Slavery to Addiction as Meaning of Dropout in Eating Disorders : Psychological Aspects among Women That Have Interrupted Treatment at a Specialized Service in Brazil 1

This study investigated the narratives of women who dropped out of treatment from a specialized outpatient unit for eating disorders (ED). Psychodynamics aspects involved were discussed regarding the issue of adherence and dropout, since it was found to be directly affected by addictive aspects. A qualitative approach was used to identify interpretative matrices that led to a broad understanding of the clinical phenomena observed. Our aim was to identify contributions to the clinical management of barriers to treatment regarding the challenge of adherence to addiction and eating disorders treatments. The sample consisted of eight in-depth interviews fully recorded and transcribed. The emergent categories indicated elements from the experience of the disorders in psychological meanings assigned to dropout by patients. Findings are discussed under a theoretical framework focused on anorexia and bulimia as addictions. The discussion aims to make a contribution to the ED treatment approach and also understand them in their addictive dimension in order to enhance compliance and retention in treatment. The psychodynamic elements from addictive experiences described by patients may contribute to supporting the discussion about The study was developed at LPCQ; the subjects interviewed used to be patients of Interdisciplinary Group of Assistance and Studies in Eating Disorders (GETA). Resulting from a Master’s Degree, research funded by scholarship to the first author from the Coordination of Improvement of the Personnel of Higher Education (CAPES-Ministry of Education, Brazil). The first and last author also received auxiliary research Grant Research from São Paulo Research Foundation (FAPESP; No.2011/20469-8) providing subsidies for the field phase of the original study following. F. M. Seidinger-Leibovitz et al. 789 eating disorders as “addictive behaviors” beyond the “addiction model of eating disorders” as a physiological model of “addictive” or “toxic” food dependence. Eating disorders experienced as addiction were found affecting the response to treatment.


Introduction
Dropout in eating disorders is a relevant issue for public health, since it is associated with high rates, high costs, disorder severity and poor outcomes.In general, unsatisfactory results in hospitalization and a tendency to rehospitalization are related to dropout.Eating disorders (ED) are severe psychiatric disorders with mental and physical impairment that are described in almost the whole world.
World Health Organization considers ED as mental illnesses that require full attention during childhood and teenage years (WHO, 2005).Smink, van Hoeken & Hoek state that the prevalence of Anorexia Nervosa (AN) and Bulimia Nervosa (BN) among women is respectively around 4% and 2% (Smink, van Hoeken, & Hoek, 2013) according to DSM-V (APA, 2013).It is also worth mentioning that mortality rates for all types of ED are much higher than among psychiatric disorders, and that suicide accounts for 20% of deaths in AN (Arcelus, Mitchell, Wales, & Nielsen, 2011).The risk of treatment dropout is known and has been reiterated in the literature since the classic study by Vandereycken & Pierloot (1983).Those authors show that dropout rates, in average, are two times greater regarding AN hospitalizations than other psychiatric disorders.Known rate range is around 50% for AN and 30% for BN, and some studies find even higher rates for patients with atypical AN (from 28% full restricting to 49%, 50% and 60% to partial, sub-threshold and without fear of weight, respectively) (Santonastaso et al., 2009), and Kahn & Pike (2001) find higher rates in purgative AN (63%).Eivors, Button, Warner, & Turner (2003) in a qualitative study warn that "the danger when patients drop out from treatment is that they may take on the role of the 'incurable anorectic'", which generally happens after dropout and worsens prognosis.Qualitative studies on treatment dropout in eating disorders are scarce.Eivors, Button, Warner, & Turner (2003) find that control is crucial for understanding treatment dropout in cases of ED.They have pointed out that the disorder function is a coping strategy for individuals dealing with broader psychological problems.They also criticized the medical/hospital model, which would increase the patients' feeling of loss of control due to its central focus on food (Eivors, Button, Warner, & Turner, 2003).Other group of authors investigated the lack of treatment adherence among ambulatory patients with ED in general, discussing non-attendance in relation to outpatient psychosocial barriers.They highlight the therapeutic relationships and traumatic experiences related to non-attendance.They suggest the need for training, supervision, and support to help staff identify the complex needs of vulnerable individuals (Leavey et al., 2011).
