Participants
The sample will be composed of relatives of patients who meet DSM-V criteria (1) for Anorexia Nervosa (both restricting and purging subtypes), Bulimia Nervosa, Binge Eating Disorder, Eating or Food Intake Disorder Not Otherwise Specified, and any PD (or PD traits). The diagnoses will be carried out by a specialist in clinical psychology or psychiatry, and the sample will be recruited in different clinical services: the University Clinical Hospital of Valencia and its Mental Health Reference Units; and the Eating Disorders and Personality Disorders Unit PREVI-ITA and its three centers in Castellón, Valencia, and Alicante.
In the case of the patients, the following inclusion and exclusion criteria will be established: 1) meet the diagnostic criteria for ED and PD (or PD traits); 2) agree to participate in the study in writing by signing the informed consent form; in the case of minors, the consent must be signed by their parents; 3) The presence of another serious pathology such as psychosis, schizophrenia, intellectual disability, etc., will be an exclusion criterion.
In the case of family members, the following inclusion and exclusion criteria will be followed: 1) being a family member of one of the patients with a diagnosis of ED and PD (or PD traits); 2) signing the informed consent; 3) The presence of any pathology in the family member that keeps the intervention from being carried out (such as psychosis, schizophrenia, intellectual disability, substance dependence, etc.) will be an exclusion criterion.
Study design
First, a pilot feasibility study will be carried out using focus groups to explore the opinions of relatives of ED-PD patients. Semi-structured interviews will be used. The research design will follow the criteria established by Cooke et al. [40], and for the development of the focus groups, the guide developed by Breen [41] for this type of methodology will be used. Second, a randomized controlled clinical trial will be conducted following the CONSORT (Consolidated Standards of Reporting Trials) guidelines http://www.consort-statement.org/ [42, 43]. This study is a superiority trial. The study design consists of a two-arm randomized controlled trial (RCT). On the one hand, there will be two conditions: Family Connections (FC-ED - PD) or Treatment as usual optimized (TAU-O), and family members will be randomized to one of the two groups. -A between-subjects design will be used with three assessment points: pre-treatment, post-treatment, and a follow-up 6 and 12 months after the end of treatment. Once the study has been explained and the informed consent has been signed, participants will be randomly assigned to one of two experimental conditions: 1) The adaptation of the FC protocol for relatives of ED-PD patients (FC: ED-PD) will be applied; 2) The Treatment as Usual Optimized Treatment (TAU-O) protocol will be applied. Randomization will be performed by an external researcher who will not participate in any of the phases of the project. The randomization sequence will be hidden from the evaluators participating in the study. The G*Power 3.1 software [44] will be used for this purpose. Assessment of participants will be carried out by a clinician other than the professional administering the treatments. Upon completion of the randomized controlled clinical trial, in-depth interviews will be conducted with family members who have successfully completed the intervention (e.g. clinically significant change) and with family members who drop out of the intervention. In all the potential groups, the aim will be to explore participants' experiences in order to learn about barriers and facilitators of the intervention.
Sample size
To calculate the sample size, effect sizes found in previous studies on the topic have been considered. Grenyer et al. [30] tested a group psychoeducation protocol for caregivers of people with BPD in a controlled study that reports medium to large effect sizes (dyadic adjustment, d = .78; family empowerment, d = 1.4). Moreover, Grenyer obtained significant improvements on measures of illness burden between post-assessment and 12-month follow-up, with medium effects (d = .45). These effects are in line with the literature on psychological treatments for relatives of people with ED (Positive Caregiver Experience, g = − .80) [14]. Given these data, an effect size of 0.50 is expected in the present study, adopting a conservative approach. Because our design includes two experimental conditions, t-tests are assumed for the statistical analyses. Considering an alpha of 0.05 and a statistical power of 0.80 in a two-tailed t-test, the total sample size needed to reach an effect size of 0.60 in loading is 90 participants (45 participants per experimental condition). To control for the maximum possible loss of subjects during treatment, based on the literature on previous programs for family members of BPD or ED patients, a dropout rate of 30% is expected [34–36]. Therefore, the required sample size should contain 124 participants in all (62 participants per group), as Fig. 1 shows. These calculations were carried out using the software program G*Power 3.1 [44].
