S-AN (31) Fam-C (98) AN-R (24) AN-B (23) BN (41) EDNOS (38) M-ED(118) F- ED (96) AN (34) M-AN (34) F-AN (34) C (49) M-C (49) F-C (49) Fam-ED (51) Fam-C (51)

AN-14.97 (12–19) C-16.2

3. Dancyger, Fornari, Scionti, Wisotsky, and Sunday (2005) USA

1. To compare FF between AN-R and AN-B. 2. To compare FF between AN and controls. 3. To compare FF between family members. 1. To compare FF in Fam-AN with published non-clinical norms. 2. To compare FF between family members. 1. To compare FF between family members. 2. To compare FF between ED subtypes.

ED-19.39 (14–24)

F

Caucasian (92.1%) Black (5.9%) Native American (2%).

The Family Implicit Rules Profile

Longitudinal– initial assessment and 1 year follow-up.

Consecutive referrals to ED unit over 15 months.

AN- (35) M-AN (35)

AN-14.9

31 F 4M

Not detailed

The FAD The McMaster's Structured Interview of Family Functioning

Cross-sectional

Consecutive admissions to an ED unit with diagnosis of AN for over 4 years and a sister without ED. An a-select sample of patients with “intact families” treated in an AN unit.

AN (31) Sis-AN (31)

F Caucasian AN-15.7 (13–18) Sis-16.2 (11–21) Not Not detailed ED stated (15–24) C (15–24)

4. Emanuelli et al. (2004) Family functioning in adolescent Italy anorexia nervosa: A comparison of family members' perceptions

1. To compare implicit family Implicit family process rules in eating disordered and non-eating process rules between Fam-ED and Fam-C. disordered families. 2. To compare FF between family members. 3. To compare FF between ED subtypes. Difficulties in family functioning 1. To investigate the relationship 6. Gowers and North and adolescent anorexia nervosa. between the severity of AN and (1999) Family functioning and life events perceived family dysfunction. UK b 2. To compare FF ratings of AN, in the outcome of adolescent North, Gowers, and M-AN and clinicians. anorexia nervosa. Byram (1997) 3. To investigate the relationship UK between outcome and FF over the course of 12 months. 1. To compare the perceptions of 7.Karwautz et al. (2003) Perceptions of family Austria relationships in adolescents with FF between AN and their sisters. anorexia nervosa and their unaffected sisters. 1. To compare cohesion, 8. Kog and Vandereycken Family interaction in eating adaptability and conflict between disorder patients and normal (1989) Fam-ED and Fam-C. controls. Belgium 2. To compare differences between ED subtypes. 1. To compare FF between M-ED 9. Ma (2011) An exploratory study of the and F-ED. Hong Kong impact of an adolescent's eating disorder on Chinese parents' well-being, marital life and 5.Gillett, Harper, Larson, Berrett, and Hardman (2009) USA

Cross-sectional

Cross-sectional

Consecutive referrals of patients with ED for family therapy at a family treatment centre.

Fam-ED (30) Fam-C (30)

M-ED (18) F-ED (15)

ED (10–29) M-ED (30–59)

15 F 5M

Chinese

The Subjective Family Image Test

Two semi-structured tasks — decision making and conflict resolution tasks. The Leuven Family Questionnaire The FAD

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

Authors (year) Country

Authors (year) Country Title

10. McDermott et al.(2002) Australia

Aims related to this review perceived family functioning in Shenzhen, China: Implications for social work practice. Parent and child report of family functioning in a clinical child and adolescent eating disorders sample.

F/M

Ethnicity

Measures of family functioning

97.5% F

Not detailed

The Family Assessment Device — General Functioning Scale (FAD-GFS)

F

Not detailed

The Family Environment Scale (FES) The Family Dynamics Survey (FDS)

F

Caucasian

The FES

F

Caucasian

The FES

F

Not detailed

The Family Relations Scale (FRS)

Not detailed

F

Not detailed

The FAD

AN-15.2 C-8.4

F

Not detailed

Not detailed

F

Not detailed

The Kinston-Loader Family Interview, The FRS The Family Climate Scale, The CRS-Turbo, The Beavers Family Competence and Family Style. 3 observed tasks: 1. Joint answering of four questions from the FRS 2. Discussing a problem 3. Puzzle. The FAM

F-ED (35–59)

Cross-sectional

Recruited through a university hospital and student health centre. Recruited through three outpatient paediatric clinics.

BN (24) AN-B (13) C (41)

1. To compare FF between AN-R, AN-B and BN. 2. To compare FF between ED and Controls. 3. To compare FF between family members. Application of the family relations 1. To compare FF between Fam-AN and Fam-BN. scale to a sample of anorexics, 2. To compare FF between bulimics and non-psychiatric Fam-ED and Fam-Controls. controls: A preliminary study Eating disorders and family 1. To compare FF between AN interaction. and BN. 2. To compare FF between ED and Controls. 1. To compare FF in Fam-AN Anorexia nervosa in teenagers: with Fam-C. Patterns of family function. 2. To compare FF ratings by Nordic FAM-AN and observers.

Cross-sectional

Not detailed.

Cross-sectional

Consecutive patients assessed at an ED unit.

Cross-sectional

Patients attending an outpatient programme and attendees at self help groups. Consecutive referrals to an ED unit over 38 months.

13. Stern et al. (1989) USA

Family environment in anorexia nervosa and bulimia.

Long-term follow-up of patientreported family functioning in eating disorders after intensive day hospital treatment.

1. To examine patient reported FF at 2 years post-intensive treatment. 2. Do the improvements in FF reported by patients at discharge persist over the course of the follow-up period? 3. Is patient report of FF at admission or discharge associated with clinical outcome at admission, discharge or 2 year follow-up?

Cross-Sectional

Cross-sectional

Longitudinal; admission, discharge and 2 year follow up.

Data was collected from participants taking part in a larger study of the outcome of treatment and long-term follow-up being carried out by a day hospital program.

