Service models in eating disorders: a scoping review

Abstract

Background

The complexity of eating disorders can be reflected in the different diagnoses, varied clinical presentations and diverse personal circumstances of people living with an eating disorder. Given this complexity, adaptable and flexible service models are essential for effective care. Service models outline a structure for health care delivery that targets the health-related needs of people. The characteristics of existing service models providing assessment and treatment for eating disorders are not well described or understood.

Objective

The purpose of this scoping review was to map and describe the different types of service models for eating disorders and their characteristics.

Method

The JBI methodology for scoping reviews was used. A search of four databases (MEDLINE Complete, EMBASE, PsychINFO, and CINAHL) and grey literature was conducted. Sources describing service models supporting individuals with eating disorders were included.

Results

After duplicates were removed, the remaining sources were screened and read in full, and 30 sources were included in the review. Most service models were eating disorder specialist, located in metropolitan areas and based in community settings. Key characteristics of eating disorder service models included person-centred care, involvement of family and carers, co-design and lived experience contribution, multidisciplinary team, accessibility, identification and management of co-occurring conditions and integration with broader service systems.

Conclusion

Findings indicate vast differences between service models for eating disorders. However, there are examples of innovative and effective service models that show promise. The characteristics of service models for eating disorders identified in this review provided insight into what constitutes an effective and high-quality service model in the sector.

Plain English summary

The purpose of this scoping review was to understand the different types of service models for eating disorders. Service models provide assessment and treatment for people with an eating disorder. In this article, the literature on service models for eating disorders was reviewed to explore what types of service models currently exist and identify their characteristics. There were 30 sources of literature, and all of these were published in academic journals. After reviewing the literature, it was found that service models for eating disorders are very different from each other. However, there were some similarities. For example, most service models for eating disorders are specialist (meaning they only assess and treat people with an eating disorder) and are mostly based in metropolitan areas. The review identified that service models for eating disorders have some key characteristics, and these are; person-centred care, involvement of family and carers, co-design and lived experience contribution, multidisciplinary team, accessibility, identification and management of co-occurring conditions and integration with the service system. Despite the vast differences between service models, innovative and effective examples exist. A clearer understanding of the characteristics of service models is useful for the development of future service models for eating disorders.

Introduction

The provision of health care is typically offered through service models. Service models are multidimensional and outline a structure for how health care should be delivered [1, 2]. A service model aims to ensure people get the health care they need at the right time and in the right place [1]. The benefit of a clearly articulated service model is consistent and effective health care that aims to sustain or improve health outcomes for a defined group of people [3]. Service models sit within contexts or environments, referred to as service settings. Service models exist across several service settings such as hospitals or the community [4].

Service models and service settings sit within a broader service system. Service systems encompass several interconnected service models across different settings. Ideally, service systems should be coordinated to optimise health outcomes through the delivery of consistent, high-quality health care that promotes the health and wellbeing of the population it serves [3]. An example of a service system being increasingly referenced in eating disorders is a stepped care service system [4, 5]. This service system consists of a range of service models and settings across a continuum of care that offers treatment and support to people with eating disorders. The central premise of the stepped care service system is that a person’s level of need is matched to the most appropriate service model within the system [5].

Eating disorder service systems are complex and vary across countries due to differences in funding, healthcare structures, and population needs. These variations often lead to inconsistent access to care, delayed assessments, and difficulties in starting treatment. As a result, individuals may experience setbacks in their recovery and overall health [6,7,8]. Given the complexity and severity of eating disorders, it is critical that people can easily access service models that provide quality, evidence-based care because delays can compromise a person’s recovery [9].

A recent rapid review identified eating disorder service models across a range of settings. Findings from this review highlighted the importance of matching the right service model and service setting to a person’s needs [4]. The review recognised the value of service models located in primary care settings to facilitate early detection of an eating disorder. In addition, the review highlighted the important role of community based service models in supporting recovery and improving quality of life. It also found that service models in the community (including day programs) can provide intensive treatment and support, helping to reduce both the number and duration of hospital admissions. While the effectiveness and versatility of eating disorder service models in the community was demonstrated, there was little information about the characteristics of these service models and how they were integrated within the wider service system [4].