Procedure
Family members of patients with ED-PD will be given the opportunity to participate in the study. Once they have signed the informed consent, an expert clinician will perform the assessment of each participant to check that the inclusion and exclusion criteria are met, and an independent investigator, unaware of the characteristics of the study, will be contacted to perform the randomization. At all times, for randomization purposes, an experimenter from outside the research team will assign each family member to one of the two study conditions (FC: ED-PD or TAU-O) using a random number software program. The randomization sequence will be concealed from the clinicians participating in the study. Independent assessors will administer the assessment protocol to the participants without being aware of the experimental condition to which they belong (FC: ED-PD or TAU-O). Patients will also be assessed before starting the intervention programs for relatives. After the assessment has been completed, all the relatives will start to receive the intervention in the condition to which they have been assigned, and the patients will participate in their current treatment or care. At the end of the intervention, the assessment protocol will be re-administered to both relatives and patients, again by independent assessors, and the same thing will be done at the 12-month follow-up.
Ethics
Researchers adhere to the Helsinki Convention and the World Psychiatric Association's Madrid Declaration on clinical research. All participants will be volunteers who have given their informed consent to participate in the study. All eligible participants will be given oral and written information about the study and the two intervention modalities. Specifically, they will be informed that they may leave the study at any time without giving an explanation, and that this decision will in no way affect their family member's regular treatment at the center. The selection and evaluation of the participants will be performed by qualified personnel who will not know to which condition a given participant has been assigned, and the treatments will be carried out by qualified and expert professionals. Furthermore, it should be emphasized that the protocols for action and custody of the information followed at the University Clinical Hospital and at the PREVI Clinical Center, where the participants will be recruited and the interventions will be carried out, comply with all the requirements of the LOPD (Organic Law 15/1999, of 13 December 2018). The Ethics Committee of the University of Valencia (Valencia, Spain) approved this study with number UV-INV_ETICA-1955599. The trial was registered at clinicalstrial.gov as NCT05404035, June 3rd 2022.
Interventions
Family Connections for relatives of people with eating disorders and personality disorders (FC: ED-PD).
The intervention lasts three months and includes 12 sessions with a weekly two-hour group format. The FC program [34]) is divided into six modules, but the research team developed the program adapted to ED and PD in seven modules, which are described below:
Module 1
Presentation of the program, objectives, structure, tasks, etc. Up-to-date information and research on ED and PD (Epidemiology, frequency, Risk factors, and protective factors).
Module 2
Psychoeducation on the development of ED and PD, explanatory theories, available treatments, and comorbidity with PD.
Module 3
Emotional regulation skills, skills of acceptance, validation, approach, and awareness, and to decrease emotional reactivity.
Module 4
Skills to improve the quality of relationships in family interactions (letting go of guilt and anger, acceptance skills in relationships).
Module 5
Communication skills and effective self-expression.
Module 6
Problem management and making safe plans for crisis management.
All the modules include practice exercises, video viewing, and homework assignments. In addition, throughout the program, with the goal of increasing social support, the FC program provides a forum where participants can stay in touch and share common problems and solutions.
Optimized treatment as usual (TAU-O).
Family members in this condition will continue to receive the usual treatment provided by their referral care center. In addition, we will optimize the treatment they are currently following, based on the recommendations of the international guidelines for the treatment of ED [45]. There will be one three-hour session in group format, with the following components:
Module 1
Up-to-date information and research on ED and PD (epidemiology, frequency, risk factors, protective factors). Psychoeducation on the development of ED and PD. Explanatory theories. Available treatments and comorbidity.
In both conditions, after each face-to-face session, the participant will be asked to review the contents addressed during the session as homework (independently of the homework corresponding to the specific module being addressed). All the interventions with family members will be performed by clinical psychologists or general health psychologists with at least a master's degree or a doctoral degree, and with previous training in administering the programs.
