The ALLIANCE study aims to understand and improve treatment relationships between staff and service users in inpatient eating disorder settings. It is NIHR funded and consists of 3 work packages (WP): WP1: Develop Programme Theory using realist literature review to identify how to effectively develop therapeutic alliance in inpatient treatment for Eating Disorders (ED), that maintains safety, whilst leading to improved treatment outcomes. WP1 consists of a literature review, using realist methods. Conducting this review will help us to understand existing evidence relating to treatment relationships in inpatient eating disorder setting. This WP is already underway. WP2: Test, expand and refine the theory via realist interviews with people who have lived experience of inpatient ED treatment and healthcare staff working in inpatient ED settings. Use the Programme Theory to develop actionable practice and policy recommendations to improve treatment outcomes, that can be operationalised (in a future study) into a complex intervention to improve alliance and treatment outcomes for ED. WP3: Codesign an animation (and supporting materials) with people with lived experience, to disseminate and communicate WP2 findings.

The ALLIANCE study aims to understand and improve treatment relationships between staff and service users in inpatient eating disorder settings. It is NIHR funded and consists of 3 work packages (WP): WP1: Develop Programme Theory using realist literature review to identify how to effectively develop therapeutic alliance in inpatient treatment for Eating Disorders (ED), that maintains safety, whilst leading to improved treatment outcomes. WP1 consists of a literature review, using realist methods. Conducting this review will help us to understand existing evidence relating to treatment relationships in inpatient eating disorder setting. This WP is already underway. WP2: Test, expand and refine the theory via realist interviews with people who have lived experience of inpatient ED treatment and healthcare staff working in inpatient ED settings. Use the Programme Theory to develop actionable practice and policy recommendations to improve treatment outcomes, that can be operationalised (in a future study) into a complex intervention to improve alliance and treatment outcomes for ED. WP3: Codesign an animation (and supporting materials) with people with lived experience, to disseminate and communicate WP2 findings