Different types of eating disorder service models have been compared with each other in some studies. The Treatment Outcome for Child and Adolescent Anorexia Nervosa (TOuCAN) randomised control trial in the United Kingdom (UK) sought to understand the clinical effectiveness of eating disorder treatment received by young people in hospital inpatient units, community based specialist eating disorder service models, and community based generalist mental health service models (service models that provide assessment and treatment for a variety of mental health conditions) [10]. The results of this study showed no differences between the physical and psychological outcomes of participants (n = 215) across the three service settings. However, a limitation was poor adherence to treatment, particularly in the hospital inpatient units [10]. Similarly, a few years later, health outcomes of young people (n = 378) with eating disorders attending community based specialist eating disorder and generalist mental health service models across 37 services in London were compared [11]. Data revealed that young people in specialist eating disorder service models were more likely to receive a correct eating disorder diagnosis, remain engaged in treatment, and were less likely to be admitted to hospital [11].

Contemporary eating disorder research and guidelines make increasing reference to a stepped care service system [4, 5]. This service system draws together different types of service models, across different settings to create a staged system. Each stage of the stepped care service system has a different capacity to respond and provide a range of evidence-based treatment options and varying levels of support that correspond to a person’s needs. Typically, service models range from least intensive (for example, primary care) to most intensive such as inpatient hospital care [5]. The available literature about stepped care service systems suggests they are more person-centred [12], cost-effective and may optimise clinical outcomes [4], however, more research is required to determine their effectiveness.

Understanding the characteristics of eating disorder service models has been the focus of some studies. One of these studies conducted by Halmi [13] identified, described and recommended the important characteristics of eating disorder service models, across different settings such as community and hospitals. The characteristics that the author recommended, such as the inclusion of a diagnostic and evaluation clinic and a multidisciplinary team, were informed by the available literature, with no primary research conducted. Halmi [13] concluded that despite some characteristics of eating disorder service models having been identified, there was insufficient research currently available to support these recommendations. Another study conducted by Escobar Koch et al. [14] explored essential characteristics of eating disorder service models from the perspectives of 294 participants from the UK and the United States of America (USA). Participants of this study were people who had accessed eating disorder service models in either country. This study found a high-quality, knowledgeable, experienced workforce in eating disorders was paramount for service models. Furthermore, participants expressed that eating disorder service models needed to be easily accessible and offer person-centred and holistic treatment options [14].

Current evidence lacks detail on the characteristics of eating disorder service models, particularly in terms of accessibility and their role in supporting recovery [4, 14]. Several existing documents, including the Australian National Eating Disorder Strategy 2023–2033 [5], Canada’s Clinical Practice Guidelines for the BC Eating Disorders Continuum of Services [15] and Ireland’s Eating Disorder Services HSE Model of Care [16] outline stepped care service system frameworks for eating disorders, detailing their structure, components and functions. These frameworks for stepped care service systems provide a high-level, broad structure on how assessment and treatment for an eating disorder can be accessed. However, these documents do not describe service models within the service system or their characteristics. Therefore, a more detailed and specific articulation of service models is warranted to provide more consistency between service models and identify what these service models need to optimise the provision of support and care to people with eating disorders.

By definition, scoping reviews explore the breadth of research on a topic [17]. Their purpose includes clarifying key concepts in the literature, identifying key characteristics related to a concept and identifying and analysing knowledge gaps [17,18,19]. Therefore, a scoping review seemed the most appropriate way to detail the different types of service models in eating disorders, highlight the key characteristics of service models and examine the variations between these. Scoping reviews are the most useful approach to map the depth and breadth of a topic or concept, especially those that are emerging or for which there is limited existing evidence [20]. In addition, scoping reviews are most appropriate for mapping the available literature [21, 22] and do so by including diverse forms of evidence [20]. Therefore, this scoping review was undertaken to explore the variety of eating disorder service models described in the literature, drawing from multiple sources to address the following questions:

• 1. What are the different types of service models for eating disorders?

• 2. What are the characteristics of eating disorder service models?

Methods

The JBI methodology for scoping reviews [21, 22] was used to describe and map the different types of service models for eating disorders and their characteristics. Further adding to the rigour of the methodology, the quality indicators from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews guidance (PRISMA-ScR) [23] were used.

Inclusion and exclusion criteria

In this review, any literature source and study design were considered for inclusion. The ‘PCC’ mnemonic guides scoping reviews and stands for ‘Person’, ‘Context’ and ‘Concept’. In this scoping review, literature sources were included if they focused on an eating disorder population (person) and described a service model (context) and its characteristics (concept). Sources were included if they were written in English and published between 2010 and 2024. These search dates were decided as 2010 was when most literature on service models for eating disorders began to be published, capturing 15 years of literature on the topic.

Any source that did not meet the inclusion criteria, had an intervention focus or did not describe a service model (for example, described a service system, such as a stepped care service system) or the characteristics of a service model were excluded.

