“I’ll do it on my own, but I’m never alone”[1 p. 402]: lived experience and practitioner perspectives of autonomy in longstanding and severe eating disorder treatment and recovery–a rapid evidence review

Abstract

Purpose

To examine autonomy within treatment and recovery from longstanding and severe eating disorders (EDs).

Background

The typically early age of onset, high incidence, and prolonged duration of EDs, has a high personal, relational, and financial burden for people who experience them. Current treatment practices rely on the exertion of external control and influence which has profound impacts on people living with EDs as well as the relationship and interactions between them and their treating professionals. Yet scant research focuses specifically on how people with EDs, especially longstanding and severe presentations, experience autonomy. Centring the lived experience voice is essential to deconstruct power imbalances and develop more effective recovery pathways.

Methods

A Rapid Evidence Review answered the research question, How do people with longstanding and severe eating disorders experience autonomy and recovery? Five electronic databases, the International Journal of Eating Disorders, and topic-specific grey literature from 2000–2021 were systematically searched. Twenty articles met the inclusion criteria. All findings and primary lived experience quotations relevant to the review topic were extracted, charted, and analysed. Reflexive Thematic Analysis (Braun and Clarke in Qual Psychol 9:3–26, 2021; Braun and Clarke in Qual Res Psychol 3:77–101, 2006;) was undertaken to inform the key findings and discussion.

Findings

The researchers identified four central themes that characterise the experience of autonomy in treatment and recovery for people with longstanding and severe EDs: Self versus ED, Self in Treatment, Self in Recovery, and Others in Relation to Self. Novel theoretical contributions to understanding longstanding and severe EDs include the ‘Lack of Autonomy’ Loop, ‘Relational Autonomy’ and the diversity of definitions of recovery.

Discussion

This review highlights the critical role of autonomy in the treatment and recovery of individuals with longstanding and severe EDs. It underscores the importance of addressing past trauma and integrating trauma-focused therapy within ED treatment. The findings advocate for collaborative, tailored, and autonomy-supportive approaches that empower individuals to break the cycle of disempowerment and promote sustained recovery.

Conclusion

The review highlights the need for conceptual and practice development to leverage desire-focused, relational approaches for effective healing and recovery.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-024-01164-z.

Plain language summary

We conducted a Rapid Evidence Review to examine the research question: How do people with longstanding and severe eating disorders experience autonomy and recovery? We systematically searched the International Journal of Eating Disorders, scholarly databases, and industry reports released between 2000 and 2021. Our findings redefine autonomy by exploring people’s relationship to themselves, their eating disorder, and others in their lives, and their experiences of agency. Our research combats the power dynamic that frequently favours researcher knowledge, illuminating shortcomings of current treatment models, providing alternate approaches for the field to consider, and offering directions for future research and policy change.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-024-01164-z.

Introduction

Conceptualisations of autonomy in longstanding and severe eating disorders (EDs) influence the development and implementation of treatment practices and programs, whether implicitly or explicitly. Current treatment practices often involve reduced choice and loss of individual control as medical stabilisation is prioritised [4]. Yet scant research specifically focuses on how people with EDs, especially longstanding and severe presentations, understand and experience autonomy including the potential intersections with recovery. The typically early age of onset, high incidence, and prolonged duration of EDs, has a high personal, relational, and financial burden for people who experience them [5, 6]. Learning from people with lived experience of EDs about the nature of autonomy and the role autonomy plays in recovery, will contribute to the evidence base informing effective approaches to early intervention and recovery, enhance person centred care, and has the potential to reduce the reliance on coercive treatment approaches. This paper reports on a review investigating how people with a longstanding and severe ED experience autonomy and recovery.

EDsare characterised by disturbances in thinking and behaviours around food, eating, exercise and body weight or shape, which become unhealthy preoccupations that significantly impair one’s quality of life [7]. This review includes the broad spectrum of ‘feeding and eating disorders’ as defined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), except for pica and rumination disorder, given their distinctive features [7]. Often accompanied by psychological distress and serious physical complications, EDs are associated with substantial comorbidities and increased health service use [8]. The lifetime prevalence of EDs is approximately 16.6% of the global population [9]. EDs have the highest mortality rate of any mental health disorder, with two large meta-analyses estimating rates of ED mortality five times higher than the general population—significantly higher than reported in other psychiatric disorders [10, 11]. The DSM-5 explicitly notes that high mortality rates for EDs are associated with physical complications and higher-than-average suicide risk [7].

Early and late adolescence are peak risk periods for the onset of an ED, though they may present at any age [12]. A history of trauma strongly links to clinically severe presentations [13–15]. There is also a significant concordance between EDs and neurodivergent conditions [16]. Other leading psycho-social-cultural contributors include thin idealisation, peer pressure, culturally entrenched body ideals and familial attitudes toward eating1 [17–21]. Most people with an ED delay seeking care for a decade or longer [5]. According to Hay [5], stigma, low health literacy, a preference for weight-loss management rather than ED treatment, shame, and poor access to evidence-based psychological therapies, contribute to delayed care-seeking. Therefore, people with an ED history require early and ongoing access to reduce the risk of future illness episodes and increase overall well-being [22, 23].

Recovery from an ED is not well defined [24, 25] with a lack of consensus on outcome measures [25]. Recovery can be a long-term process, and evidence suggests that recovery becomes less likely as illness duration increases [26]. A 2023 review showed rates of recovery for anorexia nervosa (AN) range between 18–60% with an average illness length between 6.5 and 14 years [27]. Globally, EDs contribute to an estimated 6.6 million disability-adjusted life years [28–30], with psychological comorbidity occurring in over 80% of people with EDs and over 90% of those with bulimia nervosa or binge-eating disorder [31, 32]. The typically early age of onset, high incidence, and prolonged duration, results in high financial and social burden of disease costs [6].

Longstanding and severe EDs affect up to 20% of people with EDs [23] and are characterised by lengthy illness duration combined with several unsuccessful treatment attempts [23, 27, 28]. While there is no consistent formal definition of longstanding and severe EDs [32–34], the research literature suggests a minimum of three years illness duration [29, 30, 35]. Longstanding and severe EDs, hereafter referred to as longstanding EDs, were our particular interest in this inquiry.

Weight restoration and reduction of disordered eating symptoms are often prioritised over individual choice and control in treatment protocols [24]. In contrast, people with lived experience of an ED advocate non-pathologising language and a focus on individual goals and circumstances [36, 37]. Some research advocates for the inclusion of a harm reduction approach when responding to people with EDs [38, 39]. However, there is continuing debate on “whether, when, and for whom” [40 p. 3] a harm reduction approach could be of benefit. Whilst chances of a "full recovery" are shown to decrease with time, lived experience informed understandings of recovery emphasise ‘journeys’ towards well-being. Prioritising hope and meaning are important in maintaining engagement with recovery—whatever that might mean for a person—not merely the absence of symptomology [40–43]. For example, Vos et al.’s [41] meta-analysis showed people with lived experience of EDs consider remission of their symptomology important to recovery, while the underlying dimensions of psychological well-being (autonomy, positive relationships, self-acceptance, self-resilience, personal growth, and improved ED cognitions/behaviours) were fundamental. Holistic perspectives on recovery contrast with practitioners’ and family members’ tendency to focus on symptom reduction [44], generating conflicting treatment measures and recovery goals. Recovery-focused work aims to remove opportunity barriers, create a collaborative culture of healing, and support people to reclaim their identity [42, 45].

