Introduction

Anorexia Nervosa (AN) is a psychiatric disorder, with onset typically in adolescence (Swanson et al. 2011). DSM-5 describes it with three specific criteria: intense fear of weight gain, distortion of body image and weight, and significantly low weight compared to average weight for age (APA 2013). AN is generally comorbid with anxiety disorder, mood disorders, obsessiveness, substance abuse, and it is associated with high mortality and suicide rates (Preti et al. 2009; Swanson et al. 2011; Smink et al. 2012; Krug et al. 2013; Bühren et al. 2014). As reported by Thornton et al. (2016), mortality in AN is among the highest of all psychiatric disorders, and there is elevated risk relative to that expected in the normal population (Laghi et al. 2009).

The prevalence of AN in the US is of .3%, with an incidence rate of at least 8 out of 100,000 persons per year (Hoek and Van Hoeken 2003). In Italy, Eating Disorders (ED) impact approximately two million youth: about 10 adolescents out of 100 suffer from an ED, of whom 1–2 are affected by more severe forms, such as Anorexia Nervosa or Bulimia Nervosa (BN), whereas the others exhibit transitory and incomplete clinical manifestations (Dalla Ragione 2005). These disorders affect both the psychic and the somatic level, and their management requires the intervention of different disciplines, ranging from the biological domain to the psychological and psychiatric domains. Individuals with AN have the tendency to deny their condition and do not seek professional help; it has been noted that just a small number of adolescents with the disorder receive health services specifically for eating problems (D’Onofrio et al. 2015; Laghi et al. 2015a, b, 2016; Pace et al. 2015, 2016; Tambelli et al. 2012). The current guidelines for the treatment of ED in children and adolescents recommend the integrated multidisciplinary approach as the elective intervention for this disorder (APA 2006; NICE 2004).

The plurality of interventions should be accompanied by their integration, which means sharing the objectives to be achieved, recognizing the role of the different disciplines involved, and creating a synergy. To date, there are only few published studies evaluating the effectiveness of this kind of interventions, and they deal with different therapeutic approaches. A study on the effectiveness of group psychotherapy combined with psychodynamic individual therapy showed significant changes in the disorder itself, and in the emotional containment of parents, with the decrease of family discomfort and excessive emotiveness, the improvement of family’s perception of the patient’s symptoms, and more positive family interactions both in the relationships parents-children and between siblings (Pasold et al. 2010; Prestano et al. 2008; Uehara et al. 2001). Crisp et al. (1991) found a greater weight restoration and more stable parameters in patients treated with family therapy and individual therapy.

Another study by Godart et al. (2012) compared 60 adolescents with AN divided into two groups: a control group treated with psychiatric interviews with the patient alone or with patient and parents, and a study group where couple psychotherapy was added to the treatment plan. At an 18-month follow-up, a general improvement in the study group was found, with specific regard to anorexia symptoms, social adaptation and the reduction of hospital readmission rates.

The above studies have in common the involvement of families, in particular of parents, in the treatment of adolescents with AN, for which parents are thus considered a resource with a positive outcome (Abbate et al. 2011). Family psychotherapy was found to be more effective than individual psychotherapy in the treatment of adolescents with AN, especially when the disease history was of less than 3 years (Eisler et al. 1997; Gowers et al. 2007; Lock et al. 2006a; Paulson-Karlsson 2009; Russell et al. 1987). Positive outcomes were also found with other forms of family involvement, such as short cycles of psychotherapy—of the duration of 6 months—and sessions of family therapy, involving patient and parents, both together and separately (Doyle et al 2010; Geist et al. 2000; Lock et al. 2006b; Robin et al. 1999). Lock and Le Grange (2015) proposed a model of intervention for families of adolescents with Eating Disorders, called Family Based Treatment (FBT), also referred to as Maudsley model that was based on the model originally developed by the Maudsley group (Dare 1985). FBT has been subject to several controlled studies and is now considered as one of the most recommended model for family treatment.

