Family-based treatment at home in adolescents with eating disorders and co-occurring mental health conditions: rationale and study design of a mixed methods trial | BMC Psychiatry

Study design

To evaluate the effectiveness of FBT-H, we implemented a mixed-method approach combining a single-case design and a qualitative study. In the single-case design, 10 cases are intensively monitored throughout the FBT-H process, documenting interventions and their effects [19]. The qualitative study complements this by gathering insights from adolescents, parents, and professionals about their experiences with FBT-H using structured interviews conducted separately with each group post-treatment to explore key topics. Together, these methods provide a comprehensive understanding of the impact of FBT-H.

Participants

Participants are adolescents with their families recruited from different locations in Karakter, a large mental health organization for children and adolescents in the Netherlands. Families are residents of the Twente and Salland regions in the eastern Netherlands who are receiving FBT-H.

The inclusion criteria are as follows: (1) adolescents aged 12 to 18 years; (2) meeting the criteria for AN or other specified feeding or eating disorder (OSFED), diagnosed with the Structured Clinical Interview for DSM-5 Disorders – module 10; (3) presence of co-occurring mental health conditions, determined by a clinical psychologist or child and adolescent psychiatrist; (4) sufficient command of the Dutch language by adolescents and parents; and (5) the family supports the treatment as a supportive system and is open to participating in the research.

Exclusion criteria are: (1) acute suicidal behaviour requiring hospitalization within the 4 weeks prior to assessment for this study; (2) cognitive limitations (IQ < 70); (3) participation of a sibling in the current study (to prevent transfer of therapy effects between siblings); and (4) the adolescent does not receive tube feeding at the start of FBT-H and/or is not hospitalized due to an eating disorder at the start of treatment.

Procedures

The recruitment process for the study began in November 2023 at Karakter and is expected to be completed at the end of July 2025. A multidisciplinary team comprising a licensed psychologist or psychiatrist and a family therapist conducts screenings on all patients referred for ED treatment to assess the inclusion/exclusion criteria. In cases where uncertainty exists regarding the fulfillment of these criteria, additional assessments are performed. Families that meet the criteria for participation are invited to an informed consent meeting where they receive comprehensive information about the research and a study information sheet. They are given a two-week period to contemplate their decision to participate voluntarily in the study and provide written consent, and they are offered a consultation with a co-worker with lived experiences in the ED. Due to the severity of the illness, it is preferable to start FBT-H as soon as possible, and upon agreement to participate in the study, treatment will be started immediately by two FBT-H practitioners. Families that decline participation still have access to the FBT-H.

In addition, postal letters are dispatched to general practitioners to notify them about the study. The study started in 2023 and is expected to be completed in 2026 (See Fig. 1).

Interventions

FBT

FBT is a structured treatment lasting 6–12 months, typically in an outpatient setting, with weekly sessions that gradually decrease in frequency over time [10]. It empowers parents to manage their child’s recovery from the ED. The first phase focuses on weight restoration, with parents responsible for all eating decisions and restricting physical activity to reduce ED’s influence. In the second phase, eating responsibility gradually shifts back to the adolescent. The final phase supports healthy adolescent development as ED symptoms subside. FBT emphasizes parents’ dual role of setting firm boundaries against the ED while remaining compassionate, guided by five core principles: the therapist holds an agnostic view of the ED cause, a non-authoritarian approach, parental empowerment, separating the ED from the child, and a pragmatic approach to treatment [20]. The FBT assumes that as weight normalizes, ED symptoms, anxiety, mood issues, and family dynamics improve. Unlike other treatments, FBT positions parents as central to recovery, involving them directly in meal supervision and behaviour relearning.

FBT-H

The FBT-H adapts FBT to a home setting, with treatment led by family counselors and psychologists as part of a multidisciplinary team, including therapists, psychologists, a psychiatrist and a nurse specialist, all trained in FBT by one of the authors (DLG). Weekly supervision sessions are held with an experienced family therapist, and monthly online supervision with an international FBT expert (SG, see Acknowledgements). Initially, two FBT practitioners conduct FBT-H, visiting the family twice weekly for 45–60 min. This dual approach in phase one supports the management of complex family dynamics, with a single practitioner continuing as treatment progresses. The first weekly session follows FBT guidelines, while the second provides practical support for parents in applying these strategies.

Measurements: single-case design

The study parameters are outlined in Table 1. To gather data for the research questions, adolescents are expected to spend approximately 180 min on baseline questionnaires and semi-structured interviews (T0), while parents will spend approximately 30 min. For follow-up assessments (T1 to T4), adolescents and parents each spent about 30 min completing the questionnaires. Throughout the FBT-H, weekly evaluations will be conducted for all participants to monitor weight, general well-being, and therapeutic alliance.