Search strategy

A comprehensive search strategy was conducted in May 2024 and updated in September 2024 with electronic databases MEDLINE Complete (EBSCOhost), EMBASE, PsychINFO, and CINAHL. Relevant subject headings for each of the databases were also included. A list of search terms is presented in Table 1. A search of grey literature was conducted broadly on the internet using the search terms to identify guidelines, government reports, and protocols. The authors also searched the websites of several national and international eating disorder organisations and relevant government bodies.

Table 1 Scoping review search terms

A total of 6642 individual sources were initially identified. After duplicates were removed, 3957 remained. Titles and abstracts were screened by RK against the inclusion criteria using Covidence Systematic Review Software and 3869 sources were rejected because they did not meet the inclusion criteria. If there was uncertainty about whether a source met the inclusion criteria, GP was consulted. A total of 88 sources were read in full. All full text sources were independently reviewed by two authors (RK and GP). Of these 25 were included in the scoping review. Discrepancies were discussed before reaching a consensus. Reviewing reference lists revealed a further five studies that were included because of their relevance to the review questions. No grey literature sources were located that met the inclusion criteria. Therefore, a total of 30 literature sources were included. Detailed information can be viewed in Fig. 1.

Fig. 1

PRISMA flow chart

Data extraction

In line with recommended best practice [24], a table was developed to ensure key criteria about the service models were extracted from the data in a standardised way. Data describing the different types of service models were extracted and included the country of data collection, if the service model was eating disorder specialist (or not), the service model setting and location, and the population targeted by the service model. If the literature source included research or evaluation of the service model, further data were extracted such as participant demographics, aims, study design, and outcomes. This information is detailed in Table 2. Data was also extracted on the characteristics of the service model. Determining the criteria for this part of the extraction tool was informed by the National Eating Disorders Collaboration (NEDC) National Standards for Eating Disorders practice principles [25], which form the foundation for an effective and consistent approach to eating disorder care. Characteristics of service models encompassed; person-centred care, involvement of family and carers, co-design and lived experience contribution, the multidisciplinary team, accessibility, identification and management of co-occurring conditions and integration with the service system. The characteristics are described in Table 3. During the data extraction process, iterative modifications were made to the table to ensure it captured the most relevant criteria that aligned with the review questions [19].

Table 2 Types of eating disorder service models

Table 3 Eating disorder service model characteristics

Data analysis

Descriptive statistics, a narrative summary and tables were used to present the findings in relation to the questions the scoping review aims to answer.

Results

Table 2 presents the 30 literature sources included in this scoping review. There were 11 quantitative studies, three qualitative studies and five mixed methods studies. Retrospective chart reviews or audits were commonly used in both quantitative and mixed methods studies [26,27,28,29,30,31,32]. Several studies explored the feasibility and acceptability of the service model [7, 30, 33, 34]. While some studies sought to understand the outcomes achieved by people accessing the service model [28, 29, 31, 35, 36], only one reported if the outcomes were sustained at 6 months or more [37]. Overall, there was a notable absence of high-quality quantitative studies such as comparative studies and randomised control studies. Qualitative studies focused on understanding the experience of service models [38, 39] or sought participant input into the design of a technological web-based service model [40]. A further 11 studies did not include an evaluative or research component and solely described a service model.

The literature sources originated from predominantly Western countries, including the UK (n = 11), Australia (n = 9), USA (n = 3), Sweden (n = 2), Canada (n = 2), Spain (n = 1), Israel (n = 1) and Germany (n = 1).

Most service models described were eating disorder specialist (n = 27). Only three [26, 27, 31] literature sources documented a service model that was not eating disorder specialist but provided treatment to people with an eating disorder. Table 2 provides details on the types of service models included in this review.

Table 3 summarises the key characteristics of service models included in this review, under the headings of person-centred care, involvement of family and carers, co-design and lived experience contribution, multidisciplinary team, accessibility, identification and management of co-occurring conditions and integration with the service system. A narrative summary of the different types of service models followed by key characteristics of eating disorder service models is discussed below.

Eating disorder service models: a narrative summary

Most service models included in this review were eating disorder specialist (n = 27), meaning that they solely focused on supporting the health needs of people with eating disorder diagnoses. Service models were primarily located in metropolitan areas (n = 17) [7, 26,27,28,29,30, 32, 35, 37, 41,42,43,44,45,46,47,48] with some located in regional areas (n = 7) [31, 34, 36, 38, 49,50,51] and one in a rural location [39]. An additional five service models provided health services across metropolitan, regional and rural locations [52,53,54], including telehealth service models [33, 40] (see Table 2).