Nevertheless, the above studies do not consider the role of an integrated intervention, which includes pharmacological and nutritional monitoring, in the treatment of AN. In Italian context, Onnis et al. (2012) found significant improvements in family interaction modalities correlated to a greater positive evolution of clinical parameters in the group with integrated treatment combining medical-nutritional interventions with family psychotherapy. These improvements were not found or were not as evident in the control group with the usual medical-nutritional intervention supported by psychiatric counselling, suggesting that an integrated approach in the treatment of AN is more effective than a selective approach.

Based on the above, this study aims to verify the effectiveness of our proposed Day-Hospital Multifocal Integrated Treatment (MIT) through the clinical assessments with patients and parents (illness history, duration of illness, premorbid weight, amenorrhea, current and past medical conditions, current medications, eating disorder psychopathology, and behavior and socioemotional functioning) from the beginning and at the end of treatment. We did not use a control group because in clinical setting adolescents with AN in the control group must wait to begin the treatment. Therefore, in the first part of the larger longitudinal study, we reported pre-post-quantitative results with 3, 6, and 12-month follow-up periods. The purpose of this paper is to describe the results of a short-term follow-up effectiveness study that was performed after a completed therapy program. Our hypothesis is that the Day-Hospital Multifocal Integrated Treatment (MIT) addressing at the same time the different levels of the individual and his/her family system (through the two treatments Group Therapy and Family Group Therapy) may induce an improvement in the eating psychopathology both in clinical and psychological terms.

Method

Participants

The sample consisted of 60 Eating Disorder patients and their parents admitted in a day-hospital setting to a children’s hospital specialized in the treatment of Eating Disorders with the Multifocal Integrated Treatment (MIT). All participants received information on the scope of the treatment and agreed to participate. Selection was based on the following inclusion criteria: (a) diagnosis of Anorexia Nervosa or Eating Disorder Not Otherwise Specified (EDNOS) restrictive type according to DSM IV-TR criteria; (b) BMI > 14 or less if patients are compliant to treatment stabilization of medical parameters; (c) age between 8 and 17 years. Clinical assessment collected information including duration of illness, eating disorder symptoms, and past medical history. For evaluating eating disorder psychopathology, the following instruments were used: (a) the Eating Disorder Inventory 3 (EDI–3; Garner 2004); (b) the Eating Attitude Test (EAT–26; Garner et al. 1979). The Italian version of the Schedule for Affective Disorders and Schizophrenia for School-Age Children/Present and Lifetime Version (K–SADS–PL; Kaufman et al. 1997) was used in order to determine Axis I psychiatric diagnoses. Patients diagnosed with intellectual disabilities, pervasive developmental disorders, schizophrenia spectrum disorders or associated neurological conditions were not included.

Procedure

The protocol of MIT entails three evaluation days dedicated to: a psychiatric interview, an interview for family diagnosis, and nutritional monitoring. All participants were tested individually in a quiet room. They gave their written informed consent before completing the questionnaires during a 2–3 h session. Participants’ weight and height were measured to calculate BMI. All participants were native Italian speakers.

Patients and parents were assessed at baseline and after the end of the intervention by the same study clinicians who were different from licensed clinical psychologists involved in the treatment program. BMI and pharmacological therapy were monitored at the beginning and the end of the study.

Figure 1 illustrates in detail the treatment protocol, from patients’ selection to the proposed intervention model.

Fig. 1
figure 1

The day hospital multifocal integrated treatment (MIT) for adolescents with anorexia nervosa

Full size image

This study was reviewed and approved by the Ethics Commission of the Department of Neuroscience—I.R.C.C.S., Children Hospital Bambino Gesù.

The Treatment Program

The Multifocal Integrated Treatment (MIT) follows the directives of major international guidelines, and proposes a treatment plan in three different settings, depending on patients medical and psychological conditions: outpatient, day-hospital, or impatient.

This study focuses on the treatment performed a day-hospital setting, which can be started either directly after an initial phase of diagnostic evaluation, or at the end of an inpatient stay, when the severity of psychophysical conditions has decreased.

