|Home | About | Journals | Submit | Contact Us | Français|
The purpose of this study was to explore how individuals with anorexia nervosa (AN) engage in treatment and define recovery. A mixed methods design was used to triangulate the experience of 20 women with a history of AN. Interview data were analysed thematically to explore frequency of emergent themes and current eating disorder psychopathology was assessed using standardized self-report measures. Participants’ mean age was 29.35 (SD = 12.11). Participants’ scores were indicative of persistent psychopathology. Those with more involvement in treatment choice had better motivation to change and normalized eating. Participants’ definition of recovery mapped on well to current research conceptualizations, though a substantial proportion of the group expressed some ambivalence around the concept. Results are interpreted in the context of self-determination theory of motivation and suggest that patients should be involved collaboratively in the formulation of shared goals and concepts of recovery in treatment settings.
Treatment outcome literature is plagued by a lack of consensus on what constitutes a good outcome, and particularly, on concepts of remission and recovery (Couturier & Lock, 2006a). Outcomes appear to vary widely depending on the definition applied (Couturier & Lock, 2006b), and remission status is not always equated to even short-term maintenance of treatment effects in an inpatient setting (Darcy, Dooley, Radmall-Quirke, Redmond, Keogh, & Larkin, Submitted). Being the defining feature of anorexia nervosa (AN), weight has been considered a reasonable indicator of illness status. As a consequence, treatments for AN usually employ a weight target that patients are required to work towards, and necessarily serve as, at least part of the treatment goals. While nutritional rehabilitation is an essential part of recovery (NICE, 2004), resistance to weight gain and denial of the seriousness of the problem are hallmark characteristics of the illness (APA, 1994), and thus treatment for AN can be cast as immediately at odds with the wishes of the patient.
Motivation to change is assumed to play a major role in therapeutic engagement, and various models have been applied to AN. Motivational interviewing (Miller & Rollnick, 1991; Vitousek, Watson, & Wilson, 1998) the transtheoretic model of readiness to change (Prochaska & DiClemente, 1982) and the readiness and motivation interview (Geller, 2002), all emphasize interventions that encourage the patient to take an active role in therapy and assume that the patient possesses the power for change (Vansteenkiste, Soenens, & Vandereycken, 2005). With the aim of providing a strong conceptual framework for motivational research, recent thinking proposes self-determination theory as an overarching conceptualization (Vansteenkiste et al., 2005). Within this framework, when a person experiences a sense of volition or autonomy with respect to treatment they can better engage in therapeutic change. Evidence to support this theory comes from a recent case series where patients admitted to an inpatient service under a standard procedure were compared to patients who were engaged in such a way that aimed to maximize a sense of volition and personal responsibility (Vandereycken & Vansteenkiste, 2009). Patients engaged within the latter strategy had lower rates of short-term dropout but longer-term dropout remained similar to that in the former strategy (Vandereycken & Vansteenkiste, 2009).
Dropout from treatment is a major problem inherent in treating eating disordered patients and is usually viewed as non-compliance, resistance or treatment failure (Vandereycken & Vansteenkiste, 2009). Most studies that have explored dropout have focused on patient characteristics and symptoms as predictors such as the presence of binge/purge symptoms, BMI on admission, severity of eating disorder symptomatology and greater psychiatric difficulty in general, however, findings are scare and conflicting (Wallier, Vibert, Berthoz, Huas, Hubert, & Godart, 2009). Very few studies have explored factors related to the treatment setting or therapeutic relationship (Morlino et al., 2007; Clinton, 1996; Vandereycken & Pierloot, 1983) and no study, to the authors’ knowledge, has qualitatively investigated the patients’ own reasons for dropout.
Few studies have examined the patients’ perspective of recovery and those that have, have focused on the factors that individuals with eating disorders consider important in order to achieve it. The existence of social support, motivation to change, developing an identity separate from the eating disorder and factors relating to the therapeutic alliance have been consistently reported to be important in the process of recovering from AN (Federici & Kaplan, 2008).