Children and adolescents were the focus of 11 service models, [26, 29,30,31, 34, 35, 37, 42, 45, 47, 53], while a further three service models targeted young adults (see Table 2). One service model included people from the age of 4 to 21 years [28]. Two service models accepted people over the age of 14 years [36, 39] and another two were accessible to people over the age of 16 years [43, 49]. Only four service models reported an adult focus [38, 44, 46, 52]. Another four service models were available to people of any age, [33, 40, 41, 50] and three did not report specific details on the age range of the target population [27, 48, 51] (see Table 2).

Some service models provided treatment only to people with specific diagnoses. For example, people with anorexia nervosa (n = 4) [29, 47, 49, 51], a ‘restrictive’ eating disorder (n = 1) [37], anorexia nervosa and atypical anorexia nervosa (n = 1) [34], avoidant restrictive food intake disorder (ARFID) (n = 2) [28, 42] or anorexia nervosa and autism (n = 1) [46]. Others (n = 8) explicitly stated that they accepted people with any eating disorder diagnosis into the service model [7, 26, 36, 38, 41, 45, 53, 54]. People experiencing longstanding eating disorders, or who were unresponsive to usual treatment were the focus of two service models. One of these enabled self-admission to hospital [44] and the other focused on improving a person’s quality of life [48] (see Table 2).

Predominantly, service models were based in community settings (n = 17). There were nine service models that provided community based outpatient care [31, 36, 38, 39, 41, 42, 48, 50, 53]. Another group of community based service models offered intensive support (n = 7) such as day programs, outreach or hospital in the home [29, 30, 34, 35, 37, 49, 51]. Early intervention was the focus of one community service model that was described in two separate studies [7, 54]. This service model targeted young adults within three years of their eating disorder diagnosis. Service models within hospital service settings (n = 7) encompassed one that targeted people with a diagnosis of anorexia nervosa and autism [46], one that offered a self-admission program [44], three that aimed to provide staged, consistent hospital care, [26, 28, 45] and a family admission program [47]. An additional hospital based service model sought to implement a pathway that facilitated eating disorder treatment and support for people with a primary mental health diagnosis in a psychiatric hospital with a co-occurring eating disorder [27].

There was one service model in a residential setting [32] and a further two were telehealth service models [33, 40].

Finally, two service models covered multiple service settings. One [52] included community, community intensive support and hospital service settings over a large geographical region. A service model described by Newton et al. [43] offered outpatient community based treatment and an intensive community based day program.

Service model characteristics

Person-centred care

Person-centred care was a characteristic of service models identified in 16 sources (see Table 3).

The most common reference to person-centred care was in relation to aspects of treatment (such as goals or the treatment plan). These were reported to be collaboratively developed between the clinical teams, the person with an eating disorder and sometimes, their family and carers [7, 27, 29, 30, 34, 42, 43, 51]. Only three service models [38, 44, 48] reported person-centred care that extended beyond this collaboration. One service model provided support to people who did not respond to usual treatment for eating disorders. The focus of this service model was quality of life. People using the service model could set their own treatment goals and choose their treatment intensity (how often they saw the team), highlighting a person-centred approach [48]. A second service model aimed to maximise choice and control for people by offering a person-centred ‘drop in’ self-referral [38]. This meant that people who felt they needed support for an eating disorder could attend the service and be seen immediately, instead of prearranging appointments. The third example was a service model that allowed people with an eating disorder (who were receiving treatment within the wider eating disorder service system) to request a brief self-admission to an inpatient unit when they felt they were at risk of relapse or needed additional support to manage their condition [44].

Involvement of family and carers

Family and carer involvement was identified in many service models (n = 22) (see Table 3). Family and carers were often part of the treatment approach, for example in family-based treatment or multi-family therapy [30, 31, 41, 47, 50, 53]. Additional day patient program service models (n = 2) also involved families in treatment, aiming to help families feel more confident to support their child's treatment and recovery [35, 37]. Other intensive outreach and hospital in the home service models (n = 3) also required the involvement of family or carers to facilitate treatment delivery and/or monitor a person’s safety in the community [26, 29, 30, 34]. A service model for young people hospitalised with ARFID provided families with education and guidance about meal support and involved families in discharge planning [28].

Family involvement was ‘encouraged’ by four service models (including one with two sources) [7, 32, 42, 43, 54] but limited additional details were provided. In one case, family involvement was dependent on whether the person with the eating disorder agreed to family involvement [32]. Other ways family and carers were included in service models were through clinical meetings [45], or through the provision of a structured carer-focused intervention [52]. The service model described by Tchanturia et al. [46] recognised a need to improve family involvement and noted that this was a focus of ongoing work.