The Maudsley Family Based Treatment (FBT) model requires a sequential approach to treatment, divided into three chronologically distinct, but interdependent phases. In the first phase both parents actively participate in the management of their child’s disordered eating symptoms, monitoring proper meal intake. In the subsequent phase, parents’ monitoring gradually decreases, and the capability of the child to control eating and to face personal difficulties increases. In the last phase, the therapist concentrates on the entire family unit and addresses the specific dysfunctional dynamics of the system and subsystems (Lock and Le Grange 2005). FBT has a strong evidence base (Keel and Haedt 2008), and his improvement were maintained at long-term follow-up (Grange et al. 2014). However, it is not definitive that it is the most effective treatment for mood improvement and healthy weight gain for all adolescent with AN: the drop-out rate from clinical trials is 10 to 20% and around 50% of adolescent and families do not engage with or respond well to Maudsley (Eisler et al. 2000; Lock et al. 2006a). This raises questions regarding whether alternative treatment approaches may suit some families better (Plath et al. 2016). Differently from FBT, MIT is based on simultaneous treatment model, in which group psychotherapy for the children, and their parents with a frequency of once a week, family psychotherapy sessions with a monthly frequency, and meetings for the nutritional and psychiatric monitoring every two weeks are activated in the same time. Moreover, each week all team members (psychiatrist, group psychotherapists, family psychotherapist, nurse, and nutritionist) take an interdisciplinary meeting where each patient’s clinical course (medical aspects, pharmacological monitoring, psychiatric problems, evolution of psychotherapeutic pathways of parents and children, adjustment of the treatment plan) is discussed.

At each Day-Hospital access, nurses record weight and parameters of each patient. Patients are weighed with their underwear on, and with their back to the scale, so that they cannot see their weight in order to reduce anxiety and control associated with it.

The meeting with the nutritionist is dedicated to food and to the analysis of reality related to body needs. The nutritionist gives to both patients and their parents behavioral directions, trying to restore a balanced relationship with food and a healthy and correct diet, not dominated by obsessive control. Nutritionist agrees with the patient a minimum weight to be reached, which does not cause too much anxiety to patient and that will be communicated to her only to its achievement. It is made clear to the patient that this weight may not match the final weight. Once this interim goal was achieved, the weight will become area of awareness and discussion within each nutritional encounter. This clinical approach has the goal to create a relationship of trust that would allow to lower the typical control behavior. Any difficulty to trust in nutritionist is discussed as a therapeutic topic within groups, where resolution strategies can be compared with the other girls who have passed the problem and are concluding therapy.

Groups are made of a maximum of eight patients, divided for age range. They are “open” groups, meaning that for every patient discharged, a new patient is enrolled. This approach allows to work simultaneously on the different phases of the disease. In addition to the disordered eating symptoms, this treatment addresses other issues related to transition from childhood to adolescence, such as identity, family relationships, relationships with peers, autonomy, sexuality.

At the same time, parents’ group become a space for phenomenological understanding of the disorder and its multifactorial origin as well as a containment for fear, guilt and shame often associated with AN, increasing their compliance with the treatment plan (Cottee-Lane et al. 2004). The work with parents focuses on restoring balanced generational boundaries, developing parents’ reflective capability, strengthening the parenting couple, and helping parents understand the underlying causes of eating disorders. In this regard, family psychotherapy represents an important step, where more time is dedicated to jointly discussing issues emerging during the treatment plan, in the presence of all family members, and reflecting on the developmental block in the system and on the ways to facilitate adolescents’ independence.

Finally, the psychiatric interview represents the link among the different interventions and allows the overall evaluation of the treatment course. Generally, pharmacological therapy is proposed as a last resort to avoid inpatient admission, which can however be considered in the most critical situations and it is used at low doses. We used SSRI (Fluoxetine max dosage 20 mg, Sertraline max dosage 50 mg) and atypical antipsychotics (Risperidone max dosage 1 mg, Aripripazole max dosage 7.5 mg).

Measurement

Eating disorder psychopathology and the possible presence of dysmorphophobia were investigated using Eating Disorder Inventory-3 (EDI-3; Garner 2004), Eating Attitudes Test–26 (EAT-26), and Body Uneasiness Test (BUT).

