Family-Based Treatment for Adolescents with Anorexia Nervosa: A Dissemination Study

Abstract

Objective

This open trial of Family-Based Treatment for Anorexia Nervosa was completed in order to assess the dissemination of this treatment, including effectiveness, fidelity, and acceptability.

Method

Fourteen adolescents with Anorexia Nervosa were recruited with mean age 14.0±1.5 years (range 12–17 years). Therapists were trained using a workshop, manual and weekly supervision. Sessions were videotaped and rated for treatment fidelity. Pre- and post-treatment assessments were compared.

Results

Weight was significantly increased by an average of 7.8 kg. Dietary restraint showed significant improvement, as did interoceptive deficits and maturity fears. Of the 9 participants who had secondary amenorrhea at baseline, 8 had regained menstrual function. Treatment fidelity was rated as at least considerable 72% of the time in phase I of the treatment. Adolescents and parents found the treatment to be acceptable.

Discussion

This preliminary investigation of the dissemination of Family-Based Treatment for adolescents with Anorexia Nervosa indicates that this treatment is effective not only for weight restoration, but also in improving some psychological symptoms including dietary restraint, interoceptive deficits, and maturity fears. In addition, this treatment was adopted with considerable fidelity and was acceptable to adolescents and parents.

For children and adolescents with Anorexia Nervosa, the American Psychiatric Association Guidelines for Eating Disorders (American Psychiatric Association, 2006) suggest that family treatment is the most effective intervention, citing models based on the Maudsley approach (Eisler et al., 2000; Robin et al., 1999). The Maudsley method, is now also known as, Family-Based Treatment for Anorexia Nervosa and has been manualized (Lock, le Grange, Agras, & Dare, 2001). It is an intensive, outpatient treatment which utilizes the family as a resource to re-nourish the affected child. This treatment was first developed at the Maudsley Hospital in London, England in the 1980s, and its efficacy was first described within a small randomized controlled trial published by Russell et al. (1987). In patients with an onset of illness before age 18, and duration of illness of less than three years, intensive family treatment was superior to individual treatment. At five-year follow-up, these differences between groups were maintained (Eisler et al., 1997). In a subsequent US study, when a similar type of family treatment was compared to individual therapy in adolescents with Anorexia Nervosa, both groups improved, but those in the family treatment had a more rapid recovery (Robin et al., 1999).

Family-Based Treatment has been manualized, in a form which involves approximately 9–12 months of treatment (Lock et al., 2001). One therapist is involved, along with a medical doctor to assess physical health. The efficacy of the manualized program has been demonstrated in a randomized controlled trial comparing six months of treatment to twelve months in 86 participants. There were no significant differences between six and twelve months of treatment, and improvements were maintained at long-term follow-up (Lock, Agras, Bryson, & Kraemer, 2005; Lock, Couturier, & Agras, 2006). An average weight gain of 6.8 kg was achieved, and at the end of treatment, 96% no longer met criteria for Anorexia Nervosa.

Despite the mounting evidence, and the recommendation by the American Psychiatric Association, Family-Based Treatment has not been widely adopted by clinicians. There are only two studies that have been completed outside of the centres where this treatment was developed. It has been shown to be effective in an open trial involving 20 participants at another U.S. academic centre and the authors suggest that this is a first step in providing evidence of successful dissemination (Loeb et al., 2007). Another recent pilot study examining a combination of conjoint and separated Family-Based Treatment also has shown positive results in Sweden (Paulson-Karlsson, Engstrom, & Nevonen, 2009). However, these studies did not examine aspects of dissemination such as treatment fidelity and acceptability.

The objectives of this study were to examine the dissemination of Family-Based Treatment to a Canadian pediatric eating disorders clinic with particular emphasis on three elements: effectiveness, treatment fidelity, and acceptability. It was postulated that Family-Based Treatment would be adopted within this universal health care system, but that some challenges might emerge including, therapist adherence and patient acceptability. Effectiveness, fidelity, and acceptability, were to be assessed by changes in weight, and psychological symptoms pre- and post- treatment, video taped ratings of therapist adherence to the treatment model, and dropout rates.