AN (25) C (44) CC (24) M-AN (25) M-C (44) M-CC (24) AN-R (20) AN-B (13) BN (24) M-ED (55) F-ED (2) Fam-C (57) AN (10) BN (7) Fam-ED (17) Fam-C (20) AN (12) BN (21) BN-S (8) C (27) AN (24) M-AN (24) F-AN (24) Fam-C (54)

AN (5) BN (52)

BN-20.8 AN-B21.2 C-20.9 AN-15.9 C-16.0 CC-15.8

AN-R19.4 AN-B20.9 BN-24.4 C-21.9 AN-20 BN-18.7 C-16.95

35

a Abbreviations used in Tables 2, 3 and 4: FF: family functioning, AN: participant with anorexia nervosa, AN-R: participant with anorexia nervosa restricting type, AN-B: participant with anorexia nervosa bulimic type, BN: participant with bulimia nervosa, BN-S: participant with bulimia nervosa simplex, EDNOS: participant with eating disorder not otherwise specified, ED: participant with an eating disorder, C: control participants, CC: clinical control participants, M-: mother of (AN/BN/C etc.), F-: father of …, P-: parent of … S-: sibling of …, Sis-: sister of …, Fam-C: control family, Fam-ED, eating disorder Family. b Companion paper reporting different aspects of findings.

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

ED-14.5 AN (42) (9–18) EDNOS (26) BN (12) P-ED (75) Fam-C (1462)

Family functioning and maternal distress in adolescent girls with anorexia nervosa.

17. Woodside, Lackstrom, Shekter-Wolfson, and Heinmaa (1996) Canada

Age — mean (range)

Consecutive referrals to an ED clinic over 20 months.

12. Sim et al. (2009) USA

16. Wallin and Hansson (1999) Sweden

Groups (N)

Cross-sectional

Family dysfunction in normal weight bulimic and bulimic anorexic families.

15. Waller, Calam, and Slade (1989) UK

Sample selection

1. To compare FF between family members. 2. To compare FF between ED subtypes. 3. To compare FF between Fam-ED and community norms. 1. To compare FF between BN and AN-B. 2. To compare FF between ED and Controls. 1. To compare FF between M-AN and M-C and M-CC.

11. Shisslak, McKeon, and Crago (1990) USA

14. Szabo, Goldin, and Le Grange (1999) South Africa

Design

36

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

Table 3 Methodological controls and results. Authors (year)

Controls

Significant results

Non significant and other findings

1.Casper and Troiani (2001)

Randomly selected from one school. Screened for past or present psychiatric diagnoses. All controls and clinical participants were Caucasian.

1. AN-B felt that task accomplishment, affective expression and involvement were problems within the family. Whereas problems with control were viewed as individual.

2. Cook-Darzens et al. (2005)

Participants from the validation study which established French norms for the FACES III measure. The controls were not matched in any way other than having an adolescent in the family.

3. Dancyger et al. (2005)

No controls

1. AN-B rated general FF significantly worse than AN-R and controls. 2. M-AN-B rated general FF significantly worse than M-C. 3. AN-B rated task accomplishment, communication and effective expression significantly worse in their families than AN-R and controls. AN-R paralleled controls. 4. AN-B rated themselves as more impaired for affective expression, involvement and control compared with controls. AN-R rated themselves not different from controls and better than controls for role performance and adherence to values and norms. 1. Fam-AN had significantly lower cohesion scores than Fam-C. 2. F-AN perceived significantly more adaptability than F-C. 3. Fam-AN had significantly higher ideal scores than perceived scores for cohesion and adaptability. This was the same for Fam-C. 4. In AN families fathers had the lowest levels of dissatisfaction and mothers and siblings were significantly more dissatisfied than fathers. 5. There were significant differences in the perceptions of AN family members concerning adaptability. 1. There were significant differences between family members for 4 FAD subscales; problem solving, communication, affective responsiveness and behaviour control. 2. M-ED and ED differed significantly on all subscales except roles. 3. M-ED and F-ED differed significantly for problem solving and affective responsiveness.

4. Emanuelli et al. (2004)

Control families were recruited by physicians' personal contacts with daughters between the ages of 11–18 years.

5.Gillett et al. (2009)

Control families were selected using a quota sampling method. Families were screed for eating disorder symptoms and other addictive tendencies. Families were matched for age, gender, income, race, family structure, religion and geographic region. This was systematically done.

6. Gowers and North (1999) North et al. (1997)

No control group

1. The hypothesis that AN would have higher cohesion and lower adaptability scores was not confirmed. 2. There were no trends found between FF scores and BMI, depression, number of admissions, type of treatment, age, age of onset or duration of illness. 3. There was no typical dysfunctional pattern for AN families and they ranged across a spectrum.

1. There were no significant differences between F-ED and ED on any subscales. 2. All ED scores on all subscales were in the unhealthy range, all F-ED scores except behaviour control were in the unhealthy range. For M-ED 5 subscales were in the unhealthy range except affective responsiveness and behaviour control. 3. No significant differences were found between diagnostic subgroups. 1. For communication AN perceived significantly 1. Fam-An perceived their FF as worse than Fam-C on every area of the FAD except affective worse than F-AN and M-AN whereas C perceived significantly worse than F-C but not M-C. involvement. 2. AN perceived worse FF on general functioning, problem solving and communication than both parents. 3. F-AN rated FF to be worse than M-AN on problem solving. 4. M-AN rated FF worse than F-AN on general functioning. 1. There were no significant differences between 1. For the Fam-ED the total FIRP score, for all the 3 diagnostic groups on any of the total or subscales and for the total FIRP + eating disorders scale combined were all lower (more subscales. constraining) than those of Fam-C, regardless of whose scores were being examined. All subscales were significant except for Inappropriate Caretaking of Parents. 2. ED rated their families as having fewer facilitative and more constraining family rules than did other family members, including siblings. There were significant differences for total FIRP, total FIRP + ED scale, Constraining Thoughts, Feelings and Self Subscale, Expressiveness and Connection Subscale and the Eating Disorder Rules Subscale. For the Monitoring subscale both the daughter and siblings were significantly different from mother. 1. Using the MROAS as a measures of severity 1. The families did not have a unified view of there were no significant associations between their functioning. AN were more critical than M-AN with higher mean scores on each subscale rating of FF (by any of the raters) and severity of the disorder. of the FAD. 2. In contrast to patient reports, mothers' 2. FF from the FAD-GFS and the overall score assessment of FF was not significantly associatfor the MCSIFF were strongly correlated with ed with outcome. Average Outcome Score (AOS). In both cases good outcome was associated with better judgement of FF. 3. Those with good FF maintained their relatively good outcome between 1 and 2 year follow-up. Those with poor FF improve over the

