Study setting and recruitment
The main recruitment centers are located in Berlin, Leipzig, and Bayreuth and are directed in accordance with the Medical School Berlin, the Federal Association of Bereaved Parents and Grieving Siblings in Germany (VEID e.V.), and the Association for Relatives to Suicide (AGUS e.V.). In addition, we will recruit participants through the study website (www.hilfe-nach-suizid.de) and advertisements in social media (e.g. social media platforms, self-help forums, specific websites for bereaved people), newspaper articles, flyers for self-help groups, or churches.
Participants and eligibility criteria
All participants who meet the following criteria will be included in the study: a) aged between 18 and 75 years, b) experienced a loss of the suicide of a close person, c) have access to the Internet, d) have sufficient German language skills, and e) provide a signed consent form (Appendix B). Participants who show one of the following criteria will be excluded from the study: a) acute suicidality, b) elevated depression (the score in the Beck Depression Inventory-II > 35), 3) bipolar disorder, c) current psychotic experience, d) elevated alcohol, drug, or substance use, e) suspect of Borderline Personality Disorder, and f) self-injurious behavior. The group intervention will be conducted by two group leaders a) an approved psychotherapist or a trainee in psychotherapy in the last phase of the training and b) a group leader of a face-to-face self-help group for suicide bereavement.
Sample size
We based the power and sample size estimations in the current study on results from previous studies examining the effectiveness of online bereavement interventions [34]. Assuming a between-subject effect size of d = 0.80, a power of 0.80, an alpha of 0.05 (two-sided) and a dropout rate of 30%, the within group sample size should consist of at least n = 52 participants. Hence, we aim to include N = 104 participants in total.
Procedure
The flow of all participants is depicted in Figure 1. Participants who are interested in the treatment can obtain general information about the program on the study website (www.hilfe-nach-suizid.de). This website provides widespread information on relevant topics for bereaved people after suicide. In addition, visitors of the website receive further information on helpful services (e.g. psychotherapy, counselling) as well as recommendations of useful literature.
To take part in the online intervention, participants have to register on the website first. Afterwards they receive comprehensive information about the process of the study as well as a consent form for participation. As soon as participants have signed the consent form and have sent it back (e.g. email, fax, or by post), they are invited to an online survey which contains several questionnaires for diagnostic purposes (Table 1). To further examine the inclusion and exclusion criteria, participants will take part in a clinical telephone interview. This interview tests for acute suicidality, severe depression, acute psychotic experiences, bipolar disorder, posttraumatic stress disorder (PTSD), as well as Borderline Personality Disorder (BPD). Participants who do not meet the inclusion criteria receive further information on psychotherapy, counselling or self-help groups. In case of clearly elevated scores on the measurements, participants are debriefed and are advised to see a doctor or psychotherapist. If the interviewer finds evidence for acute suicidality, he or she engages in crisis management strategies [35] or will arrange another telephone meeting to monitor the process.
Randomization
We will use block randomization without extra stratification: As soon as ten participants are included, they are considered to form a group. Then, the whole group is randomly assigned to either the treatment or the waitlist control group. One of the authors (BW) who is not involved in participant recruitment (UM, LH), clinical assessments (UM, LH), or treatment (external group leaders) executes the randomization using a generated list of www.random.org.
Blinding
Because of the high level of organizational effort involved in randomizing groups, it is not possible to blind the participants, group leaders, or study employees.
Symptom assessment
During the intervention, participants are requested to complete online surveys to assess potential changes in the therapeutic process. More precisely, there are 15 measurement points (Table 1): at the initial stage (pre-assessment, T0), after each group session (T1 to T12 including a treatment mid-point assessment after six weeks, T6, and a post-assessment at the end of the intervention, T12), after a three-months follow up (T13), and finally, after a six-months follow-up (T14).
The online-group intervention
The online group intervention is manualized and consists of 12 weekly sessions with a length of 90 minutes. Each group has a maximum of ten participants and is guided by a psychologist (i.e. a psychotherapy trainee) and a self-help counsellor, who is herself or himself bereaved by suicide. The participation is voluntary and free. Participants are allowed to skip sessions or to drop out of the program at any time.
Development of the intervention
Focus groups. In order to develop the intervention, a total of three focus groups were conducted. The composition of the groups was based on gender, age and kinship to the deceased. Also, they had to be at least 18 years old, and the time since the suicide should be at least one year. The 23 recruited participants were on average 53 years old (SD = 13, Range: 20-72) and 86% were women. Most participants had lost a child (52%), followed by a spouse or partner (24%), a parent or sibling (each 10%), or others (5%). Almost all participants received support at the time of measurement, namely from psychotherapy (86%), self-help group (48%), a doctor (29%), pastoral care (19%), or grief counselling (14%).
