Psychosocial interventions for self‐harm in adults
Abstract
Background
Self‐harm (SH; intentional self‐poisoning or self‐injury regardless of degree of suicidal intent or other types of motivation) is a growing problem in most counties, often repeated, and associated with suicide. There has been a substantial increase in both the number of trials and therapeutic approaches of psychosocial interventions for SH in adults. This review therefore updates a previous Cochrane Review (last published in 2016) on the role of psychosocial interventions in the treatment of SH in adults.
Objectives
To assess the effects of psychosocial interventions for self‐harm (SH) compared to comparison types of care (e.g. treatment‐as‐usual, routine psychiatric care, enhanced usual care, active comparator) for adults (aged 18 years or older) who engage in SH.
Search methods
We searched the Cochrane Common Mental Disorders Specialised Register, the Cochrane Library (Central Register of Controlled Trials [CENTRAL] and Cochrane Database of Systematic reviews [CDSR]), together with MEDLINE, Ovid Embase, and PsycINFO (to 4 July 2020).
Selection criteria
We included all randomised controlled trials (RCTs) comparing interventions of specific psychosocial treatments versus treatment‐as‐usual (TAU), routine psychiatric care, enhanced usual care (EUC), active comparator, or a combination of these, in the treatment of adults with a recent (within six months of trial entry) episode of SH resulting in presentation to hospital or clinical services. The primary outcome was the occurrence of a repeated episode of SH over a maximum follow‐up period of two years. Secondary outcomes included treatment adherence, depression, hopelessness, general functioning, social functioning, suicidal ideation, and suicide.
Data collection and analysis
We independently selected trials, extracted data, and appraised trial quality. For binary outcomes, we calculated odds ratio (ORs) and their 95% confidence intervals (CIs). For continuous outcomes, we calculated mean differences (MDs) or standardised mean differences (SMDs) and 95% CIs. The overall quality of evidence for the primary outcome (i.e. repetition of SH at post‐intervention) was appraised for each intervention using the GRADE approach.
Main results
We included data from 76 trials with a total of 21,414 participants. Participants in these trials were predominately female (61.9%) with a mean age of 31.8 years (standard deviation [SD] 11.7 years). On the basis of data from four trials, individual cognitive behavioural therapy (CBT)‐based psychotherapy may reduce repetition of SH as compared to TAU or another comparator by the end of the intervention (OR 0.35, 95% CI 0.12 to 1.02; N = 238; k = 4; GRADE: low certainty evidence), although there was imprecision in the effect estimate. At longer follow‐up time points (e.g., 6‐ and 12‐months) there was some evidence that individual CBT‐based psychotherapy may reduce SH repetition. Whilst there may be a slightly lower rate of SH repetition for dialectical behaviour therapy (DBT) (66.0%) as compared to TAU or alternative psychotherapy (68.2%), the evidence remains uncertain as to whether DBT reduces absolute repetition of SH by the post‐intervention assessment. On the basis of data from a single trial, mentalisation‐based therapy (MBT) reduces repetition of SH and frequency of SH by the post‐intervention assessment (OR 0.35, 95% CI 0.17 to 0.73; N = 134; k = 1; GRADE: high‐certainty evidence). A group‐based emotion‐regulation psychotherapy may also reduce repetition of SH by the post‐intervention assessment based on evidence from two trials by the same author group (OR 0.34, 95% CI 0.13 to 0.88; N = 83; k = 2; moderate‐certainty evidence). There is probably little to no effect for different variants of DBT on absolute repetition of SH, including DBT group‐based skills training, DBT individual skills training, or an experimental form of DBT in which participants were given significantly longer cognitive exposure to stressful events. The evidence remains uncertain as to whether provision of information and support, based on the Suicide Trends in At‐Risk Territories (START) and the SUicide‐PREvention Multisite Intervention Study on Suicidal behaviors (SUPRE‐MISS) models, have any effect on repetition of SH by the post‐intervention assessment. There was no evidence of a difference for psychodynamic psychotherapy, case management, general practitioner (GP) management, remote contact interventions, and other multimodal interventions, or a variety of brief emergency department‐based interventions.
Authors' conclusions
Overall, there were significant methodological limitations across the trials included in this review. Given the moderate or very low quality of the available evidence, there is only uncertain evidence regarding a number of psychosocial interventions for adults who engage in SH. Psychosocial therapy based on CBT approaches may result in fewer individuals repeating SH at longer follow‐up time points, although no such effect was found at the post‐intervention assessment and the quality of evidence, according to the GRADE criteria, was low. Given findings in single trials, or trials by the same author group, both MBT and group‐based emotion regulation therapy should be further developed and evaluated in adults. DBT may also lead to a reduction in frequency of SH. Other interventions were mostly evaluated in single trials of moderate to very low quality such that the evidence relating to the use of these interventions is inconclusive at present.
PICOs
Plain language summary
Psychosocial interventions for adults who self‐harm
We have reviewed the interventional literature regarding psychosocial intervention treatment trials in the field. A total of 76 trials meeting our inclusion criteria were identified. There may be beneficial effects for psychological therapy based on cognitive behavioural therapy (CBT) approaches at longer follow‐up time points, and for mentalisation‐based therapy (MBT), and emotion‐regulation psychotherapy at the post‐intervention assessment. There may also be some evidence of effectiveness of standard dialectical behaviour therapy (DBT) on frequency of SH repetition. There was no clear evidence of effect for case management, information and support, remote contact interventions (e.g. emergency cards, postcards, telephone‐based psychotherapy), provision of information and support, and other multimodal interventions.
Why is this review important?
Self‐harm (SH), which includes intentional self‐poisoning/overdose and self‐injury, is a major problem in many countries and is strongly linked with suicide. It is therefore important that effective treatments are developed for people who engage in SH. There has been an increase in both the number of trials and the diversity of therapeutic approaches for SH in adults in recent years. It is therefore important to assess the evidence for their effectiveness.
Who will be interested in this review?
Hospital administrators (e.g. service providers), health policy officers and third party payers (e.g. health insurers), clinicians working with people who engage in SH, the people themselves, and their relatives.
What questions does this review aim to answer?
This review is an update of a previous Cochrane review from 2016 which found that CBT‐based psychological therapy can result in fewer individuals repeating SH whilst DBT may lead to a reduction in frequency of repeated SH. This updated review aims to further evaluate the evidence for effectiveness of psychosocial interventions for people engaging in SH with a broader range of outcomes.
Which studies were included in the review?
To be included in the review, studies had to be randomised controlled trials of psychosocial interventions for adults who had recently engaged in SH.
What does the evidence from the review tell us?
Overall, there were a number of methodological limitations across the trials included in this review. We found positive effects for psychological therapy based on CBT approaches at longer follow‐up assessments, and for mentalisation‐based therapy (MBT), and emotion‐regulation psychotherapy on repetition of SH at post‐intervention. There may also be some evidence of effects for standard dialectical behaviour therapy (DBT) on frequency of SH repetition. However, remote contact interventions, case management, information and support, and other multimodal interventions do not appear to have benefits in terms of reducing repetition of SH.
What should happen next?
The promising results for CBT‐based psychotherapy at longer follow‐up time points, and for MBT, group‐based emotion regulation, and DBT warrant further investigation to understand which people benefit from these types of interventions. Greater use of head‐to‐head trials (where treatments are directly compared with each other) may also assist in identifying which component(s) from these often complex interventions may be most effective.