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

37

Table 3 (continued) Authors (year)

7.Karwautz et al. (2003)

8. Kog and Vandereycken (1989)

9. Ma (2011)

10. McDermott et al.(2002)

11. Shisslak et al. (1990)

12. Sim et al. (2009)

Controls

Significant results

second year to match the outcome of the former group. 4. Patients FF is rated as more dysfunctional at 1 year follow-up but then improves to a better level than at initial assessment by 2 year followup. This is significant for 4/7 subscales. Mothers scores are remain similar at all 3 time points. Sisters were used in place of matched controls. 1. There was a significantly lower score for Individual Autonomy for AN compared with their sisters. 2. AN had significantly higher perceived cohesion than their sisters. 3. AN perceived that they were significantly less autonomous towards their mothers compared with their sisters and mothers also reported greater lack of autonomy in their patient daughters towards them corroborating the patients' reports. Controls were selected from school records of a 1. Fam-ED showed significantly less disagreements between parents and children local high school and matched for social class, family size, sex and age of patient. They received and more stability in their behavioural interaction in the family. $45 for participating. 2. Fam-BN showed significantly less disagreement between parent and child than Fam-C. 3. Fam-AN-R reported more cohesion than Fam-BN and Fam-C. 4. When Fam AN-R and Fam-AN-B were combined together they showed significantly less disagreement than Fam-C and AN patients perceived their family as more cohesive than both BN and Controls. No control group. 1. F-ED and M-ED had mean scores above 2.00 (clinical cut-off set by Miller, Epstein, Bishop, & Keitner, 1985) on all 7 subscales of the FAD.

Non significant and other findings

1. Sisters did not differ in their perceptions of the marital relationship between mother and father and the reported scores for IA and EC between mother and father were within normal ranges. 2. When comparing sisters perceptions of their own relationships with each parent there were no significant differences between the sisters on measures of Emotional Connectedness.

1. Fam-AN-R did not significantly differ from Fam-BN or Fam-C on disagreement between parent and child.

1. F-ED three highest scoring subscales were: affective involvement, affective responsiveness and behaviour control. 2. M-ED three highest scoring subscales were: roles, behaviour control and affective involvement. 3. No significant differences between F-ED and M-ED on any of the subscales. 1. There was moderate concordance between Data was used from the Western Australia Child 1. Mean scores for the FAM-ED were significantly higher than for Fam-C. The number the parent and child scores on the FAD-GFS and Health Survey. The norms were based on data of FAM-ED falling within the clinical range was no significant differences were found. obtained from 2373 children and adolescents significantly more than the 12% for the aged 4–16 living in 1462 randomly selected 2. No significant differences were found community sample; 62.1% of the families scored between diagnostic groups on either child or households. within clinical range according to the child parent ratings. scores and 61.2% according to parents' scores. 3. When comparing symptom presentation 2. When comparing symptom presentation (restrictor vs those who had binge-purged in (restrictor vs those who had binge-purged in last 3 months) the total FAD-GFS score was not last 3 months) there were significant difference significantly different. on two items of the measure; binge-purgers reported greater difficulty in planning family activities and had greater difficulty confiding in each other. 1. There were no significant differences between 1. BN and AN-B considered their families as Controls were volunteers from university the three groups for achievement orientation, displaying significantly less cohesion, undergraduate psychology courses. Controls moral-religions emphasis, organisation, control, expressiveness and orientation toward were not matched. influencing decisions, independence and recreational activities than controls. acceptance. 2. BN families were perceived as displaying significantly more conflict than controls 3. AN-B perceived their families as discouraging independence to a significantly greater degree than BN or controls. 4. BN perceived their families as significantly less oriented toward intellectual-cultural activities than controls. 5. BN and AN-B perceived their families as significantly less emotionally supportive and as needing counselling to a significantly greater degree than controls. 6. The quality of communication in BN and AN-B families were perceived as significantly worse than controls. This study had two control groups; 1) a clinical 1. M-AN reported significantly more conflict in 1. No significant differences were found their family than M-C and M-CC. between M-AN, M-C or M-CC for family control group of adolescent girls with Insulin cohesion on family expressiveness. Dependent Diabetes Mellitus (IDDM) recruited (continued on next page)

38

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

Table 3 (continued) Authors (year)

Non significant and other findings

Controls

Significant results

from a diabetes clinics and 2) a community control group of adolescent girls from a paediatric wellness clinic.