Treatment Of Patients
With regard to the patients’ treatment, the routine treatment they receive in their centers of reference will be followed. All the interventions will be carried out by the clinical psychologists and psychiatrists working in these centers.
Measures
Family Members’ Measures (Participants)
Primary outcomes
Record of critical family-patient incidents
number of Binge eating and vomiting (purging) episodes in the past three months, number of serious arguments between patient and caregivers in the past three months, number of days of self-injury in the past three months, number of episodes of verbal/physical violence with caregivers in the past three months, number of visits to the psychiatric emergency department in the past three months, number of unscheduled therapy sessions in the past three months (face-to-face, phone calls, etc.).
Burden Assessment Scale [46]: It is a 19-item scale that assesses caregivers’ objective and subjective burden due to their loved one’s illness within the past six months. Items are rated on a 4-point Likert scale ranging from 1 (nothing) to 4 (a lot). Higher total scores indicate stronger burden. The scale shows adequate validity and reliability (Cronbach’ alpha ranges from .89 to .91) [47]
Secondary outcomes
The Eating Disorder Symptom Impact Scale [48] () in its Spanish version (Carral-Fernández et al., 2013). It is a 24-item scale that assesses eating disorder-specific caregiving experiences. Items are rated on a 5-point Likert-type scale ranging from 0 (never) to 4 (almost always). The internal consistency is good, with Cronbach’s alpha coefficients mostly superior to .70.
Family Assessment Device - Global Functioning Scale [49]). It is a 60-item self-report about family functioning in terms of problem-solving, communication, roles, affective responsiveness, affective involvement, behavior control, and general functioning. Items are rated on a 4-point Likert scale ranging from 1 (completely agree) to 4 (strongly disagree). The internal consistency is good (Cronbach's alphas between .72 and .83) for the subscales, with a Cronbach's alpha of .92 for general functioning.
Mastery and Empowerment Scale [50]. It is a 34-item scale divided into three domains: family, service system, and involvement in the community. Items are rated on a 5-point Likert scale ranging from 1 (completely false) to 5 (completely true). Different studies have demonstrated that the psychometric properties of the FES are robust in both its original and translated versions [50–52].
Multicultural Quality of Life Index [53]. The Multicultural Quality of Life Index is a culture-informed, self-rated instrument. Its 10 items cover key aspects of quality of life, from physical well-being to spiritual fulfilment. Regarding its applicability, the mean completion time was less than three minutes, and 96% of raters found it easy to use. Test–retest reliability was high (r = 0.87). A Cronbach’s alpha of 0.92 showed its internal consistency, and a factor analysis revealed a strong structure.
The Multidimensional Existential Meaning Scale [54] in its Spanish validation [55]. It is a 15-item scale that assesses existential meaning through three dimensions: comprehension, purpose, and mattering. Items are rated on a 7-point Likert scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). The three MEMS subscales showed adequate internal consistency: Comprehension (ω¯ = 91), Purpose (ω¯ = 92), and Mattering (ω¯ = 86).
Depression, Anxiety and Stress Scale [56]. It is a 21-item scale, in its short version, that measures clinical symptoms such as depression, anxiety, and stress. Items are rated on a 4-point Likert scale ranging from 0 (It did not happen to me) to 3 (It happened to me a lot, or most of the time). It shows excellent internal consistency: depression (α = .94), anxiety (α = .87), and stress (α = .91).
Difficulties in Emotion Regulation Scale [57]. in its Spanish version [58]. In its adaptation, the authors reduced the items from 36 to 28, and they considered five scales instead six. The subscales are: lack of emotional control, vital interference, lack of emotional attention, emotional confusion, and emotional rejection. Items are rated on a Likert-type scale ranging from 1 to 5 (1 = "almost never" and 5 "almost always"). Higher scores indicate more difficulties in emotional regulation. Psychometric properties were adequate, with very good internal consistency (α = .93) and good test-retest reliability (pl = .74, p < .001).
Other Pre-specified Outcome Measures
Socio-demographic data
age, sex, educational level, income, marital status, number/age of children, and history of psychological treatment.
Opinion of treatment scale, adapted from Borkovec and Nau [59].