Many authors stress the need of further qualitative research, including interviews with patients (Mahon, Winston, Palmer, & Harvey, 2001;Bell, 2003;Masson, Perlman, Ross, & Gates, 2007;Sly, 2009) and therapists (Mahon, 2000) in order to investigate psychological factors linked to dropout and its meanings (Surgenor, Maguire, & Beumont, 2004).All those authors suggest the use of qualitative methods to shed light upon understanding predictors of dropout.
This article results from a study that aims to understand the meanings of dropout in outpatient's ED treatment using a qualitative approach.The experience of the ED as an addiction emerged as central among the results, hence making the focus of this article the discussion about eating disorders as addiction behaviors (Davis & Claridge, 1998;Davis & Carter, 2009;Speranza et al., 2012;Goodman, 1990).

Method
This study uses the qualitative method defined by Malterud (2001: p. 483): it involves systematic collection, organization, and interpretation of textual material derived from talk, in the exploration of meanings of social phenomena as experienced by individuals themselves, in their natural context.The present study follows standards established by Helsinki's Declaration that regulate Health research with humans and also are approved by the National Committee for Ethics in Research2 .In order to avoid biases, the interviews were not performed in the service that was abandoned by the participants.Six of them were interviewed in their respective homes and two preferred to choose another safe and private place.All the participants read and signed a Statement of Informed Consent.

Participants and Recruitment
Dropout was defined as treatment interruption due to unilateral patient's decision.The research team received a list of patients who discontinued their treatment for at least one year after at least a month of treatment in an outpatient service that is specialized in ED.
Sampling was intentional, using the following criteria: 1) inclusion criteria-patients who were aged eighteen or older who were diagnosed with ED using criteria from DSM-IV-TR 3 (APA, 2003) and presented dropout as defined above; 2) exclusion criteria-inconclusive diagnosis or intellectual limitations that could compromise data's validity.We interviewed a total of twelve participants.Sample size was defined using saturation, which means that the sample was completed when information provided by research participants no longer contributed significantly to the analysis of data collected (Fossey, Harvey, McDermott, & Davidson, 2002;Fontanella, Ricas, & Turato, 2008;Turato, 2000).
No males were present on the dropout list supplied by the outpatient unit.Therefore, all participants are female.The limited number of participants reflects the in-depth nature of the methodological approach, in which data's detail and richness are favored by the small number of interviews.Participant's descriptive characteristics are presented in Table 1.

Procedures
The first contact was made by telephone calls informing about the study and the participants who volunteered had interviews scheduled and subsequently carried out at their homes in order to reduce biases related to the environment of the outpatient treatment service, avoiding discomfort, facilitating their participation in the research looking toward avoiding biases relatives to collect data inside the service abandoned.An adequate private place was previously ascertained.In-depth interviews guided by a semi-structured script (previously tested as a pilot version) were recorded and fully transcribed.

Data Analysis
Qualitative data analysis was performed through the following steps: listening exhaustively to audio interviews, reading the transcripts of interviews several times by first author (FMSL).After that, the emergent themes were organized in content categories.The categories were named according the meanings found in the interviews.The emerging categories were analyzed according to the chosen theoretical psychodynamic background.The findings were several times brought back to the wider group of researchers at Laboratory of Clinical Qualitative Research for further discussion and consensus about their interpretation.As this study is part of a broader analysis of the meanings of dropout for ED patients (Seidinger, 2014), only categories related to the concept of addiction were reported in this paper.