2. AN exhibited more depressive symptoms than both C and CC. 3. M-AN reported lower levels of parenting satisfaction than M-C but not M-CC. 4. M-AN reported lower levels of alliance with the child's other parent than M-C and M-CC. 5. M-AN perceived more emotional consequences of the illness than M-CC. 1. ED families on the whole rated themselves 1. There were significant differences found as less supportive of each other and less between groups on 5/10 subscales of the FES: cohesion, expressiveness, conflict, achievement encouraging of open expression of feelings than orientation and active recreational orientation. control families as well as more likely to have conflictual interactions. 2. The only significant difference between the three diagnostic categories was on achievement 2. In both clinical and control families parents view family functioning more positively than orientation where P-BN rated their families as their children. higher than P-AN-B or P-AN-R. 3. There were significant differences between parents and subjects on 5/10 subscales: cohesion, expressiveness, conflict, intellectualcultural orientation and moral-religious emphasis. In each incident the parent rated the family higher than the daughter, except on conflict. These differences were regardless of diagnosis. 1. Mothers showed no between group 1. Fathers revealed a significant difference differences on any subscales. between mean scores on two subscales: 2. There are 6 subscales therefore more flexibility and family hierarchy. For flexibility subscales had no significant differences F-C scored significantly higher than F-BN for family hierarchy F-C scored significantly higher between either ED groups and controls. than F-AN. 2. Daughters demonstrated significant differences for differentiation, BN higher than C, family hierarchy, C higher than AN and family idealisation, C higher than AN. 3. Fam-Ed and Fam-C showed significant between group differences for family hierarchy and family idealisation, C scoring higher than AN in both cases. 1. Generally women with ED rated their family 1. AN rated their families as significantly more as less healthy, particularly BN-S. unhealthy in affective involvement and behaviour control than controls. 2. BN rated their families as significantly worse for problem solving, behaviour control and affective involvement than controls. 3. BN-S rated family interaction as significantly poorer than comparisons on all scales except behaviour control. 1. AN showed no difference to controls on scores 1. M-AN and F-AN rated cohesion higher than of cohesion and F-AN and F-C did not differ on P-C. They also rated chaos as lower than P-C. closeness or chaos. 2. AN rated chaos as lower than controls. 3. M-AN rated closeness lower and chaos higher 2. Even though most AN families were rated high on cohesion there was a range of scores. than M-C. 3. Not all AN families were rated as 4. AN rated chaos higher than controls. dysfunctional. 5. Fam-AN were rated lower for family competence and family style by observers 6. Fam-AN rated higher on cohesion and hierarchical organisation. 7. Differences between self report and observer ratings were most pronounced on the chaos-rigidity scale with mother and patients rating families as chaotic whilst observers rated them as rigid. 1. Results are in the same direction as the self 1. FAM scores were consistently more rating scale for the general scale of the FAM but favourable at discharge than admission. 2. For the self rating scale of the FAM there is a they are not statistically significant. significant difference between those with good 2. Although not significant the trend for subjects or poor outcomes at 2 year follow-up but not at with poor outcome showed deterioration between discharge and 2 year follow-up but not discharge. back to admission level.

13. Stern et al. (1989)

Control subjects were matched by age, sex and race and were recruited through advertisements.

14. Szabo et al. (1999)

Control families were randomly selected on the basis of their being a daughter between the ages of 14–27 within the family who did not have a history of an eating disorder. No further detail of the selection process is given.

15. Waller et al. (1989)

Controls were volunteers with no history of psychiatric disorders. No demographic information was detailed.

16. Wallin and Hansson (1999)

The control group were from a study in which the same methods of observer ratings were being used. The control group were not matched in any way.

17. Woodside et al. (1996)

No control group

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43 Table 4 Summary of findings for comparisons of different aspects of family functioning between ED families and controls. Significantly highera than controls General family functioning Cohesion

Adaptability Conflict

Affective expression

Affective involvement

Communication

Task accomplishment Problem solving Achievement orientation Role performance Family hierarchy Behaviour control

Adherence to values and norms Constraining implicit family rules

Significantly lower than controls

No significant difference to controls

ED (10 and 15b) AN (1,4) AN mother and AN (2) AN (16) fathers (16) AN (8) BN (11,13) AN mothers (12) AN-B (11,13) BN (8) AN fathers (2) BN fathers (14) AN (16) ED (8) AN-B (11) BN (11) AN-B parents (13) AN mothers (12) AN mothers (12) BN (11) AN-R (1) AN-B (1,11) AN (4) AN (15) AN-R (1) BN (15) AN-B (1) AN (4) AN-B (1,11) AN-R (1) BN (11) AN (4) AN-B (1) AN-R (1) BN (15) AN (15) AN (4) BN (13) ED (11) AN-R (1) AN (4) AN-B (1) AN (16) AN (14) AN (15,4) AN-B (1) BN (15) AN-R (1) ED (11) AN-R (1) AN-B (1) ED (5)

a ‘Higher’ indicates a higher level of this element within the family not higher levels of functioning or dysfunction. b Numbers indicate the studies that had this finding.

3.5. Relationship between family functioning and outcome/recovery Two studies investigated the relationship between family functioning and recovery. The main finding from these studies was that perceived family functioning was strongly correlated with outcome (6). Good outcomes were associated with a more positive patient perception of family functioning. By contrast, the mothers' perception of family functioning was not significantly associated with outcome. This study found no significant association between severity of the eating disorder and ratings of family functioning, indicating that those with good

Table 5 Findings for significant differences in elements of family functioning between eating disorder subgroups. Element of family functioning

Findings

Task accomplishment Communication Effective expression Cohesion Planning activities Confiding in each other Discouraging independence Achievement orientation Problem solving Affective involvement Behaviour control Role performance

AN-B worse than AN-R (1) AN-B worse than AN-R (1) AN-B worse than AN-R (1) AN-R and AN-B higher than BN (8) AN-B and BN higher than AN-R (10) AN-B and BN lower than AN-R (10) AN-B higher than BN (11) BN higher than AN-B and AN-R (13) Bulimia simplexa worse than AN and BN (15) Bulimia simplex worse than AN and BN (15) Bulimia simplex worse than AN and BN (15) Bulimia simplex worse than AN and BN (15)

a Bulimia simplex is used to describe a patient with bulimia nervosa who has no history of restriction.