The focus group participants were asked what they believed were difficulties during their grief process and effective strategies to find meaning with the suicide. Further, they were interviewed which themes are relevant to be included in an internet-based intervention. The group conversations took place for four hours with a one-hour break and were recorded in audio format, transcribed, and evaluated with respect to the important themes for a group intervention.
Testimonial Videos. Most of the sessions start with a testimonial video in which a person bereaved by suicide talks about her or his own experiences of the session’s topic (e.g. guilt, own suicidality). The screen-plays of the protagonist’s monologue (played by an actor/actress) were developed with the excerpts of the transcribed focus-groups. In order to ensure a balanced gender ratio, both men and women were presented equally in the testimonial videos. All testimonial videos follow the same structure: 1) How did the bereaved person experience e.g. feelings of guilt in the first days and weeks after the loss? 2) How did the bereaved person experience feelings of guilt over the course of time? 3) How did the bereaved person experience feelings of guilt today? The protagonists in the testimonial videos aim to function as a role model in dealing with difficult emotions which occur after a suicide over the course of time.
Treatment modules. The selection of the relevant themes for the intervention were based on the literature on grief or treatment of traumatized people [34] as well as the results from the focus groups. Focus group interviews were analyzed and relevant themes were included in the treatment manual. Table 2 describes the complete content by session. The manual is based on cognitive-behavioral principles and consists of a) testimonial videos, b) psychoeducation (e.g. grief, guilt, suicidality), c) a guided group discussion about the topic of the session, and d) individual tasks between the sessions (e.g. writing a farewell letter to the deceased).
The internet tool. The Webinar will be held using the internet tool Adobe Connect®. Participants receive a link to log into an online-seminar room and log in with a self-chosen user name. They are connected with each other via micro- and headphones but without video-transmission. The visual transmission is disconnected to keep increased anonymity for the participants. The participants are able to see PowerPoint slides or videos, which are part of the treatment program on their computer-screen. The two group leaders can additionally access a name-list of all group members. There is also a chat option which could be used by group participants or group leaders.
Measures
Screening for eligibility criteria. The study coordinators will screen the participants for eligibility criteria using a semi-standardized clinical telephone interview. To exclude the presence of a current major depressive or manic episode, bipolar disorder, BPD, PTSD, or alcohol and substance misuse, the corresponding sections from the German version of the Structured Clinical Interview for DSM-IV [37] will be applied. The inter-rater reliabilities of the diagnoses vary between fair (e.g. Agoraphobia) and excellent (e.g. PTSD, BPD) scores [38]. In addition, the German translations [39] of the Dutch Screening Device for Psychotic Disorder [40] and the Suicide Risk Assessment Interview [41] will be used to screen for acute schizophrenia and suicidality, respectively.
Primary outcomes. The primary outcomes will be the level of suicidality and depression at post treatment, three- and six-months follow-up. Suicidality and depression will be assessed with the following instruments:
Revised Acquired Capability of Suicide Scale [ACSS-FAD; 42]. This instrument assesses fearlessness from death with seven items using a five-point rating scale (1 = “fits not at all to me”, 5 = “fits completely to me”). It showed a good internal consistency (a = .79) in a German student sample.
Beck-Depression-Inventory [BDI-II; 43]. The BDI-II is one of the most used inventories to examine the severity of depression within the last two weeks. There are 21 symptom groups and participants chose one statement out of four to seven response categories (coded 0, 1, 2, and 3) per symptom. Sum scores of 29 indicate severe depressions. Internal consistencies were a = .90 in a healthy sample and a = .93 in a client sample, respectively.
Beck Scale for Suicide Ideation [BSSI; 44, 45]. Suicidal thinking is assessed with the BSS that has 21 symptom groups, each with three different response categories, representing a three-point rating scale. The first five groups are for screening purposes and the last two groups not included into the total BSS sum score. Higher sum scores (Range 0 to 38) indicate a higher risk of suicide. The internal consistency in a non-clinical German sample was very good (a = .94).
Secondary outcomes. As secondary outcomes, prolonged grief disorder, PTSD hopelessness, anxiety, perceived stigmatization, and perceived social support will be assessed at post-treatment, 3-months follow-up, and 6-months follow-up. These variables will be assessed with the following instruments:
Short Version of the Patient Health Questionnaire [PHQ-D; 46]. The short version of the PHQ-D is a 15-item screening test for depressive disorders (nine items, four-point rating scale: 1 = “not at all”, 4 = “almost every day”), panic disorders, (i.e. panic attacks, five items, dichotomous rating scale: 1 = “no”, 2 = “yes”) and psychosocial impairment (one item, four-point rating scale: 1 = “not at all impaired”, 4 = “very impaired”). Graefe, Zipfel, Herzog, and Loewe [47] report a satisfactory Cronbach’s alpha of a = .88 for the depression scale of the PHQ-D in a German clinical sample.
