Exploratory study of the feasibility, acceptability, and effects of a multicomponent intervention for SA.

Participants were SA (18–65 years) admitted to the emergency unit of Lausanne University Hospital, Switzerland, for whom a psychiatric consultation was requested from 8:00 a.m. to 5:00 p.m. between Monday and Friday. As no specific resource was allocated for this pilot study, its implementation relied on our usual psychiatric staff. SA evaluated during the evenings, nights, and weekend could, therefore, not be considered for inclusion and only one SA per day could be included in the study. Prisoners and non-French speaking patients were excluded. Suicide attempt was defined as any non-fatal intentional act of self-poisoning or self-injury, irrespective of degree of suicidal intent or other types of motivation (28), leading to EU admission. The SA who agreed to participate in the study were invited to sign an informed consent form.

The intervention specifically addressed SA in the EU and was coordinated by a case manager specialized in mental health care and working in our Emergency Psychiatric Unit. It consisted of several components issued from a literature review of best practices (Figure 1): coordination by a case manager (CCM) (19, 29); joint crisis plan (JCP) (30–32); early meeting with relatives and existing care network (EM) (33, 34); and phone contacts during 3 months after suicide attempt (PC) (15–19, 27, 35).

Three different CCM participated in the intervention, all of them having more than 5 years experience in the psychiatric emergency setting. They aimed to offer the SA a continuous relationship with a single person during the whole intervention. Case managers first performed a clinical evaluation of the SA under supervision of a senior staff member (psychiatrist or psychologist), to whom a comprehensive assessment of the suicidal crisis, including the current and recent (8 weeks) suicidal events as well as the immediate and future suicidal thoughts (36), were presented. The context and meaning of the current crisis were also explored and a JCP was developed, together with the SA. The JCP aimed to increase (i) the patient’s implication in the identification of the stressful elements leading to suicide attempt and (ii) the exploration of potential coping strategies for a subsequent crisis and (iii) to make further therapeutic decisions easier.

After clinical evaluation, participants were discharged from the EU to inpatient psychiatric care, referred to their previous therapist or referred to a crisis follow-up clinic, which is part of our Emergency Psychiatric Unit. The case manager also contacted relatives and suggested a meeting with him and a senior clinician if the SA agreed and no major counter indications were present, with the aim to share thoughts and emotions evoked by the suicidal event and provide education on suicidal crisis issues. EM was also contacted in order to discuss the suicidal event and to reach a consensus relative to further appropriate clinical management and eventually arrange a meeting.

The final part of the intervention consisted of weekly phone contacts for 1 month, and then monthly for another 2 months. Case manager called SA and get back over the period since the last contact with him/her. They offered listening and attention to the SA and answered any relevant questions. They attempted to enhance adherence to treatment and, if required, they motivated and directed the SA to seek further assistance or care (19).

Baseline data on sociodemographic (age, gender, origin, civil status, children, education, occupation, finances, legal representation, household composition) and clinical data (person who referred, past history of treatment, past history of suicide attempt, type of suicide attempt, post suicide attempt follow-up) were assessed for 107 SA. ICD-10 diagnosis was assessed by a psychiatric fellow under the supervision of a senior clinician. In addition, hopelessness was assessed at baseline by means of the French version of the Beck Hopelessness Scale (BHS) (37), a 20-item self-report instrument designed to measure negative expectations about the future. Total scores range from 0 to 20, with a score higher than eight indicating levels of hopelessness associated with an increased risk of suicide.

Feasibility was assessed by means of regular meetings with the project team during which difficulties of the intervention were discussed.

Acceptability was measured by means of satisfaction rated at baseline and at the end of the intervention. We used the Client Satisfaction Questionnaire (CSQ-8) (38), a 8-item self-report instrument designed to evaluate the service/treatment provided to clients/patients. Each item is phrased as a four-point anchored answer without a neutral position. We added two open-ended questions to this questionnaire: “What did you most appreciate in your contacts with the psychiatric team?” and “What do you think could be improved with regard to these contacts?” This questionnaire was administered at T0 by CCM and at the end of intervention by phone by a member of the research team.

During the phone follow-up, the CCM collected other clinical variables such as drop outs, suicide reattempts, and adherence to treatment.

Table 1 summarizes the different components of the intervention and the different instruments.

