Background
Method
Inclusion and exclusion criteria
Data extraction
Quality assessment
Results
Study characteristics
Author (year)/Location | Eligibility criteria | Sample size | Age (M, SD, range) | Male/female | Time since bereavement/Relationship to the deceased | Type of intervention/Setting | Characteristics of intervention | Duration/Frequency of contact | Outcome/Instrument/Timepoints | Main results |
---|---|---|---|---|---|---|---|---|---|---|
Battle (1984) USA [33] | Intervention: suicide loss, help-seeking from a crisis centre Control: i) Non-help-seeking suicide survivors (non-group control group), ii) patients from the centre (patient control group). | Intervention N = 36 Non-group bereaved control group N = 13 Patients control group N = 31 | Range 14–66 years M = 38 Control groups: not reported | 22% /78% Control groups: not reported | 8 days – 11 years M = 2 months | Group Suicide prevention/crisis intervention service | Support group with educational component non-help seeking survivors, and patients Combination of non-directive and directive leadership Facilitated by clinicians | 1.5 h weekly sessions in first 4 months, followed by 2-weekly 1.5 h sessions 69% attend less than 10 sessions 97% attends max 14 sessions. | Psychosocial functioning incl. Problems, feelings, goals (self-constructed questions) Pre/post intervention | Participants are more emotional (more distress, pain, happiness, pleasure) than non-group controls); suicidal pain 50% vs 0%; perfect happiness: 67% vs 1%; more solutions to their problems than controls. Control patients also experienced strong emotions. Psychodynamic insights in the survivors and the suicides. No p values reported. |
Constantino & Bricker (1996) USA [34] | Widows whose spouse died by suicide | Intervention N = 16 Control N = 16 | M = 43 | Majority is female More men in intervention group than in control | Not reported | Group Setting not reported | Bereavement Group Postvention (BGP), i.e. group psycho therapy vs Social Group Postvention (SGP), e.g., socialization, recreation Directive vs non-directive leadership Facilitated by trained mental health nurses | 1.5 h weekly sessions, 8 weeks | Depression (BDI) Distress (BSI) Social adjustment (SAS) Grief (GEI) Pre/post intervention | Reduction in depression and distress in both BGP and SGP groups (both p < .05). Improvement in social adjustment in SGP group only (p = .003). Grief aspects: despair, rumination and depersonalization decreased in both BGP and SGP groups (all p < .05); anger/hostility and guilt decreased in BGP only (p < .05). |
Constantino et al. (2001) USA [35] | Widows whose spouse died by suicide Aged 18+ English speaking | N = 47 (Originally N = 60, 30 + 30) | Range 24–70 yrs. | Male/female 10/37, 21%/78% | 1–27 months M = 10.91, SD = 8.65 40% less than 6 months Widows | Group Setting not reported | Bereavement Group Postvention (BGP), i.e. group psycho therapy vs Social Group Postvention (SGP), e.g., socialization, recreation Facilitated by trained mental health nurses | 1.5 h weekly sessions, 8 sessions | Depression (BDI) Distress (BSI) Grief (GEI) Social adjustment (SAS) Pre/post intervention, 6-month, 12-month follow-up | N.s. differences between BGP and SGP groups on outcome variables at four timepoints. Analysis of both groups BGP and SGP combined: significant decrease in depression (p =. 0001), total distress (p = .0001), grief symptoms (all p < .05), except for anger/hostility and social isolation, and increase in total social adjustment (p = .0001) over the four time points. |
De Groot et al. (2007) The Netherlands [41] | First-degree relatives or spouses bereaved by suicide Aged 15+ | Intervention N = 68, 39 families Control N = 54, 31 families | Intervention M = 43, SD = 13.7 Control M = 43, SD = 13.5 | Intervention Male/female 28/40, 41%/59% Control Male/female 12/42, 22%/78% | 3–6 months after suicide Spouse (31%) Parent (31%) Child (16%) In-laws (4%) | Group / family Participant’s home | Family-based cognitive behaviour counselling programme vs treatment as usual Facilitated by trained psychiatric nurses | 2 h, 2- -3 weekly sessions, 4 sessions | Grief (ITG, TRGR2L) Depression (CES-D) Suicidal ideation (PSI) Guilt and self-blame (self-constructed questions) Baseline, 13-month follow-up | N.s. effect of intervention on complicated grief, depression, and suicidal ideation. A trend towards feeling less being to blame (p = .01) and fewer maladaptive grief reactions (p = .056). |