Dysfunctional eating. Eating disorders inventory-3

The Eating Disorder Inventory-3 (EDI-3; Garner 2004) is a self-report instrument measuring psychological traits or constructs shown to be clinically relevant in individuals with ED. This measure consists of 91 items organized into 12 primary scales, three eating disorder-specific scales (Drive for Thinness—DT; Bulimia—B; Body Dissatisfaction—BD) and nine general psychological scales (Low Self-Esteem—LSE; Personal Alienation—PA; Interpersonal Insecurity—II; Interpersonal Alienation—IA; Interoceptive Deficits—ID; Emotional Dysregulation—ED; Perfectionism—P; Asceticism—A; Maturity Fears—MF) that are highly relevant to, but not specific to, eating disorders. The reliability coefficients of the scales range from .83 and .90, and test–retest reliability coefficients for the various composite scales are between .84 and .87. The Italian version of EDI-3 (Giannini et al. 2008) has demonstrated very good day test–retest reliability, cross-informant agreement, and a good discriminating validity.

Eating attitudes test–26 (EAT-26)

The Italian version of the Eating Attitudes Test–26 (Dotti and Lazzari 1998) contains 26 items that assess characteristics and symptoms of eating disorders and has been validated as both a dimensional and categorical measure. Individuals who score 20 or higher on this measure are considered to be at risk of or likely to have clinical levels of eating disorder symptoms. Participants indicate the extent to which they endorse each item along a 6-point scale that ranges from 1 (never) to 6 (always), with higher scores reflective of more severe symptomatology.

Body uneasiness test (BUT)

Body Uneasiness Test (Cuzzolaro et al. 1999) was used for the clinical assessment of body uneasiness. The BUT-A consists of four subscales and a global severity index (GSI) that have been demonstrated to have good internal consistency and reliability: Weight Phobia (WP—Fear of being or becoming fat), Body Image Concerns (BIC—Worries related to physical appearance), Avoidance (A—Body image-related avoidance behaviour), Compulsive Self-Monitoring (CSM—Compulsive checking of physical appearance), and Depersonalization (D—Detachment and estrangement feelings toward the body). The Italian version of the instrument shows good reliability coefficients and a factorial structure congruent with the operative definition of the construct (Cuzzolaro et al. 1999).

Emotional, behavioral characteristics and psychopathological dimensions were investigated using: Multidimensional Anxiety Scale for Children (MASC), Children Depression Inventory (CDI), and Youth Self-Report (YSR).

Multidimensional anxiety scale for children (MASC)

We used the MASC (March et al. 1998), that is a 39-item self-report measure of anxiety. Items are rated on a four-point Likert-type scale ranging from zero (never true about me) to three (often true about me), such that higher scores on the subscales and the total scale indicate higher levels of anxiety. The MASC consists of four subscales that have been demonstrated to have good internal consistency and reliability: Physical Symptoms (Tense/Restless, Somatic/Autonomic; Cronbach’s α = .86), Social Anxiety (Humiliation/Rejection, Performance Fears; Cronbach’s α = .85), Harm Avoidance (Perfectionism, Anxious Coping; Cronbach’s α = .72), and Separation/Panic (Cronbach’s α = .70), and it includes a Total Scale score. Raw scores were converted into standard T-scores and a T-score more than 75 indicated the presence of anxiety symptoms.

Children depression inventory (CDI)

The Italian version of the CDI was used to assess depressive symptomatology. It is composed consists of 27 items scored on a three-point scale indicating increasing severity of symptoms (Cronbach’s α = .80). According to Italian validation criteria (Camuffo et al. 1988), 19-point cut-off indicates the ideal threshold for a child at risk of depression (Kovacs 1983). The Italian version of the scale has good psychometric properties and a convergent validity with different measures of psychopatolical dimensions (Camuffo et al. 1988).