Our analysis focuses on psychological meanings attributed by patients based on a theoretical framework that understand anorexia and bulimia as actuated addictive behaviors (Jeammet, 1999a;Brusset, 1999;McDougall & Fine, 2003).According to Jeammet (1999a: p. 45), the functioning of bulimia-understood as the basis of all eating disorders-is identical to that of addiction.Similarly, Brusset (2003) describes in these cases a direct release of instinctive load disconnected from representations, characterized by "slavery, dependency, submission, repetition, while crises repetition tends to impoverish relational, affective and imaginary life (...)" (Brusset, 2003: p. 140).The idea of "slavery" also appears in descriptions of Jeammet (1999a), Brusset (2003), McDougall & Fine (2003).It is common in the psychological perspective of these authors the view that the drug object, whether it is food or not, take place as this the helplessness determinant of psychological dimension of "slavery" and compulsive behavior.It is worth mentioning that the paradigm described by these authors offer elements that were almost forgotten in the clinical field of ED.

Results
The following content analysis categories present the meanings of dropout for participants who had as their central dimension of the experience of ED and dropout as addiction.The threads are named in two parts: the first one, a native speech statement elected due to be representative of the enunciation found in the speeches of the interviews; and a second part, its description based on the theoretical framework.
The following content analysis categories present the meanings of dropout for participants who had as their central dimension of their experiences of ED and dropout as addiction.From the main category described below slavery to addiction emerged three subcategories.They describe relevant aspects of the ED while addictions revealed by the in-depth interviews: 1.1) control versus relief; 1.2) the permanence of a relationship to slavery; 1.3) the need to be an ally-the therapeutic alliance.

Main Category: "Slavery to Addiction"-The Addictive Dimension of Anorexia and Bulimia
The participants' experience of ED resembles that of a drug addict or an alcoholic and is characterized by selfdestruction and refusing to consent to treatment, despite the awareness of its need.As long as a certain type of alliance exists-what Participant 7 described as: "an ally to disease, a drug slave", "(…) this slavery which is anorexia is like an addiction"-there will be major barriers between these patients and adherence to specialized treatment.In turn, this association makes visible the relationship between slavery and the experience of loss of control as explains Participant 5: you see that it made you lose control of your life, it is controlling you.[It's a] kind of an addiction itself, you can't control yourself.Participant 7 adds to the meaning of Slavery: [I am] at the mercy of the conditions imposed by the disease.The disease itself makes you live in slavery, in compulsive way.In the same direction, Participant 6 says: I recommend that treatment takes place before becoming addicted... It's like a drug, you start and don't want to stop!I see it as an addiction.(...) We emphasize, respectively, the location of two important points: a) subjection as a characteristic of the dynamic of the disease itself in anorexia/bulimia, with a recognition of its compulsive and repetitive nature; b) because of its addictive nature, since the disorder onset, participants believe that the relationship with treatment will be difficult.
"Treatment really should have extended longer and I ended up not letting me [emphasis added] complete the treatment" (Participant 5).This sequence shows the participants' conscious and critical assessment of the relationship with dropout.The emphasized words show a subtlety contained in this enunciation: the connection with the participant's dominance relation with its allied position to addiction.We note a sense of "fight", a kind of confrontation of the subject with oneself that says: "I ended up not letting me".Such subtlety is shown to be connected to an instance within oneself that imposes and subjects a determinant of Slavery.She struggles within this play of forces, but the side that does not allow her access to treatment ends up winning, resulting in dropout.The aspect of keeping the control also observed will be contemplated in next topic.
It should be mentioned here the known ambivalence in ED.Thus, "A slavery to addiction" shows psychodynamic processes linked to addiction as a meaning of dropout.Also it reveals a way of relating to food meaning "drug" that points to certain particularities in the relation of these participants to treatment.In addition, it reveals a relation of meaning equivalence: enslaved by the disease as if by a drug (P7).We emphasize that psychological functioning is revealed as a common denominator to the subsequent categories.In turn it leads to understand the psychological meanings of dropout as internal barriers of a psychodynamic nature that play between patient and treatment.