Self and identity are multifaceted intersecting concepts. Self-related concepts are implicated in ED theories and research, including in the development and maintenance of EDs and recovery [46]. Yet these concepts are used interchangeably and not clearly defined [47]. Recent reviews of the empirical evidence offer important findings on the role and experience of self in EDs. Kiely and colleagues’ [4] qualitative meta-analysis found that severe and enduring AN included experiences of a ‘self-in-relation’ to others; including a ‘disappearing self’, loss of identity, and a ‘global impoverishment of self’, as a response to both the experience of the ED and treatment approaches that fail to offer maximal choice and autonomy. While Croce and colleagues’ [47] recent review found that disruptions to identity may be central in both the maintenance of AN and recovery. This review preferences the term self. Self is defined in the social sciences as a person’s sense of having their own distinct and relatively consistent identity, developed through maturation and socialisation [48]. Self includes one’s sense of, or awareness of, one’s identity, and thus aligns with the lived experience of one’s self in, and journey through, the ED.

In scholarly and mainstream discourse, definitions of the construct of autonomy vary, ranging from individual constructions of autonomy as recognising the self as having agency to act and the ability to make decisions for oneself [49] to autonomy as a relational experience [50]. Directing attention to partnerships between practitioners and people with lived experiences of EDs [4] suggests that a more nuanced understanding of autonomy could help to (re)shape service delivery and improve patient outcomes [51]. The terms agency, choice, and self-efficacy are used interchangeably in the literature. This review adopts the definition of autonomy in the ED landscape used by the Butterfly Foundation, “Autonomy is about identifying with the importance of change and taking ownership over the change process. Building autonomy means fostering intrinsic motivation for change and moving away from externally imposed controls” [52 p. 35].

Autonomy and trauma-informed practice share similar principles. Trauma-informed practice recognises the importance of designing and delivering health services informed by an understanding of the impacts of trauma with two core principles being “empowerment, voice and choice” [53 p. 337]. Structural factors such as socioeconomic and political contexts limit individual autonomy [53] by influencing the opportunities presented and the range of choices available [54]. To properly understand autonomy and discern the efficacy of treatment (or journeys toward recovery) more broadly, we need to look at long-term outcomes in combination with lived experience perspectives. However, this literature is scant, with a pronounced gap between research and practice. Although collaborative person-centred approaches are increasingly mentioned in policy; for instance, Vision 2030 Blueprint for Mental Health and Suicide Prevention [55], commentators repeatedly argue that they have not transitioned into practice [55–57]. As a result, services remain risk-focused and struggle to adopt person-centred and strengths-based approaches [45]. These themes are addressed in the Australian National Research and Translation Strategy for EDs which was developed in consultation with prominent researchers, professionals, and people with lived experience, aiming to translate knowledge into practice [57].

In line with this strategy and based on the different conceptualisations of autonomy and recovery in EDs outlined in this introduction, there is a need to map the current research evidence on the nature and role of autonomy in the recovery journey from longstanding EDs. Of particular interest is to centre the voices of people with lived experience alongside the research evidence. This review responds to this need by addressing the following research aims and question.

Research aims

The purpose of this review was to systematically collate and analyse extant research on autonomy of people with lived experience of longstanding and severe eating disorders and what this means for their treatment and recovery.

Research question

How do people with longstanding and severe eating disorders experience autonomy and recovery?

Methods

Team-based reflexivity [58] was employed throughout this study in conjunction with reflexive thematic analysis [2]. The authors actively committed to reflexive processes including informal and formal team meetings, individual reflection, and reporting back to the group. They engaged in frequent discussions about positionality and what their analysis revealed about the relationship between the researchers and the data. They teased out discrepancies in interpretation of the data, engaged in discussion and developed shared understandings. This iterative process sharpened insights and enriched the depth of the thematic findings. The first author is a fierce advocate for people experiencing longstanding EDs and for harm reductive, trauma-informed practice. They exist in the dual position of practitioner and consumer. The second and third authors are social workers and tertiary educators, with expertise in human rights from below, and working therapeutically alongside survivors of violence, abuse and neglect respectively. Similar to Rankl and colleagues (58 p.1367) a team-based reflexivity model enhanced this review by facilitating the examination of our assumptions, research processes, and analytical outcomes [58].

The team adopted a Rapid Evidence Review (RER, also known as Rapid Evidence Assessment or REA) method for this study. The RER design applies robust standards to the systematic collection, appraisal, and analysis of data [59, 60]. RERs are widely used in health research as an effective method to summarise evidence and to inform future research, policy development, and practice approaches [60].

The method was modified to incorporate extraction and analysis of all relevant primary (lived experience) quotations from the included studies. This modification is both recognition of the knowledge and expertise held by people with lived experience of EDs and an attempt to redress the marginalising of lived experience voices in reviews which influences directions in policy and practice [53, 61]. This is the approach that was adopted in two recent systematic reviews centring Indigenous voices and methodologies in secondary research [62, 63] and recently in a meta-analysis of severe and enduring AN [4]. Analysis involved a synthesis of findings from primary quotations and research findings.

Search strategy

A rigorous search of allied health literature was conducted. Key search terms were applied across the electronic platforms Proquest, EBSCOhost, Web of Science, OVID and Informit. A hand-search of the International Journal of Eating Disorders and citation analyses of all included articles was conducted to reduce the likelihood of publication bias. Grey literature searches were conducted in ProQuest dissertations and theses, Emerald Insight, ED-specific data repositories (for example, Butterfly Foundation), and Google.

The first author completed the original research for this review during their Honour’s studies. Developing the search strategy in consultation with authors two and three and the institution’s librarian enhanced the reliability and validity of the study. Using team-based reflexivity all three authors met to refine the search strategy which author one then implemented in consultation with authors two and three. An online template facilitated reflexive discussions and decision making during the literature search, screening, and data collation. Detailed search information, including specific databases, search strings, filters and results are published on searchrxiv [64–69], a repository where researchers can report and share search strings and strategies [70]. Search terms are shown in Table 1.

Table 1.

Search terms

Key term	Eating disorder*a	Autonom*b	Autonomy & treatment
Synonyms and related terms	Disordered eating Search terms not expanded to include ED diagnoses, as recommended by librarian	Choice OR Agency OR Self-efficacy OR Self efficacy	Consent OR Voluntary OR Self-endorse OR Self-initiate OR Self-admi* OR Patient-Controlled OR Collaborat* OR Participat* OR Client feedback OR Client perspective OR Lived experience

^{a, b}We observed in the research design phase that Choice, Agency, and Self-Efficacy are used interchangeably with Autonomy. Freedom and Power may also be fruitful inclusions for future researchers

* placed next to a search term were used to instruct the database to search for all variations of a particular term. For example, disorder* will search for disorder, disorders, disordered

Eligibility criteria

The study employed methodologic triangulation by including qualitative, quantitative, mixed methods, and grey literature, and data triangulation by including for analysis human adults with a diagnosed or self-reported ED or in recovery from an ED, or practitioners working with people experiencing EDs [71]. The consistent application of the tabulated eligibility criteria, as detailed in Table 2, ensured a rigorous and systematic selection of articles for the review, thereby enhancing the reliability and validity of the findings. While some contend that the interpretive nature of RER can introduce bias [72], we addressed this concern by adopting an egalitarian, reflexive approach to data analysis and triangulating these categories.

Table 2.

Eligibility criteria

Criteria	Inclusion	Exclusion
Language	Available in English	All languages other than English
Location	Available online through the UNSW library	Not available online via UNSW
Date	Published January 2000- April 2021	Papers published before 2000
Status and Design	Empirical peer reviewed articles OR empirical grey literature (including doctoral dissertations) Qualitative, quantitative, and mixed methods research	Systematic reviews Unpublished dissertations Opinion pieces
Population	Human adults (18 +) with a diagnosed or self-reported ED or in recovery from an ED as defined by the broad spectrum of feeding and eating disorders in the DSM-5, excluding pica and rumination disorder Or Practitioners working with people experiencing EDs	Studies focusing on humans aged 0–17 without data separated by age group Studies focusing on participants with pica or rumination disorder as defined in the DSM-5
Topic / Content	Qualitative or mixed-methods research studies that report findings on the experiences of an ED with autonomy as a primary focus or theme of the study Quantitative research that reports on the relationship between autonomy and ED treatment, that used a standardised outcome measure	Qualitative or mixed-method studies that had a broader focus than EDs and did not have autonomy as a primary focus or theme Quantitative studies that did not use an outcome measure

Although RER methodologies often limit inclusion of grey literature, the prevalence of relevant publications by government agencies and large non-government organisations justified the importance of including grey literature in the eligibility criteria. This aligns with the approach taken by the InsideOut Institute, the Australian institute for research, translation, and clinical excellence in EDs [57, 73].