Co-design and lived experience contribution

Co-design principles were reported to be used in the development of two service models [40, 46]. In one instance, people with lived experience helped design a service model for people with anorexia nervosa and autism, with a commitment to seeking ongoing input from them [46]. In another instance, participatory design workshops provided an opportunity for people with lived experience to contribute to the service model design and development [40].

After the ‘First episode and Rapid Early intervention in Eating Disorders’ (FREED) service model was developed, a group of young people who had benefited from or supported FREED was formed. They contributed to new initiatives for promoting and expanding the service model [7, 54].

Other service models were conceptualised using feedback from families [47], consumer advisory groups [43] or used consultation from people who had previously used eating disorder services [50] to inform service model development. For a service model treating people with longstanding eating disorders, insights from focus groups with people with lived experience helped shape its design [48].

One service model reported that people with lived experience were part of the team who delivered healthcare services [33].

Multidisciplinary team

Most service models (n = 25) included a multidisciplinary team (see Table 3). However, aside from reporting on the disciplines included within the team, other detailed information about the team (such as if staff were full-time or part-time and their level of seniority) was not provided.

Some service models stated they had a multidisciplinary team [45, 46], while others specified their team included medical and non-medical clinicians with eating disorder expertise [53] or consisted of a primary therapist and physician [41]. Other service models provided more detailed information about the disciplines within the team (n = 20).

The most common disciplines included nursing (n = 19) [26,27,28,29,30,31,32,33,34,35, 37, 38, 43, 47,48,49,50,51,52], psychology and psychotherapy (n = 16) [26, 28,29,30,31, 33,34,35,36,37,38, 43, 49,50,51,52], psychiatry (n = 14) [26,27,28, 30, 32,33,34, 37, 38, 47,48,49,50,51] and dietetics or nutrition (n = 13) [26,27,28, 33,34,35,36,37, 43, 48, 50,51,52].

Less frequently reported disciplines included medical professionals (including physician, adolescent medicine specialist and gastroenterologist) (n = 6) [28, 32, 33, 38, 43, 48], social workers (n = 6) [26, 28, 36, 38, 43, 52], family therapists (n = 6) [29, 31, 37, 47, 48, 50], occupational therapists (n = 5) [34, 35, 42, 43, 50], paediatricians (n = 3) [30, 37, 47], art therapists (n = 2) [33, 37] and speech therapists [42].

In addition, an online telehealth service model was staffed entirely by trained counsellors [40]. Another included team members with lived experience of an eating disorder [33]. A third, described by Johnson et al. [36] involved general practitioners and primary care practitioners.

Training and development for team members were described in some service models. For example, a service model for people with anorexia nervosa and autism included workforce training and group huddles to boost staff knowledge and confidence [46]. Similarly, a hospital based service model introduced training modules for all staff, helping to implement the new service model and improve care for those with a primary mental health diagnosis and co-occurring eating disorder [27]. Eisler et al. [53] also described a training curriculum that facilitated coordinated upskilling of a large national workforce in England (approximately 900 people). The training program provided education on eating disorders, evidence-based treatment models, supervision, provided train-the-trainer opportunities and regular updates on new research [53]. Other service models (n = 3), recognised the value of providing high-quality supervision [41, 50] and holding regular meetings to review clinical practice and emerging research [31].

Accessibility

Most of the service models (n = 27) provided information about access to the service model, with some of these (n = 17) outlining inclusion and/or exclusion criteria (see Table 3).

Facilitating access to eating disorder service models included responding to referrals and conducting an intake call within 24 h [41] or 48 h [7, 30, 54]. Reducing the time between referral and assessment was a focus for four service models, with one model being described in two separate sources [7, 31, 36, 53, 54]. The child and adolescent service model described by Eisler et al. [53] responded to urgent referrals within 1 week, and ‘less urgent’ referrals were responded to within 4 weeks. Other service models [7, 31, 36, 54] reported a 2-week timeframe between referral and assessment for people eligible to access the service model.

Another service model overcame delays between referral, assessment and/or commencement of treatment by not having a waiting list [41]. Self-referral was possible for five service models [38, 39, 42,43,44], however, one required a medical referral before an assessment could be completed [43]. An early intervention service model described by Brown et al. [7] and later Allen et al. [54], initially did not accept self-referrals but identified that later versions of the model planned to.