Authors' conclusions
Summary of findings
| Comparison 1.1: Individual‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with CBT‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.35 | 238 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of case management on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 200 per 1000 | 80 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 1.2: Group‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with CBT‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.66 | 313 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of group‐based CBT‐based psychotherapy on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 190 per 1000 | 134 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.1: DBT compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT | |||||
| Repetition of SH by post‐intervention | Study population | OR 0.71 | 502 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate of DBT on repetition of SH at post‐intervention. The true effect is likely to be substantially different from the estimate of effect. | |
| 682 per 1000 | 604 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by two levels as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for two or more of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.2: DBT group‐based skills training compared to TAU or alternative psychotherapy for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT group‐based skills training | |||||
| Repetition of attempted suicide at post‐intervention | Study population | OR 0.66 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT group‐based skills training on repetition of attempted suicide at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 364 per 1000 | 274 per 1000 | |||||
| Repetition of NSSI at post‐intervention | Study population | OR 0.88 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT group‐based skills training on repetition of NSSI at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 576 per 1000 | 544 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.3: DBT individual therapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT individual therapy | |||||
| Repetition of attempted suicide at post‐intervention | Study population | OR 1.46 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT individual therapy on repetition of attempted suicide at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 364 per 1000 | 455 per 1000 | |||||
| Repetition of NSSI at post‐intervention | Study population | OR 1.29 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT individual therapy on repetition of NSSI at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 576 per 1000 | 636 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.4: DBT prolonged exposure protocol compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT prolonged exposure protocol | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.67 | 18 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT prolonged exposure protocol on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 333 per 1000 | 251 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 3: MBT compared TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with MBT | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.35 | 134 | ⊕⊕⊕⊕ | We are very confident that the true effect for MBT on repetition of SH at post‐intervention lies close to that of the estimate of the effect. | |
| 492 per 1000 | 253 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Comparison 4: Emotion‐regulation psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Group‐based emotion‐regulation psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.34 | 83 | ⊕⊕⊕⊝ | We are moderately confident that the true effect of group‐based emotion‐regulation psychotherapy on repetition of SH at post‐intervention lies close to that of the estimate of the effect. | |
| 775 per 1000 | 539 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgradedthis domain by one level as one study was suggestive of benefit, whilst the second trial of this intervention was not. | ||||||
| Psychodynamic psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Psychodynamic psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.45 | 170 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of psychodynamic psychotherapy on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 133 per 1000 | 65 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Case management compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Case management | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.78 | 1608 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of case management on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 114 per 1000 | 91 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Structured general practitioner (GP) follow‐up compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Structured general practitioner (GP) follow‐up | |||||
| Repetition of SH at post‐intervention (hospital records) | Study population | OR 1.01 | 143 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of structured general practitioner (GP) follow‐up on repetition of SH (according to hospital records) at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 133 per 1000 | 134 per 1000 | |||||
| Repetition of SH at post‐intervention (emergency records) | Study population | OR 2.56 | 123 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of structured general practitioner (GP) follow‐up on repetition of SH (according to emergency records) at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 72 per 1000 | 167 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.1: Emergency cards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Emergency cards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.82 | 1039 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of emergency cards on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 171 per 1000 | 145 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.2: Coping cards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Coping cards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.08 | 64 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of coping cards on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 156 per 1000 | 15 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.4: Postcards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Postcards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.87 | 3277 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for postcards on repetition of SH. The true effect is likely to be substantially different from the estimate of effect. | |
| 132 per 1000 | 117 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.5: Telephone contact compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Telephone contact | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.43 | 55 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of telephone contact on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 77 per 1000 | 35 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.6: Telephone contact, emergency cards, and letters compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Telephone contact, emergency cards, and letters | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.05 | 303 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of telephone contact, emergency cards, and letters on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 139 per 1000 | 145 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.7: Telephone‐based psychotherapy compared to TAU or alternative psychotherapy for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Telephone‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.36 | 185 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of telephone‐based psychotherapy on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 11 per 1000 | 4 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 10: Provision of information and support compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Provision of information and support | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.09 | 1853 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for information and support on repetition of SH by the post‐intervention assessment. The true effect is likely to be substantially different from the estimate of effect. | |
| 90 per 1000 | 97 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Publication bias is suspected as data from some centres have not been published. 2 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 11: Other multimodal interventions compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Other multimodal interventions | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.61 | 1937 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for other multimodal interventions on repetition of SH at post‐intervention. The true effect is likely to be substantially different from the estimate of effect. | |
| 252 per 1000 | 189 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by two levels as the I2 value indicated considerable levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 12.5: General hospital management compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with General hospital management | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.03 | 77 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of general hospital management on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 51 per 1000 | 53 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 12.8: Long‐term therapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults Intervention: Long term therapy | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Long term therapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.00 | 80 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of long‐term therapy on repetition of SH at post‐intervention is limited.The true effect may be substantially different from the estimate of the effect. | |
| 225 per 1000 | 225 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
Background
Description of the condition
Self‐harm (SH), which includes all intentional acts of self‐poisoning (such as intentional drug overdoses) or self‐injury (such as self‐cutting), regardless of degree of suicidal intent or other types of motivation (Hawton 2003), has been a growing problem in most countries. In Australia, for example, it is estimated that there are now more than 26,000 general hospitalisations for SH each year, or a rate of 116.7 per 100,000 persons (Harrison 2014), similar to rates observed in a number of other comparable countries (Canner 2018; Griffin 2014; Morthorst 2016; Ting 2012; Wilkinson 2002). However, it is notable that rates of emergency department presentations for SH are often higher than hospitalisations (Bergen 2010; Corcoran 2015). In the UK, for example, higher rates of emergency department presentations for SH in both females (442 per 100,000) and males (362 per 100,000) have been reported (Geulayov 2016). There are also many more episodes of SH occurring in the community that do not come to the attention of clinical services. Worldwide, for example, the World Health Organization (WHO) estimates that the rate of SH may be as high as 400 per 100,000, according to self‐report data (WHO 2014).
In contrast to suicide rates, rates of hospital‐presenting SH are higher in females than in males in most countries (Canner 2018; Griffin 2014; Masiran 2017; Morthorst 2016; Ting 2012; Wilkinson 2002), with rates peaking in younger adults up to 24 years of age (Perry 2012). However, this difference decreases across the life cycle (Hawton 2008). SH is less common in older people, but tends to be associated with higher suicidal intent (Hawton 2008), with consequent greater risk of suicide (Murphy 2012).
For those who present to hospital, the most common method of SH is self‐poisoning. Overdoses of analgesics and psychotropics, especially paracetamol or acetaminophen, are common in some countries, particularly high‐income countries. Self‐cutting is the next most frequent method used by those who present to hospital. However, in the community, self‐cutting and other forms of self‐injury are far more frequent than self‐poisoning (Müller 2016).
SH is often repeated. Up to one‐quarter of those who present to hospital with SH return to the same hospital within a year (Carroll 2014; Owens 2002), although some individuals may present to another hospital. Others may not present to hospital at all given that studies identifying SH repetition via self‐report suggest as many as one in five report further SH episodes following a hospital presentation (Carroll 2014). Repetition is more common in individuals who have a history of previous episodes of SH, personality disorder, psychiatric treatment, and alcohol or drug misuse (Larkin 2014). Risks of repeat SH may also be associated with method. Rates of repetition are higher among those who present to hospital following self‐injury alone (Carroll 2014; Lilley 2008), or combined self‐injury and self‐poisoning (Perry 2012), compared to those who present for self‐poisoning alone.
SH is associated with suicide. The risk of death by suicide within one year among people who present to hospital following SH varies across studies from nearly 1% to over 3% (Liu 2020; Carroll 2014; Owens 2002). This variation reflects the characteristics of the population, and the background national suicide rate. In the UK, for example, during the first year after an episode of SH, the risk of suicide is around 50 times that of the general population, with a particularly high risk in men (Carroll 2014; Geulayov 2019). One quarter of these deaths are estimated to occur within one month after discharge, and almost 50% by three months (Forte 2019), although the risk of suicide appears to remain elevated for a number of years (Geulayov 2019). A history of SH is the strongest risk factor for suicide across a range of psychiatric disorders. Repetition of SH further increases the risk of suicide (Zahl 2004).
SH and suicide are the result of a complex interplay between genetic, biological, psychiatric, psychosocial, social, cultural, and other factors. Psychiatric disorders, particularly mood and anxiety disorders, are associated with the largest contribution to the risk of both SH (Hawton 2013), and suicide in adults (Ferrari 2014). Personality disorders, including borderline personality disorder, are also associated with SH, particularly frequent repetition. Alcohol use may also play an important role (Ferrari 2014). Both psychological and biological factors appear to further increase vulnerability to SH. Psychological factors may include difficulties in problem‐solving, low self‐esteem, impulsivity, vulnerability to having pessimistic thoughts about the future (i.e. hopelessness), and a sense of entrapment. Biological factors include disturbances in the serotonergic and stress response systems (Van Heeringen 2014).