Using phenomenological qualitative analysis, one study interviewed patients who had and had not maintained a healthy weight about the factors that had lead to their successful or unsuccessful weight maintenance (Federici & Kaplan, 2008). Salient themes that differentiated both groups were motivation (internal vs. ambivalence), treatment specific factors (satisfied with discharge planning versus dissatisfaction with treatment), social support (being well supported vs. socially isolated), affect difficulties (awareness and tolerance of negative emotion) and self-validation versus self-criticism (Federici & Kaplan, 2008). Both groups described recovery as an ongoing process. In another study, frequency analysis of themes that emerged from interviews with 69 former patients revealed that the existence of a supportive relationship (partner or therapist) and maturational processes were the most commonly cited factors leading to recovery (Tozzi, Sullivan, Fear, McKenzie, & Bulik, 2003). Also commonly cited as fuelling the process of recovery was the desire to become pregnant or start a family.
While these studies help to unravel the complex process of recovery, there remains a wide gap in our knowledge about how individuals with AN themselves would operationalize recovery. We still have little or no understanding of why patients dropout of treatment, and little research has investigated why individuals participate and engage in treatment. An understanding of shared goals of treatment is essential to the development of interventions that are better able to mobilize the resources of individuals themselves to foster engagement, avoid dropout, and strive for full recovery.
This study aimed to triangulate the perspective of former AN patients in terms of how they engage with treatment and define recovery using qualitative and quantitative data. The study was largely exploratory, with one exception; according to self-determination theory (Ryan & Deci, 2000; Vansteenkiste et al., 2005), we hypothesized that there would be a significant positive relationship between autonomy in treatment and motivation.
A total of 24 women with a lifetime history of AN were recruited through various community sources and Internet postings. Potential participants were screened when they first contacted the clinic by phone for eligibility, the criteria for which was age above 18, a lifetime history of AN, and to have attended some type of related treatment. Twenty-four women were deemed eligible and gave their consent to be interviewed and to complete self-report questionnaires. Out of the 24 who volunteered their time, 23 were interviewed individually (n =12) or in focus groups (n =11). Three participants did not complete quantitative measures and one participant completed questionnaires but did not attend for interview. For the purposes of this study, these four cases were omitted from the analysis.
To quantitatively contextualize participants’ level of eating disorder-specific psychopathology, we used the eating disorders evaluation questionnaire (EDE-Q) (Fairburn & Beglin, 1994)—a 36-item self-report version of the semi-structured interview yielding four subscale scores that are comparable to the interview—restraint, eating concerns, shape concerns and weight concerns. To assess motivation, we administered the anorexia nervosa stages of change questionnaire [ANSOCQ; (Rieger, Touyz, & Beumont, 2002)]—a 20-item measure that places respondents in one of the five operationally defined stages of change based on Prochaska and DiClemente’s model (Prochaska & DiClemente, 1982) along three-dimensions; eating concerns; weight and shape concerns; and ego-alien aspects. The eating disorders quality of life instrument (Engel, Wittrock, Crosby, Wonderlich, Mitchell, & Kolotkin, 2006) was also administered. The EDQOL is a 25-item multidimensional self-report measure that assesses the extent of eating disorder-specific quality of life.
The semi-structured interview schedule consisted of open-ended and probe questions designed to facilitate an exploration of participants’ perspective of treatment, dropout and their conceptualization of recovery. The style of the interview was one of ‘directed conversations’ (Pidgeon & Henwood, 1996) in which the schedules were applied flexibly in order to allow for exploration of any new issues that might be raised by participants, and asking of additional questions where appropriate.
Ethical approval for the study was granted by our institutional review board.
Data were analysed thematically using an inductive (data-driven) approach. Data analysis was guided by the procedure outlined by Braun and Clarke (Braun & Clarke, 2006). Data codes were generated systematically, then collated into ‘thematic maps’ and applied to the entire data set to generate frequencies.