The criteria that determined access to each service model varied in detail. For example, an early intervention service in the UK required people to have received a diagnosis of an eating disorder within the last three years [7, 54]. Weight-based criteria such as rapid weight loss or weight below a particular threshold were reported for two service models [37, 51]. Another two service models specified that there were no weight criteria for access [35, 44], but had other criteria determining access. For example, the service model described by Strand et al. [44], required people to be engaged in eating disorders treatment and to have had an inpatient admission in the previous 3 years. An additional three service models reported no criteria for access at all [38,39,40], meaning anyone who thought they had an eating disorder could access the service model for assessment. A primary eating disorder diagnosis (any eating disorder) was explicitly stated as an inclusion criterion for many service models (n = 12, including one with two sources). A specific type of diagnosis, for example ARFID, was a requirement for other service models [28, 42]. One service model targeted specific co-occurring conditions (anorexia nervosa and autism) [46]. One service model required an eating disorder diagnosis but excluded individuals with extreme eating disorder compensatory behaviours (such as purging or refusing to eat) [30] and another was unable to accept people with a ‘severe’ eating disorder [34]. There were two service models requiring people accessing the service model to be able to eat food (i.e. not being fed via a nasogastric tube) [30, 47].

Being medically unstable was an exclusion criterion for five service models [7, 30, 32, 35, 47]. Other service models stipulated that a person needed to ‘be safe’ (without defining what safe meant) for treatment in the community [7, 51] or required a person to be psychiatrically stable [32]. People with a planned admission to a residential unit were unable to access one service model [34].

Other exclusion criteria reported by service models included suicidal ideation [30, 44], people with a high risk of aggression [30], substance use or self-injurious behaviour [44]. Co-occurring conditions, including learning disabilities, low intelligence or organic brain disorders, prevented access to two service models [7, 34].

An unwillingness to commit to treatment or a lack of readiness to change were other exclusion criteria [30, 38, 47]. For people to access to some service models, a parent or carer needed to commit and be available to support treatment [30, 34, 47]. Furthermore, living outside a designated geographical area prevented access to three service models [30, 34, 43].

Co-occurring conditions

Co-occurring conditions were reported for participants of six service models [26, 28, 34, 35, 42, 43]. However, it was unclear whether the service model provided treatment for co-occurring conditions. Two service models offered treatment for (some) co-occurring mental health conditions alongside an eating disorder, but did not include further detailed information [41, 53]. In contrast, in a hospital service model for people with general mental health conditions, eating disorders were treated as the co-occurring condition [27].

Another service model recognised a different approach was required for a specific set of co-occurring conditions (anorexia nervosa and autism) [46].

Integration with the service system

Fifteen (including one with two sources) service models were integrated within a service system. Most (n = 12, including one with two sources) eating disorder specialist service models were part of a wider eating disorders service system (see Table 3). One service model sat within a paediatric inpatient service setting and reported connections with child and adolescent mental health service models [45]. Other service models (n = 2) reported established partnerships with external organisations or parts of the health system (such as between private and public service models, or with academic institutions) [41, 43].

Some service models (n = 3) aimed to improve cohesion and connection between other service models and/or settings within the service system. For example, enhancing pathways between primary care providers (who often identify or diagnose eating disorders) with service models that could provide appropriate, evidence-based treatment [36, 39, 52]. One service model aimed to integrate treatment and support in an inpatient setting for people with co-occurring mental health conditions and eating disorders to optimise outcomes [27]. Providing navigation and help-seeking support to people seeking help for an eating disorder was a key function of a national telehealth service model [40]. Another service model facilitated knowledge sharing between experienced and less experienced clinicians, to improve treatment for people with eating disorders [33].

Integrating two existing service models created a new one, allowing children, youth and adults to receive treatment in one place, improving continuity of care [50].

Discussion

This scoping review aimed to describe the types of service models for eating disorders and their characteristics. A range of service models for eating disorders were included showing significant differences in design, setting, purpose and target population.

Most service models for eating disorders were based in the community and were specialist eating disorder services, providing assessment and treatment exclusively to people with eating disorders. While several advantages of specialist eating disorder service models have been identified in the literature, their capacity to offer assessment and treatment for co-occurring conditions, often seen alongside eating disorders, remains unexplored. This is concerning when a recent rapid review found people with eating disorders have high rates of medical and psychiatric co-occurring conditions [55].

There were 17 service models within this review that outlined eligibility and/or exclusion criteria people had to meet to gain access. Extensive eligibility criteria can hinder access to service models [56] and exclude people with complex eating disorder presentations [57]. Given the stringent eligibility criteria of most specialist eating disorders service models, it is likely that many individuals with eating disorders are prevented from accessing the care they need.

Simplified access to service models is key to ensuring early intervention of eating disorders. Early intervention is important because it improves the likelihood of recovery from an eating disorder [9, 58]. Early intervention is crucial for reducing the ‘duration of untreated illness’, which is the time between the onset of an eating disorder and the start of treatment [9]. Research suggests the average duration of untreated illness is 5.28 years [59]. Barriers such as ‘inaccessible treatment’ [59] or strict and rigid eligibility criteria may contribute to this delay [56, 57]. Because delayed treatment leads to poorer health outcomes for people with eating disorders [4], improving access to service models is essential.