Description of the intervention
Psychological approaches used to treat adults who engage in SH typically involve brief individual‐ or group‐based psychological therapy. Treatment may vary in terms of initial management, location of treatment, continuity, intensity, and frequency of contact with therapists. There is also considerable variation among countries in the availability of services to provide such interventions. Consequently, there is no standard psychosocial treatment of SH in adults. However, in high‐income countries, treatment generally consists of a combination of assessment, support, and individual psychological therapies. In lower and middle‐income countries, aftercare more usually involves various forms of support, both face‐to‐face and via digital means.
How the intervention might work
Psychosocial interventions may address some of the underlying psychological risk factors associated with SH. The mechanisms of action of these interventions might help people improve their coping skills and tackle specific problems, manage psychiatric disorders, improve self‐esteem, increase a sense of social connectedness, and reduce impulsivity and harmful reactions to distressing situations. What follows is a description of the psychosocial interventions that are typically available for adults who engage in SH behaviours.
Cognitive behavioural therapy‐based psychotherapy
Cognitive behavioural therapy (CBT)‐based psychotherapy helps people identify and critically evaluate the ways in which they interpret and evaluate disturbing emotional experiences and events, and aims to help them change the ways in which they deal with problems (Westbrook 2008). This is achieved in three steps: first, people are helped to change the ways in which they interpret and evaluate distressing emotions; second, they learn strategies to help them change the way in which they think about the meanings and consequences of these emotions; finally, with the benefit of modified interpretation of emotions and events, they are helped to change their behaviour and develop positive functional behaviour (Jones 2012).
Problem‐solving therapy (PST) is an integral part of CBT, although it can be delivered as a therapy in and of itself. PST assumes that ineffective and maladaptive coping behaviours that drive SH might be overcome by helping the person to learn skills to actively, constructively, and effectively solve the problems he or she faces in their daily lives (Nezu 2010). PST typically involves identification of the problem, generation of a range of solutions, implementation of chosen solutions based on appraisal, and the evaluation of these solutions (D'Zurilla 2010). Treatment goals include helping people to develop a positive problem‐solving orientation, use rational problem‐solving strategies, reduce the tendency to avoid problem‐solving, and reduce the use of impulsive problem‐solving strategies.
Dialectical behaviour therapy
In contrast to CBT and PST, which focus on changing behaviour and cognitive patterns, the focus of dialectical behavioural therapy (DBT) is to provide people with the skills to develop an awareness and acceptance of thoughts and emotions, including painful or distressing internal experiences, without judgement or attempts to alter, suppress, avoid, or otherwise change these experiences (Lynch 2006). The primary treatment goals of DBT are three‐fold: to reduce SH, reduce behaviours that interfere with the success of treatment, such as treatment non‐adherence, and reduce any other factors that may adversely affect the person's quality of life (e.g. frequency or duration of psychiatric hospitalisations) (Linehan 1993).
Mentalisation‐based therapy
Mentalisation refers to the ability to understand the behaviour of both one's self and others in terms of motivational and emotional states (Allen 2008). Maladaptive and impulsive coping behaviours, including SH, are presumed to arise from a disrupted ability to engage in these processes. In mentalisation‐based therapy (MBT), the goal is to help people understand their emotions and behaviours, and develop strategies to regulate them to minimise the risk that they will engage in SH during times of distress.
Emotion‐regulation psychotherapy
Emotion‐regulation psychotherapy is based on DBT, Acceptance and Commitment Therapy (ACT), and emotion‐focused therapy. It includes psychoeducation, identification of emotions, distress tolerance, emotional acceptance, behavioural activation, developing alternative coping strategies, impulse control, and identifying and clarifying valued directions.
Psychodynamic psychotherapy
Psychodynamic approaches focus on affective experiences, exploring and understanding the unconscious meaning and function of SH, and exploring and resolving difficulties in interpersonal relationships, and resulting emotional difficulties, within a therapeutic relational framework (Yakeley 2018).
Case management
"In its simplest form...case management is a means of coordinating services. Each...person is assigned a 'case manager' who is expected to assess that person's needs, develop a care plan, arrange for suitable care to be provided, monitor the quality of the care provided, and maintain contact with the person" (Marshall 2000a). Case management might have a significant role in the aftercare of people who engage in SH because many of them demonstrate poor treatment adherence, and because of the varied nature of the problems these individuals are often facing (Lizardi 2010).
Remote contact interventions
Remote contact interventions, which may include letters, brief text messages delivered by telephone, telephone calls, and postcards, are low‐resource and non‐intrusive interventions that seek to maintain long‐term contact with people. These interventions provide a sense of ongoing concern, and may mitigate the sense of social isolation reported by many people who engage in SH. They may also help to improve their knowledge about triggers and warning signs for SH, provide them with information on alternative coping behaviours to SH, and where they can access help (Milner 2016).
These interventions may also be combined with emergency card interventions, which encourage people to seek help when they feel distressed, and offer on‐demand emergency contact with psychiatric services or inpatient care. The aim is to reduce the risk of SH by facilitating rapid access to care.
Other multimodal interventions
Any of the above active psychosocial interventions may also be combined into a multimodal approach (Högberg 2008).
Why it is important to do this review
SH is a major social and healthcare problem. It represents significant morbidity, is often repeated, and is linked with suicide. Many countries now have suicide prevention strategies, all of which include a focus on improved management of people presenting with SH (WHO 2014). SH also leads to substantial healthcare costs (Sinclair 2011). In the UK, the overall median cost per episode of SH has been estimated to be £809, although costs are significantly higher for cases of combined self‐injury and self‐poisoning, compared to either self‐injury of self‐poisoning alone. These costs are mainly attributable to health‐service level contact (i.e. inpatient stay or admission to intensive care; Tsiachristas 2017).
In the UK, the National Collaborating Centre for Mental Health (NCCMH) produced the first guideline on the treatment of SH behaviours in 2004 (NCCMH 2004). This guideline focused on the short‐term physical and psychological management of SH. This guidance was updated in 2011, using interim data from a previous version of this review as the evidence base, and focused on the longer‐term psychological management of SH (NICE 2011). Subsequently, similar guidelines have been published by the Royal College of Psychiatrists (Royal College of Psychiatrists 2014), the Royal Australian and New Zealand College of Psychiatrists (Carter 2016), and German Professional Associations and Societies (Plener 2016), amongst others (Courtney 2019).
In 2021, the guidance contained in the 2011 NICE guidelines for the longer‐term management of SH will be due for updating. Therefore, we are updating our review (Hawton 2016), in order to provide contemporary evidence to guide clinical policy and practice.
Objectives
To assess the effects of psychosocial interventions for self‐harm (SH) compared to comparison types of care (e.g. treatment as usual, routine psychiatric care, enhanced usual care, active comparator) for adults (aged 18 years or older) who engage in SH.
Methods
Criteria for considering studies for this review
Types of studies
We considered all randomised controlled trials (RCTs) of specific psychosocial interventions versus treatment as usual, routine psychiatric care, enhanced usual care, active comparator, or a combination of these, in the treatment of adults with a recent (within six months of trial entry) presentation for self‐harm (SH). All RCTs (including cluster‐RCTs [cRCTs] and cross‐over trials) were eligible for inclusion regardless of publication type or language; however, we excluded quasi‐randomised trials.
Types of participants
While exact eligibility criteria often differ both within and between regions and countries (Witt 2020a), we included participants of both sexes and all ethnicities, who were 18 years and older, with a recent (i.e. within six months of trial entry) presentation to hospital or clinical services for SH.