39

outcome and better perceptions of family functioning were not necessarily less severe at admission (6). Those with good family functioning maintained their relatively good outcome between one and two year follow-up. Those with poor family functioning improved over the second year to match the outcome of those with better family functioning (6). In Study 17 family functioning was significantly different between those with good or poor outcomes at two year follow-up but not at discharge. The patients' perception of family functioning significantly improved from admission to discharge (17); however, they then perceived the family as becoming more dysfunctional at one year followup, but improving to above that at initial assessment by two year follow-up (6). This was significant for four of the seven subscales of the FAD. Mothers' scores however remained similar at all three time points (6). 3.6. Study quality 3.6.1. Samples 3.6.1.1. Sample size. A potential source of bias was the sample sizes and the power to detect significant differences should they exist. The majority of studies did not report power calculations, therefore, for the purpose of the current review, samples were considered to be underpowered if they were smaller than 50, for those using ANOVA or t-tests and smaller than 25 in each group, for those using correlations. These numbers are based on a medium effect size of 0.5 and power of 0.8 (Clark-Carter, 2004). Using these criteria, studies 1, 2, 6, 7, 8, 9, 11, 14, 15 and 16 were underpowered for at least one of the statistical calculations they conducted. In two studies this limited the statistical procedures that could be done (6,8). 3.6.1.2. Recruitment, response rates and demographics. The majority of studies used participants who were treatment seeking and where they and their families had completed measures at assessment, biasing towards those with higher motivation and more supportive family members. Response rates varied across the studies ranging from 53% (3) to 100% (4) and those with lower response rates would therefore be biased towards patients willing to participate. Studies predominantly included families where the participant with the eating disorder was the child in the family, limiting the generalisability of findings to families where the person with the eating disorder is the partner or parent. Whilst the family member with the eating disorder was predominantly the child, the lowest mean age was 14.5 years indicating that the studies did not include those with early onset eating disorders. It could be that the families where the eating disorder is present at age 10–12, for example, have different patterns of family functioning to those where it develops later. Other variations in the participants that could potentially impact on family functioning, but were not considered by all papers, were co-morbidities or specific family stressors. Of the studies that reported gender and ethnicity the samples were predominantly Caucasian females, questioning the generalisability of the results to other ethnicities and male patients with eating disorders. There were studies that gave no demographic information thus limiting the ability to assess generalisability. 3.6.1.3. Family members included. The studies varied in the family members included. Five studies included mothers, fathers and siblings (2,5,7,8,14), seven included mothers and fathers (1,3,4,9,10,13,15), two included mothers (6,12) and three did not include any family members (11,15,17). Family Systems Theory suggests the need to look at the family as a whole and so studies using information from multiple family members would provide a better picture of the family's functioning.

40

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

3.6.1.4. Control groups. Thirteen studies used a control or comparison group. These ranged from having unmatched controls (1,4,11,14–16) to using sisters as a control group (7) or matching controls on a number of demographic factors (5,8,13). Two studies used norms or community samples (2,10) and one study included a clinical control and a community control group (12). Studies using matched controls on factors such as number and age of siblings, ethnicity and socio-economic status would be more successful at controlling for these potentially confounding variables. 3.6.2. Measurements All of the studies used validated measures with known reliability and validity. Ten studies used measures based on a theoretical framework that had previously been shown to distinguish between families with healthy or unhealthy functioning (1–6,9,10,15,17). However, the reliance on self-report measures is not desirable due to the impact of denial or social desirability on responses. Only two studies supplemented the self-report measures with observations (8,16). Study 8 did not detail the observational tasks; however Study 16 used The Beavers Family Competence and Family Style (Beavers & Hampson, 1990) observational measure which has been shown to differentiate families with and without clinical difficulties. 3.6.3. Study design The majority of the studies (15 out of 17) used a cross sectional design. This gives a static view of family functioning, which may be misleading. Some of the studies did not detail at what point of treatment the participants were at, whether they still lived with family members or whether they were completing the measures retrospectively. If the measures were completed retrospectively this could reduce the reliability of the responses. Another potential difficulty of using participants at the assessment stage of treatment is that patients may be affected by starvation effects prior to any normalisation of eating. 4. Discussion The current review explored the research evidence on family functioning in families where a member has an eating disorder. The findings will be discussed in relation to the specific aims of the review. 4.1. Differences in family functioning between eating disorder families and controls When general family functioning was considered, it was found that ED families had worse family functioning than controls and a higher proportion were rated above clinical cut-offs for family functioning. However, when different components of family functioning were considered separately, the evidence was conflicting. 4.1.1. Findings in relation to models of family functioning According to Minuchin et al. (1978) psychosomatic families were thought to be enmeshed, over-protective, rigid, avoidant of conflict and lacking conflict resolution skills. The studies in the current review did not measure enmeshment as a construct; however it was thought that this might be represented as having high cohesion and less defined roles within a family. Findings that supported the notion of the ‘psychosomatic’ family were that, compared with controls, AN families were observed to have more cohesion and lower family hierarchy and ED families had more constraining family rules and less conflict. Participants with AN lacked individual autonomy compared with their sisters and BN families were less flexible than controls. However, there were findings that did not support the notion of the ‘psychosomatic’ family. In the majority of studies, ED families reported lower or similar cohesion to controls. AN families scored higher on role performance and a number of studies found no differences between ED

families and controls on measures of adaptability and rigidity. Studies found BN and AN families to have more conflict than controls and other studies reported no differences. The majority of the current findings did not support the notion of a ‘psychosomatic’ family and the general picture indicated that these characteristics were variable across ED families and between diagnostic groups. This was similar to that reported by Eisler (2005) and Kog and Vandereycken (1985) who concluded that there was no consistent pattern of family functioning in ED families and that there was a lack of support for the ‘psychosomatic’ family model. The aspects of family functioning believed to be relevant to eating disorders according to the McMaster model (Epstein et al., 1978) and the process model (Steinhauer et al., 1984) are; problem-solving/task accomplishment, communication, affective responsiveness, affective involvement, behaviour control, role performance and adherence to values and norms. Findings that supported these models were that AN-B and BN participants rated task accomplishment significantly worse than controls, as did BN and AN participants for problem solving. The quality of communication within the family was rated lower by participants with AN, AN-B and BN. Both BN, AN and AN-B families were rated poorer for affective responsiveness and BN and AN-B for affective involvement. AN-R families scored higher for role performance and adherence to values and norms and AN and BN families scored lower for behaviour control. Findings that did not fit with these models were that a number of studies found no significant differences between at least one ED subgroup and controls for task accomplishment, affective expression, affective involvement, communication, problem solving, role performance, adherence to values and norms and behaviour control. Overall, the findings were supportive of these models; however, in the majority of studies the scores on different aspects of family functioning were variable with little consistency between family members. 4.2. Differences in family functioning between eating disorder subgroups Findings relevant to family functioning and eating disorder diagnoses were also mixed. 4.2.1. General family functioning When considering general family functioning, two studies found no significant differences between the diagnostic groups. The one study that reported differences found those with binge–purge behaviours rated family functioning significantly worse than those with restricting behaviours. The explanation was that AN-R participants may have been influenced by denial, idealisation and conflict avoidance when reporting a lack of dysfunction within the family (Casper & Troiani, 2001). 4.2.2. Specific elements of family functioning Focussing on specific areas of family functioning, AN families reported significantly more cohesion but less achievement orientation than BN families and BN participants reported greater difficulty in planning family activities and confiding in each other than AN-R participants. These findings were similar to those of Kog and Vandereycken (1985) who found that AN families reported higher levels of cohesion; however they also reported differences between AN and BN families on aspects of family functioning that were not observed in the current review, for example conflict. Possible explanations for this and other inconsistent findings were the range of diagnoses included and how they were defined, for example whether studies distinguished between AN-R and AN-B or BN and bulimia simplex. Also, eating disorder diagnoses are not fixed and patients can move between diagnoses over time (Braun, Sunday, & Halami, 1994; Bulik, Sullivan, Fear, & Pickering, 1997). People with a diagnosis of EDNOS were included in two of the studies and this diagnosis includes elements of both AN and BN. These two studies did not find significant differences between the diagnostic groups but one found differences between those with restricting compared with binging–purging behaviours. It may therefore be misleading to try and