Baseline

Somatic data – T0

The somatic data of the adolescents are collected by a professional, including weight, height, medication status, and menstrual status. An overview of hospital admissions, if applicable, is also provided. Somatic data will be collected at T0 and T4 in 5 min.

Diagnosing psychological problems; scid-5-junior – T0

The SCID-5-junior is a semi-structured interview aimed at diagnosing DSM-5 disorders [21]. It consists of 17 modules with questions about the feelings, thoughts, and behaviours of adolescents regarding certain parts of the DSM-5 (such as depression, anxiety disorders, or eating disorders). Each module indicates whether an adolescent meets the diagnostic criteria for the corresponding disorder. The assessment utilizes dichotomous (yes/no) responses regarding the presence of symptoms associated with specific disorders. Preliminary results of studies investigating the psychometric properties of the SCID-5 junior show adequate parent–child agreements (except for externalizing problems and insomnia) and adequate convergent validity (especially in clinical samples) [21]. This interview is conducted in its entirety at T0 and has a duration of ca. 90 min.

Depression and anxiety; depression, anxiety and stress scale (DASS-21) – T0, T4

The DASS-21 is a questionnaire that serves as a screening tool for general mental complaints related to depression, anxiety, and stress ([22]; translated by de Beurs et al., 2001). The DASS-21 consists of three scales (depression, anxiety, and stress) with seven questions each. Scores above 10 on the depression scale indicate depressive symptoms, scores above 8 on the anxiety scale indicate anxiety symptoms, and scores above 15 on the stress scale indicate stress symptoms. Coker and colleagues indicated that the reliability of the DASS-21 is good, with Cronbach’s alpha values of 0.81, 0.89, and 0.78, respectively [23]. The adolescents are asked to fill out this questionnaire twice: at T0 and T4, and it requires 10 min to complete.

Parenting family questionnaire (Gezinsvragenlijst; GVL) – T0

The GVL is a questionnaire used to evaluate the quality of parenting conditions for children between the ages of 4 and 18 years. It consists of 45 statements that assess five different aspects: responsiveness, communication, organization, partner relationship, and social network [24]. Each statement can be answered on a 5-point scale, varying from totally disagree to totally agree. The reliability of the GVL has been demonstrated to be good (van der Ploeg & Scholte, 2008). Parents are asked to complete this questionnaire once at the initial assessment point (T0), which takes 15 min.

Primary outcomes

Weight – each session

The weight of the adolescent in kilograms (kg) is measured weekly to monitor progress. The weight is measured by an FBT-H practitioner in an FBT-H session. The weight is measured in the same way every week (e.g., in underwear). This is a standard measurement associated with the FBT intervention and takes approximately 5 min.

Secondary outcomes

Outcome rating scale (ORS) – each session

The ORS is a questionnaire in which participants indicate how they are doing in four components of their lives: personal, interpersonal, social, and general ([25]; translated by Hafkenscheid et al., 2003). These four components form the four items in this questionnaire, which can be answered on a scale from 0 to 10. There is a child and adult version of the ORS. A score of more to the left (0) indicates a negative score on the corresponding component, and a score of more to the right [10] indicates a more positive one. The internal consistency, measured by Cronbach’s alpha, is 0.97 [26]. Adolescents and their parents participating in this study will be asked to fill out this questionnaire at the beginning of each session, which requires approximately 1 min.

Session outcome rating scale (SRS) – each session

The SRS is a questionnaire in which adolescents express how they have experienced a session [27]; translated by Hafkenscheid et al., 2003). The four items covered are relationship/contact, goals and topics, approach and method, and overall satisfaction. The SRS has the same setup as the previously mentioned ORS. The internal consistency, measured by Cronbach’s alpha, is 0.88 [27]. Adolescents and parents participating in this study will be asked to fill out this questionnaire at the end of each treatment session, which takes approximately 1 min.

Assessing eating disorders; eating disorder examination questionnaire (EDE-Q) – T0 to T4

The EDE-Q is a self-report questionnaire used to assess the behaviours and characteristics of ED [28]; translated by van Furth, 2001). The EDE-Q consists of a global scale (22 items) and four subscales: restraint, eating concern, weight concern, and shape concern. Each question can be answered on a scale from 1 to 6 or with a yes/no response. Satisfactory levels of internal consistency were observed [29]. Adolescents in this study will complete the EDE-Q five times (T0 to T4). It takes 10 min to complete the questionnaire.

Quality of life; kidscreen-27 – T0 to T4

Kidscreen-27 is a questionnaire measuring the quality of life of children in five dimensions: physical well-being, mental well-being, parent relations and autonomy, social support (from peers), and school situation [30]. It consists of 27 questions, each of which has five answer options. Cronbach’s alpha for all scales varies between average and good [30]. Adolescents and their parents are asked to complete this questionnaire five times (T0 to T4) and it requires 10 min to complete.