We defined SH as all intentional acts of self‐poisoning (such as intentional drug overdoses) or self‐injury (such as self‐cutting), regardless of the degree of suicidal intent or other types of motivation (Hawton 2003). This definition includes acts intended to result in death ('attempted suicide'), those without suicidal intent (e.g. to communicate distress, to temporarily reduce unpleasant feelings; sometimes termed 'non‐suicidal self‐injury'), and those with mixed motivation. We did not distinguish between attempted suicide and non‐suicidal self‐injury in this review, because there is a high level of co‐occurrence between them, and the two cannot be distinguished in any reliable way, including on levels of suicidal intent (Klonsky 2011). Lastly, the motivations for SH are complex and can change, even within a single episode (De Beurs 2018).
We excluded trials in which participants presented to clinical services for suicidal ideation only (i.e. without evidence of SH).
Types of interventions
Categorisation of the interventions in this review was informed by the trials themselves, and based on consensus discussions among members of the review team, who have considerable experience in both research and clinical practice related to SH. However, based on the previous version of this review (Hawton 2016), we anticipated the following groupings:
Interventions
These included the following:
-
Cognitive behavioural therapy (CBT)‐based psychotherapy (e.g. CBT, problem‐solving therapy [PST]) versus TAU or another comparator;
-
Dialectical behaviour therapy (DBT) versus TAU or another comparator;
-
Mentalisation‐based therapy (MBT) versus TAU or another comparator;
-
Emotion‐regulation psychotherapy versus TAU or another comparator;
-
Psychodynamic psychotherapy versus TAU or another comparator;
-
Case management versus TAU or another comparator;
-
Structured general practitioner (GP) follow‐up versus TAU or another comparator;
-
Brief emergency department‐based interventions versus TAU or another comparator;
-
Remote contact interventions versus TAU or another comparator;
-
Provision of information and support versus TAU or another comparator;
-
Other multimodal interventions versus TAU or another comparator;
-
Other mixed interventions versus TAU or another comparator.
Comparators
Treatment‐as‐usual (TAU) is likely to vary widely between settings and between studies conducted over different time periods (Witt 2018). Following previous work, we defined TAU as routine clinical care that the person would receive had they not been included in the study (i.e. routine care or 'standard disposition'; Hunt 2013). Other comparators could include no specific treatment, or enhanced usual care, which refers to TAU that has, in some way, been supplemented, such as by providing psychoeducation, assertive outreach, or more regular contact with case managers, and standard assessment approaches.
Types of outcome measures
For all outcomes, we were primarily interested in quantifying the effect of treatment assignment to the intervention at baseline, regardless of whether the intervention was received as intended (i.e. the intention‐to‐treat effect).
Primary outcomes
The primary outcome measure in this review was the occurrence of repeated SH over a maximum follow‐up period of two years. Repetition of SH may be identified through self‐report, collateral report, clinical records, or research monitoring systems. As we wished to incorporate the maximum data from each trial, we included both self‐reported and hospital records of SH, where available. Preference was given to clinical records over self‐report where a trial reported both measures. We also reported proportions of participants repeating SH, frequency of repeat episodes, and time to SH repetition (where available).
Secondary outcomes
Given increasing interest in the measurement of outcomes of importance to those who engage in SH (Owens 2020b), we planned to analyse data for the following secondary outcomes (where available) over a maximum follow‐up period of two years:
Treatment adherence
This was assessed using a range of measures of adherence, including: pill counts, changes in blood measures, and the proportion of participants that both started and completed treatment.
Depression
This was assessed as either continuous data, by scores on psychometric measures of depression symptoms, for example, total scores on the Beck Depression Inventory (BDI; Beck 1961), or scores on the depression subscale of the Hospital Anxiety and Depression Scale (HADS; Zigmond 1983), or as dichotomous data as the proportion of participants who met defined diagnostic criteria for depression.
Hopelessness
This was assessed as either continuous data, by scores on psychometric measures of hopelessness, for example, total scores on the Beck Hopelessness Scale (BHS; Beck 1974), or as dichotomous data as the proportion of participants reporting hopelessness.
General functioning
This was assessed as either continuous data, by scores on psychometric measures of general functioning, for example, total scores on the Global Assessment of Functioning (GAF; APA 2000), or as dichotomous data as the proportion of participants reporting improved general functioning.
Social functioning
This was assessed as either continuous data, by scores on psychometric measures of social functioning, for example, total scores on the Social Adjustment Scale (SAS; Weissman 1999), or as dichotomous data as the proportion of participants reporting improved social functioning.
Suicidal ideation
This was assessed as either continuous data, by scores on psychometric measures of suicidal ideation, for example, total scores on the Beck Scale for Suicidal Ideation (BSSI; Beck 1988), or as dichotomous data as the proportion of participants reaching a defined cut‐off for ideation.
Suicide
This included register‐recorded deaths, or reports from collateral informants, such as family members or neighbours.
Search methods for identification of studies
Electronic searches
An information specialist searched the following databases (to 4 July 2020), using relevant subject headings (controlled vocabularies) and search syntax as appropriate for each resource: Cochrane Common Mental Disorders Specialised Register (Appendix 1), Cochrane Library (Central Register of Controlled Trials; CENTRAL), Cochrane Database of Systematic Reviews (CDSR), MEDLINE Ovid, Embase Ovid, and PsycINFO Ovid (Appendix 2).
A date restriction was applied to the search as the search was to update an earlier version of this review (Hawton 2016). However, we did not apply any further restrictions on language or publication status to the searches.
We searched for retraction statements and errata once the included studies were selected.
We also searched the World Health Organization International Clinical Trials Registry Platform, and the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov to identify ongoing trials.
The search was based on population only, participants who self‐harm (all ages). Records were screened to identify trials which were relevant to this review and two others (Witt 2021a; Witt 2021b).
Searching other resources
Conference abstracts
In addition to conference abstracts retrieved via the main electronic search, we also screened the proceedings of recent (last five years) conferences organised by the largest scientific committees in the field:
-
International Association for Suicide Prevention (both global congresses and regional conferences);
-
Joint International Academy of Suicide Research and American Foundation for Suicide Prevention International Summits on Suicide Research.
Reference lists
We also checked the reference lists of all relevant RCTs, and the reference lists of major reviews that included a focus on psychosocial interventions for SH in adults (Bornheimer 2020; Briggs 2019; Calati 2016; D'Anci 2019; DeCou 2019; Ghanbari 2015; Gøtzsche 2017; Hawton 2016; Hanratty 2019; Hetrick 2016; Inagaki 2019; Krysinska 2017; Meerwijk 2016; Milner 2015; Milner 2016; Padmanathan 2020; Riblet 2017; Tighe 2018).
Correspondence
We consulted the corresponding authors of trials, and other experts in the field to find out if they are aware of any ongoing or unpublished RCTs on the psychosocial treatment of adults who engage in SH that were not identified by the electronic searches.
Data collection and analysis
Selection of studies
Review authors KW, KH, and one of either SH, GR, TTS, ET, or PH, independently assessed the titles of reports identified by the electronic search for eligibility. We distinguished between:
-
eligible or potentially eligible trials for retrieval, in which any psychosocial intervention was compared with a comparator (i.e. TAU, routine psychiatric care, standard disposition, no specific treatment, enhanced usual care);
-
ineligible, general treatment trials, not for retrieval (i.e. where there was no control treatment).
All trials identified as potentially eligible for inclusion then underwent a second screening. Pairs of review authors, working independently from one another, screened the full text of eligible or potentially eligible trials to identify whether the trial met our inclusion criteria. We resolved disagreements in consultation with the senior review author (KH). Where disagreements could not be resolved from the information reported in the trial, or where it was unclear whether the trial satisfied our inclusion criteria, we contacted corresponding trial authors for additional clarification.
We identified and excluded duplicate records, and collated multiple reports of the same trial, so that each trial, rather than each report, represented the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram (Liberati 2009), and completed a 'Characteristics of excluded studies' table.
Data extraction and management
Review author KW and one of either SH, or GR independently extracted data from the included trials, using a standardised extraction form. Where there were any disagreements, they were resolved in consensus discussions with KH.