1.1 "I do not like things get out my control"-dropout as relief While patients perceived themselves as being monitored and under surveillance due to treatment, the act of dropping out could mean relief.Through withdrawing from treatment they try recover the control that has been lost, as show the excerpts: (P1): If you are in treatment, then you don't have privacy (…) you are watched all the time (...) "your daughter has lost weight" (...) Then more persecution starts [emphasis added]; (P2) I felt quite suffocated too [… There] was a lot of pressure, the main reason I quit [...] I do not like my mother's demands [... it] oppressed me a bit.The relief found in dropout meant a way of breaking free from surveillance from their family and professional team, since this is experienced as invasive.In this sense, also the use of medicaments may be felt as a bad experience: P6: It's very important for me to have control of the situation.(P5) Antidepressant ... looks like you're not more controlling over your emotions.(...) I do not like things get out my control.Locus of control is situated on the team.This is experienced as disturbing and this feeling increases while being in treatment.Dropout is revealed as an act that aims to reverse the sense of loss control.Such mechanism points the issue of control, a known psychological conflict in ED that affect the interpersonal relationship.This relationship will be further explored below in the discussion with the literature.
1.2 "Anorexia is for life"-attachment/adherence to the symptom The dimension of dropout as relief reappears and reveals its connection to a permanent character of the ED symptom: (P6) I can't see myself eating and not vomiting.It looks like it will suffocate me or kill me (…) I think that's why I have always dropped out.The discourses of participants convey a "naturalized" understanding of eating symptoms.A participant refers as getting used to, as something that becomes part of her identity.She is conscious of the inversion of the normality parameter.So, there is an impossibility of seeing herself without such symptoms: P6: (...) I see it as a lifestyle that I'm going to die with it.(...) I watch people who eat normally, I think that they are different (…) and I'm normal.Is it so crazy, but it is how I feel!This core sense is related to dropout as relief in the attitude of resignation at their permanence in accordance to the psychological aspect of addiction's alliance.This connection with the symptoms' maintenance, with no exception in the interviews, reflects a resignation to anorexia or bulimia as something for life.They report to be aware of this paradox, but unable to leave.This ends up competing with treatment, winning the fight by taking control and leading to relief in keeping it.
We chose to describe this relationship with ED symptom as attachment or adherence because what happens is from the same order of the addictive relationship with food.The sub-categories themes point barriers linked to slavery to addiction that stresses psychodynamic mechanisms.This is also manifest in the ambivalence towards treatment and been cured.Thus, it is understandable that participants rule out the representation of a treatable disease, similar to what happens with the alcoholic or addict.The participants describe disbelief that arises when the treatment has achieved some improvement and leads them to the verge of no longer showing symptoms.They revealed facing impasses at this moment.Once improvements are considered to be sufficient, they retreat on the edge of abandoning the symptoms, and end up dropping out of treatment instead of leaving "addiction".Different degrees of improvement are supported against such impasses, and thus considered sufficient.Being "too well" (P5) conveys the idea of not being allowed to be well under their psychodynamic functioning of slavery.So, it becomes necessary to drop out.Anorexia and bulimia are represented as having no cure-"for life" -is a compromise to keep the coping strategy to their psychological conflicts.Control also plays an important role in this game.(P4): if I'm managing to control myself, I will run out of medicaments, which is better.
1.3 "Ally to the slavery"-from the condition of being ally of the disease to the condition of therapeutic alliance A participant highlights an interesting symbolic effect of this interaction by sharing the "subjective weight" in the support found when talking with the physician with whom she had an alliance.She points to relevant psychological aspect that in the case of people who experience body as excessively heavy.In their speech, the term "lighter" is connected to a subjective relief (P2): As I talked a lot with Dr. X, I felt lighter by not needing to be alone inside me and myself4 .
They picture themselves resigned to anorexia and no cure for life, as revealed by the experience reported by Participant 6: "dropout to avoid the risk of quitting vomiting".This meaning of accepting the domination, such as surrender, surrendering to the victorious "anorexia for life", is typical in a psychological functioning of slavery to addiction.