Parent-child interaction; parent-child interaction questionnaire-revised (Ouder kind interactie Vragenlijst; OKIV-R) – T0 to T4

The OKIV-R consists of questionnaires that measure how parents view the relationship with their child and how children view the relationship with their parents [31]. The child-mother, child-father, and parent-child versions are available. It has 21 statements, and each statement has five answer options varying from never [1] to always [5]. The revised version of the OKIV (OKIV-R) has high internal consistency [31]. The OKIV-R is administered a total of five times to both adolescents and their parents (T0 to T4). The duration to complete the questionnaire is approximately10 minutes.

Parental burden; parenting burden questionnaire (Opvoedingsbelasting Vragenlijst; OBVL) T0 – T4

The OBVL is a questionnaire for parents about experiencing stress in parenting (Veerman et al., 2014).The OBVL consists of five scales (34 questions): parent-child relationship, parenting competence, depressive moods, role limitations, and health complaints. Each question has four answers that the parent can choose, varying from do not agree to totally agree. The scales of the OBVL have a reliability between 0.90 and 0.96 (McDonald’s Omega) [32]. The OBVL is administered a total of five times to parents participating in this study (T0 to T4), and it requires 15 min of their time.

Data analyses

Data will be analysed using SPSS V.24.0 for Windows (SPSS Incorporated) and the R package scan [33] for analyzing single-case data. Descriptive statistics will be performed for baseline characteristics of the study population. Parametric data will be presented as means with standard deviation (SD) and non-parametric distributed variables as median with interquartile ranges (IQRs). For the primary outcome we will combine visual analyses, randomization tests and effect sizes for a comprehensive analysis of the intervention based on state-of-the-art knowledge on analysing a single-case design [34].

Secondary outcomes will be analyzed using reliable change indexes (RCIs) to determine statistically significant individual changes, accounting for instrument reliability and baseline variability. RCIs are well-suited to single-case studies, allowing for individualized analysis of clinically meaningful changes, even in a small, non-randomized sample of 10 participants. They complement time series analyses by providing a detailed view of intervention effects at the individual level, aligning with the study’s single-case design goals.

Measurements: qualitative research

Upon the conclusion of the FBT-H, separate interviews will be conducted with adolescents, parents, and FBT-H practitioners to evaluate their experiences and outcomes. We use a self-developed topic list for this purpose.

Data analyses

Interviews will be transcribed into written form, and individual transcripts will be analysed using the grounded theory approach. Data will be coded using three types of coding: open coding, axial coding, and selective coding. To ensure inter-coder reliability, at least two coders will conduct the coding process separately. Researchers will compare and discuss selected text fragments until a consensus is reached on the most plausible interpretations and corresponding codes. Qualitative data analysis software (Atlas.ti 8.3 Windows) will be used to analyse the transcripts, with data collection and analysis occurring alternately. The results will comprise the researchers’ findings, selections, and interpretations of the data.

In this qualitative study, we aim to recruit participants divided across three groups: adolescents (12–18 years) with AN or OSFED and co-occurring mental health conditions who have undergone FBT-H, their parents, and FBT-H practitioners. We will monitor for data saturation throughout the analyses process and will continue participant recruitment until data saturation is reached. Similar qualitative studies have shown that saturation can often be achieved within small, focused samples, particularly when studying specific populations with shared experiences [35].

Data collection and management

Confidentiality is strictly maintained in this study, with personal data handled in accordance with the European General Data Protection Regulation (AVG). Participants will be identified by unique codes assigned at inclusion, stored digitally on Karakter’s secured, password-protected drive, and accessible only to project researchers. Non-anonymous data, such as informed consent forms, will be digitized and kept in password-secured folders with restricted access. All local databases are protected with password systems, and CASTOR EDC software will securely collect and store questionnaire data. Paper documents will be stored in a locked cabinet at Karakter Child and Adolescent Psychiatry, accessible only to authorized researchers.

Adverse events (AE)

AEs are defined as any undesirable experiences occurring during the study. All AEs reported by participants or observed by researchers will be recorded according to Dutch legislation. Serious adverse events (SAEs) include those that (1) result in death, (2) are life-threatening, (3) require or extend hospitalization, (4) cause significant disability, and (5) involve congenital anomalies. SAEs will be reported to the accredited MREC by the coordinating investigator using ToetsingOnline. Life-threatening or fatal SAEs will be reported within seven days, with a final report completed within eight days. AEs will be monitored until they are resolved or stable. Follow-up may include medical procedures, tests, or referrals. SAEs must be reported until the study’s completion.