Data extracted from each eligible trial included the following:
-
Participant information: number randomised, number lost to follow‐up or withdrawn, number analysed, mean or median age, sex composition, diagnoses, diagnostic criteria, inclusion criteria, and exclusion criteria;
-
Methods: trial design, total duration of the trial, details of any 'run‐in' period (if applicable), number of trial centres and their location, setting, and date;
-
Intervention(s): details of the intervention, including dose, duration, whether concomitant treatments were permitted and details of these treatments, and any excluded treatments;
-
Comparator(s): details on the comparator, including dose, duration, whether concomitant treatments were permitted and details of these treatments, and any excluded treatments;
-
Outcomes: raw data for each eligible outcome (see Types of outcome measures), details of other outcomes specified and reported, and time points at which outcome were reported;
-
Notes: source of trial funding, and any notable conflicts of interest of trial authors.
We extracted both dichotomous and continuous outcomes data from eligible trials. As the use of non‐validated psychometric scales is associated with bias, we extracted continuous data only if the psychometric scale used to measure the outcome of interest had been previously published in a peer‐reviewed journal, and was not subjected to item, scoring, or other modification by the trial authors (Marshall 2000b).
We planned the following main comparisons:
-
Cognitive behavioural therapy (CBT)‐based psychotherapy (e.g. CBT, problem‐solving therapy [PST]) versus TAU or another comparator;
-
Dialectical behaviour therapy (DBT) versus TAU or another comparator;
-
Mentalisation‐based therapy (MBT) versus TAU or another comparator;
-
Emotion‐regulation psychotherapy versus TAU or another comparator;
-
Case management versus TAU or another comparator;
-
Remote contact interventions versus TAU or another comparator;
-
Other multimodal interventions versus TAU or another comparator.
Assessment of risk of bias in included studies
Highly biased studies are more likely to overestimate treatment effectiveness (Moher 1998). Review author KW and one of either SH, or GR independently evaluated the risk of bias for the primary outcome (i.e. repetition of SH at post‐intervention) by using version 2 of the Cochrane Risk of Bias tool, RoB 2 (Sterne 2019). This tool encourages consideration of the following domains:
-
Bias in the randomisation process;
-
Deviations from the intended intervention (assignment to intervention);
-
Missing outcome data;
-
Bias in the measurement of the outcome;
-
Bias in the selection of the reported result.
For cluster‐RCTs, we also evaluated the following:
-
Bias arising from the timing of identification and recruitment of participants.
Signalling questions in the RoB 2 tool provided the basis for the tool’s domain‐level judgements about the risk of bias. Two review authors independently judged each source of potential bias as low risk, high risk, or some concerns. An overall 'Risk of bias' judgement was then made for each study by combining ratings across these domains. Specifically, if any of the above domains were rated at high risk, the overall 'Risk of bias' judgement was rated at high risk. We reported this overall judgement, which could be low risk, high risk, or some concerns, in the text of the review, and in the 'Risk of bias' tables.
Where inadequate details were provided in the original report, we contacted corresponding trial authors to provide clarification. We resolved disagreements through discussions with KH.
We entered and organised our RoB 2 assessments on an Excel spreadsheet (Microsoft Excel RoB2 Macro), and made them available as electronic supplements.
Measures of treatment effect
Dichotomous outcomes
We summarised dichotomous outcomes, such as the number of participants engaging in a repeat SH episode, or number of deaths by suicide, using the summary odds ratio (OR) and the accompanying 95% confidence interval (CI), as the OR is the most appropriate effect size statistic for summarising associations between two dichotomous groups (Fleiss 1994). Time to SH repetition was summarised using the hazard ratio (HR) and its accompanying 95% CI.
Continuous outcomes
For outcomes measured on a continuous scale, we used mean differences (MDs) and accompanying 95% CI where the same outcome measure was used. Where different outcome measures were used, we used the standardised mean difference (SMD) and its accompanying 95% CI.
We aggregated trials in a meta‐analysis only where treatments were sufficiently similar. For trials that could not be included in a meta‐analysis, we provided narrative descriptions of the results.
Hierarchy of outcomes
Where a trial measured the same outcome, for example, depression in two or more ways, we planned to use the most common measure across trials in any meta‐analysis. We also planned to report scores from other measures in a supplementary table.
Timing of outcome assessment
The primary end point for this review was post‐intervention (i.e. at the conclusion of the treatment period). We also reported outcomes for the following secondary end points (where data were available):
-
Between zero and six months after the conclusion of the treatment period;
-
Between six and 12 months after the conclusion of the treatment period;
-
Between 12 and 24 months after the conclusion of the treatment period.
Where there was more than one outcome assessment within a time period, we used data from the last assessment in the time period, unless different outcomes were assessed at different time points. For treatment adherence, we also planned to use within‐treatment results.
Unit of analysis issues
Zelen design trials
Trials in this area are increasingly using Zelen's method, in which consent is obtained subsequent to randomisation and treatment allocation (Witt 2020a). This design may lead to bias if, for example, participants allocated to one particular arm of the trial disproportionally refuse to provide consent for participation or, alternatively, if participants only provide consent if they are allowed to cross over to the other treatment arm (Torgerson 2004).
We identified five trials that used Zelen's design (Carter 2005; Gysin‐Maillart 2016; Hatcher 2011; Hatcher 2015; Hatcher 2016). Where possible, we therefore analysed data for these trials using data for all randomised participants as this is consistent with Zelen’s original intention (Zelen 1979), and preserves randomisation. This was typically possible for our primary outcome, repetition of SH, as this was generally ascertained from clinical, hospital, and/or medical records in these five trials. However, for certain self‐reported secondary outcome measures, data were only reported on the basis of those who consented to participation. We therefore planned to conduct sensitivity analyses to investigate what impact, if any, the inclusion of these trials may have on the pooled estimate of treatment effectiveness.
Cluster‐randomised trials
Cluster randomisation, for example by clinician or general practice, can lead to overestimation of the significance of a treatment effect, resulting in an inflation of the nominal type I error rate, unless appropriate adjustment is made for the effects of clustering (Donner 2002; Kerry 1998).
One trial included in this review used cluster randomisation (Bennewith 2002). We had planned to follow the guidance outlined in Higgins 2019a. Specifically, where possible, we planned to analyse data using measures that statistically accounted for the cluster design. Where this was not possible, we planned to analyse data using the effective sample size. However, the trial authors were unable to provide values for either the inter‐cluster correlation coefficient or the design effect, and further, there was no similar cluster‐RCT of this intervention approach from which these values could be approximated. We were therefore unable to statistically account for the effects of clustering in this trial.
In future updates of this review, should we be able to obtain information on either the inter‐cluster correlation coefficient or the design effect, we will follow the guidance outlined in Higgins 2019a.
Cross‐over trials
A primary concern with cross‐over trials is the carry‐over effect, in which the effect of the intervention treatment (e.g. pharmacological, physiological, psychological) influences the participant's response to the subsequent control condition (Elbourne 2002). As a consequence, on entry to the second phase of the trial, participants may differ systematically from their initial state, despite a wash‐out phase. In turn, this may result in a concomitant underestimation of the effectiveness of the treatment intervention (Curtin 2002a; Curtin 2002b).
One trial included in the current review used cross‐over methodology (Marasinghe 2012). To protect against the carry‐over effect, we only extracted data from the first phase of this trial, prior to cross‐over.
Studies with multiple treatment arms
Five trials in the current review included multiple treatment arms (Andreoli 2015; Armitage 2016; Linehan 2015; Stewart 2009; Wei 2013. For two of these trials (Andreoli 2015; Stewart 2009), it was possible to combine data from the two intervention arms given their similarity. For the remaining three, however, two different psychotherapies were compared (Armitage 2016; Linehan 2015; Wei 2013. Therefore, we included information from these trials in both categories of intervention using data from the relevant intervention arm. As we did not combine data from the intervention arms in any meta‐analysis, we used the same comparator arm data for both analyses.
Studies with adjusted effect sizes
Where trials reported both unadjusted and adjusted effect sizes, we included only observed, unadjusted effect sizes.
Dealing with missing data
We did not impute missing data, as we considered that the bias that would be introduced by doing this would outweigh any benefit of increased statistical power that may have been gained by including imputed data. However, where authors omitted standard deviations (SD) for continuous measures, we contacted corresponding authors to request missing data. Where missing data could not be provided, we calculated missing SDs using other data from the trial, such as CIs, based on methods outlined in Higgins 2019b.