METHOD

Participants

Participants were recruited consecutively through a pediatric eating disorders clinic. Inclusion criteria were: a diagnosis of Anorexia Nervosa, age between 12–17 years, and at least one parent willing to participate who speaks English. Participants were excluded if they wanted to continue to receive other forms of psychotherapy or dietary counseling. The diagnosis of Anorexia Nervosa was confirmed by clinical interview and by the Eating Disorder Examination version 15.0 (Fairburn, 2005; Fairburn & Cooper, 1993). This study was approved by the Research Ethics Board at the University of Western Ontario, and was funded by a small internal grant. Consent was obtained from both parents and the adolescent. Treatment was provided free of charge and no financial incentives were given for participation.

Assessment

Ideal body weight was calculated based on a growth history using growth curves provided by the Centers for Disease Control. Weights were measured on an electronic scale in one layer of street clothes, without shoes. Weight at baseline for this study was defined as weight at the first therapy session. Outcome measures for effectiveness included weight, and percent ideal body weight, menstrual status, the Eating Disorder Examination 15th Edition (Fairburn, 2005; Fairburn & Cooper, 1993) and the Eating Disorder Inventory-3 (Garner & Olmsted, 1986). All therapy sessions were videotaped and sessions 1, 2, 3, and a session from each of phase II and phase III were rated for treatment fidelity using a measure designed by the third author (JL). In general, this measure of treatment fidelity looks at how closely the therapist adheres to the true treatment model according to the manual. Some examples of items include “the therapist charges the parents with the task of refeeding”, or “the therapist helps the parents to convince their child to eat one more bite”. An independent observer blinded to outcome (LI) rated the sessions for fidelity. Acceptability was measured using the Outcome Effectiveness Scale (Krautter & Lock, 2004) looking at parent and adolescent satisfaction at the final session.

Treatment and Training

Family-Based Treatment was delivered in three phases as described in the Treatment Manual for Anorexia Nervosa: A Family-Based Approach (Lock et al., 2001). This is an outpatient therapy that generally involves 20 sessions over a one year period. In the first phase, sessions occur weekly and the focus is on parental control of weight restoration. The second phase involves gradually handing control back over to the adolescent, and meetings occur biweekly. In the final phase, meetings occur monthly and the focus is on adolescent developmental issues. Each phase lasts about 3–4 months. Treatment completion for this study was defined as progressing through the first two phases of the treatment or attending 10 or more sessions. This definition of treatment completion is supported by the studies by Lock et al. (Lock et al., 2005; Lock et al., 2006) showing that outcomes from six months (ten sessions) of treatment were no different from outcomes from twelve months (20 sessions) of treatment. As per the treatment manual, participants and parents were not seen by a dietician during the study, but had appointments with a medical doctor as deemed necessary by that physician in order to monitor medical stability. Consultation with a psychiatrist was also available for management of psychotropic medication. No concurrent individual therapy was permitted.

Study therapists (one psychiatrist, one psychologist, and one social worker) were trained using the Treatment Manual for Anorexia Nervosa: A Family-Based Approach, (Lock et al., 2001) along with a workshop conducted by the third author (JL). This workshop consisted of two days of instruction on the treatment model outlining the key interventions within each phase of this treatment. Role play was also used to demonstrate key concepts. Weekly group supervision was provided by the primary author (JC). This supervision consisted of one hour meetings and lasted for the duration of the study. The primary author (JC) completed a one year training fellowship with the third author (JL) prior to this study, and thus was qualified to supervise others in this treatment model.

Data Analysis

The software package SPSS version 15 was used to analyze data. Pre- and post-intervention data was collected and compared using paired t-tests. An intent-to-treat analysis was used, with the last observation carried forward. All tests were two-tailed. Significance was set at 0.05. Correction of the p-value for multiple testing was not done as this was an exploratory study.

RESULTS

Participants

Fourteen adolescents and their families were recruited to participate. All participants met full criteria for Anorexia Nervosa (12 restricting, 2 binge-eating/purging type). Only one participant was diagnosed with a co-morbid condition at baseline which was Body Dysmorphic Disorder. Mean age was 14.0 ± 1.5 years (range 12–17 years) and all participants were Caucasian and female. The majority of the participants (86%) were living with both biological parents, whereas 7% were living with a single biological parent, and 7% were in a reconstituted family. The average duration of illness at assessment was 8 months. At baseline, 9 participants had secondary amenorrhea, 3 participants had primary amenorrhea, and 2 participants were taking the birth control pill. Six participants were hospitalized prior to initiating the study.

Twelve out of fourteen participants (85.7%) completed treatment, and only one participant refused to complete post-intervention measures. No participants were hospitalized during the course of the study. Thirty-six percent were treated with medication concurrently with Family-Based Treatment including 29% taking an atypical antipsychotic, and 7% taking both an atypical antipsychotic and a selective serotonin reuptake inhibitor. Average number of sessions attended was 13 (range 4–26).

