Participants
The sample of relatives consisted of 10 relatives (age range 33 - 58 years; M = 51.50, SD = 7.26; 70% female, n = 7). Regarding the caregiver role, the majority were mothers (60%, n = 6), 30% were fathers (n = 3), and 10% were siblings (n = 1). With regard to the living situation, 60% of the relatives lived with the patient (n = 6), and 40% did not (n = 4). The rest of the variables can be seen in Table 1, which shows the socio-demographic data of the relatives.
------------------------------------Insert Table 1 ---------------------------------------
As Table 2 shows, the patient sample consisted of nine patients with a mean age of 23 years (age range 17 - 33 years; M = 23.11, SD = 5.88). Regarding gender, 88.89% were women (n = 8), and 11.11% were men (n = 1). Of these patients, 33.33% had primary education (n = 3), 22.22% had secondary education (n = 2), and 44.44% had higher education (n = 4). All the patients were single (n = 9). In terms of the diagnosis, all of them had borderline personality disorder, two had comorbidity with an eating disorder (22.22%, n = 2), and one had comorbidity with attention deficit hyperactivity disorder (11.11%, n = 1). Finally, all of them were receiving psychological treatment, most of them combined with pharmacological treatment (88.89%, n=8).
--------------------------------------Insert Table 2 ---------------------------------------
The inclusion criteria for relatives were: a) The patient had to have made a suicide attempt in the past two years. It had to be a real attempt; that is, the act could not be non-suicidal self-harm, initiated in a state of delirium/confusion, or be carried out for political or religious purposes; b) They had to understand Spanish, both written and oral; c) sign an informed consent document; and d) be of legal age. The exclusion criteria were: a) the presence of a serious mental disorder (SMD) in the family member, such as psychosis, schizophrenia, substance addiction, or high suicidal ideation, and b) the presence of a physical disability that interferes with participating in the programme.
The inclusion criteria for patients were: a) having made at least one suicide attempt in the past two years; b) having signed the informed consent form regarding their participation in the study. In the case of minors, the document also had to be signed by their legal guardians. The exclusion criteria were: a) having another SMD, such as psychosis, schizophrenia, bipolar disorder, etc.; and b) the presence of a physical disability that interferes with participating in the programme.
Measurement Instruments
Sociodemographic data
Sociodemographic Measures of Relatives and Patients. The sociodemographic characteristics of the relatives were recorded: age, sex, relationship with the patient, living with the patient, educational level, current employment status, marital status, number of children, and current or past psychological/psychiatric problems.
Measures of primary results
Burden Assessment Scale (BAS; Reinhard et. al, 1994). Using three subscales (Disruptive activities, Social and personal dysfunction, Worry, guilt, and overload) and a total score, this scale assesses the burden of illness in the past six months. Cronbach's alpha values ranged from .74 to .87 in this study. Higher values on this scale correspond to greater burden in the past six months.
Depression, Anxiety, and Stress Scale (DASS-21; Lovibond & Lovibond, 1995; Daza et. al, 2002). This scale assesses different negative emotional symptoms related to depression, anxiety, and stress through 21 items distributed in three subscales and a total score. It has high reliability, with Cronbach's alphas ranging from .87 to .94, and from .78 to .95 in this study. Negative symptoms have greater intensity when the score obtained on this scale is high.
Measures of secondary results
Family Empowerment Scale (FES; Koren et. al, 1992). This scale assesses the feeling of empowerment through 34 Likert-type items that form three subscales: Family, which assesses the management of everyday family situations; Service system, which assesses active involvement with the service systems used by the patient; and Participation in the community and in social policies, which assesses the demand for and defence of improvements in the services available in the community, and the total score. The scale has good internal consistency, with coefficients between .87 and .88. Cronbach's alpha values ranged from .77 to .90 in this study. Higher scores indicate greater empowerment.
Treatment Expectations Scale (Borkovec & Nau, 1972). This is a six-item Likert scale ranging from 0 (not at all) to 10 (very much), and it is filled out before starting the intervention. The scale measures the relatives' expectations of the intervention they will receive, the rationale for the intervention, whether they think the treatment will be aversive, whether they think it will be useful, and whether they would recommend it to other relatives of patients with the same problem.