Assessment of heterogeneity
Between‐study heterogeneity can be assessed using either the Chi² or I² statistics. However, in this review, we used only the I² statistic to quantify inconsistency, as this is considered to be more reliable (Deeks 2019). The I² statistic indicates the percentage of between‐study variation due to chance, and can take any value from 0% to 100% (Deeks 2019).
We used the following values to denote relative importance of heterogeneity, as per Deeks 2019:
-
unimportant: 0% to 40%;
-
moderate: 30% to 60%;
-
substantial: 50% to 90%;
-
considerable: 75% to 100%.
We also took the magnitude and direction of effects and strength of evidence for heterogeneity into account (e.g. the CI for I²).
Where substantial levels of heterogeneity were found, we explored reasons for this heterogeneity (see Subgroup analysis and investigation of heterogeneity for details).
Assessment of reporting biases
Reporting bias occurs when the decision to publish a particular trial is influenced by the direction and significance of the results (Egger 1997). Research suggests, for example, that trials with statistically significant findings are more likely to be submitted for publication and, subsequently, be accepted for publication, leading to possible overestimation of the true treatment effect (Hopewell 2009).
To assess whether trials included in any meta‐analysis were affected by reporting bias, we planned to enter data into a funnel plot when a meta‐analysis included results of at least ten trials. Should evidence of any small study effects be identified, we planned to explore reasons for funnel plot asymmetry, including the presence of possible publication bias (Egger 1997).
Data synthesis
For the purposes of this review, we calculated the pooled odds ratio (OR) and accompanying 95% CI using the random‐effects model, as this is the most appropriate model for incorporating heterogeneity between studies (Deeks 2019). We used the Mantel‐Haenszel method for dichotomous data, and the inverse variance method for continuous data. We conducted all analyses in Review Manager 5.4 (Review Manager 2020).
Subgroup analysis and investigation of heterogeneity
Subgroup analyses
We planned to undertake the following subgroup analyses where there were sufficient data to do so:
-
sex (males versus females);
-
repeater status (first SH episode versus repeat SH episode).
Given the increasing use of enhanced usual care rather than TAU in trials in this field (Witt 2020a), we also planned to undertake subgroup analyses to determine whether comparator choice influenced the pattern of results observed.
Formal tests for subgroup differences were undertaken in Review Manager 5.4 (Review Manager 2020). However, it is only possible to undertake these subgroup analyses if randomisation was stratified by these factors, otherwise, there is the risk that doing so could lead to confounding.
Randomisation was stratified by sex in five trials (Carter 2005; Hvid 2011; McAuliffe 2014; Walton 2020; Van der Sande 1997). Six trials also stratified randomisation by repeater status (Gratz 2014; Hatcher 2015; Hatcher 2016; McAuliffe 2014; Morthorst 2012; Mouaffak 2015).
Investigation of heterogeneity
Several meta‐analyses were associated with substantial levels of between‐study heterogeneity (i.e. I² ≥ 75%). For these analyses, KW and KH firstly independently triple‐checked data to ensure these were correctly entered. Next, we investigated the source of this heterogeneity using a formal statistical approach as outlined in Viechtbauer 2020.
Sensitivity analysis
We planned to undertake the following sensitivity analyses, where appropriate, to test whether key methodological factors or decisions may have influenced the main result:
-
Where a trial made use of Zelen's method of randomisation (see Unit of analysis issues);
-
Where a trial contributed to substantial between‐study heterogeneity (see Subgroup analysis and investigation of heterogeneity).
Five trials used Zelen's design (Carter 2005; Gysin‐Maillart 2016; Hatcher 2011; Hatcher 2015; Hatcher 2016). We were therefore able to undertaken sensitivity analyses, excluding these trials, to investigate what impact, if any, Zelen's design had on the pooled estimate of treatment effectiveness. We reported results of these sensitivity analyses in the text.
Additionally, several meta‐analyses were associated with substantial levels of between‐study heterogeneity. We also reported results of these sensitivity analyses in the text, alongside discussion of the likely causes of these differences.
Summary of findings and assessment of the certainty of the evidence
For each comparison, we planned to construct a 'Summary of findings' table for our primary outcome measure, repetition of SH post‐intervention, following the recommendations outlined in Schünemann 2019. These tables provide information concerning the overall quality of the evidence from all included trials that measured the outcome. We assessed the quality of evidence across the following domains:
-
'Risk of bias' assessment;
-
Indirectness of evidence;
-
Unexplained heterogeneity or inconsistency of results;
-
Imprecision of effect estimates;
-
Potential publication bias.
For each of these domains, we downgraded the evidence from high certainty by one level (for serious) or by two levels (for very serious). For risk of bias, we downgraded this domain by one level when we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for any of the studies included in the pooled estimate, or by two levels when we rated multiple studies at high risk for any of these sources. For indirectness of evidence, we considered the extent to which trials included in any meta‐analysis used proxy measures to ascertain repetition of SH; we downgraded this domain by one level if one study used proxy measures, and by two levels if multiple studies used proxy measures. For unexplained heterogeneity or inconsistency of results, we downgraded this domain by one level where the I² value indicated substantial levels of heterogeneity, or by two levels where the I² value indicated considerable levels of heterogeneity. For imprecision, we downgraded this domain by one level where the 95% CI for the pooled effect included the null value. Finally, for the potential publication bias domain, we considered any evidence of funnel plot asymmetry (if available), as well as other evidence such as suspected selective availability of data, and downgraded by one or more levels where publication bias was suspected.
We then used these domains to rate the overall quality of evidence for the primary outcome according to the following:
-
High certainty: further research is very unlikely to change our confidence in the estimate of effect;
-
Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate;
-
Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect, and may change the estimate;
-
Very low certainty: we are very uncertain about the estimate.
We constructed 'Summary of findings' tables using GRADEpro GDT software (GRADEpro GDT 2015).
Reaching conclusions
We based our conclusions only on findings from the quantitative or narrative synthesis of the studies included in this review. Our recommendations for practice and research suggest priorities for future research, and outline remaining uncertainties in the area.
Results
Description of studies
Results of the search
For this update, a total of 7186 records were found using the search strategy as outlined in Appendix 1 and Appendix 2. Five further records were identified following correspondence and discussion with researchers in the field. After deduplication, the initial number was reduced to 4678. Of these, 4454 were excluded following title/abstract screening, whist a further 157 were excluded after reviewing the full texts (Figure 1).
Study Flow Diagram
Included studies
In the previous version of this review (Hawton 2016), 55 trials of psychosocial interventions for self‐harm (SH) in adults were included. The present update located 21 new trials of psychosocial interventions for SH in adults. The present review therefore includes 76 non‐overlapping trials (see the Characteristics of included studies tables for further information on these trials).
Unpublished data were obtained from the trial authors for 42 of these trials (Bateman 2009; Beautrais 2010; Bennewith 2002; Brown 2005; Carter 2005; Cedereke 2002; Clarke 2002; Crawford 2010; Davidson 2014; Dubois 1999; Fleischmann 2008; Gratz 2006; Gratz 2014; Guthrie 2001; Gysin‐Maillart 2016; Harned 2014; Hassanian‐Moghaddam 2011; Hatcher 2011; Hatcher 2015; Hatcher 2016; Husain 2014; Hvid 2011; Kapur 2013; Kawanishi 2014; Linehan 1991; Linehan 2006; Marasinghe 2012; McAuliffe 2014; McMain 2009; Owens 2020; Patsiokas 1985; Priebe 2012; Slee 2008; Sreedaran 2020; Stewart 2009; Tapolaa 2010; Turner 2000; Tyrer 2003; Vaiva 2006; Walton 2020; Wei 2013; Weinberg 2006).