Study selection

The search of literature published between 2000–2021 identified 1563 studies. The preliminary scan and removal of duplicates left 445 studies. Texts were then screened by title and abstract. When it was unclear if an article met the inclusion criteria, at least two authors independently assessed against the eligibility criteria and discussed. The full-text review included 24 studies. A total of 20 empirical peer-reviewed studies met the eligibility criteria for inclusion in this review. Of the 98 grey literature articles found, none met all the eligibility criteria. The search, identification, and selection process is illustrated by Fig. 1 ‘Preferred Reporting Items for Systematic Reviews and Meta-Analyses’ (PRISMA) [74].

Fig. 1.

PRISMA flowchart

Quality appraisal procedure

A quality assessment was performed using the 2018 Mixed Methods Appraisal Tool (MMAT) [75]. The MMAT is designed to facilitate inclusion of a diversity of research designs, with methodologies grouped into five categories: qualitative research, randomised controlled trials, non-randomised studies, quantitative descriptive studies, and mixed-methods studies [75]. Use of this tool supported equivalence in appraising articles with differing research designs. Sixteen of the twenty articles assessed met all specified MMAT criteria, and the remaining four articles met most. Quality appraisal results, including MMAT outcomes and attrition rates, are presented in Appendix 1.

Data extraction and analysis

Data extraction and analysis followed Braun and Clarke’s [2, 3] six-step framework of Reflexive Thematic Analysis (TA). This method facilitated analysis of the studies as a whole, and of primary quotations. The first author took the lead role in data extraction and analysis, supplemented by weekly team-based reflexive meetings with authors two and three. Familiarising, conducted by the first author, involved reading and re-reading each article, documenting initial ideas and representations of autonomy, treatment and recovery, and extracting key characteristics of the studies to chart them in Excel (step one). Coding was performed manually and systematically identifying features and repeated patterns. Data were then organised electronically using NVivo 12 (step two). Initial themes were generated over a period of six months by iteratively reviewing the data and exploring meaning within and across the dataset (step three). Coded data and themes were reported to the group and reviewed by all authors for consistency and coherence, to accurately represent the data (step four). Refining, defining, and naming themes was supported by journaling, mind-mapping, critical reflection, and reflexive discussions (step five). Finally, themes were collated and refined for publication (step six).

These steps were not performed linearly but were reflexive, with back and forward movement as needed [3]. Consistent with Rankin and colleagues [76], this process comprised a descriptive phase and an interpretive phase [2], facilitating the differentiation of themes and sub-themes while also appreciating the interconnection within and across themes.

Limitations

By modifying the RER method to include the extraction, analysis, and synthesis of primary quotations from the included studies the authors attempted to redress the absence of the lived experience knowledge from secondary research. However, one of the risks of using secondary data is that lived experience quotes are already mediated by the original authors [63]. As this is an exploratory approach with few precedents, it calls for further research and evaluation.

Future intersectional research will prioritise sourcing other than English language studies. Noting that minority populations face an elevated risk of developing longstanding EDs [33], our inability to examine the data using an intersectional lens and the exclusion of potentially relevant empirical research published in diverse languages is unsatisfactory [60]. Included studies also unfortunately contain insufficient detail about diversity markers.2 Furthermore, concepts of’freedom’ and’power’ have the potential to elevate this analysis and should be considered as future search terms for research in this area.

Ethical considerations

Engaging critical social theory as our ideological foundation fostered critical self-awareness, keeping ethical considerations at the forefront of the research. Our commitment to procedural ethics was evident through our use of team-based reflexivity which emphasised honesty, care, respect and purposeful critique. Atem and Higgins [77] describe reflexive dialogue as a commitment to disrupting established power dynamics and bringing forward alternative, often devalued, epistemologies. Accordingly, our systematic preparation, reflection-on-action, reflection-in-action, and thorough documentation of processes enhanced the validity of our findings [78].

To emphasise primary quotes from participants with EDs as active agents with dignity and unique perspectives, we used pseudonyms to report our findings rather than codes. We adopted pseudonyms used in the original studies [79], and where pseudonyms were not provided, we assigned pseudonyms that mirrored the gender of the participants. In mixed-gender cohorts where gender was not specified, we allocated gender-expansive pseudonyms. Selected pseudonyms reflected the locations of the studies, with popular names selected for use according to available information about the country from which the data originated.

Descriptive summary of included studies

Of the 20 articles included in this review, ten were qualitative (semi-structured, narrative, exploratory, and ethnographic interviews), eight were quantitative (observational, cross-sectional, comparison, and preliminary outcome designs), and two were mixed-method designs (concept mapping and generic). Author/date details, citation numbers, and study locations are provided in Table 3 alongside demographic data, methodological features, and the representation of the higher-order themes across the studies, highlighting their prevalence and significance in the included literature.

Table 3.