Potential solutions to improve service model access were identified in this review. Solutions included promoting informal, self-referral entry points [38, 39, 42,43,44] and prioritising intake and assessment when a person has a suspected eating disorder [7, 42, 43, 54]. Removing waiting lists was another strategy identified [41], but unless a person can easily access help from another service model this may still result in difficulty with access and treatment delay. Some evidence suggests that people put on a waiting list may be less likely to commence treatment or more likely to drop out of treatment [60, 61].

Recent evidence suggests that effective integration of service models also promotes access to healthcare [62], improves a person’s outcomes, and enhances the efficiency of the service system [63]. Within this review, some service models had existing pathways between other eating disorder service models, supporting continuity of care and forming a service system. Coordinated integration of multiple service models across different service settings can create a stepped care service system [5]. This is promising, as advantages of stepped care service systems have been described in the literature [4] and are a recommended framework for eating disorders service systems in Australia, Ireland and Canada [5, 15, 16].

Only a few service models had established care pathways and integrated with service models outside the eating disorder sector, for example with paediatric service models [45] or with primary care providers [36]. Cross-sector service model integration is complex and difficult to do [63] with no clear or transferable guidance [64]. However, cross-sector integration can offer a solution to effectively addressing eating disorders and co-occurring conditions. While limited examples of cross-sector service model integration were found in this review, they are promising as they demonstrate that overcoming challenges with level of integration is possible [12].

This review highlights the benefits of community based service models, which aligns with recent evidence showing that contemporary mental health and eating disorder service models are primarily community based [65]. There are multiple benefits to eating disorder service models in the community. For example, a person can remain at home, remain connected to family and friends and participate in their usual routines [4, 65]. Recommended psychological therapy approaches for eating disorders (such as Cognitive Behaviour Therapy-Enhanced and Family Based Treatment) are mainly designed for delivery in community settings [66, 67]. Health care provided in the community is also more cost-effective and promotes recovery [4, 65].

Over half of the service models described in this scoping review mentioned person-centred care as an essential characteristic of the model. Person-centred care is treating people as individuals with unique needs and circumstances [12]. While person-centred care is more akin to an approach to service delivery than a characteristic of a service model, this was a key finding of this scoping review. Literature on person-centred care suggests that incorporating the approach into service models can be challenging yet crucial to enhance a person’s experience of service models and their health outcomes [12, 68]. In this review, it was difficult to ascertain how person-centred care was embedded into many service models. However, three service models included a detailed explanation of how person-centred care was prioritised [38, 44, 48]. One of these was the community outreach eating disorder specialist service model described by Williams, Dobney and Geller [48]. The focus on quality of life and autonomy given to people accessing this service model aligns with person-centred care for serious illness, identified as a construct in a systematic review by Giusti et al. [69]. In that systematic review, person-centred care included person and family empowerment and autonomy, and treatment that has a quality of life focus [69]. Further emphasising the value of person-centred service models, a recent meta-synthesis of 22 studies from eight countries found participants with a lived experience of an eating disorder viewed person-centred care as crucial in treatment [70]. Research conducted on a person-centred eating disorder service model included in this review [48] found improvements in participants body mass index, eating disorder symptoms and feelings of hopelessness. These results suggest that person-centred service models can lead to improved outcomes for eating disorders.

Interestingly, family-empowerment and autonomy, identified by Giusti et al. [69] as part of person-centred care, was mentioned in 22 service models included in this review. This suggests that involving family and carers is a key characteristic of service models in eating disorders. The use of family-centred approaches, such as family-based treatment, was a common way families and carers were included in models described in this review. This aligns with findings from a recent Australian report evaluating the impact of the Medicare Benefits Schedule items. This report found that families and carers should be involved in treatment for a loved one with an eating disorder [71]. Other ways family and carers were incorporated in service models was through the provision of education and coaching, and collaboration on treatment goals and discharge planning. Involving family and carers in decisions about treatment and recovery from an eating disorder is crucial [72]. This is because family and carers are often essential drivers of treatment [73], and frequently have a role in supporting a person emotionally and financially [74, 75]. A recent scoping review confirmed that providing education to families and carers of adults with eating disorders may improve their loved one’s outcomes [76]. Involving families and carers in service models is often necessary, however, there is scope to better understand and recognise carers’ needs, experiences and perspectives beyond their involvement in treatment and provision of education.

