Self-admission is a novel treatment tool whereby patients who are well-known to a service who have high previous utilization of health care are offered the possibility of self-admission to the inpatient ward for up to seven days without having their motive for admission questioned. Patients are free to admit themselves because of deteriorating mental health, acute stress, lack of structure in their everyday life, loneliness, boredom, or any other reason. The patients decide when they want to admit themselves and can discharge themselves at any time. The purpose behind the self-admission model is to increase the availability of inpatient care for severely ill patients, to avoid stressful and possibly destructive visits to the emergency service, and to decrease total inpatient care utilization. Patients offered a contract for self-admission usually have a history of repeated and prolonged hospitalizations. By encouraging them to monitor their own mental health status and allowing them to seek help swiftly when they are feeling poorly, the delay from first signs of deterioration to admission can be minimized and full-blown relapse can be avoided, ultimately reducing the total time spent in hospital. Until now, projects of self-admission have mainly targeted patients with long-standing psychotic disorders, such as schizophrenia, or bipolar disorder. Starting in August 2014, a four-year clinical project at the Stockholm Centre for Eating Disorders began offering self-admission to patients with severe and enduring eating disorders. The purpose of this study is to determine whether this model is viable in a specialized eating disorders treatment setting, if it does lead to increased patient participation and agency and a reduction of the total time spent hospitalized for this particular patient group, and if it is cost-effective.

There are approximately 100 000 individuals in Sweden who suffer from an eating disorder such as anorexia nervosa or bulimia nervosa. The course of these disorders is often prolonged and can lead to enduring disability for many years. Some patients require lengthy periods of inpatient care, and relapse after discharge is common. Unfortunately, there is still little evidence concerning the optimal model of inpatient care for patients suffering from a severe and enduring eating disorder (SE-ED) and unlike psychiatry in general, the hospitalization rate for SE-ED patients has not been markedly reduced in the past decades.

In Norway, a novel treatment concept for psychiatric patients has been in place for about ten years: self-admission. In self-admission, patients who are well-known to a service who have high previous utilization of health care are offered the possibility of self-admission to the inpatient ward for up to seven days without having their motive for admission questioned. Patients are free to admit themselves because of deteriorating mental health, acute stress, lack of structure in their everyday life, loneliness, boredom, or any other reason. The patients decide when they want to admit themselves and can discharge themselves at any time. The purpose behind the self-admission model is to increase the availability of inpatient care for severely ill patients, to avoid stressful and possibly destructive visits to the emergency service, and to decrease total inpatient care utilization. Patients offered a contract for self-admission usually have a history of repeated and prolonged hospitalizations. By encouraging them to monitor their own mental health status and allowing them to seek help swiftly when they are feeling poorly, the delay from first signs of deterioration to admission can be minimized and full-blown relapse can be avoided, ultimately reducing the total time spent in hospital.

The Norwegian results are very promising. Self-admission has led to increased patient participation and compliance, strengthened the patients' abilities to handle their symptoms and their everyday life, and clearly reduced the total time spent in inpatient care. The total time spent in non-voluntary inpatient care and the number of violent or threatening episodes at the ward was also reduced.