De Groot et al. (2010) The Netherlands [42] | First-degree relatives or spouses bereaved by suicide Suicide < 8 weeks Aged 15+ | Total N = 122 Intervention N = 68, 39 families Control N = 54, 31 families | Intervention M = 43, SD = 13.7 Control M = 43, SD = 13.5 | Intervention Male/female 28/40, 41%/59% Control Male/female 12/42, 22%/78% | 3–6 months after suicide Spouse (31%) Parent (31%) Child (16%) In-laws (4%) | Group / family Participant’s home | Family-based cognitive behaviour counselling programme vs treatment as usual Facilitated by trained psychiatric nurses | 2 h, 2–3 weekly sessions, 4 sessions | Grief (ITG, TRGR2L) Depression (CES-D) Suicidal ideation (PSI) Guilt and self-blame (self-constructed questions) Clinical assessment (SCAN2.1) Baseline, 13-month follow-up | Participant with suicidal ideation compared with non-ideators: N.s. decrease of complicated grief in suicide ideators Reduction in maladaptive grief reactions (p = .03) and risk of suicidal ideation (p = .03) among ideators. |
Farberow (1992) USA [36] | Loss by suicide Aged 18+ | Intervention N = 60 Control N = 22 | Range 18–60+ | Intervention Male/female 18/42, 30%/70% Control 5/17, 23%/77% | Less than 3 to 24+ months 77% between 6 and 8 months after death Sibling (35%) Child (23%) Parent (20%) Spouse (13%) Sweathearts and other | Group Suicide prevention centre | Bereavement group with therapeutic and educational aspects vs refusers, dropouts Quasi-experimental design Facilitated by mental health professional and trained peer | 1.5 h weekly sessions, 8 sessions, followed by open monthly sessions | Health, impact of loss, coping, major changes, feelings (self-constructed questions) After suicide (T1), pre/post intervention (T2/T3) | Intervention group: decreased scores from T1 to T2 to T3. Control group: decreased scores from T1 to T2; only in anxiety for T2 to T3. Intervention group at T3 significant higher on depression and puzzlement compared to Control group. No p values reported. |
Hazell & Lewin (1993) Australia [43] | Students selected by school staff on basis of close friendship with deceased student | Intervention N = 63 Control N = 63 | School A: M = 15.1 School B: M = 14.4 | Not reported | Within 7 days after suicide Fellow students of same school | Group School setting | Group counselling and information vs no counselling Facilitated by child psychiatrist or trainee psychiatrist with assistance of senior school staff | 1.5 h session One session | Behaviour (YSR) Risk behaviour (RBQ) Suicidal ideation/behaviour Drug/alcohol use 8 months after suicide | N.s. differences between intervention and control group on internalizing, externalizing, depression, risk behaviour, suicidal ideation/behaviour or drug/alcohol use. |
Kovac & Range (2000) USA [37] | Undergraduate students who had a close person die by suicide in the past 2 years and were upset by the death | Total N = 42 Intervention N = 20 Control N = 22 N = 30 completed follow-up tests | Range 18–46 M = 23.98 SD = 7.34 Intervention M = 23.16, SD = 6.99 Control M = 25, SD = 7.98 | Male/female 9/33, 21%/79% Intervention 5/14, 26%/74% Control 3/18, 14%/86% | Intervention M = 13.26 months, SD = 9.32 Control M = 11.95, SD = 6.54 Not reported | Individual Experimental/laboratory setting | Writing task: profound, death-related writing vs trivial writing Facilitated by researchers | 15 min sessions, 4 sessions over 2 weeks | Grief (GRQ, GEQ) Impact of grief (IES) Essay evaluation form Experiment follow-up form Pre/post-test (6 weeks) | Reduction in impact of grief (p < .05), and general GRQ grief levels (p < .05) in intervention and control group. Suicide-specific grief GEQ more reduced in intervention than control group (p < .05). No difference in self-reported health visits between groups. |