Eating disorders (EDs), such as anorexia nervosa, bulimia nervosa or binge eating disorder are serious psychiatric disorders marked by abnormal eating and/or weight-control behaviours. They have the highest mortality rates of any psychiatric disorder in the UK. EDs are devastating to quality of life and expensive to treat - making them 'disabling, deadly and costly'. EDs are also prevalent: it is estimated that 1 in 5 women may have an ED, although males, trans and non-binary individuals are also increasingly affected. Inpatient treatment is reserved for the most unwell and high-risk patients, admissions for EDs are increasing, but rates of recovery remain low and long-term recovery following inpatient treatment is inconsistent. Intensive inpatient treatment is expensive yet drop-out rates and disengagement are extremely high (up to 60%) . Given the prevalence of EDs, their high mortality rates and the lack of successful treatment options for acute patients, understanding why inpatient treatment is often ineffective, and developing novel responses to this, are desperately needed. Due to the risk of serious illness and death, inpatient treatment of EDs prioritises physical safety and weight gain over psychological needs. This is operationalised into policy that dictates statutory frameworks that have a direct bearing on how care is delivered (such as inspections delivered by the Care Quality Commission). This policy is enacted through interactions between front-line staff and service users leading to practices such as monitoring (e.g., all food and drink intake), withholding (e.g., not allowing people to have time alone) and surveillance (e.g., constantly watching people, including in the toilet). Whilst important from a safety perspective, these practices are distressing to both service users and staff and have been found to undermine trust and the development of positive therapeutic relationships. This is critical as poor therapeutic alliance has been found to be related to drop-out from treatment and is a significant predictor of longer-term recovery. The literature on ED in-patient treatment highlights that inpatient practices designed to prioritise physical safety have a negative impact on the development of therapeutic alliance. What is not currently understood is how inpatient treatment undermines the therapeutic alliance (that is, the mechanisms of change underlying the relationship between therapeutic alliance, disengagement from treatment and recovery from the ED) or how this could be improved in clinical practice. It is crucial to improve therapeutic alliance, as this has been associated with engagement in treatment and treatment outcomes. Relationships between healthcare staff and service users are impacted by a wide range of complex systemic factors, such as, policy and management systems, the anxiety of clinical responsibility for managing high levels of risk, and the dissonance caused by delivering practices are counter to personal values of collaborative care. The ALLIANCE project therefore also integrates a systems-based approach to understand what support systems are required to enable staff to effectively establish therapeutic relationships, whilst safely managing high levels of risk. Therapeutic alliance is related to treatment outcomes for EDs and is a significant predictor of disengagement from treatment. Research suggests that inpatient practices, erode service users' trust in, and engagement with, healthcare staff and ultimately undermine factors that are associated with personal recovery, such as identity, confidence, and agency. Additionally, it has been found that inpatient practices, via complex mechanisms (discussed below), can cause staff to participate in behaviour that can further result in the breakdown of therapeutic alliance. What is not currently known is how staff can deliver inpatient treatment, that maintains physical safety but does not do this at the cost of the therapeutic alliance. There is a range of research on experiences of inpatient treatment for EDs which has largely focused on anorexia nervosa and bulimia nervosa (due to the greater likelihood of these diagnosis leading to inpatient care). This research highlights the difficulty of staff and service users working together in inpatient settings which can be oppressive, stressful and (often) ineffective. Healthcare is ideally regarded as a partnership between professionals and their patients, but with regard to EDs, this dynamic has been understood as complex and often fraught. The importance of therapeutic alliance to recovery from ED stresses that honesty, empathy, commitment and a non judgement attitude are the linchpins of therapeutic alliance. As observed by Ramjan et al (2004), 'without a trusting, therapeutic … relationship, the treatment and recovery of people with anorexia can be unnecessarily impeded and prolonged'. However, achieving trust between patient and treatment provider is often especially challenging. Service user trust has been associated with being listened to, treated as a person/ individual rather than an illness or diagnosis, and perceiving that staff are knowledgeable, competent and well-meaning. Trust is also related to systemic factors (e.g., trust that healthcare organisations, their priorities and protocols, have their best interest in mind). ED service users have emphasised the importance to engagement of feeling that they are actively involved in treatment. Help-seeking behaviours and engagement are also highly influenced by trust in staff members and the service user's perception that they are receiving person-centred care. Despite existing clinical research which suggests that treatment rationales should always be clear for service users, research indicates that they are routinely denied such explanations in ways which impact trusting relationships. As identified by a service user in the Holmes (2021) study "There wasn't often a rationale. It was often just "you're ill" or "you're angry" ... it made me feel untrusted, and like I wasn't worth listening to". This in turn engenders antagonism toward staff and treatment as it fosters the perception that some rules or restrictions are punitive rather than being driven by a safety rationale, which can also result in disengagement from treatment. Graham et al (2020) conducted a meta-ethnography of 37 studies of staff perspectives of working in ED services, which sheds light on why therapeutic alliance breaks down from the staff perspective. This found a primary theme of "caring without curing", which identified that a combination of chronicity of illness, treatment refusal, and systemic challenges confronted staff with the fact that their care and treatment might be unwanted and/or insufficient. The result was a mismatch between the desire to help and the reality of what they were able to offer. Together, these factors resulted in expressions of moral distress as staff experienced a conflict between systemic practices (such as surveillance) that were perceived as a violation of their core values. This moral distress manifested as both self criticism and the development of dehumanising narratives about service users. Staff were found to adopt coping strategies to defend against dissonance such as avoiding, blaming, and battling with service users. Poor treatment outcomes may also arise from systemic frameworks involving rigid systems of monitoring and surveillance, which deny patient autonomy and control, and which are enacted in the relationship between staff and service users, replicating and intensifying some of the structures of ED itself. In terms of current practice, relapse rates remain high, and inpatient stays may result in 'poorer psychosocial outcomes than community-based treatment in the longer term. The experience of inpatient treatment itself, including the relationships with healthcare staff, may actually work against the wider goal of recovery from an ED. Additionally, research has shown that the experience of restrictive inpatient practices, and the subsequent mistrust of services, is likely to persist following discharge from inpatient units, as it can undermine trust in health services in general. This can lead to a lack of engagement with community services following discharge, which may in turn lead to deterioration in the community and subsequent readmission to the inpatient context. Achieving good therapeutic alliance is likely to overcome some of these problems as factors found to be associated with the experience of the therapeutic relationship, such as sense of identity, agency and social connection, are also domains of personal recovery from ED. Additionally, self compassion and self-acceptance have been found to be critical domains of personal recovery from ED that can be fostered by a compassionate, kind and non judgemental therapeutic relationship. We have conducted an initial scoping search that has identified a range of literature that starts to explain the observed relationship between therapeutic alliance and treatment outcomes for ED. However, a systematic understanding of the mechanisms of change underlying this, and how this can be improved in practice, is not currently available to service providers, policymakers and researchers. Additionally, to date, the research has largely lacked a systems-based perspective, to identify how factors such as healthcare policy, local risk management protocols, and clinical supervision structures impact on therapeutic alliance. Lastly, patients and health professionals tend to be studied separately. This arguably reflects a wider 'research-practice' gap in the field of understanding how patient and staff experience, sense-making and behaviour interact to develop of undermine therapeutic alliance and effective inpatient treatment for EDs.