Scale of Opinions about the Treatment (Borkovec & Nau, 1972). This scale contains six items rated on a Likert scale ranging from 0 (not at all) to 10 (very much), and it is filled out after the intervention has ended. It measures the same variables as the previous scale, but at the end of the treatment.
Critical Incident Log. This is an ad hoc questionnaire completed by the family member that includes questions about the patient: number of suicide attempts, number of self-injuries, number of episodes of verbal or physical violence towards the family member, and number of times the patient has gone to the psychiatric emergency room. This is an objective record that is presented at the beginning and end of the intervention to find out the number of times the patient has engaged in these disruptive behaviours in the previous three months.
Procedure
After obtaining the approval of the Ethics Committee of the University of Valencia, the research group offered family members who attended a Unit specialised in personality disorders in the Valencian Community and the Valencian Association of personality disorders (ASVA- TP) the possibility of participating in the FC-S intervention. Interested relatives were interviewed individually by a psychologist to verify that they met the inclusion criteria. For approximately one hour, the patient's clinical history and the family member's way of handling it, the presence of clinical symptomatology in the family member, and dysfunctional personality traits or personality disorder were assessed. Coping strategies and problem management skills were also evaluated, in addition to other clinically relevant issues.
Relatives who met the necessary characteristics to participate signed the informed consent form and completed the pre-intervention protocol online through the Lime survey platform (the platform accepted by the University of Valencia). As described above, the protocol consisted of the following questionnaires: BAS, DASS-21, FES, the scale of Expectations (before) and Opinions (after) about the treatment, and the critical incident log. The assessment protocol was completed before starting the intervention and after receiving it. After completing the assessment, the intervention began in online mode. It consisted of 12 weekly two-hour sessions over a period of three months.
As Figure 1 shows, the initial assessment was carried out by a total of 14 family members, 2 of whom were excluded because they did not meet the inclusion criteria. Twelve participants began the intervention, and two dropped out due to scheduling problems. Finally, 10 participants finished the intervention and filled out the post-assessment.
--------------------------------------Insert Figure 1 ---------------------------------------
Once the intervention was over, a focus group was held with the relatives to find out their subjective view of the intervention (Vilaregut et. al, 2021). This was a qualitative interview in an open and group format where the relatives gave their opinions of the programme they had received. This session lasted two hours and included open questions about their opinion of the FC-S, the skills they found more and less useful, aspects of the programme they would improve, and whether they would recommend it to others, etc.
Intervention
Family Connections for relatives of people with suicidal behaviours (FC-S)
FC-S is an adaptation of FC (Hoffman et. al, 2005). It is a group training intervention that contains 12 weekly two-hour sessions, and it has several objectives: first, to provide family members with up-to-date information about the disorder; second, to provide training in family and individual coping skills; and third, to facilitate the creation of a support network among the participants (Hoffman et al, 2005; 2007). The intervention is divided into six modules, as Figure 1 shows.
Likewise, in the adaptation for relatives of people with suicidal behaviour, the interpersonal theory of suicide (Joiner, 2005) is introduced in the second psychoeducation module of the programme. Joiner's theory is presented as a challenge that family members can begin to work on and use to build bridges. The three categories of Joiner's theory are presented from a therapeutic perspective. In the factors of burden to others (perceived burden), frustrated belonging (not feeling connected to a social or family group), and acquired capacity for suicide (noticing, identifying risky events or situations), participants are taught the importance of not blaming and learning to create an environment of mutual support and respect. The relatives watch numerous scenarios and videos with actors, so that they can see how the situation turns out without applying the skills and when applying the new skills.
--------------------------------------Insert Figure 2 -------------------------------------
Data Analysis
The data were analysed using SPSS.25 (IBM Corp, 2017). First, the reliability of each instrument for the present study sample was tested, and subscales with Cronbach's alpha values below ,60 were eliminated. Based on this criterion, the Social and Personal Dysfunction subscale (BAS, α = .27) was eliminated. Second, Student's t-test for related samples was conducted to examine whether there were differences between before and after the intervention, and Cohen's d was calculated as an effect size measure.