Pfeffer et al. (2002) USA [38] | Families where child’s parent or sibling died by suicide Children aged 6–15 No psychiatric disorders | Total N = 75 children, 52 families Intervention N = 39, 27 fam Control N = 36, 25 fam | Intervention M = 9.6, SD = 2.9 Control M = 11.4, SD = 3.5 | Male/female Intervention 16/23, 41%/59% Control 12/24, 33%/67% | Within a year after death Siblings (11/39), children (28/39) and parents | Group / family Clinical setting | Manual based bereavement group intervention for children grouped by age Psycho-educational, support group for parents No treatment control Facilitated by trained psychologists | 1.5 h weekly sessions 10 sessions | Children: Posttraumatic stress symptoms (CPTSRI) Depression (CDI) Anxiety (RCMAS) Social adjustment (SAICA) Parents: depression (BDI) Pre/post intervention (12 weeks) | Children: Significantly greater reduction in anxiety and depressive symptoms in intervention vs. control group (p ≤ .01). N.s. differences in posttraumatic stress or social adjustment. Parents: N.s. differences in depression between groups. |
Sandor et al. (1994) USA [39] | Members of a church-related youth group | Intervention N = 15 Control n = 19 | Intervention Range 14–17, M = 15.73 Control range 14–18, M = 16.37 | Intervention Male/female 5/15, 33%/67%r Control Male/female 6/13, 32%/68% | A few days after the death Relationship: peer group | Group Church youth group | Supportive community intervention; Survivor Group (SG) vs Comparison Group (CG; no intervention) Quasi experimental design Facilitated by church youth group leaders | 3 meetings: 2 h open session with youth and parents, after two days one closed psycho-educational session with youth, a day later a memorial service in church | Problem solving (APSA) Self-perception (HSP) Self-efficacy (SES) Baseline (T1), 2 months after suicide (T2), 2-month follow-up (T3) | Greater self-efficacy at T2 and T3 compared to T1 in SG vs. CG group (p < .01). Greater social acceptance and job competence at T2 in SG vs. CG (p < .05), but not at T3. SG vs CG group not compared on problem-solving appraisal, scholastic competence, and global self-worth. |
Wittouck et al. (2014) Belgium [44] | Suicide of a significant other 3–24 months before participation Aged 18+ Dutch speaking | Intervention N = 47 Control N = 36 | Intervention M = 49.3, SD = 13.8 Control M = 47.6, SD = 12.8 | Intervention male/female 9/38, 19%/81% Control 11/25, 31%/69% | Intervention 9.8 months, SD = 5.7 Control M = 12.4, SD = 6.3 Deceased: child (n = 20; 42%), partner (n = 12, 25%), parent (n = 1, 2%), sibling (n = 8, 17%), other (n = 6, 13%) | Group / family Participant’s home | Cognitive-behavioral therapy-based psycho-educational intervention vs no treatment Facilitated by clinical psychologist | 2 h sessions, 4 sessions Frequency not reported | Depression BDI-II) Hopelessness (BHS) Grief (CGQ, ITG) Coping (UCL) Baseline, 8 (months | N.s. decrease in depression, hopelessness and grief in intervention vs. control group. Decrease in intensity of grief, depression, passive coping style, social support seeking and behavioural expression of (negative) feelings in intervention group only (all p < .05). |
Zisook et al. (2018) USA [40] | People bereaved by suicide SB), accident, homicide (A/H), and natural causes (NC) with ≥30 ITG score Exclusion: Being suicidal, other psychiatric disorders except depression, other treatments | Total N = 395 SB N = 58 A/H N = 74 NC N = 263 SB Medication N = 14 Placebo N = 17 CGT + Medication N = 17 CGT + Placebo N = 13 | Range 18–95 SB M = 47.2, SD = 14.1 A/H M = 51.6, SD = 14.8 NC N = 54.6, SD = 14.2 | Male/female SB 10/48, 17%/82% A/H 18/56, 24%/76% NC 59/204, 22%/78% | Time since death SB M = 3.9 yrs. SD = 4.6 A/H M = 6.6 SD = 9.1 NC 4.3 SD = 7.1 SB deceased is a partner 18 (31%), parent 7 (12%), child 19 (33%), other 14 (24%) | Individual Clinical setting | Manual-based structured Complicated Grief Therapy (CGT) Therapists, including social workers, psychiatrists, psychologists Antidepressant medication (citalopram)with individual follow-up Facilitated by Trained researcher | CGT: 16 sessions over 20 weeks Medication: 12-week with 2–4 weekly visits until week 20 | Grief (CG-CGI-I, GRAQ, ICG, SCI-CG, TBAQ) Suicidal ideation (C-SSRS-R) Work/social adjustment (WSAS) | Lower improvement on clinician-rated CG-CGI-I in SB vs. A/H and NC groups (p < 0.5). N.s. differences on other measures of grief, suicidal ideation or work/social adjustment between SB, A/H and NC groups. Low rates of post treatment active suicidal ideation in SB, A/H and NC groups. |