Outcome

Weight was significantly increased by an average of 7.8 kg following the intervention (t=−5.3, df=13, p<0.001) (Table 1). Percent ideal body weight was also significantly improved from 80.4% at baseline to 95.7% post-intervention (t=5.5, df=13, p<0.001). The restraint subscale on the Eating Disorder Examination showed significant improvement post-intervention (t=2.9, df=12, p<0.014). In addition, the interoceptive deficits and maturity fears subscales of the Eating Disorder Inventory-3 were significantly improved post-intervention (t=2.8, df=12, p<0.015 for interoceptive deficits; t=3.3, df=12, p<0.006 for maturity fears). In terms of weight restoration, 86% (12/14) were at or above 85% of ideal body weight at their final session, and 57% (8/14) were at or above 95% of ideal body weight. In terms of psychological measures, 54% (7/13) were within 2 standard deviations of a normal adolescent population (Wade, Byrne, & Bryant-Waugh, 2008) on the global Eating Disorder Examination score. All of these 7 participants were also above 85% of ideal body weight, although only 4 of them were above 95% of ideal body weight. Of the 9 participants who had secondary amenorrhea at baseline, 8 had regained menstrual function. The other 5 participants were either continuing to take the birth control pill (2 participants), or continued to have primary amenorrhea (3 participants). The two participants who were engaged in binge/purge behaviour at assessment continued to exhibit these behaviours at final assessment (one participant had worsened from 4 episodes of purging per month at baseline to 15 episodes at final assessment, and the other participant had improved from 14 purges per month at baseline to 5 per month at final assessment). Co-morbid conditions at final assessment included the same participant with Body Dysmorphic Disorder, one participant with Post-traumatic Stress Disorder (from a trauma that occurred prior to treatment), and one with Major Depressive Disorder.

Table 1.

Pre- and post-intervention effectiveness outcome data

	Pre-intervention Mean±SD	Post-Intervention Mean±SD	t-statistic	df	p-value
Weight (kg)	43.2±6.6	51.0±10.1	5.29	13	0.001*
Percent Ideal Body Weight	80.4±3.6	95.7±10.6	5.53	13	0.001*
Eating Disorder Examination
Restraint	3.20±1.89	1.52±1.84	2.86	12	0.014*
Eating Concern	1.55±1.51	1.09±1.25	1.37	12	0.195
Weight Concern	2.05±1.86	1.98±1.95	0.16	12	0.871
Shape Concern	2.81±2.12	2.04±1.96	1.16	12	0.269
Eating Disorder Inventory – 3
Drive for Thinness	17.08±7.90	10.92±11.33	1.90	12	0.082
Bulimia	3.54±4.81	3.77±7.45	0.18	12	0.860
Body Dissatisfaction	19.54±9.94	14.69±17.29	1.05	12	0.312
Low Self Esteem	6.46±4.52	5.15±7.40	0.91	12	0.380
Personal Alienation	6.92±5.50	6.54±7.64	0.21	12	0.838
Interpersonal Insecurity	7.62±7.94	7.15±7.90	0.23	12	0.822
Interpersonal Alienation	4.85±4.47	4.92±5.27	0.06	12	0.951
Interoceptive Deficits	14.38±10.76	7.38±6.45	2.84	12	0.015*
Emotional Dysregulation	5.46±4.67	5.54±6.45	0.09	12	0.931
Perfectionism	9.23±4.68	10.23±5.40	0.93	12	0.373
Asceticism	6.92±4.61	7.62±7.49	0.35	12	0.735
Maturity Fears	12.77±6.39	8.46±3.86	3.34	12	0.006*

^*p<0.05

Treatment fidelity was rated as 5 or above on a 7 point scale (where 5 was defined as considerable) 72% of the time in phase I of the treatment, 47% of the time in phase II, and 54% of the time in phase III (Table 2). This cut point was chosen as it seems to represent adequate fidelity. These fidelity results involve all participants (not just those who completed treatment). Treatment fidelity during the first session appeared to show some trends toward higher means in those who completed treatment, although no statistically significant results were found (Table 3). Participants and parents generally found the treatment to be acceptable (Table 4). As expected, those who completed treatment generally had higher ratings than those who dropped out of treatment, however, both adolescents and parents who dropped out of treatment found the parental control of refeeding aspect to be effective, or highly effective.