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

compare diagnostic groups as separate entities and unlikely that any distinct pattern of family functioning would be found. The recent addition of specifiers in the DSM-5 for qualification for the different subtypes of AN and the length of time symptoms are present could lead to less migration between diagnoses. These changes are recent and therefore the impact at present is unclear. 4.3. Differences in the perspectives of family members The review by Kog and Vandereycken (1985) described very few differences between family members as they were often not included in studies up to that time. However, they did report that mothers and daughters perceived more dysfunction than fathers. The most consistent finding from the current review was that patients had a more critical view of their family's functioning than one or both of their parents. The same was true for control families but not to the same extent. There is a possibility that the patients' responses reflected cognitive misrepresentations or were being influenced by starvation effects (Stern, Dixon, Jones, & Lake, 1989). There were discrepancies in the views of different parents although the nature of the discrepancy varied between studies. One study reported that AN mothers and siblings reported more dissatisfaction with family functioning than fathers. Another study found that fathers and daughters agreed on the family's difficulties but mothers rated less dysfunction. Two studies reported no significant differences in the perceptions of family members. However one of these studies did not separate the perspectives of mothers and fathers but combined these into one overall parent score. Given that other studies found differences in the perspectives of mothers and fathers, it may be that combining scores masked any potential differences. In the studies that found differences in the perspectives of family members but did not use control groups it is not possible to identify whether these differences are specific to ED families. The differences in the perspectives of family members may have implications for treatment as they may act to maintain the disorder to some extent. It was suggested that parents may not be aware how their child perceives the family's functioning and this could be addressed in family therapy. In addition, if there are specific elements of family functioning, such as communication or problem solving, where the differences in perceptions are significant these could be targeted in treatment. 4.4. Summary of family functioning in eating disorder families The findings of the current review suggested that ED families perceive themselves, and are observed to be, more dysfunctional than control families. However, the areas of dysfunction varied and there does not appear to be a consistent pattern of family dysfunction for ED families as a whole or for different types of eating disorder. This could suggest that the specific areas that a family function less well in are present prior to the onset of the eating disorder and become more pronounced when the family has to cope with and adjust to having a member with a potentially life threatening illness. These particular areas of family dysfunction may vary from family to family although there is some evidence that the areas identified by the McMaster and Process models of family functioning have particular relevance in ED families. However, this is not to suggest that such areas of dysfunction were causal nor intended to blame the family. 4.5. Family functioning and outcome/recovery The relationship between family functioning and outcome or recovery from eating disorders has not, to our knowledge, been addressed in previous literature reviews. The findings of these two studies were that a more positive view of family functioning by the patient was associated with better outcome, irrespective of the severity of the disorder.

41

However, the mother's perception of family functioning was not significantly associated with outcome. No other family members were included in these studies. Both studies found that patients' perceptions of family functioning improved from admission to discharge. Those with good perceptions of family functioning at assessment maintained their relatively good outcome between one and two year follow-up. Those with poor family functioning at assessment had poorer outcome at one year follow-up but improved over the second year to match the outcome of those with better family functioning. This could suggest that worse perceptions of family functioning may potentially act as a maintenance factor for the eating disorder. For those with poorer treatment outcomes, perceptions of family functioning deteriorated between discharge and two year follow-up but did not return to the level it was at admission. However the mothers' perceptions remained similar at all time points. This could suggest that the families' functioning had not changed; however having active symptoms of an eating disorder may influence the patients' perceptions of family functioning. Another possibility is that the mothers were less aware of changes in family functioning or were being influenced by social desirability when completing the measures. The current review included two studies with this as a focus and it is recommended that further research be conducted in the area. 4.6. Quality assessment Ten of the 17 studies had sample sizes limiting the power of the statistical analyses used. This would have increased the chance of a Type 2 errors and may have accounted for some of the variability in the findings. The samples were also biased due to the recruitment process to those seeking treatment. The measures used to assess family functioning (e.g. FAD, FED, FACES) were reported as having good reliability and validity; however Folse (2007) has criticised the use of these measures in eating disorder research for their lack of reliability and validity testing and for not being designed specifically for eating disorders. Folse (2007) stated they do not consider relevant factors such as family concerns about shape, weight and eating, consistency in fostering open expression of emotions and maintaining boundaries. As these measures are not specific to eating disorders, they can be used to compare family functioning across different psychiatric diagnoses. However, for future research in family functioning and eating disorders, it could be more informative to use a measure which includes these other factors, such as the Family Experience with Eating Disorders Scale (FEEDS; Folse, 2007). Other than Studies 5 and 7 the quality of reporting was often limited, particularly regarding the characteristics of the sample and potentially confounding variables. 4.7. Further investigation and clinical implications Kog and Vandereycken (1985) highlighted the need for more systematic studies, preferably using both observational and self-report measures. However, since their recommendations, studies have continued to rely on self-report measures. There remains a need for researchers to consider using observational methods in order to reduce the impact of the biases of self-report measures affected by factors such as denial or social desirability. Future research using self-report measures might also consider using measures of family functioning developed for an eating disorder population, such as the FEEDS (Folse, 2007). Kog and Vandereycken (1985) recommended using reports from different family members and more recent studies have began to do this. However there is still a need for studies to assess family functioning over time and to consider the maintaining influence of family functioning and its relationship with outcome and recovery.