Co-design is one way that people with a lived experience of eating disorders (and their families and carers) can be involved in the design, development and/or delivery of a service model [77]. The advantages of co-design include more effective resource allocation, innovative solutions to complex problems and the development of new knowledge [77]. Given these advantages, it is surprising that only two of the service models identified in this review incorporated principles of co-design [40, 46]. This may be explained by challenges associated with co-design in the field of eating disorders, such as stigma and potential increased risk of relapse by being exposed to eating disorder related content [78]. More broadly in the field of eating disorders, co-design appears to be gaining traction with the emergence of co-designed treatment approaches [79] and co-designed educational programs for the eating disorder workforce [80]. Given the lack of co-designed service models, it is a timely opportunity for the sector to invest and promote the involvement of people with lived experience in this area.

The recent introduction of a lived experience workforce to the multidisciplinary team in eating disorder service models is a significant advancement in the field. This workforce was part of only one telehealth service model included in this review [33], despite the documented recognition of the valuable contribution the lived experience workforce can make [81, 82]. Furthermore, the composition of the multidisciplinary team, and disciplines within it, varied significantly across eating disorder service models. Multidisciplinary teams are well established as best practice in eating disorders [4, 66, 83]. However, recent research in eating disorders has indicated that the value of multidisciplinary teams is enhanced with interprofessional collaboration. According to WHO, interprofessional collaborative practice occurs when members from a multidisciplinary team work together with people, their families and others to deliver high quality care [84]. A qualitative study conducted in Australia found that interprofessional collaborative practice provided to people with eating disorders improved their treatment satisfaction, engagement and outcomes [85]. Multidisciplinary teams of service models in this review predominantly included nurses, psychologists and psychiatrists, suggesting there is an opportunity to increase the diversity and expertise within teams, to offer holistic, comprehensive care that addresses the breadth of a person’s recovery needs.

Summary

In summary, service models for eating disorders are diverse. However, some important characteristics have been identified and discussed in this review. These characteristics are evident at a person level (person-centred care, involvement of family and carers), service model level (accessibility, the multidisciplinary team, co-design and lived experience contribution and identification and management of co-occurring conditions) and service system level (integration of service models). In the future, service models should be configured and designed with these fundamental characteristics in mind as the sector moves towards service models and a coherent integrated service system that is responsive and effective at providing the right support, at the right time, for people with eating disorders.

Limitations, future directions and clinical implications

Thorough description, development and evaluation of service models are complex [86], which could explain why it is not frequently completed and reported in publications. A key focus for future research should be thorough evaluation of existing service models to understand their characteristics and how service models promote access to high quality assessment, treatment and recovery outcomes.

Service models in this review predominantly targeted children and young people, while significantly less focused on people with longstanding eating disorders, or people with co-occurring conditions. This signals a need to develop and evaluate service models that can respond to the diverse clinical populations requiring assessment and treatment for an eating disorder.

Completing a scoping review ensured that the breadth and extent of the literature on eating disorder service models and their characteristics were captured and not restricted to primary research studies [21]. However, some service models may not have been captured in this review. While every effort was made to develop a comprehensive search strategy, there may be search terms or terminology omitted. For example, some terms may differ between different countries, regions and healthcare systems. Furthermore, some service models may have been missed because their design and characteristics have not been documented or made publicly available. There were no studies included in this review from Africa, Asia or South America, which is a limitation of the review.

Of the studies included in this review, many collected data retrospectively from existing medical records. This indicates that future research is needed to capture outcomes before, during and after a person accesses an eating disorder service model.

There are several clinical implications relating to the findings of this review. Firstly, the importance and value of clinicians adopting and advocating for person-centred care is paramount. This includes treatment that addresses not only the eating disorder, but also any co-occurring conditions. Interprofessional collaboration can also improve person-centred care by leveraging the diverse perspectives and expertise of a multidisciplinary team. As clinicians, it is crucial to advocate for the wider adoption of co-designed service models and to include people with lived experience as key members of the eating disorder workforce.

Conclusion

The results of this scoping review underscore the vast differences between eating disorder service models. These differences indicate the absence of agreed shared characteristics for service models in the sector. While acknowledging these differences, it is important to note that some key characteristics of eating disorder service models were identified. These were person-centred care, involvement of family and carers, co-design and lived experience contribution, a multidisciplinary team, accessibility, identification and management of co-occurring conditions and integration with the service system. In addition, and interestingly, this review uncovered innovative and, where a research component was included, effective service models that support recovery from eating disorders.

In this context of diversity, differences, and variations, this scoping review has contributed to the limited research available on eating disorder service models and their key characteristics.