Screening for Complex PTSD [SkPTBS; 48]. The SkPTBS is a screening instrument for complex PTSD with 34 items including traumatic events (self of witness) and reactions to the most burdening event on a seven-point rating scale (1 = “not at all”, 7 = “completely”). The internal consistency was a = .91 in a German clinical sample.
Revised Impact of Event Scale [IES-R; 49]. We will assess posttraumatic stress reactions within the last seven days with the IES-R, which consists of the three subscales intrusion (seven items), avoidance (eight items), and hyperarousal (seven items) and uses a four-point rating scale (1 = “not at all”, 4 = “often”). Internal consistencies for each subscale were reported for patient groups and vary between = .71 and a = .90.
Posttraumatic Cognitions Inventor [PTCI; 50]. The German version of the PTCI is a 33-item questionnaire that assesses dysfunctional cognitions of traumatic events on a seven-point rating scale (1 = “totally disagree”, 7 = “totally agree”). In a clinical sample, the three subscales negative cognitions about the self (21 items), negative cognitions about the world (seven items), and self-blame (five items) showed satisfactory to very good internal consistencies: a = .97, a = .88, and a = .86, respectively.
Inventory of Complicated Grief [ICG-D; 51]. The German version of the ICG [52] measures the extent of complicated grief symptoms [synonyms: prolonged grief, persistent complex bereavement disorder; 53] with 19 items on a five-point rating-scale (1 = “never”, 5 = “always”). Cronbachs Alpha was a = .87 in a clinical sample.
Grief Experience Questionnaire [GEQ; 54]. The GEQ is a questionnaire that measures reactions to grief and symptoms that are typical for suicide bereavement. In the current study, only the stigmatization subscale is taken. This scale has ten items on a five-point rating scale (1 = “never”, 5 = “almost always”) with a good Cronbach’s Alpha (a = .86) in a sample with bereaved students. The original items were translated into German by the last author and another research assistant using back and forth translation (see Appendix C).
Stigma of Suicide and Suicide Survivor [STOSASS; 55]. To evaluate the level of perceived stigmatization towards completed suicide, we applied the STOSASS, which comprises the two subscales “stigma towards the suicidal person” and “stigma towards the suicide survivor”. Participants indicate their agreement with 17 items on a four-point rating scale (1 = “strongly disagree”, 4 = “strongly agree”). Internal consistencies for the subscales in non-clinical and clinical samples were good (between .79 to .83). The original items were translated into German by the last author and another research assistant using back and forth translation (see Appendix D).
H-Scales [56]. The H-Scales are a German translation of the hopelessness scale by Beck, Weissman, Lester, & Trexler [57]. The short version with ten items to measure the participants’ negative expectations of themselves, the environment and their future life on a six-point rating scale (1 = “completely wrong”, 6 = “completely correct”) will be used. Cronbach’s alpha ranged between a = .74 and a = .92 in clinical and non-clinical samples.
Berlin Social Support Scales [BSSS; 58]. The BSSS assesses social support via five subscales: perceived available social support (eight items, a = .83), actually received social support (eleven items, a = .83), (3) need for support (four items, a = .63), mobilization of social support (five items, a = .819, and protective buffering (six items, a = .82). The questionnaire was validated in a clinical sample and uses a four-point rating scale (1 = “is not correct”, 4 = “is completely correct”). In the current study, only the subscales perceived available social support, need for support, and mobilization of social support will be used.
Questionnaire on Police Delivery of Death News (QPDDN; Hofmann & Wagner). The PDDN is a newly constructed questionnaire to examine how mourners perceive the behavior of the officers who transmitted the death news. The questionnaire consists of 34 items with items of perceived stigmatization, experienced emotional support and perceived behavior of the police (see Appendix E) with 2 items using a six-point rating (1 = “not correct at all”, 6 = “completely correct”). 7 items assess additional information regarding received information material, how the message was delivered, if there was any support (e.g. doctor, pastor), which behavior they found most disturbing and what they would have wished for. Information on the validity and reliability will be provided once the data have been analyzed.
Process measures. In addition to the primary and secondary outcomes, we are also interested in variables that relate to the therapeutic process and the development of suicidal or depressive markers from one treatment session to another. We expect that the depressive and suicidal symptoms will reduce over the course of the intervention. After each session, depression will be assessed with the depression subscale from the PHQ-D [46]. Suicidality will be measured with the “suicide item” from the BDI-II [43].