Study characteristics

Author , country	Participant characteristics	Duration of illness	Study design	Outcome measures	Theme
Darcy et al. [80] Canada	20 women, aged 19–52, with a lifetime history of AN (85% Caucasian, 15% biracial)	Lifetime history	Mixed Method: Inductive thematic analysis, semi-structured interview and self-report measures	EDE-Q, ANSOCQ, EDQOL	Themes 1, 2, 3, 4
Dawson et al. [81], Australia	8 women, aged 31–64, who have been in AN recovery for 5 + years (Ethnicity NR)	Mean 15.5 years	Qualitative Narrative Inquiry: Inductive paradigmatic analysis, narrative interview and self-report measures	EDE and BMI	Themes 1, 2, 3, 4
Espíndola et al. [82], Brazil	15 women, aged 22–32, who have been in AN recovery for 5 + years (Ethnicity NR)	NR	Qualitative Ethnographic interviewing: Deductive content analysis, ethnographic interview	N/A	Themes 1, 2, 3, 4
Geller et al. [83], Canada	198 inpatients (97% female, 2% transgender, 1% male), mean age 32.4, with AN (72), BN (49) and OSFED (25) (89% Caucasian, 11% NR)	Mean 15.93 years	Quantitative study: Statistical analysis, Diagnostic Clinical Interview and self-report measures	BSI, RMQ, EDI-3, EC, EDQLS, CSQ, NNRS and BMI	Themes 2, 3, 4
Gorse et al. [84], France	63 female inpatients, mean age 27.6, with severe AN (Ethnicity NR)	NR	Qualitative Narrative interview study: Inductive content analysis, narrative interview	BMI	Themes 1 and 3
Lindgren et al. [85], Sweden	5 women, aged 23–26, self-report as recovered from BN for 2 + years (Ethnicity NR)	NR	Qualitative Narrative interview study: Content analysis, narrative interview	N/A	Themes 1, 2, 3, 4
Mansour et al. [86], Canada	155 women, mean age 27.3, with bulimia spectrum ED and part of a specialised ED program (Ethnicity NR)	NR	Quantitative study: Statistical analysis, self-report measures	ACMTQ, EDE-Q, BASIS-32, CES-D	Theme 2
Nordbø et al. [87], Norway	18 women, aged 18–39, with AN treated in 1 of 3 clinical institutions in Norway in the last 2 years (Ethnicity NR)	Mean 7.6 years	Qualitative Exploratory study: Inductive semantic and contextual analysis, in-depth-interview	N/A	Themes 1, 2, 3, 4
Olofsson et al. [88], Norway	11 Norwegian women, aged 28–59, with BN (5) , OSFED (subthreshold AN) (5), and BED (1) , with a history of childhood trauma	Mean 22.4 years	Qualitative Exploratory study: Deductive Interpretative Phenomenological Analysis, Diagnostic assessment, interview, and self-report measures	Diagnostic assessment: EDE-16, MINI, SCID-II Self-report: EDE-Q, CTQ-SF	Themes 1, 2, 3, 4
Paquin Hodge et al. [89], Canada	82 female inpatients, mean age 26.5, with AN (+ subtypes) (Ethnicity NR)	NR	Quantitative Observational Comparison study: Statistical analysis, self-report measures	EDE-Q, BASIS-32, and BMI	Theme 2
Patching et al. [90], Australia	20 women, aged 24–51, who self-identify as recovered from AN (6) , BN (2) , or both AN & BN (12) (Ethnicity NR)	Range 4–17 years	Qualitative Inquiry: Life-history approach: Thematic analysis, life-history interview	N/A	Themes 1, 2, 3, 4
Sansfaçon et al. [91], Canada	770 outpatients (97% female), mean age 27.5, in a specialised ED Program between 2005 and 2016, with AN (216), BN (282), and OSFED (272) (Ethnicity NR)	NR	Quantitative Naturalistic observation: Statistical analysis, self-report measures	EDE-Q, ACMTQ	Themes 1, 2, 3
Smart et al. [92], U.S.A	11 female and 1 male ED practitioner, aged 33–56, with an average of 11 years in practice (Asian American (7) , European American (4) , Latina (1)	N/A	Qualitative study: Combined grounded theory and consensual qualitative analysis, semi-structured interview	N/A	Themes 1, 2, 3, 4
Smith et al. [93], Scotland	21 female inpatients, aged 18–41, with AN (Ethnicity NR)	Mean 6.3 years	Qualitative study: Inductive thematic analysis, semi-structured interview	N/A	Themes 1, 2, 3, 4
Steiger et al. [94], Canada	97 female outpatients, mean age 28.6, with AN (30) , BN (34) , OSFED (30) , and ARFID (1) . (Ethnicity NR)	NR	Quantitative: Statistical analysis, self-report measures	ACMTQ, EDE-Q, HCCQ	Themes 1, 2, 4
Strand et al. [1], Sweden	15 women and 1 man, aged 18–56, mean age 31 years, with AN and self-admission contracts with SCA (Ethnicity NR)	Mean 15 years	Qualitative study: Inductive content analysis, semi-structured interview	N/A	Themes 1, 2, 3, 4
Strand et al. [95], Sweden	29 adults with AN and self-admission contracts with SCA and two comparison groups (113 [low-utilisation] and 27 [moderate-utilisation] participants attended ‘treatment as usual’ specialised ED services in Sweden. The mean age across groups is 29.2, and 95.9% are female (Ethnicity NR)	Mean 13.4 years	Quantitative Comparison cohort study: Statistical analysis, self-report measures	EDE-Q, CIA, GAF, and BMI	Themes 1, 2, 3
Tomba et al. [96], Italy	234 female and 11 male outpatients, mean age 27.8, with AN (57) , BN (105), and BED (83) , and 60 matched controls (Ethnicity NR)	NR	Quantitative Cross-sectional Comparison study: Statistical analysis, Diagnostic Clinical Interview and self-report measures	Structured Clinical Interview for DSM-IV Self-report: EAT, GHQ-30, PWB, BSS	Themes 1, 2, 3, 4
Walker et al. [97], International	50 multidisciplinary ED practitioners, with an average of 14 years in practice (Gender NR, Ethnicity NR)	N/A	Mixed Method; Concept Mapping study: Statistical analysis. Group brainstorming, sorting, and frequency/impact ratings	N/A	Themes 1 and 2
Williams et al. [98], Canada	31 outpatients, mean age 31, with AN (15) , BN (3) , and EDNOS (13) in a Community Outreach Partnership Program (Gender NR, Ethnicity NR)	Mean 15.23 years	Quantitative preliminary outcome research: Statistical analysis, self-report measures	BMI, BSI, BHS, SAWBS, HIQ, QOLI	Themes 1, 2, 3, 4

Diagnoses: AN, anorexia nervosa; BN, bulimia nervosa; BED, binge eating disorder; BSED, bulimia-spectrum eating disorders; OSFED, other specified feeding and eating disorder; ARFID, avoidant/restrictive food intake disorder; EDNOS, eating disorder not otherwise specified

Other: N/A, not available; NR, not reported; DSM-IV, diagnostic and statistical manual of mental disorders, fourth edition; SCA, a public specialist ED clinic in Stockholm Sweden

Outcome measures—psychiatric symptoms: BASIS-32, Behavior and Symptom Identification Scale; BHS, Beck Hopelessness Scale; BSI, Brief Symptom Inventory; CES-D, Center for Epidemiological Studies—Depression Scale; CIA, Clinical Impairment Assessment Questionnaire; PSS-SR, PTSD Symptom Scale–Self-Report version; CTQ-SF, Childhood Trauma Questionnaire – Short Form; SCID-15, Structured Clinical Interview for DSM-5

Outcome measures—eating disorder symptoms: BMI, body mass Index; ANSOCQ, anorexia nervosa stages of change questionnaire; BSS, body satisfaction scale; EAT, eating attitude test; EDE-Q, eating disorders evaluation questionnaire; EDQOL, eating disorders quality of life instrument; EC, experience of collaboration; EDI-2, eating disorder inventory-version 2; EDI-3, eating disorder inventory-version 3; EDQLS, eating disorder quality of life scale; HIQ, health inventory questionnaire; SAWBS, shape and weight based self-esteem inventory

Outcome measures—autonomy and quality of life: ACMTQ, autonomous and controlled motivations for treatment questionnaire; GHQ-30, general health questionnaire-30 items; GAF, global assessment of functioning; HCCQ, health care climate questionnaire; PWB, psychological well being; QOLI, quality of life inventory; RMQ, readiness and motivation questionnaire

Studies were conducted in ten countries, including Canada (seven), Sweden (three), Norway (two), Australia (two), and one each from Brazil, France, Italy, Scotland, USA, and international. Participants in 18 of the articles were individuals with lived experience of an ED, where the analysis focused on treatment (n = 13) or recovery (n = 5). The two remaining articles researched practitioners who worked with clients experiencing EDs. Half of the 18 articles researching participants with lived experience of EDs reported mean illness duration or an illness range consistent with longstanding EDs. Mean illness duration ranged from 6.3 years [93] to a mean of 22.4 years [88]. Patching and colleagues [90] reported an illness range between 4–17 years. This shows that the experience of participants with lived experience of longstanding EDs is well represented in the dataset.

The combined dataset contained the voices of 2066 participants3; 1944 participants had lived experience of an ED, 62 were practitioners, and 60 were controls. Participants were 94.8% female. There were four male and four transgender participants. The gender of 100 participants was not stated. The age range of participants was 18–64 years. These characteristics of the included studies are summarised in Table 3.

Definition of autonomy

Despite selecting the articles for their focus on autonomy, a concerning lack of definitional consistency emerged, with only one article providing a standalone definition and another defining it in the context of familial autonomy. A notable majority (seven) utilised it within the framework of self-determination theory (SDT), two mentioned SDT without defining it, and eight articles were entirely devoid of definition.

Findings

Reflexive thematic analysis [2, 3] of the data produced four connected themes contributing to understanding autonomy of people with lived experience of longstanding EDs and what this means for their treatment and recovery. The four themes, Self versus ED; Self in Treatment; Self in Recovery; and Self in Relation to Others are represented in Table 4, alongside thirteen subthemes. Following a description of each theme, the relational connections across these four themes are presented as a conceptual finding; the Lack of Autonomy Loop. Reflexive analysis of the four themes show being stuck in a lack of autonomy loop acts as a barrier to treatment and recovery while paradoxically asking for help, and freely surrendering control may facilitate increased autonomy and self-agency, facilitating treatment and recovery. Importantly, the findings identify features of autonomy and relationships that support treatment and recovery through fostering increased safety, trust, and collaboration.