The importance of speech and bonding, as allies of treatment is enhanced by findings that point to participants' perception of mental distress, criticizing the focus on diet, weight and body image.The interviews suggest the need to find support to the opposite path: disinvesting the "object food", excessively invested by the displacement typical of ED.As this participant suggests (P1)-To talk about other things unrelated to food.(…) these makes us very frustrated (...) I wanted to talk about why I'm sad.(...)You hold sadness, you hold anger, you hold grudges, hurt, anything, because you do not have the courage or cannot talk to anyone (...) we are sad, or we are insecure, or we are afraid (...) who has difficulty [to speak], has more desire than anyone else.
The richness of psychological aspects of ED brought by this excerpt point the importance of the talking experiences into the treatment and the alliance to make it possible.Participants emphasize the importance of their need for treatment of psychological aspects and emotional support, focusing on speech as a means of treatment, information and support in the alliance with professionals.(P7): I liked to talk, solve my doubts and to know that I was not alone in my fight.That was good!
The participants' discursive sequences from all the categories reveal a relation of subjection to addiction as it described by theoretical frame as an element that characterizes ED as addictions.At the level of "acted out conduct" we highlight the mode of relation to food as similar to the relation to a drug.Therefore, for the studied group, this relation is something for life.In turn will lead to the treatment dropout, circuit well illustrated by the speech of Participant 6: "I dropped out not to stop vomiting".

Discussion
It is important to keep in mind that these results cannot to be generalized, since this is a qualitative study with an intentional sample.It would be interesting to see new studies that investigate the phenomenon of treatment dropout in ED considering other types of dropout, such as non-adherence, or other scenarios, such as inpatient care, and using different methods.Also, the application of addiction scales in individuals with BN but inclusive AN, both restrictive and purgative subtypes.Also, it is worth mentioning that our results also don't apply to psychological factors regarding patients that haven't interrupted treatment.The non-generalizability of the study does not preclude the examination of the responses of the participants enables greater knowledge about the topic addressed.

Eating Disorders and Addiction
After the Russell's description of Bulimia Nervosa in 1979, BN was considered "the addiction of the 80s" by Brisman & Siegel (1984), who quoted Casper, Eckert, Halmi, Goldberg, & Davis (1980), Halmi, Falk, & Schwartz (1981: p. 113) examining what they described as "a psychologically addictive cycle of binge eating and purging of food".[emphasis added] From this sentence it's worth highlighting the "psychological" dimension in addictive cycles to introduce something present in our qualitative findings: the existence of a psychological functioning that is named by the participants as "slavery".From standpoint of their experiences this psychological functioning preexists, leaving them at the mercy of it (P7), and is reinforced by the food or thinking on food (P1; P7) cycle of addiction.In the same sense, the authors cited above describe as "addictive" this "psychological dimension" of binge eating and purging that characterizes ED.
The number of publications on the intersection between eating and addictive disorders increased steadily during the 1990s, after the milestone definition of "addictive disorders" by Goodman in 1990(Katz, 1990;Krahn, 1991;Wilson, 1991).More recently, comprehensive reviews have investigated the evidences of the addictive model to "compulsive overeating" (Davis & Carter, 2009), and BN and other ED Umberg, Shader, Hsu, & Greenblatt (2012), approaching the ED's field to a discussion formerly linked exclusively to obesity.Meanwhile, binge eating disorder (BED) became a specified ED in the DSM 5 (APA, 2013).It could be interesting to investigate the hypothesis that this may indicate that the distance between these fields is narrowing due to the fact that BED was first acknowledged as among the addictions, and now it is a specified ED.