To determine reliability of the coding scheme, a random 20% sample of interviews was selected for analysis by an independent coder. Chance-adjusted Cohen’s Kappa coefficients were between 0.64 and 1.0 for all of the themes.
Data from a total of 20 participants (all female) were analysed. Participants ranged in age from 19–52 (mean age = 29.35; SD = 12.11). All participants’ had either childhood or adolescent onset of illness. Eighty-five per cent were Caucasian, with the remaining 15% Biracial. Demographic details are presented in Table 1.
The group’s mean scores on the four measures are presented in Table 2 below. As a means of contextualizing the health and quality of life status of the group, mean scores were compared to published scores from other clinical and community samples, where available. Just less than half of the sample had scores on the EDE-Q greater than 2 SDs from community norms for all subscales, suggesting persistent, significant impairment. Mean scores on the ANSOCQ were not significantly different from a sample of inpatients with AN with the exception of readiness to change ego-alien aspects of the illness. Overall scores indicated that the sample was at the preparation stage for change. Participants’ scores on the EDQOL did not significantly differ from those of a sample of eating disorders patients (Engel et al., 2006) suggesting that overall, the quality of life of participants remained impaired at the time of assessment.
Participants were asked how the decision had been reached to avail of their various treatment options. Figure 1 shows a thematic map of responses.
Most participants had no involvement in any decisions pertaining to treatment. Only three participants reported being involved in the decision to seek treatment, and one participant reported that seeking treatment had been their choice. Themes emerged from the data that broadly fall into the categories of external decisional forces and familial involvement. The majority of participants reported influences that were external to themselves and their families. The two most common subthemes to emerge in terms of external decisional forces were that in seven cases, the decision to treat was against the participant’s will, and the same number reported that doctors had been the major decision makers. In four cases, health insurance was cited as being the primary deciding factor in choice of treatment.
Ten participants cited a family member or loved one as being important in the decision making process. Five participants noted that their parents had made the decisions regarding treatment, three cited another loved one, and as noted above, three described a process that was joint with their parents.
In order to test the hypothesis that a sense of autonomy would be related to better motivation, participants’ responses on treatment choice were considered a crude measure of autonomy and collapsed into mutually exclusive themes—either family/self (8; 40%) or external factors (12; 20%). Given the low numbers in each cell, a Fisher’s exact significance test was carried out on treatment choice and ANSOCQ total score. Perceiving the decision to seek treatment as originating within oneself or family was significantly related to more advanced stage of change (preparation; action and maintenance) [χ2 (df = 1) = 7.17; p <.01]. Reporting external factors as primarily important was related to current motivation in the precontemplation and contemplation stages of change. In addition, having more autonomy in treatment choice was significantly related to having an EDE score within 1 SD of a community mean [χ2 (df =1) = 7.5; p <.01] (see Tables 3 and and44).
Participants were asked what their goals had been upon entering into treatment. Results of the frequency analysis are displayed in Figure 2. Their responses fell broadly into two themes—past goals and those that they would have if they were seeking treatment now. In terms of what their goals had been at the time they sought treatment (Former), seven participants noted that the treatment goals had been their parents’ and not their own. The same number of participants (seven) noted personal goals for treatment. Four participants cited that they wanted acknowledgement of their symptoms, or ‘to be heard’, and eight participants described other miscellaneous goals such as to address symptoms of obsessive–compulsive disorder, or to feel less depressed (see Figure 2).
There were a variety of subthemes that emerged from the data in terms of what goals participants might have in treatment if they were to seek it at this point in their life. Six participants cited that they would aim to work on disordered thinking, or other psychological elements of their illness. Four participants spoke about goals that pertained to social interactions and relationships, three participants spoke about goals that involved family therapeutic work, and another three participants felt that they would work on accepting themselves for who they are. Three participants either said they would not have any goals or did not know what they might be. Finally, eight participants described other goals such as learning to care for oneself, and learning effective ways of coping.