42

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43

The findings from this current review have a number of clinical implications. The specific areas of family dysfunction and differences in perspectives between patients and parents could be involved in the maintenance of the eating disorder and might be addressed in treatment offered. Treatments empowering parents to support their child through recovery and improve the functioning of their families can change the focus of blaming the family to supporting them to be part of the solution. The lack of any clear pattern of family dysfunction highlighted the need to assess each family individually and not make assumptions about the areas of difficulty. The findings from the longitudinal studies highlighted the need to also consider family functioning after initial recovery, particularly following discharge from inpatient admission, in order to prevent relapse.

5. Conclusion The current review aimed to systematically review the literature on family functioning in ED families, including the impact of family functioning on recovery and outcome. There was evidence to suggest that ED families have worse family functioning than controls; however there was little evidence found for a typical pattern of dysfunction for ED families or the diagnostic categories. The evidence suggested that ED families vary in their family functioning and there are often discrepancies in the perceptions of the different family members. This should encourage clinicians to assess each family thoroughly and explore the perceptions of each member to establish a greater understanding of the family functioning as a whole. The evidence suggested links between good perceptions of family functioning and better outcomes. This supported the notion that poor family functioning may be maintaining the disorder and emphasised the importance of working with the families with the aim of improving their functioning in a way that is supportive to the patient's initial and sustained recovery.

Appendix A. Definitions of elements of family functioning Achievement orientation — the emphasis the family places on achievements. Adaptability — the extent to which a family system is flexible and able to change and adapt to different circumstances or stressors. Measures of adaptability range from rigid to chaotic. A middle range is considered to be most well functioning. Adherence to values and norms — the emphasis the family places on behaving in a way that is acceptable to their values and norms. Affective involvement — the degree to which the family shows an interest in and values its individual members. The level of concern for each other. Affective responsiveness/expression — the family's ability to respond to a range of stimuli with the appropriate quality and quantity of feeling and emotion. Appropriate expression of emotion. Behaviour control — the way in which a family establishes and maintains standards for the behaviour of its members. The clarity of the rules within the family. Cohesion — the emotional bonding that family members have towards each other and the extent that boundaries are maintained. Measure of cohesion range from enmeshed to disengaged. A middle range is considered to be most well functioning. Communication — how well information is verbally exchanged within the family. The clarity and directness of verbal interactions. Conflict — in relation to this paper conflict is viewed in consideration to conflict avoidance or the family's ability to resolve conflict. Cross-generational blurring — lack of clear roles of the different family members, coalitions between parents and children. Dependency — lack of individual autonomy and independence.

Enmeshment — an extreme form of proximity and intensity in family interactions. An extreme level of cohesion, over-involvement, blurring of boundaries and roles. Family Hierarchy — the extent that there is a hierarchy or power structure for taking responsibility for leadership and setting limits. The extent to which there are internal boundaries or parent–child coalitions within the family. Flexibility — a high amount of adaptability to new or stressful situations. Implicit family rules — unwritten family norms that govern the family members' behaviours. Organisation and structure — the strength of the roles, hierarchy and structure within the family. Over-protectiveness — a high degree of concern for each other's welfare. Can result in a lack of autonomy for individual family members. Problem solving/task accomplishment — the family's ability to resolve problems or complete tasks and maintain an effective level of family functioning. Rigidity — inflexibility within the family and a commitment to keeping things the same. A lack of adaptability. Role performance — the differentiation of tasks to different family members. The clarity and appropriateness of roles within the family and the distribution of responsibility and accountability.

References⁎ Beavers, R., & Hampson, R. (1990). Successful families. Assessment and intervention. New York: Norton. Braun, D. L., Sunday, S. R., & Halami, K. A. (1994). Psychiatric co morbidity in patients with eating disorders. Psychological Medicine, 24, 859–867. Bulik, C. M., Sullivan, P. F., Fear, J. L., & Pickering, A. (1997). Predictors of the development of bulimia nervosa in women with anorexia nervosa. The Journal of Nervous and Mental Disease, 185, 886–895. *Casper, R. C., & Troiani, M. (2001). Family functioning in anorexia nervosa differs by subtype. International Journal of Eating Disorders, 30(3), 338–342. Clark-Carter, D. (2004). Quantitative psychological research, a student's handbook. Hove: Psychology Press. Commonwealth of Kentucky (2001). Retrieved from http://www.jjab.ky.gov/terms.htm *Cook-Darzens, S., Doyen, C., Falissard, B., & Mouren, M. (2005). Self-perceived family functioning in 40 French families of anorexic adolescents: Implications for therapy. European Eating Disorders Review, 13(4), 223–236. *Dancyger, I., Fornari, V., Scionti, L., Wisotsky, W., & Sunday, S. (2005). Do daughters with eating disorders agree with their parents' perception of family functioning? Comprehensive Psychiatry, 46(2), 135–139. Eisler, I. (2005). The empirical and theoretical base of family therapy and multiple family day therapy for adolescent anorexia nervosa. Journal of Family Therapy, 27(2), 104–131. *Emanuelli, F., Ostuzzi, R., Cuzzolaro, M., Baggio, F., Lask, B., & Waller, G. (2004). Family functioning in adolescent anorexia nervosa: A comparison of family members' perceptions. Eating and Weight Disorder, 9, 1–6. Epstein, N.B., Baldwin, L. M., & Bishop, D. S. (1983). The McMaster Family Assessment Device. Journal of Marital and Family Therapy, 9, 171–180. Epstein, N.B., Bishop, D. S., & Levin, S. (1978). The McMaster model of family functioning. Journal of Marital and Family Therapy, 4(4), 19–31. Folse, V. N. (2007). The family experience with eating disorder scale: Psychometric analysis. Archives of Psychiatric Nursing, 21(4), 210–221. *Gillett, K. S., Harper, J. M., Larson, J. H., Berrett, M. E., & Hardman, R. K. (2009). Implicit family process rules in eating-disordered and non-eating-disordered families. Journal of Marital and Family Therapy, 35(2), 159–174. *Gowers, S., & North, C. (1999). Difficulties in family functioning and adolescent anorexia nervosa. British Journal of Psychiatry, 174, 63–66. Greenhalgh, T. (1997). How to read a paper: Getting you bearings (deciding what the paper is about). British Medical Journal, 315, 243–246. Jack, S. (2001). Working with families. In J. E. Mitchell (Ed.), The outpatient treatment of eating disorders — A guide for therapists, dieticians and physicians. Minneapolis: University of Minnesota Press. *Karwautz, A., Nobis, G., Haidvogl, M., Wagner, G., Hafferl-Gattermayer, A., Wöber-Bingöl, C., et al. (2003). Perceptions of family relationships in adolescents with anorexia nervosa and their unaffected sisters. European Child & Adolescent Psychiatry, 12(3), 128–135. Kog, E., & Vandereycken, W. (1985). Family characteristics of anorexia nervosa and bulimia: A review of the research literature. Clinical Psychology Review, 5(2), 159–180. *Kog, E., & Vandereycken, W. (1989). Family interaction in eating disorder patients and normal controls. International Journal of Eating Disorders, 8(1), 11–23.