The model of "food addiction" defined by Ronel & Libman (2003) characterizes binge eating as a physiological addiction food and recommends abstaining from certain foods considered addictive or toxic.The theme addressed as "food drug", unlike the physiological conception of "food addiction" from Ronel & Libman is related to our findings by the emphasis in the psychological aspect of psychological dependency, by addictive behavior, and not of the "addictive potential" of the substance, drug or food, by itself.Umberg, Shader, Hsu, & Greenblatt (2012) support the definition of BN as addictive disorder through a comprehensive review, collecting neurobiological evidences under the hypothesis that the same neurobiological substrates affected by drug addictions perpetuate the binge-purge cycle in BN.Furthermore, this study's definition of BN is based on the central symptom of binge eating accompanied by a specific psychological aspect: "feeling of loss of control".In the same direction we emphasize the description of bulimia nervosa adopted by the authors for this study: "(...) a psychological addiction to the binge and purge."(Brisman & Siegel quoted by Umberg, Shader, Hsu, & Greenblatt (2012: p. 376).Umberg, Shader, Hsu, & Greenblatt (2012: p. 380) conclude by stating the pertinence of redefining BN within a paradigm of addiction, criticizing the general statement "food is addictive" by affirming that it is the compulsive pattern found in BN what confers addictive potential for food.Similarly, the women from our study talk about their experiences with ED and treatment giving a stressed sense of relevance to the psychological functioning of addiction towards eating behaviors.
The study by Speranza et al. (2012) cites an author included in our theoretical framework.They quote Fenichel (1945: p. 182) pointing his view of "addiction not as a simple property of the drug, but as resulting from an intemperate relationship between the person and object of his addiction".By applying personality scales the authors (Speranza et al., 2012: p. 188) conclude that a subgroup of subjects with ED experience their disorders as addictions and, among these subjects, patients with BN and purging AN show higher scores of addictive disorders, followed by restrictive AN.This point deserves emphasis since it was rarely considered in previous studies, which usually linked BED to obesity, disregarding AN and non-purgative variants.Such studies also failed to address personality traits that add risk to drug addiction, including the restrictive subtype next to purging AN and BN in the addiction field.
It is interesting to collate these conclusions with the conceptions theorized by Jeammet (1999a), Brusset (1999), Selvini-Palazzoli (1965) quoted by Jeammet (1999a), McDougall & Fine (2003) regarding the compulsive and addictive nature of all ED, including AN, understood as a form of denial of bulimic binge.This view considers that both disorders share the same addictive nature.Jeammet says: "the bulimic ghost is always present behind any anorexic behaviour5 (Jeammet, 1999a: p. 45)".The author, in accordance to Brusset and Selvini-Palazzoli understands Bulimia as primary.This was present in our empirical findings, as illustrated by Participant 1 (P1) in the control-related category: in order to be compulsive and avoiding control over its relationship with the "food object", she manages to solve the lack of control of her "bull's hunger" through extreme control, becoming anorexic as a formative reaction.The narrative in question shows that it is not a lack of appetite, but rather, a compulsive, addictive and "extreme" appetite that leads the subject to "eat nothing".The category slavery to addiction shows that it works under the logic of extremes regarding control, revealing compulsion and lack of control as a psychological mechanism.This trait, also found at the basis of restrictive AN, reveals that the same "addictive nature" exists in the compulsive and purgative subtypes.