Participants were asked how they would define recovery. Five major themes emerged from the data. Figure 3 displays the thematic map with quotes from participants. The majority of participants (65%) defined recovery in terms of a cessation of eating disordered symptoms. Within this theme, the most frequent subtheme was eating comfortably (eight participants). Restoration of normal function of the physiological symptoms pertaining to AN (weight and menses) featured in seven participants’ definitions of recovery. Five participants spoke about recovery in terms of having no distorted thinking patterns, and two participants spoke about a resolution of body image or self-esteem deficits. Finally, two participants spoke about recovery in terms of general overall healthy functioning.
Five participants (25%) believed that recovery does not exist. Seven participants (35%) gave responses that were coded as ambiguous; four participants were unable give a description of what they felt recovery might be; two participants spoke about recovery as an ongoing process, and another two stated that the illness was easier than recovery. Interestingly, both of these were the only participants that had been diagnosed in the late 1970s.
Twenty per cent of participants spoke about recovery in social and/or interpersonal terms. Four participants spoke of recovery meaning the ability to develop and/or maintain loving and meaningful relationships with others. For another two participants, recovery involved a personal sense of confidence and the ability to be assertive in relationships. Thirty-five per cent of participants noted a wide range of miscellaneous factors that did not map on to existing themes.
Almost all participants had dropped out of therapy at some point. Participants were asked what factors had contributed to their decision to dropout. Two themes emerged from the data (see Figure 4). The majority of participants cited aspects about the treatment setting. For example, eight participants (45%) spoke about a specific clinician who had influenced their decision to dropout. Five participants (25%) spoke about the punishment/reward system in some treatment settings and the same number spoke about food options, and in particular, about a lack of flexibility on the part of the treatment program. Finally, three participants (15%) reported that being admitted to a place that they deemed inappropriate for their needs such as a general psychiatric ward or a ward primarily housing patients of a different age groups, had impacted their decision to dropout. Almost half of the group (45%) also mentioned unique setting-specific factors that did not fit with any other subtheme. Some examples were being unhappy with medications prescribed, dissatisfied with the therapeutic approach, and feeling pushed into making change.
Seven participants (35%) described factors that were personal or were related to the nature of the illness as having a major impact on their decision or desire to dropout of treatment. Four participants (25%) reported that they had grown tired of the treatment setting or it had become overwhelming. Three participants (15%) reported being homesick, while for three participants (15%) the fear of weight gain was a major contributing factor to dropout. Two other participants (10%) dropped out for other reasons, related to their illness.
The study aimed to explore the views of former AN patients on how they engage with treatment and how they define recovery. This study joins the small number that recruited from the community to consult former patients (e.g. de la Rie, Noordenbos, Donker, & van Furth, 2006, 2007, 2008), and as such, the sample represented a wide range of treatment approaches, modalities and settings. Mean scores for this sample indicated a motivation to change and a quality of life that is comparable to those observed in inpatient samples. In addition, about half of the participants scored in the clinical range for eating symptomatology. This is both context and caveat for the qualitative findings since this level of persistent symptomatology may be unusually high as other studies with similar follow-up periods report recovery rates of about 85% (Couturier & Lock, 2006b).
The hypothesis that a lack of a sense of volition in engaging in treatment may relate to poor motivation was supported by this study, which found evidence of a significant relationship between autonomy and stage of change. There was also a significant relationship between having more volition in treatment choice and psychopathology. These findings fit with a self-determination perspective on motivation, however, numbers are very small and so the finding should be interpreted with caution and replicated within a larger sample, ideally with measures that quantitatively assess these factors. Interestingly, the inclusion of close family members in the broad categorization of autonomy as distinct from external factors appears to be appropriate and suggests that family involvement may have less impact on an individual’s sense of autonomy or sense of volition in treatment choice. Recalling their treatment history, most participants reported that they had not been involved in decisions regarding treatment. Since all but one participant had child or adolescent onset this may have been a necessary step in seeking treatment in many cases. Individuals may be able to internalize this reasoning more readily than the recommendations of clinicians or constraints of health insurers.