Comparisons between groups were performed with independent t-tests for continuous variables and Mann–Whitney U-tests for ordinal variables. For nominal variables, analyses were performed with Pearson’s Chi-Square tests or Fisher Exact tests when appropriate. All statistical tests were two-tailed and significance was determined at the 0.05 level. All statistical analyses were performed with IBM SPSS version 22.

Clinical and demographic characteristics of included patients (N = 19) were compared to those who were not approached (N = 51), excluded (N = 22), and those who refused (N = 15) (all referred to as non-included) (see Table 2). An important proportion (47%) of SA was not included because of the intervention scheduling (only available during working days). Included patients significantly differed from non-included patients on several dimensions: they were more often female [84 vs 50%, χ²(1) = 7.4, p = 0.006], more often referred by a relative (68 vs 40%, p = 0.041), they more often had financial problems [78 vs 52%, χ²(1) = 3.8, p = 0.049], and were more often under curatorship (26 vs 3%, p = 0.005).

Table 3 shows the comparison between included patients (N = 19) and those who refused (N = 15): more first-time attempters refused the intervention (60 vs 26%; p = 0.044). Only SA with self poisoning accepted the intervention, hence 27% of those who refused used several other methods (drug and other substances poisoning, cutting and jumping from a height; p = 0.029). People who had children also more frequently refused the intervention (79 vs 37%, p = 0.017). All other sociodemographic characteristics were similar between groups. The average score of the BHS, completed by the 19 included SA, reached 8.2 (SD 4.2) on average.

CCM was perceived most of the time as satisfying by the research team. However, JCP was observed to be difficult to complete during EU stay, because of the acute psychiatric state and the numerous other tasks required during the EU hospitalization (comprehensive clinical assessment, information on study, baseline data collection). Therefore, the JCP’s documentation was rarely completed.

Six joint meetings with the SA and their relatives were organized, four after discharge and two in the EU. The relatives were motivated to come in EU, but reported difficulties in participating in a meeting a few days after the suicide attempt. Meetings with relatives were rare and could not be organized as systematically as expected. A meeting with the existing care network was organized only once.

The phone follow-up revealed to be feasible. Only 20 of the 102 planned calls could not be realized because patients did not answer despite three attempts. Case Managers reported that a trustful relationship with an adequate therapeutic distance could be built, fostering the development and continuity of follow-up.

The CSQ-8 was completed by 14 SA after the intervention in the EU. The average score of the total of points (max sum score = 32) of the questionnaire was 27.8 (SD 2.8). Open-ended questions revealed that elements, such as “empathy,” “availability,” “listening,” and “attention” were valued by patients. Answers to the question about what could be improved revealed a perceived lack of time and of intimacy during the intervention and difficulties of collaboration with EU staff.

The CSQ-8 was completed by phone for 15 SA at the end of the intervention: the average score was 25.4 (SD 4.9). The intervention was described as supportive and respectful, facilitating EU discharge and providing structure and support during the 3 months following the suicide attempt. The weekly rhythm of phone contacts was described as satisfying, but the diminished rhythm between phone contacts at 1 and 2 months was perceived as abrupt.

Among the 19 SA included in the study, 14 terminated the intervention, four dropped out (one after the meeting in the EU, one after 1 month and, two after 2 months), and one (5%) died by suicide. Twelve of the 19 included patients did not reattempt suicide during the intervention, three (16%) reattempted suicide (one once, one twice and one three times). The SA who died by suicide was diagnosed with major depression. He accepted crisis follow-up, but declined inpatient psychiatric care. He missed no appointment during follow-up. After reporting a favorable evolution during a last meeting, he hung himself. BHS score was three for the patient who died by suicide, and 11 for the three reattempters.

This pilot study was limited by the small sample size and the large proportion of SA who presented out of the inclusion times (nights and weekends). Furthermore, some components of intervention, in particular, family meetings proved to be much more difficult to organize than initially expected. 24/7 availability of our intervention would greatly enhance patient recruitment and engagement of patients in our proposed treatment.

Another limitation is the definition of the term of “suicide attempt.” Silverman et al. (60) showed that the variability in terminology and definitions used in the suicide literature strongly affects the collected data and hampers comparison, extrapolation, and generalization. The term “suicide attempt” can encompass a wide range of non-fatal self-inflicted behaviors. In this study, we considered all intentional self-harm as suicide attempt, following the approach favored in the UK (28). Suicidal intent is regarded as a dimensional rather than a categorical concept within the definition of “self-harm” (28). The choice of this definition of suicide attempt may have had a considerable influence on the recruited patients.