MUMMIEBODIES is a research study when women with eating disorders encounter pregnancy, childbirth and postpartum. These are some condensed months that represent major challenges for this group. There are both a scientific and clinical basis for the mother's mental health to influence care and interaction with the child. In the case of eating disorders, there is also a risk of harmful consequences for the fetus as well as for more complicated pregnancies and births. This is a very important field of knowledge, but also a field about which there is far too little scientific knowledge. There is too little scientific knowledge about how women with eating disorders physically and mentally relate to their bodies and food during and after pregnancy. There is a great need for research that promotes expertise in how to help and meet women with eating disorders in these important phases of life. The aim of the study is to bring out the users' experiences. The investigators interview face-to-face pregnant women and mothers with eating disorders about their subjective experiences both during pregnancy and during childbirth. Recruitment of participants will be via health stations. The investigators will identify the course, experiences and solutions when women with eating disorders encounter pregnancy, childbirth and postpartum. The investigators focus on three themes: 1) experience of course and change, 2) experience of emotional, cognitive and relational core experiences, and 3) women's own perceptions of what is the best help. The three themes have in common that they are fundamental for later development of help for this vulnerable group. The investigators have a clear idea that some of the best preventive work can be done before life really starts.

The epistemological approach is empirical realism (not constructivist) where the investigators assume that experiences are real and that the informants is an expert on her own experience. The strategy is to utilize the informant's experience as a source of knowledge. The point of view is the informants's, where the investigators try to formulate, categorize, characterize and analyze dimensions in eating problems in the face of pregnancy and childbirth as the informants sees it.

There are three reasons for this choice: (1) The woman's experience is an underused source of knowledge in this area. (2) Therapeutic and indicated preventive measures are dependent on an alliance with the woman. Alliance formation is normally based on the woman's experience. (3) The woman's subjective presentation of the course may deviate from what an external observer would have reported. The measures the women themselves launch may run counter to what research and experienced clinicians recommend. But objective measures that neglect the user's own experience are unlikely to have a particularly lasting effect.

The issues require two qualitative research designs. For the first problem, the investigators will use the method Grounded Theory (25, 26) because the purpose is to model processes. Grounded Theory is a method for developing concepts and theoretical understandings that are rooted in data through theoretical sampling, identification of similarities and differences between categories, open coding and focused coding. The aim is to construct a theoretical model for characteristic processes, anchored in empirical data from informants in the study. For the other two issues, the method Interpretative Phenomenological Analysis (IPA) will be used. IPA is a qualitative research method that is particularly suitable for psychological topics (27). The aim of IPA is to analyze and systematize the informants' experience and understanding of a phenomenon. Through this method, overall meaningful themes can be identified. At the same time, one can bring out nuances and variation in the material and look for any dimensions in the experiences.

Informants. The investigators will use a heterogeneous sample. This has several causes. We currently know too little scientific to identify one scientifically and clinically interesting homogeneous group. Our clinical experience indicates that there can be great variation between different women in this life situation. In this round, the investigators seek to identify different experiences one may have. Varied diagnoses have been chosen because the investigators want to describe different courses, because these diagnoses are often unstable, and because the investigators want to identify possible measures with different experiences and courses. Such diversity has great clinical relevance and has not been previously described scientifically. In practice, it facilitates the recruitment of informants. A disadvantage of using a heterogeneous sample is that each individual informant becomes relatively important. This is a small problem as long as the purpose is to identify different experiences. The limitation lies in the cases where a supposedly important experience is described in only one or very few informants and at the same time there is basically little depth and nuance in the descriptions. This is compensated to some extent by very high demands on the quality of the way the investigators interview the informants. The investigators define the group of informants in pregnancy and follow longitudinally with two measurement times: in pregnancy and after birth.

The number of informants for problem 1 will be determined on the basis of a saturation criterion where we include informants until after at least ten interviews, three subsequent interviews follow which do not add anything significantly new to the answer to the problems (30). Based on previous experience with similar surveys, the number of informants for each issue is expected to be somewhere around 20 informants, ie about 40 informants. From these, a strategic choice will be made on issues 2 and 3. Strategic choice can mean looking for differences in order to obtain different prototype experiences. Recruitment of informants will take place continuously from the start of the project via a network.The project period has been extended to four years to get a sufficient number of informants.

Before further inclusion in the study, eating disorders are mapped with a revised version of the Eating Disorders Examination Questionnaire, EDE-Q adapted to pregnant women and women immediately after birth (16). The form is scored by the research fellow before further inclusion. EDE-Q is a self-completion form based on the "Eating Disorder Examination" (EDE) interview. EDE-Q measures the core symptoms of eating disorders. EDE-Q adapted to pregnant women and women immediately after birth is based on factor analysis. If the informant satisfies diagnostic criteria for a cut-off equivalent to ≥ 2.8 for an eating disorder, we will conduct a researcher-based interview, Eating Disorder Examination (EDE), for diagnosing the informant, and obtain standard background information. EDE is designed to assess the current condition, as well as to generate an operationalized defined eating disorder diagnosis. The interview is conducted after the experience interview.

