Participants will receive the usual treatment in their mental health service and will also receive a psychoeducation session of approximately 3 hours where the predictive and protective factors of suicidal behavior will be explained.

We have used the short, validated Spanish version with 21 items on the frequency of negative emotional symptoms in the past week. The items are rated on a Likert scale (0-3) where the higher the score, the higher the frequency of symptoms of depression, anxiety, and/or stress. The internal consistency of the scale was excellent, with Cronbach's alphas for the DASS-21 subscales: Depression (α = .94), Anxiety (α = .87) and Stress (α = .91)

We used the Spanish validation containing 28 items. This questionnaire is divided into five subscales: (1) Lack of emotional control, (2) Life interference, (3) Emotional inattention, (4) Emotional confusion, and (5) Emotional rejection. The items are rated on a Likert scale (1-5) where higher scores indicate greater difficulty in regulating emotions. Psychometric properties are excellent, with an internal consistency of α = .93 and test-retest reliability of pl = .74, p < .001

It consists of a 10-item index of perceived quality of life. It refers to physical and emotional well-being, functioning at work, satisfaction with personal relationships and self-independence, support in the community and from an emotional point of view, spiritual well-being, and perceived overall quality of life. The items are rated on a Likert scale (0-10) where higher scores indicate higher perceived quality of life. The psychometric properties are good for both internal consistency (α = .89) and test-retest reliability (r = 0.87).

This scale was designed and developed by members of the research team and constructed from an adaptation of another opinion and expectations questionnaire [49]. The constructs this scale assesses are: opinion, acceptance and satisfaction with the skills training program, and the changes in the participants after the completion of each module. The questions refer to the rationale for the intervention, recommendation of the program, satisfaction with the program, usefulness and expectations of the skills training. The items are rated on a Likert-type scale ranging from 0 "Not at all" to 10 "Very much".

his register was developed ad hoc for this study. The questions recorded are the following: frequency of suicide attempts in the past six months, number of days of self-harm in the past six months, number of episodes of verbal/physical violence with caregivers in the past six months; frequency of visits to the psychiatric emergency room in the past six months, frequency of therapy sessions conducted out of schedule in the past six months (face-to-face, phone calls, etc.).

We use the Spanish version of this 15-item questionnaire that assesses the degree of dissatisfaction with their need to belong (frustrated belonging) and the degree to which they perceive themselves as a burden to others (perceived burden). The items are rated on a Likert-type scale (1-7) where higher scores indicate higher levels of frustrated belonging and perceived burden to others. Psychometric properties were good: scale reliability was very good (perceived burden, α = 0.96; and frustrated belonging, α = 0.78).

It consists of a nine-item questionnaire that assesses depressive symptoms in the past two weeks. Specifically, it includes the DSM-IV diagnostic criteria A for major depressive disorder (American Psychiatric Association, 2000). The items are rated on a Likert scale (0-3) where higher scores indicate higher frequency of depressive symptoms. The severity of depression on this questionnaire is measured through the total score, which can be categorized as none or minimal, mild, moderate, moderately severe, and severe. Validity has been adequate, with a sensitivity of 88% and a specificity of 88% for major depression

We use the Spanish version of this questionnaire, which consists of a five-item instrument that assesses the frequency and intensity of anxiety symptoms in the past week. In addition, it measures interference in work and academic, social, and daily life domains, as well as avoidance behaviors. The items are rated on a Likert-type scale (0-4). The psychometric properties are good in terms of internal consistency (α= 0.86), convergent and discriminant validity, and sensitivity to change (α= 0.86).

It consists of a 16-item scale on the validation and invalidation of family members' responses about their loved ones. It is divided into two subscales (validation and invalidation), and the items are rated on a Likert scale (0-4) where higher scores indicate higher perceived validation or higher perceived invalidation (depending on the subscale).

This 15-item questionnaire measures the perceived emotional availability of primary caregivers. The items are rated on a Likert scale (1-6) where higher scores indicate greater emotional availability of caregivers. Psychometric properties were excellent for both subscales (mother, α = .9; and father, α = .93). In addition, test-retest reliability was also adequate for the mother's subscale (r = .92) and the father's subscale (r = .85).