Table 4.

Breakdown of review results by theme

Themes	Subthemes	Number of articles the sub-theme occurs
Self versus eating disorder	ED in Charge Léa: “…this wretched thing has taken over my life…” [84 p. 4]	9
	Feeling stuck in the ED Olivia: “…I had dozens of hospital admissions with no success…” [81 p. 499]	10
	Self is called to action Ruby: “…I made a conscious decision to get better…” [81 p. 501]	6
Self in treatment	Coercion and control Elsie: “…everyone is deciding what is happening and nobody is asking me” [93 p. 21]	6
	Autonomy support Kalle: “…it’s possible to do it by your own free will…” [1 p. 404]	16
	Awareness of Self and Disorder Susan: “…the main reason for my [ED] is trauma, because the [ED] is only really a symptom…” [88 p. 61]	5
Self in recovery	Feeling stronger than the ED Bonnie: “…it is me in charge here” [93 p. 21]	14
	(Re)defining recovery Maisie: “I want to get better, but…it is still difficult. I still struggle” [93 p. 23]	7
	Letting go of the ED and moving forward	16
Others in relation to self	Formation of healthy attachments	6
	Experiencing understanding and acceptance from others Mia: “[My therapist] had had an [ED] herself, and she was by far the most important therapist I had ever seen…She understood it and she instilled hope that I could get over it” [81 p. 501]	3
	Encouragement and companionship Misha: “…I’ll do it on my own but I’m never alone” [1 p. 402]	7
	Positive results of recovery-promoting relationships	7

Access to qualitative, quantitative, and mixed-methods data sets was a strength of the study. The first author developed theme names to succinctly summarise the themes, while lived experience participant quotes extracted from the selected studies, were provided to exemplify and describe each theme. As themes three and four were derived mainly from a thematic synthesis of the research findings of the selected studies, these themes used less extracted lived experience quotes. Each theme begins with a description of the theme and discussion of the findings of qualitative and lived experience data and concludes with a summary of the statistical support provided by the quantitative studies included in that theme.

Self versus eating disorder

Self versus ED appeared across 17 of the included articles [1, 80–82, 84, 85, 87, 88, 90–98] (ten qualitative, five quantitative, and two mixed-methods) and personifies a struggle between the Self and the ED. This theme portrays the internal conflict between two opposing forces: the individual's yearning for a fulfilling life and the EDs relentless pursuit of control.

ED in charge appeared across nine studies [1, 80–82, 84, 85, 87, 90, 93] and refers to feeling disconnected from the self and life; imprisoned by the ED. It includes findings where the ED contributes to experiences of suicide, isolation, and distress.

…disgust with myself, endless suffering…it is awful when you cannot stand yourself, who you are, the struggles, the continual losing battles to binges (Inès) [84 p. 3].

I was waiting for the right moment when I could commit suicide...I knew for sure that I could not go on living if things were to continue to be like this (Linnéa) [85 p. 864].

…this wretched thing has taken over my life, destroying me, cutting me off from the world, from my family and from life (Léa) [84 p. 4].

Feeling stuck in the ED appeared across ten studies [80, 81, 83, 84, 86, 89, 91, 92, 94, 97], describing participants’ ambivalence towards change and coinciding fluctuations in motivation for recovery, their relationship with the illness, and lengthy treatments where no progress is made. Other people may assume control and arrange hospitalisation when they see the person suffering; however, this leads to increased powerlessness, isolation, and a further lack of connection.

…I had dozens of hospital admissions with no success…Treatment felt like banging my head against a brick wall (Olivia) [81 p. 499].

…I want to be able to be happy, only I’m afraid for myself because I’m so ambivalent. I have the constant feeling that two beings are constantly struggling inside me: Marie and the anorexic (Marie) [84 p. 4].

Self is called to action appeared across six studies [80, 81, 83, 84, 86, 94] and refers to internal or external processes that trigger change, with a different and more autonomous quality of motivation.

That day a man died from [ED] in the hospital. It became very clear to me then that you could die from [ED]...It hit me that I wasn’t normal, and I was so sick of battling my brain. I had to [decide]: Am I going to let myself die, or am I going to beat this and make a go of my life? I made a conscious decision to get better. That day was a turning point (Ruby) [81 p. 501].

As Dawson and colleagues [80] explain the illness can change from being perceived as something of potential value to something destructive and futile due to the ED’s severity and a growing awareness of the long-term impact. Life with ED can become untenable, with change the only option:

It was something I had to make a choice to do, and I made that choice because I didn’t want to be a prisoner anymore (Emily) [81 p. 501].

Findings from five quantitative [91, 94–96, 98] and two mixed-methods [80, 97] studies are consistent with these qualitative findings, highlighting persistent barriers to change even when individuals are ready to embark on recovery. Tomba and colleagues [96] found significant differences in psychological well-being between 245 outpatients and 60 community controls, underscoring the distress permeating across the lives of those with EDs. Strand and colleagues [1] reported that while self-admission contracts reduced hospitalisation by 50% at 12-month follow-up, they were not associated with improvements in ED morbidity or BMI.

Self in treatment

Self in Treatment emerged as a major theme across 19 articles [1, 80–83, 85–98] (nine qualitative, eight quantitative, two mixed-methods) and refers to the experiences and impacts of treatment delivery and how people living with an ED experience the process of treatment.

Coercion and control appeared across six studies [80, 81, 88, 93, 97, 98]. This encompasses a lack of collaboration and involvement in treatment decisions, the importance of relationships with the treatment team and promoting client agency. In their mixed-methods study, Darcy and colleagues [80] report that the majority of participants were not involved in any treatment-related decisions.

I feel I am out of the loop and I find that it makes me feel really anxious…everyone is deciding what is happening and nobody is asking me (Elsie) [93 p. 21].

Although coercive practices and therapeutic alliances were not the focus of this study, their pervasive presence in the autonomy literature signals the systemic and widespread nature of such practices in ED treatment. The analysis revealed a spectrum of experiences, from feelings of exclusion and powerlessness to encounters with coercive and punitive treatment processes. These experiences often left individuals clinging to their EDs as a last resort for maintaining a sense of control in an environment where their autonomy was repeatedly undermined:

But the punitive thing…that just fuels the disease, that does nothing helpful (Madison) [80 p. 267].

If the clinicians told me what to do I’d think, ‘Well screw you. I’m not doing that.’...It became more about the ‘us’ and ‘them’ (Amelia) [81 p. 499].

Autonomy Support appeared across 16 articles [1, 80–82, 85–91, 93–95, 97, 98] and refers to delivering treatment in a manner that is informative and free of threats, coercion and/or force to achieve compliance [91]. Being heard, feeling like your experience matters and someone is listening, promotes a sense of collaboration and autonomy:

The key factor in motivating me to change my outlook and increase my propensity to challenge anxious and depressive traits has been the relinquished agenda of forced treatment in exchange for a non-threatening and comprehensive, interactive treatment plan..... that seeks nothing more than to provide support, one that meets me where I’m at (Julie) [98 p. 94].

Freely choosing to engage in treatment, being supported to do so, and having individualised treatment options available and accessible like self-admission, can significantly enhance a person’s sense of autonomy:

Even if I don’t do everything with joy when I’m [at the ward], I know that I’m doing it because I’ve chosen it myself. And then maybe I can show others that have even greater difficulties that it’s possible to do it by your own free will…[sic] (Kalle) [1 p. 404].

…it affects my everyday life 24/7, just knowing that I have the opportunity [to self-admit if needed] (Misha) [1 p. 402].