Until now, projects of self-admission have mainly targeted patients with long-standing psychotic disorders, such as schizophrenia, or bipolar disorder. Starting in August 2014, a three-year clinical project at the Stockholm Centre for Eating Disorders (SCÄ) began offering self-admission to patients with SE-ED (mainly anorexia nervosa, which is the most common diagnosis at the SCÄ inpatient ward). The overall goal of the self-admission project at SCÄ is to create a high-quality treatment approach that is also cost-effective. The relatively brief nature of these self-admissions will of course not allow for full-scale weight restoration or achievement of other long-term treatment goals. Instead, this admission concept is probably best used as booster opportunity or for providing a short respite from stressful life situations where the risk of relapse is elevated. In this way, offering brief inpatient treatment as a preemptory tool to be made use of rather than as a "necessary evil" when severe mental breakdown has already occurred can increase patient participation, agency, self-awareness, and autonomy. Furthermore, bypassing the traditional clinician-led decision-making in evaluating whether inpatient treatment should occur could also foster reflection on subjective motives and barriers in asking for help. Here, self-admission promotes partnership between patient and clinicians instead of distrust. The approach could enhance self-knowledge and motivation and, by extension, foster skills and strategies that can influence the course and prognosis of the disorder. It has been shown that patients suffering from anorexia nervosa do request inpatient treatment after the initial period of reinforcement inherent to the disorder has passed and that this is part of a personal process of increased awareness. As they gradually become more aware of their loss of control over eating disorder symptoms and of their ambivalence towards change, they also become less reluctant to ask for assistance in regaining control. For patients with long-standing illness this process may take several years. It is hypothesized that such a process could be solidified and hastened by offering patients a concrete tool in the form of self-admission, by which they are also required to practice self-monitoring and develop agency.

The current program is the first of its kind to offer self-admission to patients with an eating disorder. If the model proves to be successful, it can hopefully be established as a standard tool in the treatment of patients with SE-ED. However, a health economics analysis is critical to determining whether the model merits broader implementation. The potential health economic benefits are evident: lower health care costs for the specific patient group, increased availability of much sought after hospital beds for "regular" patients in the queue to the ward, improving patients' chances of maintaining everyday activities with functioning family life and work, lowered rates of sick leave, etc.

For a preliminary duration of three years, two out of ten beds at the inpatient ward at SCÄ will be reserved for self-admission; i.e. they will not be occupied by regularly admitted patients but available only for patients with a self-admission contract, who may at will admit themselves for a maximum of seven days at a time. Participants are also free to discharge themselves at will. Upon arrival at the ward, they will see a nurse for a brief discussion of current status and treatment goals but they will not be assessed by a medical doctor at this point (unless the ward nurse deems this as necessary due to acute medical impairment, suicidal communication etc.). There will be no restriction in the total number of admission episodes and when regarded as necessary patients with a contract can also be admitted for longer periods of time for a regular admission. The aim is to include a total of 20-40 patients in this program. Inclusion criteria are simple: to be offered a contract, patients must be well-known to the clinic, they must have received inpatient treatment at the ward at least once in the previous three years, and they must be willing and able to follow the basic treatment structure at the ward. Exclusion criteria are active suicidal or self-injurious behavior and active substance use disorder; however, no body mass index (BMI) criteria will be applied. In effect, this means that patients considered for inclusion suffer from SE-ED and that many of them will have been hospitalized on several previous occasions. Most patients in the program will probably present with a clinical picture of treatment-refractory anorexia nervosa but it is not ruled out that patients with severe bulimia nervosa may be able to make use of the concept as well.

In order to evaluate the clinical self-admission project thoroughly, a broad number of methods will be employed. The following data will be collected:

At inclusion:

• RAND-36 (the RAND Corporation version of SF-36 (Short Form Health Survey));quality of life questionnaire.
• EQ-5D (EuroQol Health Questionnaire); health experience questionnaire.
• Two questions on inpatient care experiences and expectations of project participation.

At six months:

• Semi-structured recorded patient interview concerning subjective experiences of self-admission. The interview manual, designed specifically for this study, consists of a set of open-ended questions on several areas of interest regarding participant experiences. The interview is conducted as an "open" discussion of these topics and the interviewer continuously checks that all these areas are covered. The recorded interviews are then transcribed and computer software NVivo is used to conduct a qualitative content analysis, grouping the interview material into similar nodes and themes.

At one year:

• RAND-36.
• EQ-5D.
• Number of days spent in inpatient care (at SCÄ as well as at other clinics).
• Number of days spent in involuntary inpatient care (at SCÄ as well as at other clinics).
• Number of outpatient visits (at SCÄ as well as at other clinics).
• Diagnoses giving rise to inpatient care/outpatient visits.
• Current medication prescriptions.
• Number of days with social security benefits (sick leave from work/school, disability pension and similar models).