Design
Most trials (97.4%) randomised at the individual‐level employing either simple randomisation (Allard 1992; Armitage 2016; Beautrais 2010; Brown 2005; Clarke 2002; Crawford 2010; Davidson 2014; Dubois 1999; Evans 1999a; Evans 1999b; Fleischmann 2008; Gibbons 1978; Gratz 2006; Hawton 1981; Hawton 1987; Kapur 2013; Liberman 1981; McLeavey 1994; Morgan 1993; Mousavi 2015; Mousavi 2017; Naidoo 2014; O'Connor 2015; O'Connor 2020; Patsiokas 1985; Sahin 2018; Salkovskis 1990; Slee 2008; Stewart 2009; Tapolaa 2010; Torhorst 1987; Torhorst 1988; Turner 2000; Van Heeringen 1995; Wang 2016; Waterhouse 1990; Wei 2013; Weinberg 2006; Welu 1977), or a restricted randomisation scheme such as adaptive minimisation (Linehan 2015), blocked randomisation (Amadéo 2015; Andreoli 2015; Grimholt 2015; Guthrie 2001; Hassanian‐Moghaddam 2011; Husain 2014; Lin 2020; McMain 2009; McMain 2017; Priebe 2012; Sreedaran 2020; Vaiva 2018), or minimisation procedure (Bateman 2009, Harned 2014, Kawanishi 2014; Linehan 1991; Linehan 2006; O'Connor 2017; Owens 2020) procedure. One trial used a matched pair randomisation procedure (Cedereke 2002). Nine trials used stratification (Gratz 2014; Hvid 2011; McAuliffe 2014; Morthorst 2012; Mouaffak 2015; Tyrer 2003; Vaiva 2006; Van der Sande 1997; Walton 2020). Five used Zelen's design (Carter 2005; Gysin‐Maillart 2016; Hatcher 2011; Hatcher 2015; Hatcher 2016).
In one trial, cluster randomisation was used (Bennewith 2002). One trial was a cross‐over RCT (Marasinghe 2012).
Setting
Of the 76 independent RCTs included in this review, one‐quarter (25.0%) were from the UK (Bateman 2009; Bennewith 2002; Clarke 2002; Crawford 2010; Davidson 2014; Evans 1999a; Evans 1999b; Gibbons 1978; Guthrie 2001; Hawton 1981; Hawton 1987; Kapur 2013; Morgan 1993; O'Connor 2017; Owens 2020; Priebe 2012; Salkovskis 1990; Tyrer 2003; Waterhouse 1990), 14 were from the USA (Brown 2005; Gratz 2006; Gratz 2014; Harned 2014; Liberman 1981; Linehan 1991; Linehan 2006; Linehan 2015; O'Connor 2015; O'Connor 2020; Patsiokas 1985; Turner 2000; Weinberg 2006; Welu 1977), four from France (Dubois 1999; Mouaffak 2015; Vaiva 2006; Vaiva 2018), four from New Zealand (Beautrais 2010; Hatcher 2011; Hatcher 2015; Hatcher 2016), three from Australia (Carter 2005; Stewart 2009; Walton 2020), three from Canada (Allard 1992; McMain 2009; McMain 2017), three from Iran (Hassanian‐Moghaddam 2011; Mousavi 2015; Mousavi 2017), two from Denmark (Hvid 2011; Morthorst 2012), two from Germany (Torhorst 1987; Torhorst 1988), two from the Netherlands (Slee 2008; Van der Sande 1997), two from the Republic of Ireland (McAuliffe 2014; McLeavey 1994), two from Switzerland (Andreoli 2015; Gysin‐Maillart 2016), two from Sweden (Cedereke 2002; Sahin 2018), two from Taiwan (Lin 2020; Wang 2016), and one was from each of Belgium (Van Heeringen 1995), China (Wei 2013), Finland (Tapolaa 2010), French Polynesia (Amadéo 2015), India (Sreedaran 2020), Japan (Kawanishi 2014), Malaysia (Armitage 2016), Norway (Grimholt 2015), Pakistan (Husain 2014), South Africa (Naidoo 2014), and Sri Lanka (Marasinghe 2012). One was a multicentre trial conducted in a number of countries (Fleischmann 2008).
In the majority of trials, participants were identified following a clinical presentation for SH. In 16 trials (21.1%), participants were identifying following referral to outpatient mental health and/or specialist personality disorder treatment services (Bateman 2009; Gratz 2006; Gratz 2014; Harned 2014; Liberman 1981; Linehan 1991; Linehan 2006; Linehan 2015; McMain 2009; McMain 2017; Priebe 2012; Sahin 2018; Salkovskis 1990; Walton 2020; Wang 2016; Weinberg 2006). All participants in these trials had a history of SH resulting in presentation to clinical services within six months preceding trial entry.
For most trials (85.5%), treatment was delivered in an outpatient setting or in the participants' home environment. In four trials, acute‐phase treatment was delivered in an inpatient setting followed by outpatient follow‐up appointments (Amadéo 2015; Torhorst 1987; Torhorst 1988; Van der Sande 1997), whilst in two older trials, treatment was delivered in an inpatient setting (Liberman 1981; Waterhouse 1990). In the remaining five trials, the intervention was delivered while individuals were receiving treatment in hospital and/or the emergency department (Armitage 2016; Crawford 2010; O'Connor 2015; O'Connor 2017; O'Connor 2020).
Participants and participant characteristics
The included trials comprised a total of 21,414 participants. All had engaged in at least one episode of SH prior to trial entry. A history of SH prior to the index episode (i.e. a history of multiple episodes of SH) was a requirement for participation in 17 trials (Bateman 2009; Evans 1999b; Gratz 2006; Gratz 2014; Harned 2014; Liberman 1981; Linehan 1991; Linehan 2006; McMain 2009; Mousavi 2015; Mousavi 2017; Priebe 2012; Sahin 2018; Salkovskis 1990; Torhorst 1988; Tyrer 2003; Weinberg 2006). In 20 trials, around half of the sample had a history of multiple episodes of SH (Allard 1992; Brown 2005; Cedereke 2002; Evans 1999a; Guthrie 2001; Gysin‐Maillart 2016; Hatcher 2011; Hatcher 2015; Hatcher 2016; Kapur 2013; Kawanishi 2014; Lin 2020; Morthorst 2012; Mouaffak 2015; O'Connor 2017; Owens 2020; Slee 2008; Torhorst 1987; Van der Sande 1997; Welu 1977) Only one trial excluded those with a history of multiple episodes of SH prior to trial entry (Morgan 1993).
Information on the methods of SH for the index episode was not reported in half (k = 38; 50.0%) of the trials (Allard 1992; Amadéo 2015; Andreoli 2015; Armitage 2016; Bateman 2009; Cedereke 2002; Davidson 2014; Dubois 1999; Evans 1999b; Fleischmann 2008; Gratz 2006; Gratz 2014; Hvid 2011; Kapur 2013; Liberman 1981; Lin 2020; Linehan 1991; Linehan 2006; Linehan 2015; Marasinghe 2012; McMain 2009; Morthorst 2012; Mousavi 2017; Naidoo 2014; O'Connor 2020; Patsiokas 1985; Priebe 2012; Sahin 2018; Salkovskis 1990; Sreedaran 2020; Stewart 2009; Tapolaa 2010; Turner 2000; Tyrer 2003; Walton 2020; Wang 2016; Wei 2013; Weinberg 2006). Full details of the methods used at the index episode for the remaining trials is provided in Table 1. Whilst the predominance of participants engaging in self‐poisoning in the majority of these trials is reflective of the typical pattern observed in those who present to hospital, or in the community, SH more often involves self‐cutting and other forms of self‐injury (Müller 2016).