Youth self-report (YSR)

To assess the adolescents’ view of their behavior and socioemotional functioning, the Italian version of the Youth Self-Report (YSR; Achenbach and Rescorla 2001) was used. This questionnaire has to be completed by the 11- to 18-year-old adolescent and contains 112 problem items covering behavioral, emotional, and social problems that occurred during the past 6 months. The YSR can be scored on syndrome scales: Anxious/Depressed, Withdrawn/Depressed, Somatic Complaints, Social Problems, Thought Problems, Attention Problems, Aggressive Behavior, and Rule-Breaking Behavior. The Internalizing scale can be derived from the first three syndrome scales, and the Externalizing scale from the last two. This measure, in its validated Italian version has demonstrated very good day test–retest reliability, cross-informant agreement, and success in discriminating between referred and no referred adolescents (Frigerio and Monterosso 2002).

Parents completed the Child Behavior Checklist (CBCL; Achenbach and Rescorla 2001), a questionnaire allowing to investigate a wide range of characteristics of the developmental age, with regards to both competencies and involvement in the activity, and the presence of emotional and behavioural problem. Reliability and validity of the Italian version have been shown to be satisfactory (Frigerio and Monterosso 2002).

Morgan-Russel outcome assessment scale

Patients’ clinical course has been checked at 3, 6, and 12-month follow-up with the Morgan-Russel Outcome Assessment Scale (MROAS; Morgan and Hayward 1988), in the version modified by Jeammet et al. (1991). The MROAS is a scale for the biopsychosocial assessment of the outcome, compiled by the clinician at the end of treatment, based on information received by the patient or observed during treatment. The Morgan-Russell Average Outcome Score (MRAOS) is derived from a guided interview assessing core clinical features of AN including food intake, menstrual state, mental state, psychosexual adjustment and vocational adjustment. Scores are rated on a continuous scale by only one experienced clinicians, although the team was composed by other two clinician who participated. The clinician rated each item with a score from 1 (satisfactory) to 6 (very unsatisfactory). Immediately after the end of the assessment for each patient, the clinicians shared the evaluation, and a final decision was taken as to which code should apply. For the final score assignment, they discussed each score upon which they disagreed, until full agreement was reached.

Following this procedure, patients are divided into three groups, depending on the score: good when at least 8 items have been rated 1 or 2; intermediate, if 4 to 7 items have been rated 1 or 2; poor if 3 items or less have a score of 1 or 2.

Data Analyses

The Statistical Package for the Social Sciences (SPSS 22) was used to conduct bivariate and univariate analyses relating to independent variables. Preliminary analyses were carried out for univariate outliers. We conducted Paired-sample t-tests to investigate differences between pre and post treatment regarding the key variables used in the present study, and Chi-square for categorical variables. Cohen’s d values were calculated as a measure of effect size, and results were interpreted using Cohen’s (1988) guidelines for determining small (<.30), medium (.30−.80), and large (>.80) effects.

Results

Demographic and Clinical Characteristics

93.3% of patients were females. Patients’ mean age upon evaluation was 14.8 years (SD = 1.8; range 11–17.7), but the onset of the disorder was approximately 8 months before (Mean = 14; SD = 1.9; range: 10–17.5). In the majority patients were first-born of two children (53.3%) coming from Rome and its province (71.7%). Mothers’ median age was 46 years (SD = 7.2), and fathers’ median age was 49 years (SD = 6.5). The majority of parents were together, with only 20% of separated parents. Their level of education and social status were average. The diagnosis of ED follow the DMS-IV criteria. However, all subjects would be included in the diagnosis of Anorexia Nervosa if considering the current updated criteria of the DSM-5 (APA 2013). Only 38.3% of patients had a previous inpatient admissions due to the deterioration of clinical conditions, or were hospitalized in the course of the treatment. Treatment median duration was of 8 months (SD = 2.27). All enrolled patients and their parents completed the treatment program.

Diagnostic and Medical Parameters

Results revealed that patients reported greater improvement immediately after treatment with significant increase of BMI and a good clinical course, as reported in Table 1.

Table 1 Changes in diagnostic and medical parameters
Full size table

We verified the possible incidence of pharmacological treatment on the variables pre-post treatment. Paired-sample t-tests showed significant differences only for BMI, perfectionism scale, and externalizing disorders as assessed with the Youth-Self-Report, as reported in Table 2.