Table 2.

Treatment fidelity scores (scale of 1 to 7, with 1 being “not at all” and 7 being “very much”)

	N	Mean	SD
Phase I
Session 1
Greet family in sincere but grave manner	14	4.64	1.55
Take a history that engages each family member	14	5.29	0.73
Separate the patient from the illness	14	5.50	1.34
Orchestrate an intense scene around seriousness	14	4.57	1.09
Charge parents with task of refeeding	14	5.64	0.84
Session 2
Take history and observe family patterns around food	12	4.83	1.53
Help parents convince patient to eat one more bite	12	5.83	0.94
Align patient with siblings	12	4.75	1.54
Session 3
Focus therapeutic discussion on food and eating	13	6.00	1.08
Help parental dyad’s efforts at re-feeding	13	5.38	1.19
Discuss and evaluate siblings efforts	13	3.77	2.01
Continue to modify parental and sibling criticism	13	4.92	1.44
Continue to distinguish patient’s interests from those of AN	13	4.92	1.26
Phase II
Continue to support parents in management of ED symptoms	11	5.09	1.22
Assist parents in negotiating return of control to adolescent	11	4.82	1.60
Encourage family to examine relationships between adolescent issues and the development of AN	11	2.91	1.45
Continue to modify parental and sibling criticisms	11	3.55	1.97
Continue to assist siblings in supporting their ill sibling	11	1.00	0.00
Continue to highlight difference between adolescent’s own ideas and those of AN	11	4.18	1.72
Phase III
Review adolescent issues with family to model problem solving	8	4.88	0.99
Involve the family in “review” of issues	8	4.50	0.93
Check with parents how much they are doing as a couple	8	3.75	2.19
Explore adolescent themes	8	4.63	1.60
Plan for future issues	8	3.63	1.65
Manage termination	8	4.25	1.67

Table 3.

Treatment fidelity scores from the first session of those completing treatment, and those who dropped out

	Completers			Dropouts
Phase I, Session 1	N	Mean	SD	N	Mean	SD
Greet family in sincere but grave manner	12	4.83	1.47	2	3.50	2.12
Take a history that engages each family member	12	5.42	0.67	2	4.50	0.71
Separate the patient from the illness	12	5.50	1.31	2	5.50	2.12
Orchestrate an intense scene around seriousness	12	4.58	1.16	2	4.50	0.71
Charge parents with task of refeeding	12	5.67	0.89	2	5.50	0.71

Table 4.

Parent and adolescent acceptability scores (scale of 1 to 5, with 1= highly ineffective and 5 = highly effective)

	Adolescent Total Sample (n=13)		Adolescent Completers (n=11)		Adolescent Dropout (n=2)		Parent Total Sample (n=14)		Parent Completers (n=12)		Parent Dropout (n=2)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Therapeutic rapport	3.69	1.03	3.82	1.08	3.00	0.00	4.29	0.83	4.42	0.79	3.50	0.71
Psychoeducation about AN	3.92	0.86	4.09	0.70	3.00	1.41	4.50	0.65	4.58	0.67	4.00	0.00
Inclusion in treatment	3.92	1.32	4.18	1.25	2.50	0.71	4.64	0.74	4.75	0.62	4.00	1.41
Separation of the illness from the patient	2.75	1.22	3.00	1.15	1.50	0.71	3.86	0.66	4.00	0.60	3.00	0.00
Parental refeeding	3.23	1.48	3.00	1.48	4.50	0.71	3.93	0.92	3.92	0.90	4.00	1.41
Adolescent control of refeeding	3.00	1.68	3.36	1.57	1.00	0.00	3.00	1.41	3.17	1.47	2.00	0.00
Fight against anorexic thinking	2.62	1.76	2.91	1.76	1.00	0.00	3.21	1.25	3.33	1.30	2.50	0.71
Encouraging sibling support	2.62	1.12	2.73	1.19	2.00	0.00	3.43	1.09	3.67	0.98	2.00	0.00
Addressing adolescent issues	2.54	1.13	2.64	1.12	2.00	1.41	3.07	1.38	3.25	1.42	2.00	0.00
Overall success	3.38	1.50	3.55	1.57	2.50	0.71	3.57	1.45	3.83	1.40	2.00	0.00