Furthermore, the therapeutic relationships within the intervention groups were assessed using the German version of the Group Questionnaire [GQ-D; 59]. This inventory consists of the subscales positive bonding relationship ( = .92), positive working relationship ( = .89), and negative relationship ( = .79) and differentiates between group members, group leaders, and the group as a whole. Agreement with the statements is indicated on a seven-point rating scale (1 = “is not correct at all”, 7 = “is completely correct”).
Data management and storage
Personal Data. As part of the registration process, participants provide their name, telephone number, email, and home address. This information is pseudonymized using encrypted codes. The pseudo-anonymization will be saved in a coding list, which exists on paper and only in one version. This list will be kept locked up with no third-party access. As long as the coding list exists, the study participants are able to request the deletion of all data collected from them. The coding list will be destroyed after data collection. Then all data of the study will be completely anonymous and the personal data of participants can no longer be identified.
Furthermore, we assess socio-demographic data, symptoms of general psychopathology, grief or suicidality within the telephone interview and the online surveys. We conduct all surveys using the program Unipark®. The storage and evaluation of the study-related data is carried out according to legal provisions without the participants name or IP address. The data will be transferred to the statistical programs SPSS 20. We keep the data in a place that only employees of this study have access to. We store the digital data on secured servers at the Medical School Berlin for at least ten years. After expiration of the retention obligation, we will delete the data.
The audio recordings that were made in the context of the focus groups will be deleted after the transcription has been completed. The transcription has been anonymized.
Statistical analysis
We will use a repeated measures ANOVA to analyze the changes in the primary endpoint variables (suicidality and depression) and the secondary outcomes from baseline to post-measurement and the two follow-up assessments (3-months, 6-months), whereby the dependent variable will be the treatment condition (treatment group vs. waitlist control group). In addition, we will control for the baseline levels of depression and suicidality using an ANCOVA. The predictor analyses will be conducted using multiple regressions.
Handling of Missing Data
We will follow the principles of intention-to-treat analyses. Participants who drop out of the study after randomization are asked to state their reasons for doing so and are asked to continue the assessments. Missing values will be imputed using the Last Observation Carried Forward approach.
Monitoring, ethical considerations, and safety
The group leaders record each session so that the study employees can randomly examine the adherence to the intervention manual. In addition, the group leaders document each session (e.g. attendance, mood of participants, disturbances) and contact participants individually in case of absence, problematic behavior, suicidality or deterioration of symptoms via the study online portal. In addition, the study employees receive an automatic E-Mail when participants report increased levels of suicidality in the questionnaires after each session. If suicidality or adverse effects are suspected, the study employees will call the participants once a week to monitor and assess the mood and offer support.
The randomized controlled trial was approved by the ethics committee of Medical School Hamburg (March 3, 2018). Participants are informed in a written form (i.e. study information, consent form) about the procedure of the program, the aims of the study, the inclusion and exclusion criteria, risks and costs of participation, confidentiality and data storage. They have the opportunity to drop out of the program at any time without giving reasons. They are also allowed to use other assistance (e.g. psychotherapy) in addition to the study participation.
The participants are also informed about the chance to detect abnormalities in the clinical interview (before treatment) or during the data analysis. In that case, we contact participants for a more detailed clinical assessment. If we find clinically relevant aspects or diagnoses, we inform participants about it. However, we point out that study assessments do not substitute for clinical diagnostics by a medical doctor or psychotherapist.
During treatment, participants may experience a worsening in mood and sadness at first when they deal with the suicide of their loved ones. Such mood aggravations are normal in an intensive grief process and usually exist only for a short time. In case of acute suicidality (i.e. corresponding hints in the suicidality item of the BDI-II), we contact the participant by telephone and offer counselling. He or she may be excluded from the program and referred to local professionals. However, the scientific evidence so far suggests that the proposed online program can reduce difficulties in coping with grief reactions and symptoms. Hence, we do not expect any unwanted side effects or lasting psychopathological worsening at the end of the program. External supervisors will accompany the process of the treatment at regular intervals (every 4th session).
There will be no external Data Monitoring Committee because the sponsor as well as cooperation partners will function as instances where procedures, results and modifications have to be communicated to on a regular basis (at least twice a year). Although the execution of the study and the data analyses are independent of these institutions, we will write several interim reports about the achievement of project aims, results, and finances and justify deviations. If adverse events arise in the suggested procedures (e.g. eligibility criteria, randomization strategy, parts of intervention, etc.) the study coordination will discuss the necessity and impact of possible changes to the study protocol. An audit is not scheduled in our study.