Partnering to reduce harm refers to people with EDs collaborating with care providers to reduce the negative impact of the disorder and improve their quality of life [51]. One example from this review is the Community Outreach Partnership Program (COPP) discussed by Williams and colleagues [98]. In this approach, the individual is at the centre of all decision-making and the team works alongside them to achieve their personally identified goals, which may or may not align with traditional treatment objectives [98]. A second example is Strand and colleagues’ [1] study of freedom to self-admit which demonstrably reduces the length of time as an inpatient and prevents deterioration. This harm-reductive approach affords access to treatment before deterioration, and participants reported that simply knowing this option was available provided an “enormous sense of security” [1 p. 402]. Two further studies [83, 89] showed the benefits of non-coercive treatment approaches and demonstrated movement towards autonomy-supportive treatment.

Awareness of Self and Disorder appeared across five studies [84, 88, 92, 93, 98] and refers to increased insight into the disorder's underlying cause and function. For many, EDs develop as a response to traumatic experiences:

I started to become aware that the [ED] wasn’t a choice— it was a reaction. As a teenage girl, the only thing I could control was my body because I had no power. Exploring the issues behind the [ED] was helpful for me—knowing where my need to be perfect came from and realizing that I achieved perfection through eating (Jessica) [81 p. 500].

A memory that I had completely blocked as a child was uncovered. During…therapy, I suddenly remembered that I had been sexually abused by my brother. This revelation was very distressing and intense for me. My need for control made a lot more sense. It seemed clearer to me why I wanted to torture myself and why I felt so controlled and angry as a child (Chloe) [81 p. 500].

Addressing traumatic experiences in treatment is essential in the recovery process:

I feel that the main reason for my [ED] is trauma, because the [ED] is only really a symptom. And if I had started working with trauma…I think it would’ve been easier to work with the [ED], and not the other way around, the way they did it at the ward (Susan) [88 p. 61].

Recovery is supported when participants begin to externalise the disorder. Externalising involves separating the ED from the self [98]:

I started to wonder that maybe a disease was controlling me and that, perhaps, I wasn’t to blame for it (Charlotte) [81 p. 500].

Autonomy-support highlights the important role of care providers in helping people to externalise their disorder, shifting our understanding of autonomy from individualistic to relational. In autonomy-supportive treatments, the approach is collaborative, inclusive, and the client is in control [88]. According to Geller and colleagues [83] mandatory treatment components and autonomy-supportive treatment can co-exist.

Quantitative findings [91, 94–96, 98] support that choice and autonomy play a significant positive role in motivation and treatment outcomes. Inpatient treatment prioritising autonomy may yield equivalent [89] or better [91] treatment outcomes than longer, more controlling and costlier approaches. Geller and colleagues [83] found that a collaborative environment was statistically associated with improved psychological functioning, motivation, and reduced ED symptomology in 146 inpatients. Analysis demonstrated a strong positive correlation between people with EDs experience of collaborative treatment and five of the six treatment domains central to collaborative practice (providing rationale, choices, consistency, fairness and overall satisfaction) [83]. Mansour and colleagues’ [86] study of 155 women with bulimia spectrum ED found higher pre-treatment autonomous motivation reduces ED severity and improves outcomes post-treatment. However, Walker and colleagues [97] demonstrate that the largest challenges faced by practitioners include limited access to appropriate evidence-based care and balancing client autonomy with duty of care concerns.

Self in recovery

Self in Recovery appeared across 18 studies [1, 80–88, 90–96, 98] (ten qualitative, seven quantitative, one mixed-method), and captures the nuanced journey of recovery.

Feeling stronger than the ED appeared across 14 of the studies [1, 81–85, 87, 88, 90–93, 96, 98] and is about seeking and accepting help, feeling able, and strengthening the self. According to Lindgren and colleagues [85], wanting help is the most important factor in distinguishing between attending professional support unwillingly versus engaging in recovery through empowerment and self-agency, and accepting help.

Feeling able refers to having faith in one’s capacity for change and belief in one's own ability to do the hard work of recovery, with a different and more autonomous quality of motivation. Dawson and colleagues name this shift as “making a conscious decision to recover" [81 p. 501].

Strengthening the self involves separating the self from ED and (re)finding one’s own identity outside the illness [80, 89]. As Smith and colleagues [92] explain, this leads to “increased feelings of self-awareness and self-confidence” [93 p. 23]. Similarly, Gorse and colleagues [84] delineate the movement from an ED-defined self to a self capable of managing the illness and an awareness of the changes occurring to the disease-based identity.

These shifts indicate that the self is becoming stronger than the ED, and, in turn, contribute to increased autonomy in treatment and recovery:

…as you actually manage without using [ED] behaviours it is so empowering…I can take a normalised control over my well-being and it is me in charge here (Bonnie) [93 p. 21].

(Re)defining recovery appeared across seven studies [80, 82, 85, 88, 90, 93] and includes critiques of current definitions, and alternate definitions of recovery.

Critiques of current definitions emphasise that people with lived experience and treatment providers disagree about what constitutes recovery. For example, physical restoration of weight and menstruation is not necessarily indicative of improved psychological symptoms or reduced disordered behaviour [93]. Similarly, improved behavioural symptoms do not equate to an absence of struggle:

I want to get better, but I still have the negative thoughts…it is still difficult. I still struggle (Maisie) [93 p. 23].

Typically, a single modality, or treatment working in isolation, is not sufficient, as different parts of the self often need different kinds of treatment:

[EDs] can’t be treated in a single way. There’s no magic solution. Anything, however good it is, is not enough. So, you gotta try everything you can, all therapies and treatments in use. Only with all professionals working together and caring for each part of you at the same time can you pull it off…(Clara) [82 p. 3].

Surfacing these misconceptions about what recovery is heightens awareness about alternate definitions of recovery within the literature. As Table 5 demonstrates, recovery is described in multiple ways by people with lived experience of longstanding EDs and there were several distinct framings of recovery identified in the reviewed studies. Described in this sub-theme (re)defining recovery) and the next (letting go of the ED and moving forward), these perspectives are critical of current (bio-medical) definitions and position the self as central to and constitutive of recovery.

Table 5.

Study findings that (re)define recovery

Recovery definition	Behaviour	Influence on recovery	Evidence
Feeling stronger than the ED	“Making a conscious decision to recover”	Tipping point	Dawson et al. [81]
“A long process of hard work”	Doing the work	Descriptive	Lindgren et al. [85]
(Re)finding one’s own identity outside the illness	“Increased feelings of self-awareness and self-confidence”	Supportive	Smith et al. [93]
“A radical change in [participants’] lives”	No behaviour indicated	Descriptive	Lindgren et al. [85]
“A sense of freedom that had not been possible during their illness”	Exercising self-determination	Supportive	Lindgren et al. [85] Patching et al. [90]
Not a fixed state	No behaviour indicated	Descriptive	Dawson et al. [81]
Different for everyone	No behaviour indicated	Descriptive	Dawson et al. [81]
An “end state”	No behaviour indicated	Descriptive	Dawson et al. [81]
An ongoing process that require[s] a continual awareness to identify potential future signs and triggers of an [ED]	Cultivating awareness and ongoing monitoring of signs and triggers (by self and others)	Descriptive	Dawson et al. [81]
Letting go of the ED	Working on goals specific to life desires Participating in self-help groups Exercising self-determination	Supportive	Williams et al. [99] Patching et al. [90] Espindola and Blay [82]
Holistic concept informed by critical theory	“Fostering independence, increasing hope and enhancing social skills… in the context of the client’s economic, social and physical living environment”	Supportive	Williams et al. [98]
Associated with self-esteem	Valuing relationships more Valuing shape and weight less	Supportive	Williams et al. [98]
Ongoing engagement with treatment providers	Setting personally meaningful goals	Supportive	Williams et al. [98]
Readiness and motivation for change at the end of treatment	Collaborative decision-making Developing shared treatment goals	Supportive	Geller et al. [83]

The study by Lindgren and colleagues [85 p. 863] describes recovery as “a long process of hard work”. To these authors, recovery also represents “a radical change in [participants’] lives… a sense of freedom that had not been possible during their illness” [85 p. 866]. Dawson and colleagues [81 p. 502] emphasise that recovery is not a fixed state and is different for everyone, noting that some see recovery as an “end state” while others perceive it as “an ongoing process that require[s] a continual awareness to identify potential future signs and triggers of an [ED]" .