DISCUSSION

This preliminary investigation of the dissemination of Family-Based Treatment for adolescents with Anorexia Nervosa to a Canadian population indicates that this treatment is effective in this population not only for weight restoration, but also in improving some of the psychological symptoms such as dietary restraint, interoceptive deficits and maturity fears. In addition, this treatment was adopted with considerable fidelity and was acceptable to adolescents and parents. The dropout rate in this study was similar to other studies. In those families who did not complete treatment, parents were likely to advocate for their child being admitted to hospital, or attending the day treatment program which were both available at no cost to the families (covered by provincial health insurance). It is of interest that the two patients with the binge-eating/purging type of Anorexia Nervosa were not abstinent from these behaviours at the conclusion of the study. Although parents were instructed to try to normalize these behaviours as well as to refeed their child, this is often very difficult for parents to accomplish simultaneously.

This is the first study to examine fidelity of this treatment in a dissemination context. Interestingly, fidelity appeared to decrease in phases II and III of the treatment, with the highest ratings of fidelity in the first phase of treatment. Perhaps this occurred because interventions within the first phase are more highly structured. There were particular interventions which were performed with high fidelity such as charging the parents with the responsibility of the refeeding, and others where a much lower fidelity was observed, such as the interventions directed at siblings. Further study should examine whether these interventions are critical to the success of the treatment as participants appeared to do well with this treatment even though fidelity was low in terms of engagement of siblings.

In terms of acceptability, parent scores were generally higher than adolescent scores. This is similar to results found by Krautter & Lock (2004). Parents were most pleased with the aspect of being involved in the treatment. Adolescents appreciated the psychoeducational aspect of the treatment. In their written comments, adolescents and parents voiced concerns about not having exposure to a dietician, and not receiving individual therapy. Despite these concerns, many adolescents did very well in terms of effectiveness outcomes.

Limitations of this study include its small sample size and lack of control group. In dissemination studies such as this involving smaller scale clinics, it would be hard to obtain a large sample size unless a multi-centred trial were pursued. This study was prospective in nature, but did not involve a control group. Thus, it is difficult to determine whether improvements would have been seen without treatment at all, solely with the passage of time. In addition, participants were treated concurrently with medication. This is in keeping with more naturalistic effectiveness studies, but also raises questions of how effective the treatment would have been if it had been used alone, without any medication. However, no studies to date have systematically shown that Anorexia Nervosa is responsive to medication. In addition, medication was not a mediator or moderator of treatment outcome in a recent study on Family-Based Treatment for Anorexia Nervosa (Lock et al., 2005). Generally co-morbidity was low within this study which makes the results difficult to generalize, however, this was a sample representative of patients attending the clinic and co-morbidity was not an exclusion criteria.

In addition, this study was voluntary, and parents who agreed to participate were likely more motivated to help their child. Of course, Family-Based Treatment does not work for all families. More research needs to be devoted to determining which family factors are predictive of good outcome. Some studies have indicated that patients who have severe obsessive-compulsive features surrounding their eating disorder, or who have non-intact families need a longer course of Family-Based Treatment (Lock et al., 2005). Furthermore, other studies indicate that families who have high expressed emotion should be treated in a “separated” fashion, so that parents are seen separately from the adolescent while still following the principles of the Family-Based Treatment model (Eisler et al., 2000). Clinically, parents who are actively suffering from an eating disorder, or those with severe mental illness may not be able to be responsible for refeeding their child. However, these conditions would not preclude an initial attempt at Family-Based Treatment. This treatment is contraindicated in families where there is active abuse and/or violence.

Despite these limitations, this study adds to the literature on Family-Based Treatment. This is the first study to examine the training of clinicians from a variety of disciplines using treatment fidelity as an outcome. This research suggests that more study be directed at how to train clinicians appropriately. Further study is also needed to examine the key interventions in all phases of the intervention for treatment success, and the types of family factors that are related to outcome. Future studies might also examine the barriers and facilitating factors to the adoption of this treatment modality in eating disorder programs in the community. The philosophy of parents taking charge of weight restoration is in contradiction to more traditional models of family therapy which may affect clinicians’ willingness to adopt it within their practices. There may be other significant issues such as access to ongoing training that impede transfer of this treatment to clinical practice. Despite these potential barriers, this study demonstrates that transfer of this treatment to a clinical setting is quite possible with adequate training and supervision. Furthermore, this study of the dissemination of Family-Based Treatment to a Canadian setting shows that the treatment was effective, adopted with considerable treatment fidelity, and was acceptable to adolescents and parents; comparable outcomes to those found in randomized controlled trials conducted at large research centers.