At three years:

• Same data as collected at one year.
• Semi-structured recorded patient interview concerning subjective experiences of self-admission.

Data on health care consumption (other than at SCÄ), diagnoses, medication, and social security benefits will be collected from the National Board of Health and Welfare and from Statistics Sweden, respectively.

In addition, all participants will be assessed with the following psychiatric scales/questionnaires at inclusion and at one- and three-year follow-up:

• SCID-5-CV (Structured Clinical Interview for DSM-5 Disorders - Clinician Version); assessment of psychiatric co-morbidity.
• SEDI (Structured Eating Disorder Interview); assessment of the occurrence of a clinical eating disorder.
• GAF (Global Assessment of Functioning); assessment of global function, the previous axis V in DSM-IV.
• PSRS (Psychiatric Status Rating Scale for Eating Disorders); assessment of eating disorder severity.
• EDE-Q (Eating Disorder Examination Questionnaire); self-assessment of central diagnostic components of an eating disorder.
• CIA (Clinical Impairment Assessment); assessment of secondary disability due to an eating disorder.
• CPRS (Comprehensive Psychiatric Rating Scale); self-assessment of psychiatric symptoms.
• SASB (Structural Analysis of Social Behavior); self-assessment of self-image.

There is no randomization; instead, every single patient will function as her or his own control, in that their individual data on health care consumption etc. during the study period will be compared with their retrospective data for the same amount of time (thus, up to three years) proceeding inclusion in the self-admission project. Since all patients in the study are well known to the clinic, retrospective data are ready available in the patient files.

Regarding the health economics evaluation, a cost-benefit analysis will be performed based on the collected data on quality of life, disease progression, health care consumption, medication, social security benefits, etc. Changes in availability of hospital beds at the inpatient ward at SCÄ for "regular" patients without a self-admission contract will also be analyzed. Hypothetically, by reserving two beds for self-admission and thus reducing the number of regular beds, overall availability may also be negatively affected. On the other hand, by establishing a "fast lane" for high consumers of inpatient care and thereby removing them from the general admission queue, waiting lists may ultimately be reduced.

An estimation of statistical power, assuming a standard deviation based of the effects sizes found in a previous review of the Norwegian data (multiplied by a factor of 3 in order to achieve a more conservative estimate), suggests that power is well above 80% (approx. 98%). As is customary, no separate power estimation has been made for the health economics evaluation at this point as it will largely be based on the clinical effect data.

The time plan for the clinical self-admission project is as follows:

• Autumn 2014: The clinical pilot project is started and the first patients are included.
• 2015: Patient interviews at six months. One-year follow-up of the first patients in the project.
• 2016: Half-time evaluation at the end of the year.
• 2018: The clinical pilot project ends and a temporary clinical evaluation is performed.

Project participants will, however, be included successively during this entire period and thus, the research part of the project (data collection and analysis) will proceed for an additional three years, i.e. until 2020:

• 2017: Analysis of combined 12-month quantitative data. Qualitative content analysis of 6-month interviews.
• 2018: Cost-benefit analysis of 12-month data.
• 2019-2020: Analysis of combined quantitative 36-month data. Qualitative content analysis of 36-month interviews. Cost-benefit analysis of 36-month data.

This study will produce four separate sub-studies:

1. Qualitative interview study about patient experiences in the self-admission program at 6 months
2. Analysis of changes in health care consumption patterns, eating disorder severity and quality of life at 12 months
3. Health economics analysis at 12 months
4. Analysis of changes in health care consumption patterns, eating disorder severity and quality of life, including health economics analysis, at 36 months

Inclusion criteria:

• Having been admitted to the inpatient treatment ward for adults at the Stockholm Centre for Eating Disorders at least once before.
• Being able to follow the routines and rules and partake in the treatment at the inpatient ward.

Exclusion criteria:

• Active and untreated substance use disorder.
• Active suicidal ideation.

No BMI criteria will be applied.