Letting go of the ED and moving forward appeared across 16 studies [1, 80–85, 87, 88, 90–93, 95, 96, 98] and refers to the desire for a life without ED and contributors in this process.

Working on goals specific to life desires exemplifies letting go of the ED and can in and of itself be recovery. In a study by Williams and colleagues [98], a participant, Jane,4 wanted to attend a three-hour art class once a week. Initially due to her poor nutritional status, she was unable to focus for the duration of the class. Moving towards a life without ED and in line with her values, Jane made an empowered choice to increase her caloric intake to improve her art class experience [98].

Findings emphasise that unsettling the strong link between recovery and conventional treatments contributes to letting go of the ED and moving forward [82, 90]. Espíndola and Blay [82] stress that while qualified treatments have a role in ED recovery, data suggests the contribution is often “partial or incomplete” leading them to conclude that “treatments can and should be adjusted [and tailored] to each patient” (82 p. 5). Similarly, Patching and colleagues highlight self-determination, and self-help groups as the most noteworthy contributors to recovery from an ED [90].

Quantitative studies [91, 95, 98] also (re)define recovery beyond a medical frame. Collaboration, which involves the patient in decisions about treatment, may be a way to realise the benefits of autonomy in recovery [83, 98]. Williams and colleagues [98 p. 90] suggest a definition that encompasses fostering independence, hope and social skills "in the context of the client’s economic, social and physical living environment” [98]. They also associate self-esteem and recovery, emphasising the increased importance of relationships, and decreased importance of shape and weight [98]. Their evaluation of COPP demonstrates that reconceptualising recovery can increase the motivation of the person with ED by aligning with personally meaningful goals. Sansfaçon and colleagues [91] redefine recovery by emphasising ongoing engagement with treatment providers. Geller and colleagues [83] similarly frame recovery as readiness and motivation for change at the end of treatment. These findings affirm the value of collaborative decision-making and shared treatment goals that respect diverse lived experiences.

Others in relation to self

Others in Relation to Self, represents findings that invite a reinterpretation of autonomy as relational. Appearing across 14 studies [1, 80–83, 85, 87, 88, 90, 92–94, 96, 99] (nine qualitative, four quantitative, one mixed methods), it concerns the features of relationships that support recovery.

Formation of healthy attachments appeared across six studies [81, 87, 88, 90, 93, 98] and refers to the healing that is experienced through and in relationship with others. Findings support a healing that is experienced through growth in trusting relationships with parents, peers and people in professional support roles [80, 81, 94]. The expression and experience of support and empathy are common attributes in all these relationships [81]. This relational growth contributes to changes in perception of the ED, improves existing relationships, and opens the door for new connections [81, 82, 94].

Experiencing understanding and acceptance from others appeared across three studies [81, 82, 93]. and refers to feelings of shared struggle, finding community, and acceptance from peers. While experiencing acceptance and understanding during treatment is important across all relationships [91], findings support that mutuality (mutual acceptance and understanding) contributes to people in treatment placing a higher value on peer mutuality than acceptance and understanding found in relationships with staff [91]. Shared lived experience in therapy can enhance hope and empowerment:

I eventually found two amazing therapists who I stayed with until the end of treatment. One of them had had an [ED] herself, and she was by far the most important therapist I had ever seen. It felt like she was reading my mind. She understood it and she instilled hope that I could get over it (Mia) [81 p. 501].

Similarly experiences of acceptance and understanding, that are separate to success or failure in treatment, are valued in “feeling free of [ED]” [81 p. 502].

Encouragement and companionship appeared across seven studies [1, 81, 83, 85, 87, 88, 93] and refers to experiencing relationships of meaningful support and camaraderie—knowing you are not alone:

The ward can be sort of a walking companion; I don’t have to drag everything along all by myself. I’ll do it on my own but I’m never alone (Misha) [1 p. 402].

Relationships of support enhance faith in one’s capacity for change and ability to cope [1, 80, 82].

I started to feel more understood by my dad. After that point I knew I had the support of my parents and that was really all I needed as long as I knew they were prepared to go on the journey with me (Emily) [81 p. 500-1].

Positive results of recovery-promoting relationships appeared across seven studies [82, 85, 87, 88, 92, 96, 98] and refers to the positive impact of supportive relationships on the severity and dominance of the ED. Findings highlight that as relationships strengthened, greater separateness from the ED became possible, and the power of the ED diminished [98]. Similarly, women in Olofsson and colleagues’ [88] study reported that reduced shame and increased self-knowledge contributed to recovery. Study participants linked relationships with meaningful support persons to a sense of satisfaction and a sign of healing [88].

Quantitative [83, 94, 96, 98] and mixed-methods [80] findings support the influence of relationships on autonomy and recovery from an ED. Steiger and colleagues’ [94] study reported autonomy support from therapists, relatives, and peers was linked to increased motivation and better treatment outcomes for 97 women.

This theme demonstrates the centrality of relationships and relational influences on autonomy and recovery from an ED. When relationships of support interact in nuanced and purposeful ways, it supports the person to build wisdom about trauma experiences, prompts recovery-oriented conversations and decisions, and the individual experiences an increased sense of autonomy, power, and control.

Conceptual finding: lack of autonomy loop

The themes of Self versus Eating Disorder and Self in Treatment (themes one and two) demonstrate that developing, living with, and being in treatment for an ED is a battle for control. Data shows the ED can produce a paradox whereby the person in treatment experiences being in control while simultaneously being controlled by the illness [81, 84, 90, 93]. We theorise the inner struggle of living with an ED, the battle in your head and treatments that override autonomy or exert control, as a Lack of Autonomy Loop (Fig. 2). This reflects a cycle of helplessness where people do not understand their choices, the root causes behind their behaviour, or how to exercise autonomy—which is dangerous when combined with a lack of autonomy.

Fig. 2.

Lack of autonomy loop

The Lack of Autonomy Loop, developed by author one, emphasises the cyclical nature of control dynamics. Beginning with a lack of control, an individual turns to an ED as a means of regaining control, which temporarily increases their sense of control. However, eventually control shifts, resulting in the individual feeling out of control and controlled by the disorder. This can bifurcate into two possible outcomes: helplessness or an increased ability to ask for help. Helplessness itself can also lead to increased help-seeking.

When an individual feels governed by the ED, this often triggers external control by others (e.g. family, caregivers, professionals), which can lead to heightened feelings of powerlessness and a lack of connection to both self and others, which has implications for one’s connection to life. These feelings feed back into the initial sense of having no control, thus perpetuating the Loop. This juncture suggests a critical intervention point exists between the stages of feeling out of control and the imposition of external control. Here, the Loop can be interrupted if the individual chooses to freely hand over control, underscoring the complex interplay between control, autonomy, and external influences in the context of EDs. This emphasises the importance of voluntary choice in disrupting the destructive cycle.

Relational, trauma-informed, and contextual strategies can break this Loop at the intersection above, as demonstrated by our analysis. Service providers can break the Loop by introducing autonomy-supportive treatments, exemplified by interventions like self-admission [95] and COPP [98]. Autonomy-supportive treatments are underpinned by joint decision-making and working towards client-identified goals. They share the onus and responsibility for recovery and offer people with longstanding EDs dignity and choice. People with a longstanding ED do not engage in autonomy-supportive relationships instrumentally, for example, to prove they can manage their recovery. Instead, they are supported to consciously build their relationships, focusing on sustaining, repairing, and promoting them, perceiving that these relationships are pivotal for recovery.