Work Package 1 will develop programme theory using realist methods, and informed by a systems-based approach. Evidence will be collected from peer reviewed research, and used to develop an initial programme theory. Further iterative searches will be conducted, including grey literature and policy documents. Expert stakeholder will also contribute to programme theory development. A systems approach will be used to map personal, service, organisational, local and national level contexts that influence treatment relationships and subsequent outcomes. Review stage 1 output will be the initial programme theory. WP2: The programme theory will be tested, expanded and refined using iterative realist interviews. We will conduct 10 service user, and 10 staff realist interviews across our recruitment pathways (total n = 20). This sample size is considered adequate to ensure enough data is collected to test, expand and refine the programme theory. The literature synthesis will explore the impact of attributes such as gender, ethnicity, neurodiversity and duration of illness. We will recruit in line with testing emerging programme theory relating to these factors. To ensure diversity, there will be targeted recruitment and purposive sampling from the clinical sites and from the charity and third sector contexts. Staff interviews will include staff from across different levels of the system (e.g., nurses, including agency staff as suggested by lived experience collaborators, clinical psychologists, psychiatrists, managers, and the board). Interview topic guides will be developed from the initial programme theory, with input from key stakeholders. Interviewees will be presented with observations identified from our programme theory, and asked if they agree or disagree, as well as to expand on the issue based on their own experiences. These interviews can provide new evidence not already identified in the literature. WP3: We will co-design a video animation with accompanying material to communicate the programme theory and recommendations. The animation will be co-designed with service users and staff interviewed in WP2 and the lived experience collaborators. The animation will bring to life the findings of the realist synthesis, foregrounding the lived experiences of service users and staff by representing scenes where trust and therapeutic alliance is enhanced or undermined in ED in-patient treatment. This will include the key recommendations for practice to improve therapeutic relationships between healthcare staff and service users. To develop the animation, we will recruit service-users and staff (n=8-10, total) for co-design groups. We will conduct two 1.5-2 hour co-design workshops. The workshop will initially present the findings of the realist synthesis, then the recommendations derived from the programme theory. We will explore the needs of both service users and staff in relation to the animation and the accompanying materials. Participants will identify key issues that are and are not working in ED inpatient treatment and how this could be improved in practice. These will then be prioritised (as to what should be included in the animation). After the workshop we will collate the suggestions and share them with participants via email. The second co-design workshop will be organised online via secure video-call. Participants from both in person groups will attend together. The workshop will generate, refine and select ideas from previous workshops for scenes and topics to be included in the animation, as well as for a preliminary storyboard. After the workshop, a storyboard created by designers will be shared with participants via email for feedback. To finalise practice and policy recommendations we will run two feedback groups to present the animation to stakeholders, including healthcare staff and lived experience advisors and collaborators. The research team will use this stakeholder discussion to finalise recommendations. We will also collect stakeholder feedback on the animations in these events through a short questionnaire both on the materials and their use. As the content of the questionnaire is dependent upon the outcome of WP1 & 2, this will be developed later on in the study

Inclusion Criteria:

• Clinical population: All service users (18 years and over), with capacity to give informed consent, admitted within the previous 5 years to inpatient ED units will be eligible for inclusion in the study Staff: All staff with clinical roles in inpatient ED units.

Exclusion Criteria:

• No additional exclusion criteria

Adults who have a prior diagnosis of ED and who have lived experience of being admitted onto an inpatient ED ward. Participants are identified from community ED teams and advertised through national ED support charities.

Healthcare professional participants who have a professional experience of working on inpatient ED wards will be approached through participating NHS sites and national ED charities.

Anonymised qualitative data about service experiences and experiences of individual practitioners will not be shared publicly by agreement with ethics board due to sensitive nature of responses.