Supporting ED as addictive behavior, but underlining the discussion to include AN, more recent studiessuch as Kaye, Wierenga, Bailer, Simmons, Wagner, & Bischoff-Grethe (2013)-review evidence supporting the hypothesis of shared neurobiology between ED and addictive disorders (AD), including starvation in AN as a form of addiction, alongside BN.Evidence supporting ED as an addictive behavior is also found in AN, including high scores in Eysenck Personality Questionnaire's Addiction Scale, and self-imposed hunger strike and abuse of physical exercise as an endogenous opioid addiction (Davis & Claridge, 1998).This underlines the imperative psychological character of "self-imposition" which is crucial in eating and addictive symptoms.Ac-cording our findings and the literature may shed light on this polemic theme and maybe point to neglected aspects in the treatment of ED that could play a role in dropout.Davis & Carter (2009) evaluate the clinical and behavioral parallelism and risk factors they found in the literature and a series of evidence of neurological mechanisms of dopaminergic system the is common to addiction and compulsive overeating.They argue that both disorders share the same diathesis.Discussing the loss of control, tolerance, abstinence and relapse, the authors argue in favor of modifying treatment for compulsive overeating by recognizing the similarities to "conventional drug addictions" (Davis & Carter, 2009: p. 6).Those authors conclude in favor of understanding food addictive behavior as analogous to addictions such as gambling, shopping, in the same sense as argued by Goodman to Addictive Disorders (Goodman, 1990).This is also aligned with our own results.In contrast, Cassin & von Ranson (2007), despite finding evidence of dopaminergic mechanisms classified as addiction and binge eating disorders (BED) for 92.4% of the participants (DSM-IV) (APA, 2003) and 40.5% for the more conservative criterion of Goodman, they advise against the redefinition of BED as an addiction, arguing that the disorder shares with AN and BN the fundamental trait of concern about weight and shape and includes binge eating.This is one of several arguments of the critical analysis by Wilson (1991) to the additive model of ED.In contrast, our results find in symbolic meanings attributed by participants the relationship with weight and shape as resulting from the primary addictive psychodynamic.

Therapeutic Alliance, Dropout and Addiction
The slavery to addiction refers to a need of an object that is essential for living but is transformed in its function: food, whose transformation into a 'drug object' converts the subjects "allied" to it into slaves.The theoretical framework anchoring our study helps to understand and explain the shift of the "food object" that is invested with the character of a drug, and it is found as fulfilling the function of "supporting" voraciousness in the face of helplessness.Our findings are aligned with Brisman & Siegel statement that considers bulimia and alcoholism as two sides of the same coin: "However, it is probable that substance abusing populations have a core personality structure in common that is deficient in emotional and self-regulatory functioning.Without a change in this structure, abused substances may be used as alternatives in search for self-care (Brisman & Siegel, 1984: p. 117)".An epidemiological study by Hay, Mond, Buttner, & Darby (2008) also stresses the increase of purgative behaviors.It is worth to comment this finding under the light of social and cultural issues related to consumption of food, drug and consumption in general.It is interesting to add a sociocultural dimension widely studied and discussed by Gracia-Arnaiz (2010) to the debate.The author points to the cultural consequences of dietetic norms affected by the market logic in industrial capitalist societies with impact on obesity and ED, a fact that may go beyond individual psychological facts.The social discourses and practices that incite consumerism are components of the "addictive behavior", manifest in the distress referred by participants: they are self-described as living under the imperative of consumption since that is the main feature of the slavery to addiction reported.
Goodman claims that the concept of addictive disorders "provide the basis for a more comprehensive and effective approach to the understanding and treatment of people experiencing these disorders" (Goodman, 1990(Goodman, : p. 1407)).When the addictive dimension of ED is considered, it can make a difference in the treatment, especially if we consider that effective and comprehensive approaches are still lacking, and dropout rates are high.
In Brusset and other authors, we found reported as "addictive" the "anorexic/bulimic object relationship".(Jeammet, 1999a;Brusset, 1999;McDougall & Fine, 2003;Brusset, 2003;Jeammet, 1999b).This aspect seems to be connected to dropout in the content categories that emerged in this study.The relationship to treatment appears to suffer interference from the psychological aspects of addiction.The meaning "anorexia is for a life" departs the disorders' experience from the representation of a treatable disease, experienced as an alcoholic or addict who despite of the disease that "only destroys" (P8).By assuming anorexia/bulimia forever, participants describe a feeling of disbelief about their improvements achieved by pointing out that, when they have to abandon certain symptoms, not being cured of the "slavery to addiction", they decide to dropout.