Denotes studies included in the systematic review.

A. Holtom-Viesel, S. Allan / Clinical Psychology Review 34 (2014) 29–43 Le Grange, D., Lock, J., Loeb, K., & Nicholls, D. (2010). Academy for eating disorders position paper: The role of the family in eating disorders. International Journal of Eating Disorders, 43(1), 1–5. *Ma, J. L. C. (2011). An exploratory study of the impact of an adolescent's eating disorder on Chinese parents' well-being, marital life and perceived family functioning in Shenzhen, China: Implications for social work practice. Child and Family Social Work, 16, 33–42. *McDermott, B.M., Batik, M., Roberts, L., & Gibbon, P. (2002). Parent and child report of family functioning in a clinical child and adolescent eating disorders sample. Australian and New Zealand Journal of Psychiatry, 36(4), 509–514. Miller, I. W., Epstein, N.B., Bishop, D. S., & Keitner, G. I. (1985). The McMaster Family Assessment Device: Reliability and validity. Journal of Marital and Family Therapy, 11, 345–356. Minuchin, S., Rosman, B.L., & Baker, L. (1978). Psychosomatic families: Anorexia nervosa in context. Cambridge MA: Harvard University Press. Mosby's Medical Dictionary, 8th edition (2009). Retrieved from http://medical-dictionary. thefreedictionary.com/family+functions"Nfamilyfunctions NHS CRD (). Systematic reviews — CRD's guidance for undertaking reviews in health care. York Publishing Services Ltd. (Retrieved from http://www.york.ac.uk/inst/crd/pdf/ Systematic_Reviews.pdf) *North, C., Gowers, S., & Byram, V. (1997). Family functioning and life events in the outcome of adolescent anorexia nervosa. British Journal of Psychiatry, 171, 545–549. Polivy, J., & Herman, C. P. (2002). Causes of eating disorders. Annual Review of Psychology, 53, 187–213. *Shisslak, C. M., McKeon, R. T., & Crago, M. (1990). Family dysfunction in normal weight bulimic and bulimic anorexic families. Journal of Clinical Psychology, 46(2), 185–189.

43

*Sim, L. A., Homme, J. H., Lteif, A. N., Vande Voort, J. L., Schak, K. M., & Ellingson, J. (2009). Family functioning and maternal distress in adolescent girls with anorexia nervosa. International Journal of Eating Disorders, 42, 531–539. Skinner, H., Santa-Barbara, J., & Steinhauer, P. D. (1983). The family assessment measure. Canadian Journal of Community Mental Health, 2, 91–105. Steinhauer, P. D., Santa-Barbara, J., & Skinner, H. (1984). The process model of family functioning. Canadian Journal of Psychiatry, 29(2), 77–88. *Stern, S. L., Dixon, K. N., Jones, D., & Lake, M. (1989). Family environment in anorexia nervosa and bulimia. International Journal of Eating Disorders, 8(1), 25–31. *Szabo, C. P., Goldin, J., & Le Grange, D. (1999). Application of the family relations scale to a sample of anorexics, bulimics and nonpsychiatric controls: A preliminary study. European Eating Disorders Review, 7(1), 37–46. Treasure, J., Sepulveda, A.R., MacDonald, P., Whitaker, W., Lopez, C., Zabala, M., et al. (2008). The assessment of the family of people with eating disorders. European Eating Disorders Review, 16(4), 247–255. *Waller, G., Calam, R., & Slade, P. (1989). Eating disorders and family interaction. British Journal of Clinical Psychology, 28(3), 285–286. *Wallin, U., & Hansson, K. (1999). Anorexia nervosa in teenagers: Patterns of family function. Nordic Journal of Psychiatry, 53(1), 29–35. Whitney, J., & Eisler, I. (2005). Theoretical and empirical models around caring for someone with an eating disorder: The reorganization of family life and inter-personal maintenance factors. Journal of Mental Health, 14(6), 575–585. *Woodside, D. B., Lackstrom, J., Shekter-Wolfson, L., & Heinmaa, M. (1996). Long-term follow-up of patient-reported family functioning in eating disorders after intensive day hospital treatment. Journal of Psychosomatic Research, 41(3), 269–277.

A systematic review of the literature on family functioning across all eating disorder diagnoses in comparison to control families.

The objectives of this review were to systematically identify and evaluate quantitative research comparing family functioning (a) in eating disorder f...
435KB Sizes 0 Downloads 0 Views