Participants’ goals were diverse and most notable for their lack of alignment with those commonly proposed by treatment providers (e.g. weight targets, restoration of menses, etc.). A sizeable portion of participants articulated that they did not have any goals upon entering treatment that any desired achievements had been their parents’ and not their own. This suggests the extent of individuality in the perceived needs of the patient at commencement of treatment and at the potential clinical usefulness of investing in collaborative approaches to enhance engagement through listening to the patient’s concerns, and guiding them through the rational and potential usefulness of treatment. Twenty per cent of the sample was explicit in citing that to be listened to was what they had wanted most from treatment.
While the majority of participants incorporated at least one symptom-focused factor in their conception of recovery, the variation in responses was unexpected, with more than a third identifying unique (‘other’) factors. The relatively frequent miscellaneous responses when asked about goals and recovery suggest the individual nature of participants’ motivation to engage in the treatment process. This has implications for motivational enhancement interventions, as well as for the therapeutic process itself and suggests that individuals may be more engaged in treatment by being engaged in a discussion around aims that are tangible and meaningful to them.
It is perhaps noteworthy that few of the factors that previous reports have found to be important in the process of recovery—e.g. the development of a separate identity, emotional factors, desire to have children or the development of a supportive relationship (Federici & Kaplan, 2008; Keski-Rahkonen & Tozzi, 2005)—were important in participants’ conceptualizations of recovery. In many cases this would be expected. For example, while several studies have found that motivation to change is an important factor in the process of recovery (Federici & Kaplan, 2008), it does not necessarily follow that the state of recovery includes an element of motivation. The findings here suggest that how patients operationalize the end point of recovery is distinct from the factors they report to be important to the process. A reasonable question is to what extent participants’ operationalizations of recovery map onto definitions of outcomes commonly applied in the literature? Participants most commonly cited the resolution of eating-related anxiety and cognitive disturbance, as well as physiological symptoms. This appears to fit well with commonly applied definitions of remission and recovery—a combination of dietary restraint (measured by the EDE) and weight within a normal range (Couturier & Lock, 2006a,b) though there are many other factors that form the overall concept.
The extensive amount of ambivalence around the concept of recovery is a concern as it suggests that it may remain intransigent even after having experienced treatment where one would expect at least some thinking around this topic. It may be reflective of an ambivalence that is a result of the illness state itself since a similar proportion of participants’ EDE scores were indicative of severity warranting a diagnosis of AN. Nonetheless, this suggests that exploring this topic itself, in addition to motivational enhancement, may benefit patients in providing a personal and meaningful focus to strive toward in the long term. In particular, some participants may benefit from learning that people can and do recover.
The most commonly cited reason for dropping out or the desire to do so was a specific therapist or therapeutic rupture. This is similar to one previous study that implicated patient–therapist factors as being important (Clinton, 1996). Participants generally cited factors that were specific to the treatment setting, rather than personal factors. This is contrary to the literature on dropout that has tended to report patient characteristics such as demographic, clinical features or personality factors (Vandereycken & Vansteenkiste, 2009). Future studies should include variables related to therapeutic alliance and the treatment setting in future investigations of treatment engagement.
The strength of the study is also a limitation. Relying on participant’s recall and impression is subject to bias but provides rich data from which we may draw hypotheses for further study. Future research could minimize the impact of bias when assessing reasons for dropout, for example, by limiting the study to those that have dropped out in the past year. In general, the large amount of unique and individual responses that failed to form into coherent themes could be a reflection of a recovery state—whereby behaviour and experience look less similar than when in an illness state when variation tends to narrow into definable processes—the range of the morbid being narrower than the range of the non-morbid (Palmer, 2003). However, it may also reflect a true variation in the needs of individuals at the time of their treatment and implicates the need for individualized care, and collaborating with the patient to foster a sense of autonomy in standard clinical practice.