The central psychological mechanisms that were revealed in the categories presented in our study deserve to be highlighted: compulsion to repetition, loss of control and relapse, noncompliance manifested as dropout despite awareness of its necessity, ambivalence, and reluctance to improve.It is worth to mention that these are all included in Goodman's construct of addictive disorders (Goodman, 1990(Goodman, : pp. 1404(Goodman, -1406)).Goodman has proposed modifying the nosographic entity "addiction".According to him, it was "defined as a process whereby a behavior that can function both to produce pleasure and to provide relief from internal discomfort is employed in a pattern characterized by (1) recurrent failure to control the behaviour (powerlessness) and (2) continuation of the behaviour despite significant negative consequences (unmanageability) (Goodman, 1990(Goodman, : p. 1404)".
The analysis of in depth interviews suggests that adherence6 and retention to treatment reveals to be tolerated by participants to the extent that everyone could benefit from treatment in the reason to achieved improvements in the field of psychological treatment of the psychodynamic.Participants indicate that other strategies for normalization of diet and weight and the extinction of compulsive and purging behaviors are powerless, while they remain at the mercy of slavery to addiction (P7).They thus refer to, the understanding of the ED as a defense, a kind of strategy that the subject cannot give up easily.
The category attachment to the symptom indicates psychological functioning as "protective of the self", as described by McDougall & Fine (2003: p. 190) as we have found.Nordbø, Espeset, Gulliksen, Skårderud, Geller, & Holte (2012) name this category in AN as "reluctance to recover".It is also worth mentioning the parallel with Goodman's description of the relationship of addicts to treatment as the "the tenacity of addictions and the resistance of addictive behaviour to modification (Goodman, 1990(Goodman, : p. 1406)".

Clinical Implications
Psychological meanings of Dropout reveal a relationship to slavery to disease as a drug (P7).Thus, the psychodynamics involved should be considered in studies of the dropout phenomenon, as well as the implications of the addiction dimension for treatment of Eating Disorders.It is possible that the patient in treatment may not desire to, or may not have reached the emotional conditions of giving up the symptoms and to begin functioning under a different strategy.It's worth to consider the need of a psychological approach of ED symptoms as addictions.Thus, we interpret from participants' speeches: the need to help them to build conditions to turn off the addictive psychodynamic functioning, the holding to "helplessness" that drove them to food as a drug is in itself, the psychological treatment.
In general, current treatment approaches in ED are not designed for dealing with addictions' psychodynamics and symptoms related to addictive behaviors.Based on our findings, we can suggest following points to the management of aspects revealed as important in relationship to treatment, when the ED are experienced as AB: • Psychological interview in admission, step-structured treatment, prioritizing the building of a link with the team, emotional support, in order to avoid the strictest nutritional approach to following steps; • Managing team to avoid expectations for complete abstinence; • The creation of relief devices of the relationship with the treatment team, since it is probably going to be experienced as invasive.Thus, it is necessary to prioritize investment and management of specific therapeutic alliances.It should take into account that the interpersonal relationship of these patients and with the service will be inevitably permeated by ambivalence, control and counter-control mechanisms participating in "addictive" relations; • To build conditions of internal support to move from an alliance to addiction (ally to disease) to a therapeutic alliance would imply: for patients-a phase of "addiction to the treatment"; for the team, to hold and manage the acts addressed to the staff and therapists.Since the addictive psychological functioning and enacted behaviors are prevailing, the difficulties in interpersonal relationship are expected.

Conclusion
We found elements that indicate the relevance of the experience of ED as addictions in the discourses of a subgroup of patients-those who had dropped out of treatment.The reported relationship of ED dropout with the psychodynamics of addictive disorders may foster a better understanding of ED phenomenon and treatment non-compliance-including dropout-for such disorders.One should keep in mind the possibility that, for these patients, an approach taking into account the addictive dimension may be important to cover psychological aspects that need to be addressed in therapeutics.It is also worth considering that adherence and retention in treatment for ED patients may be improved if the ED professional team offers support and appropriate clinical management in order to achieve a proper therapeutic alliance.This can be a way leading to the relief of psychological conflicts through treatment and not through dropout.

Table 1 .
Descriptive characteristics of the sample.