The informants' background will be thoroughly documented. Background information will include data on age, gender, diagnosis, type of eating disorder, weight before, during and after pregnancy, onset of illness, first contact with the support system, relationships, first-time or multiple births. Background information is filled in by interviewer after the experience interview.

Information about the issues is collected through a semi-structured, interactive face-to-face interview. The aim of the interviews is to get descriptions as precise and close to the participants' own experiences as possible. A semi-structured interview guide is designed in advance. The Interview Guide is developed in close collaboration with user representatives in the reference group and ensures that relevant input from the reference group is covered in the interview. The interview guide contains all the topics we want the interview to cover in advance. The interview guide is primarily a thematic guide and aims to prepare interviews as an instrument. It is also used as a summary and checklist for interviews at the end of the interview to ensure that all relevant topics are touched upon. The informants are thoroughly informed in advance about what the interview will be about and why, and how it will proceed. The better informed the informant is in advance, the better founded the voluntary participation. The actual interview takes the form of improvisation on the topics as it is natural from the dynamics that develop in the conversation. The informant can introduce their own relevant topics independently of the interview guide. The supervisors have published a double-digit number of scientific articles and trained thousands of clinicians and researchers with this method, which is informally called the "Experience Interview" developed by Arne Holte (19, 21-23). The interviews are quality assured by a professional who is not directly involved, continuously listening through randomly selected sections of randomly selected interviews and providing guiding comments. The purpose of this is to ensure that consistency and validity in the interview format is maintained.

Data analysis. All interviews are transcribed verbatim, stored, organized, coded and analyzed in the electronic software NVivo version 11. Data for problem 1 will be analyzed according to the Grounded Theory method (25, 26). The analysis involves eight phases such as categories, theoretical sampling, identification of similarities and differences between categories, open coding, focused coding, negative case analysis, theoretical sensitivity, theoretical sampling, theoretical saturation, memo writing. The analysis will be regularly discussed in the research group and continuously monitored by the main supervisor to ensure reliability and credibility. Data for issues 2 and 3 will be analyzed according to the IPA method (27). Repeated listening and reading of each interview is carried out to get an overview and context. Each text is then explored using open thematic coding according to the "bottom-up principle". Each interview is divided into text excerpts that are given a definition based on the content of meaning and distributed in a tree structure. The definitions of potential constructions are finally tested against the text by confirmation and selective coding according to the "top-down principle".

Validation of the results will take place through focus group interviews with informants from the survey with varying stories. The number and selection of focus group participants depends on the results. The participants will be briefly presented in advance with the most important results in writing. The results will then be presented orally to the group. The group will be asked to comment on the presentations. The purpose is to investigate whether the informants recognize their experiences as they have been analyzed, presented and illustrated, and possibly adjust and correct this. We make audio recordings of the focus group interview so that a person who is not directly involved in the project will listen through and check the conclusions the research fellow makes from the focus group interview.

Procedure. The informants will be asked orally and in writing for consent to participate as described above: EDE-Q, EDE interview, "Experience interview" with audio recording, summary check according to interview guide and background information. Based on previous experience, the experience interview is expected to take 90-120 minutes. Total time spent per informant is calculated for a total of 140-180 minutes of interview hours (EDE interview, experience interview and background interview). The text is printed verbatim and stored electronically in NVivo 11. Audio recordings are deleted immediately after electronic storage. The electronically stored text is deidentified so that names and other clear characteristics cannot be recognized. After the data has been analyzed and before the completion of articles, validation is carried out using focus group methodology. All data is deleted when the survey is completed in accordance with the agreement with REK.

Inclusion criteria:

• Woman
• age between 20-40 years
• Diagnostic diagnosis anorexia nervosa or bulimia nervosa
• Must be pregnant
• Must have been interviewed in pregnancy and be in labor/post partum.

Exclusion Criteria:

• Ongoing psychosis
• Drug addiction
• Organic brain damage / disease.

Women experience of pregnancy and post partum period.

Women with eating disorders who are pregnant and in labor (the child's first year of life). In this round, we seek to identify different experiences one may have.

Pregnant women reported before participating the study a history of self-reported eating disorder for anorexia nervosa, bulimia nervosa, overeating disorder or unspecified eating disorder

14 months. Two time points for each participant: During pregnancy and between 4-6 months after birth. From two to one group of participants that were interviewed at two time points: Pregnancy and postpartum. There were sufficint pregnant women that were recruited, and this made it possible to follow the same group of women longitudinal. No one were recrutied to the other group at this time.

There were mad a contingency plan for all 24 participants in case of needed specialized from pregnancy to after birth at the two time points.

Representative of women with ED. Past history of ED, including diangosis severity, as well as treatment history was self-reported with no verification with health records. Multi-diagnostic approach. The same person have conducted the semi-structured interview and EDE clinical interviews that might have influenced the respons or outcome of the diagnosis. Different interview points during pregnancy. Lack of corroborating information.