Quality assessment
Quality Criteria | Battle (1984) USA [33] | Constantino & Bricker (1996) USA [34] | Constantino et al. (2001) USA [35] | De Groot et al. (2007; 2010) | Farberow (1992) USA [36] | Hazell & Lewin (1993) Australia [43] | Kovac & Range (2000) USA [37] | Pfeffer et al. (2002) USA [38] | Sandor et al. (1994) USA [39] | Wittouck et al. (2014) Belgium [44] | Zisook et al. (2018) USA [40] |
---|---|---|---|---|---|---|---|---|---|---|---|
A. Selection bias | |||||||||||
Representativeness | Not likely | Not likely | Not likely | Somewhat likely | Somewhat likely | Somewhat likely | Somewhat likely | Somewhat likely | Can’t tell | Not likely | Not likely |
Percentage agreed | Can’t tell | 80–100% | Can’t tell | < 60% | Can’t tell | 60–79% | 60–79% | 60–79% | Can’t tell | 80–100% | Can’t tell |
Rating | Weak | Weak | Weak | Weak | Moderate | Moderate | Moderate | Moderate | Weak | Weak | Weak |
B. Study design | |||||||||||
Study design type | Other: 3 groups comparison | RCT | RCT | RCT | Cohort analytic | Case-control | RCT | RCT | Cohort analytic | RCT | RCT |
Described as randomized? | No | Yes | Yes | Yes | No | No | Yes | Yes | No | Yes | Yes |
Method of randomization described? | N.a. | No | Yes | Yes | N.a. | N.a. | No | Yes | N.a. | Yes | Yes |
Method appropriate? | N.a. | No | Yes | Yes | N.a. | N.a. | No | Yes | N.a. | Yes | Yes |
Rating | Weak | Strong | Strong | Strong | Moderate | Moderate | Strong | Strong | Moderate | Strong | Strong |
C. Confounders | |||||||||||
Pre-intervention differences? | Can’t tell | No | No | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes |
Percentage confounders controlled for | Can’t tell | N.a. | N.a. | 80–100% (most) | Can’t tell | 80–100% | 60–79% (some) | 80–100% (most) | 80–100% | 80–100% | < 60% (few or none) |
Rating | Weak | Strong | Strong | Strong | Weak | Strong | Moderate | Strong | Strong | Strong | Weak |
D. Blinding | |||||||||||
Outcome assessors were blinded? | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | No | Can’t tell | Yes | Can’t tell | No | Yes |
Participants were blinded? | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Yes | Can’t tell | Can’t tell | Can’t tell | Yes |
Rating | Weak | Weak | Weak | Weak | Weak | Weak | Moderate | Moderate | Weak | Weak | Strong |
E. Data collection methods | |||||||||||
Valid measures? | Can’t tell | Yes | Yes | Yes | Can’t tell | Yes | Yes | Yes | Yes | Yes | Yes |
Reliable measures? | Can’t tell | Yes | Yes | Yes | Can’t tell | Can’t tell | Yes | Yes | Yes | Yes | Yes |
Rating | Weak | Strong | Strong | Strong | Weak | Moderate | Strong | Strong | Strong | Strong | Strong |
F. Withdrawals and drop-outs | |||||||||||
Numbers and reasons reported per group? | Yes | Can’t tell | No | Yes | Can’t tell | No | No | No | Yes | Yes | No |
Percentage completing study? | N.a. | 80–100% | 60–79% | 80–100% | Can’t tell | 80–100% | 60–79% | < 60% | 80–100% | 80–100% | < 60% |
Rating | N.a. | Strong | Moderate | Strong | Weak | Weak | Weak | Weak | Strong | Strong | Weak |
Total A-F: | WEAK | WEAK | WEAK | WEAK | WEAK | WEAK | MODERATE | MODERATE | WEAK | WEAK | WEAK |
Number of ‘strong’ ratings | 0/6 | 4/6 | 3/6 | 4/6 | 0/6 | 1/6 | 2/6 | 3/6 | 3/6 | 4/6 | 3/6 |
G. Intervention integrity | |||||||||||
Percentage participants received intervention? | Can’t tell | 80–100% | 60–79% | 80–100% | 80–100% | 80–100% | 60–79% | 80–100% | Can’t tell | 80–100% | 60–79% |
Intervention consistency measured? | Can’t tell | Yes | Yes | Yes | Yes | No | Yes | Yes | Can’t tell | Can’t tell | Yes |
Confounding unintended intervention? | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell | Can’t tell |
H. Analyses | |||||||||||
Unit of allocation | Individual | Individual | Individual | Individual (family) | Individual | Individual | Individual | Individual (family) | Individual | Individual | Individual |
Unit of analysis | Individual | Individual | Individual | Individual | Individual | Individual | Individual | Individual | Individual | Individual | Individual |
Appropriate statistical methods? | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Analysis by intention-to-treat status | No | Can’t tell | No | Yes | No | Can’t tell | Can’t tell | No | Can’t tell | No | Can’t tell |