Table 2 Effect of drug therapy for BMI, perfectionism, and externalizing disorders after treatment
Full size table

Eating Disorder and Psychopathological Symptoms: Differences between Pre and Post Treatment Evaluations

Paired-sample t-tests did not show significant differences for Dysmorphophobia Total score (BUT). Significant differences were obtained for EAT-26 Total score and for some psychological dimensions, as reported in Table 3. Additionally, paired-sample t-tests revealed that patients reported greater improvement in overall dimensions on MASC and CDI Total score, as reported in Table 3.

Table 3 Pre-post treatment: significant differences on EAT-26, BUT, EDI-3, MASC and CDI dimensions
Full size table

Behavior and Socioemotional Functioning: Change in Patients and Parent Report Evaluations Pre and Post Treatment

Paired-sample t-tests revealed that patients and their parents reported greater improvement in overall dimensions on Child Behavior Checklist (CBCL) and on Youth Self-Report (YSR), as reported in Table 4.

Table 4 Pre-post treatment scores on CBCL dimensions completed by patients and their parents
Full size table

3, 6 and 12-month follow up

Scores obtained from the Morgan Russel scale were grouped in three subscales, as provided by the author: poor, intermediate and good. The majority of patients were included in the groups good and intermediate at 3,6, and 12-month follow-up (Table 5). However, patients’ drop-out increases with the increase of time of follow-up.

Table 5 Outcome at 3, 6 and 12 months from discharge
Full size table

Discussion

This study represents an attempt to explore and operationalize, using psychometric instruments, the Day Hospital Multifocal Integrated Treatment (MIT; Vicari et al. 2014; Zanna and Castiglioni 2014) with two diagnostic groups: Anorexia Nervosa restrictive type and Eating Disorder Not Otherwise Specified. In accordance with the treatment’s objectives, patients’ weight restoration was significant, achieving a BMI of 19.4 (healthy weight). At start point, most female patients presented with amenorrhea (83.3%), and at the end of treatment, the percentage of patients having a regular menstrual cycle increased to 61.7%. As regards the psychiatric diagnosis: despite the remission of anorexia symptoms in 63.3% of patients, there was an increase of EDNOS, going from 30 to 33%. The transition from a well-defined diagnosis as Anorexia Nervosa to a more vague Eating Disorder may be mainly explained by the fact that the restoration of physical parameters (i.e. weight) is faster than the remission of some psychopathological characteristics and of dysfunctional eating behaviors. In this context, pharmacological therapy may contribute to a faster resolution of clinical parameters, in particular of BMI, without impacting in a significant way the underlying psychopathological variables, such as dysmorphophobia and tendency to perfectionism. So, it is possible that pharmacological treatment contribute in unlocking an initial emergency situation, restoring the weight fast, but does not completely solve alone the underlying psychological symptoms.

Comparing scores from patients self-reports before the treatment with the ones after the treatment, we found a substantial decrease of disordered eating symptoms, as well as of internalizing and externalizing problems. In particular, we observed a decrease of the risk of depression and of the different anxiety symptoms (physical symptoms, avoidance, social anxiety, and separation anxiety), and an increase of self-esteem and interpersonal confidence, as well as emotional regulation and interoceptive skills. So, it is possible that patients’ group therapy participation contributes to reduce social anxiety, because patients acquire, through an empathic relationship with the psychotherapist and with peers, a greater confidence in relationship management in social situations (Knijnik et al. 2004; Wersebe et al. 2013; Barkowski et al. 2016).

On the other hand, results suggest that the treatment has no effect on the EDI-3 scores, relevant to personal and interpersonal alienation, namely the difficulty to establish relationships, and the resulting belief to be better alone. In other words, the treatment can be useful in decreasing the anxiety caused by social situations in which patients seem to feel more at ease; however, it seems to be ineffective on their tendency to prefer solitary contexts rather than group contexts and on their difficulty in building and keeping relations. Thus, they are better able to manage social contexts, but maintain a more introversive attitude compared to peers.