Measures ‑ Patients
Primary outcomes
The Eating Attitudes Test-26 [60]. in its Spanish version [61]. This scale is used to evaluate the patients. It is a 26-item self-report that assesses attitudes toward eating. Items are rated on a 6-point Likert scale ranging from 1 (never) to 6 (always). The reliability analysis indicated good internal consistency [60, 62–65].
Patient Health Questionnaire [66]. in its Spanish version [67]. This scale is used to evaluate the patients. It assesses each of the nine DSM-IV criteria for depression through nine items. Items are rated on a 4-point Likert scale ranging from 0 (not at all) to 3 (nearly every day).
Secondary outcomes
Overall Anxiety Severity and Impairment Scale [68]. in its Spanish version [69]. This scale is used to evaluate the patients. It is a 5-item scale that assesses the severity and frequency of anxiety symptoms, behavioral avoidance, and the functional impairment related to anxiety. Items are rated on a 4-point Likert scale ranging from 0 to 4. It shows good psychometric properties in terms of test-retest reliability and internal consistency (α = .80).
Validating and Invalidating Responses Scale [70]. This scale is used to evaluate the patients. It is a 16-item scale that assesses the levels of perceived validation and invalidation of caregivers’ responses, divided into validating and invalidating responses. Items are rated on a 5-point Likert scale ranging from 0 (never) to 4 (almost always). Higher scores indicate greater perceived validation or invalidation of the responses of the caregiver being assessed.
Lum Emotional Availability of Parents [71]. This scale is used to evaluate the patients. It is a 15-item questionnaire that assesses the emotional availability of mothers and fathers as perceived by the person assessing their relatives. Items are rated on a 6-point Likert scale ranging from 1 (never) to 6 (always). The psychometric properties are very good; for a clinical sample, excellent internal consistency is observed (for mother α = .92, and for father α = .93). In addition, it has adequate test-retest reliability for the mother's form (r = .92) and the father's form (r = .85).
Data Analyses
Regarding data analysis, the results of the pilot feasibility study using focus groups will be analyzed using thematic analysis in order to identify the topics that emerge from the focus groups [72]. Data reporting will be carried out following the COREQ guidelines [73]. In addition, qualitative research quality criteria recently developed by Levitt et al. [74] will be pursued in order to achieve what the authors refer to as methodological integrity in the qualitative field.
In the controlled clinical trial, the CONSORT guidelines [43] will be followed. First, participants' scores in the two conditions will be compared before the intervention to check that there are no significant differences between them on the outcome measures and that they are, therefore, comparable after randomization. ANOVAs will be conducted for continuous variables and Chi-square tests for categorical variables. For outcome measures at post-treatment, we will study whether the assumption of homoscedasticity is met with Levene's test. If this assumption is met, repeated-measures ANOVAs and F-tests will be used to compare the two experimental conditions. If the homoscedasticity assumption is not met, the Brown-Forsythe test will be applied. F-tests for statistical significance will be followed by post hoc comparisons. In particular, Tukey will be used when the homoscedasticity assumption is met, and Games-Howell if the homoscedasticity assumption is not met. Appropriate analyses will also be carried out to calculate intervention effect sizes and confidence intervals. The intention-to-treat principle will be used when analyzing pre- and post-treatment data and at 12-month follow-up, using mixed-effects models with full information and maximum likelihood estimation. To complement the results of ANOVAs and post hoc comparisons, effect sizes will be calculated using the standardized mean difference proposed by Cohen [75]. These effect sizes will be calculated to assess changes within and between groups, all based on a pooled standard deviation.
For the qualitative study, semi-structured in-depth interviews will be used. These interviews will follow the guidelines of Knox and Burkard [76]. The research design will be carried out following the criteria established by Cooke, Smith and Booth [40]. Data will be analyzed using the consensual qualitative research (CQR) method. Data reporting will be carried out following the COREQ guidelines [73]. Furthermore, we will attempt to meet the criteria for quality in qualitative research recently developed by Levitt et al. [74] in order to pursue what the authors refer to as methodological integrity in the qualitative field.