Discussion

Findings of this review depict living with an ED, being in treatment, and recovering from an ED. Treatment and recovery features are explained both by people with lived and professional experience. The Lack of Autonomy Loop begins with ED in charge and illustrates the beginnings of externalisation processes where the self is called to action. Data from this study illuminates how people with a longstanding ED relate to their illness, how and when they differentiate themselves from it, and their advances into spaces of agency that promote recovery. Several factors that power the Lack of Autonomy Loop are implicated, including ignoring past trauma, coercive treatment practices, and loss of self. Data also shows how people with a longstanding ED are supported (or undermined) by caregivers, clinicians, peers, and friends.

Past and present trauma fuels EDs [81, 84, 90] and implicates autonomy [81, 90, 92, 98]. While most therapists and ED scholars recognise the salience of trauma in treatment, the complexity of how focusing on trauma, directly or indirectly, translates to practice, especially in the early stages of treatment, needs to be resolved. The effects of the prevailing silence about trauma in treatment are highlighted in our review and represented in the conceptual model, the Lack of Autonomy Loop. Lived experience perspectives in this study clearly demonstrate that delving into the initial causes of the ED can be key to understanding what maintains them for extended periods and is vital to recovery. The integration of trauma-focused therapy with ED treatment going forward is crucial because untreated trauma fuels the Loop and hinders recovery. Alongside continued attention to tailoring the approach to the individual’s specific needs, it appears that striving to understand an individual’s trauma history and life experience is central to effective treatment.

EDs represent a profound assault on the self, stripping individuals of their autonomy and distancing them from their potential to reclaim it. Coercive and trauma-blind treatment regimens may inadvertently reinforce this disempowerment, deepening the sense of helplessness that characterises these disorders. Autonomy, though complex and paradoxical, is central to recovery: individuals may temporarily relinquish control to others, but they never willingly cede it entirely. The desire to retain and eventually regain autonomy is critical to breaking the cycle of disempowerment. Within such frameworks, individuals are consistently granted voice and agency, even as their autonomy ebbs and flows, offering a more hopeful path to recovery.

The findings point toward reconceptualising autonomy in ED treatment and recovery as relational. Relational autonomy is grounded in feminist theory [50, 99] and is increasingly advocated in moral psychology [51] and biomedical ethics [100]. Relational autonomy challenges the individualist view of autonomy and of the self, instead situating it as necessarily interrelated with others while recognising that relationships occur in context [50, 99]. According to Nedelsky [50], in this way, a “culture that values autonomy and understands autonomy in relational terms will be better equipped to treat everyone with respect. People will be better able to see themselves with respect (which in turn enhances autonomy)” [50 p. 65].

Increasing awareness of the relational aspects of healthcare have not yet exerted influence within longstanding ED treatment and recovery literature, making this research an innovative contribution. In the relational model of autonomy, the more one processes and understands the causes and drivers of the ED, the more it is possible to exert autonomy, strengthening relationships with caregivers and the self, and fortifying efforts to attain and maintain recovery. Relational autonomy can reinvigorate the therapeutic relationship, offering an alternative to top-down models where autonomy is earned from or titrated by practitioners. This aligns with Kiely and colleagues [4 p. 19] who showed that maximising autonomy and a person-centred treatment approach allows people living with EDs to engage in recovery “on their own terms” [4]. Such an empowered position could foster hope in the possibility of recovery and in turn enable and empower a person experiencing an ED to move toward collaborative relationships supporting recovery.

Our findings about autonomy are deeply rooted in the idea of ‘desire narratives’. Currently explored in research literature and academic discourses, ‘desire narratives’ are yet to be translated into desire-focused practice. As Tuck [101] describes, desire-focused (decolonial) research represents a shift away from documenting the pain story and towards how communities wish to talk about and remember themselves. A ‘thicker’ description [102] of autonomy signifies recovery as re-engagement with desire and will. Although twelve of the twenty articles in the dataset refer to desire in some form, there is a paucity of research into desire-focused treatments for longstanding EDs. The absence of explicit attention to these narratives suggests that further research is warranted. By highlighting studies that gear treatment more toward consumer will and desire than compliance, and by elevating collaborative approaches to treatment and goal setting, this review contributes insight into aligning expectations of service providers with those of people with longstanding EDs [82, 98].

Recovery from an ED is not a single, linear process but rather a multifaceted journey that hinges on the interplay between autonomy and the individual's commitment to change. This paper distinguishes between the initial tipping point—when an individual decides to commit to change—and the broader, more complex definition of recovery itself. This distinction is pivotal, as recovery is not merely about symptom cessation but also about reclaiming and sustaining autonomy in the long term. By tabulating the various definitions of recovery, this paper makes a novel contribution, highlighting how these definitions converge on the importance of autonomy. In this context, recovery involves more than just overcoming disordered behaviours; it’s about restoring a sense of self, where the individual retains the ability to make autonomous decisions and actively participate in their healing process. Autonomy-supportive, collaborative approaches become essential, as they enable individuals to navigate the fluctuating journey of recovery with their sense of agency intact. This finding aligns with Downs’ [103 p. 180] analysis that pathways to recovery for people experiencing longstanding EDs, as opposed to “managed decline”, requires shared understandings of illness, treatment, and recovery developed in collaboration with people with lived experience and their supporters [103]. Extending this collaboration to define and position autonomy’s role in recovery could help strengthen pathways to recovery.

Piloting trauma-informed, relational, and autonomy-supportive treatment models within a supportive service environment is an imperative that emerges from this review. This will require a whole-of-system reflection on how practitioners incorporate and share their understandings of the impact of past and present trauma with consumers. It also invites (re)consideration of when and how much space is provided for people experiencing longstanding and severe EDs to disclose their traumas as part of their recovery journey.

Conclusion

This review redefines autonomy and recovery by examining qualitative and quantitative evidence about the lived experience of treatment from the perspectives of people with longstanding EDs and service providers. Our findings highlight that most participants have experienced multiple traumas and demonstrate the need for a holistic assessment of treatment modalities in light of a sustained increase in longstanding and severe EDs. A critical focus on relational autonomy and trauma-informed practice is key to more positive outcomes for individuals, families, and broader society. Further research on and incorporation of collaborative and desire-focused therapeutic methods would respond to the rich insights of people with lived experience of EDs presented in this review.

Abbreviations

AN

Anorexia nervosa

BED

Binge eating disorder

BMI

Body mass index

BN

Bulimia nervosa

BSED

Bulimia spectrum eating disorder

COPP

Community outreach partnership program

DSM

Diagnostic and statistical manual of mental disorders 5th revision

ED

Eating disorder

EDNOS

Eating disorder not otherwise specified

MMAT

Mixed methods appraisal tool

NSW

New South Wales

OSFED

Other specified feeding and eating disorder

PRISMA

Preferred reporting items for systematic reviews and meta-analyses

PWB

Psychological well-being

RER

Rapid evidence review

SDT

Self-determination theory

SE-AN

Severe and enduring anorexia nervosa

Author contributions

FB conceived of the study. FB undertook the search, extraction and thematic analysis of data for the requirements of an honours thesis (outlined in the methods section of the article). MH and MM contributed during this phase through their role in supervision. FB composed the original manuscript and developed the conceptual diagrams. All authors contributed to the revised manuscript. All authors contributed to the conceptual development and all authors read and approved the final manuscript.

Funding

Not applicable.

Availability of data and materials

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

As a review of published literature, ethics approval was not necessary. Not applicable.

Consent for publication

Not applicable. All quotes and data referenced have already been published in the peer-reviewed included studies, so therefore, not applicable.

Competing interests

The authors declare no competing interests.