Moreover, total scores on body dissatisfaction (dysmorphophobia), tendency to perfectionism and fear of maturity remain unchanged, suggesting that the treatment is not effective on these areas. These characteristics actually represent the deep core of anorexia and, as such, need more time to be eradicated (Bearman et al. 2006; Friederich et al. 2010; Hrabosky et al. 2007; Mohr et al. 2010; Schneider et al. 2009). Future interventions could better focus also on the treatment of these specific aspects, for example through specific experiential workshops.

Parents reported an overall improvement of their children’s psychopathology, generally in those psychiatric-comorbidities may occur prior to, but also coincidental with, the Anorexia Nervosa, such as affective problems, anxiety and depression (Dalle Grave 2011; Dalla Ragione 2012). Moreover, there was evidence of a difference between parents’ and patients’ perception of the severity of symptoms: patients tended not to recognize dysfunctional areas in their social relationships or internalizing behavior, whereas these dysfunctional areas were evident in parents’ description. This seems to confirm the typical attitude of anorexic patients, who tend to deny their problems and compromised health, and let their body express their discomfort (Bruch 1978, 1982; Selvini Palazzoli et al. 1998; Zappa 2009), as demonstrated by the scores obtained from self-report questionnaires that are rarely clinically significant, apart from the areas of eating behavior.

At 3, 6, and 12-month follow-up, despite a moderate drop-out percent, the improvement of clinical and psychological parameters was confirmed, with a high percentage of patients showing adequate functioning. Drop-out data, however, may be interpreted in the same direction, as patients needing help are generally the ones who meticulously attend follow-up appointments (Di Pietro et al. 2002).

The data collected in this study confirm the results of similar studies in the literature (Onnis et al. 2012; APA 2006; NICE 2004), where integrated treatment is considered to be the only way to obtain a long term improvement of Eating Disorders. The adoption of a bio-psycho-social model (Engel 1977) seems to be a precondition in the design of an effective treatment plan which, using pharmacological therapy, a correct eating education, group and individual psychotherapy, can decrease the factors determining the maintenance and cronicization of the disorder.

This study shows that the therapeutic objectives of the Multifocal Integrated Treatment (MIT) have been achieved, confirming the efficacy of this program, as an example of integrated treatment as opposed to single therapies. FBT has been found to work for around two third of adolescent with AN who remain for the duration of clinical trials (Lock and Le Grange 2005) and around 50% of adolescent and families do not engage with or respond well to Maudsley. So it is necessary to think about alternative treatment approaches that may suit some family better (Plath et al. 2016). MIT simultaneous model gives the opportunity to meet the specific needs of patient and family, for example by enhancing individual interventions, and using the group intervention, gives an adequate therapeutic response to the high demand for care in institutional contexts that often have limited resources.

The fact that achieving a global improvement takes more time than what we can do in our Hospital is not surprising. A global improvement should not include only the decrease of dysfunctional eating behaviors, the reassessment of thinness ideals, and the decrease of the psychopathology underlying the disorder, but also the increase of self-esteem and maturity and the decrease of the tendency to perfectionism and interpersonal relationships problems. Therefore, after a first phase aimed at improving the eating symptoms, restoring control on weight and body and understanding the specific psychopathological problems that the patient has not managed to overcome, the treatment should be continued in a second phase in which the above problems shall become the psychotherapeutic targets to be addressed specifically.

The major limitation of the study is the nature of our sample, which is a non-randomized clinical sample, and the absence of a treatment as usual condition to compare results. In fact, the modest size of the sample and the impossibility to offer an alternative treatment have not allowed us to design a research study with experimental group. Given the fact that parents and adolescents are active participants in the intervention and might be susceptible to bias, additional third party assessments like those obtained through the Morgan Russel scale in the current study or behavioral observations, would be beneficial toward establishing the validity of the findings.

These limitations, however, are characteristic of a an open label clinical pilot study, which represents an attempt to systematize data and prepare a manual on the various phases of the proposed treatment. Future studies will have to be aimed to a greater control of all the existing variables, in particular concerning the change of familiar and parental relationships, and will need to have a homogeneous and adequately comparable experimental group.