| Reference | Method | |||
| Self‐poisoning n (%) | Self‐injury n (%) | Combined self‐poisoning and self‐injury n (%) | Unspecified n (%) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 250 (76.7) | 64 (19.6) | ‐ | 15 (4.6) | |
| 7,733 (89.7) | 158 (8.2) | ‐ | 41 (2.1) | |
| 70 (58.3) | 33 (27.5) | ‐ | 17 (14.2) | |
| 772 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 442 (94.6) | 25 (5.3) | 8 (1.70) | ‐ | |
| 74 (71.8) | 25 (24.3) | ‐ | 4 (3.9) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 808 (97.7) | ‐ | ‐ | 19 (2.3) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 400 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 149 (100) | ‐ | ‐ | ‐ | |
| 119 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | 26 (100) | ‐ | ‐ | |
| 2,300 (100) | ‐ | ‐ | ‐ | |
| 471 (85.3) | 81 (14.7) | ‐ | ‐ | |
| 532 (77.8) | 125 (18.3) | 27 (3.9) | ‐ | |
| 115 (68.9) | 41 (24.5) | 11 (6.6) | ‐ | |
| 96 (100) | ‐ | ‐ | ‐ | |
| 80 (100) | ‐ | ‐ | ‐ | |
| 217 (98.2) | 4 (1.8) | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 707 (77.3) | 332 (36.3) | ‐ | 42 (4.6) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 161 (37.2) | 57 (13.2) | ‐ | 4 (0.9) | |
| 39 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 207 (97.6) | ‐ | ‐ | 5 (2.4) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 55 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 504 (97.3) | ‐ | ‐ | 14 (2.7) | |
| ‐ | ‐ | ‐ | ‐ | |
| 40 (64.6) | 12 (19.3) | 10 (16.1) | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 73 (89.0) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 141 (100) | ‐ | ‐ | ‐ | |
| 80 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 605 (100) | ‐ | ‐ | ‐ | |
| 926 (93.8) | ‐ | ‐ | ‐ | |
| 232 (84.7) | ‐ | ‐ | 42 (15.3) | |
| 463 (89.7) | ‐ | ‐ | 53 (10.3) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 77 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | 120 (100) | ‐ | ‐ | |
1 Percentages greater than 100% as participants could have used multiple methods.
2 Methods of self‐harm for the remaining 221 (48.7%) participants were not reported.
3 The majority of participants engaged in 'non‐violent' methods (n = 258; 85.2%), followed by 'violent' methods (n = 45; 14.8%). The methods considered to be 'non‐violent' or 'violent' in this trial were not reported, however.
4 The majority of participants engaged in 'low lethality' methods (n = 57; 89.1%), followed by 'high lethality' methods, including hanging, charcoal burning, or jumping from a height (n = 7; 10.9%). The methods considered to be 'low lethality' in this trial were not reported, however.
Whilst most trials included both male and female participants, the majority of participants in the 71 trials that reported information on sex were female (61.9%), reflecting the typical pattern for SH (Hawton 2008). Of the 64 trials that reported information on age, the weighted mean age of participants at trial entry was 31.8 years (SD: 11.7 years).
In the 38 trials that reported information on psychiatric diagnoses (Allard 1992; Amadéo 2015; Andreoli 2015; Bateman 2009; Brown 2005; Carter 2005; Cedereke 2002; Clarke 2002; Crawford 2010; Davidson 2014; Dubois 1999; Evans 1999a; Evans 1999b; Gibbons 1978; Gratz 2006; Gratz 2014; Gysin‐Maillart 2016; Harned 2014; Kawanishi 2014; Lin 2020; Linehan 2006; Linehan 2015; McLeavey 1994; McMain 2009; McMain 2017; Morgan 1993; Mouaffak 2015; Priebe 2012; Sahin 2018; Slee 2008; Sreedaran 2020; Turner 2000; Tyrer 2003; Vaiva 2018; Van der Sande 1997; Van Heeringen 1995; Walton 2020; Wei 2013), participants were most commonly diagnosed with major depression (56.7%), followed by any other mood disorder (37.6%), any personality disorder (31.6%), any anxiety disorder (26.7%), and substance use disorder (26.7%). A total of 14 trials focused specifically on participants diagnosed with borderline personality disorder (Andreoli 2015; Bateman 2009; Gratz 2006; Gratz 2014; Harned 2014; Linehan 1991; Linehan 2006; Linehan 2015; McMain 2009; McMain 2017; Sahin 2018; Turner 2000; Walton 2020; Weinberg 2006), and three focused on participants diagnosed with any personality disorder (Davidson 2014; Evans 1999b; Priebe 2012). Only two trials reported information on the proportion of participants without psychiatric diagnoses at trial entry (Evans 1999a; Sreedaran 2020). In these two trials, a small proportion (15.1%) were not diagnosed with a major psychiatric disorder.
Information on comorbid diagnoses was reported in eight trials (Andreoli 2015; Bateman 2009; Davidson 2014; Gratz 2014Harned 2014; Mouaffak 2015; Sahin 2018; Turner 2000). Under one‐half (40.9%) were diagnosed with comorbid psychiatric diagnoses, typically comorbid mood or anxiety disorders and personality disorders, although the nature of these comorbidities was not always clearly reported in these trials. In three further trials (Carter 2005; McMain 2009; Slee 2008), the median number of psychiatric diagnoses was greater than two, indicating that most participants in these trials were also diagnosed with more than one psychiatric disorder; however, none of the three reported further information on specific diagnoses.
Interventions
The trials included in this review investigated the effectiveness of various forms of psychosocial interventions:
-
Cognitive behavioural therapy (CBT)‐based psychotherapy (e.g. CBT, problem‐solving therapy [PST]) versus TAU or another comparator;
-
Dialectical behaviour therapy (DBT) versus TAU or another comparator;
-
Mentalisation‐based therapy (MBT) versus TAU or another comparator;
-
Emotion‐regulation psychotherapy versus TAU or another comparator;
-
Psychodynamic psychotherapy versus TAU or another comparator;
-
Case management versus TAU or another comparator;
-
Structured general practitioner (GP) follow‐up versus TAU or another comparator;
-
Brief emergency department‐based interventions versus TAU or another comparator;
-
Remote contact interventions versus TAU or another comparator;
-
Provision of information and support versus TAU or another comparator;
-
Other multimodal interventions versus TAU or another comparator;
-
Other mixed interventions versus TAU or another comparator.
Comparators
Of the 76 RCTs included in this review, the majority (76.3%) compared the intervention to TAU (Allard 1992; Amadéo 2015; Andreoli 2015; Armitage 2016; Bateman 2009; Beautrais 2010; Bennewith 2002; Brown 2005; Carter 2005; Cedereke 2002; Clarke 2002; Crawford 2010; Davidson 2014; Dubois 1999; Evans 1999a; Evans 1999b; Fleischmann 2008; Gibbons 1978; Gratz 2006; Gratz 2014; Grimholt 2015; Guthrie 2001; Gysin‐Maillart 2016; Hassanian‐Moghaddam 2011; Hatcher 2011; Hatcher 2015; Hatcher 2016; Hawton 1987; Husain 2014; Hvid 2011; Kapur 2013; Lin 2020; Linehan 1991; McAuliffe 2014; McMain 2009; Morgan 1993; Mouaffak 2015; Mousavi 2017; O'Connor 2015; O'Connor 2017; O'Connor 2020; Owens 2020; Patsiokas 1985; Priebe 2012; Sahin 2018; Salkovskis 1990; Slee 2008; Stewart 2009; Tapolaa 2010; Tyrer 2003; Vaiva 2006; Vaiva 2018; Van der Sande 1997; Van Heeringen 1995; Wang 2016; Weinberg 2006; Welu 1977). The remaining trials compared the effectiveness of the intervention to alternative forms of psychotherapy (Harned 2014; Hawton 1981; Liberman 1981; Linehan 2006; Linehan 2015; McLeavey 1994; Mousavi 2015; Naidoo 2014; Sreedaran 2020; Torhorst 1987; Torhorst 1988; Turner 2000; Walton 2020), waiting list (Marasinghe 2012; McMain 2017), no treatment (Waterhouse 1990; Wei 2013), or enhanced usual care (EUC; Kawanishi 2014; Morthorst 2012).
Outcomes
Primary outcome
Information on the primary outcome, repetition of SH, was available for all but two (3.0%) of the included trials (Patsiokas 1985; Sreedaran 2020).
In the majority of trials, repetition of SH was ascertained either from self‐reported information alone (Amadéo 2015; Armitage 2016; Brown 2005; Davidson 2014; Fleischmann 2008; Gratz 2006; Gratz 2014; Harned 2014; Hassanian‐Moghaddam 2011; Husain 2014; Liberman 1981; Linehan 1991; Linehan 2006; Linehan 2015; McMain 2009; Mousavi 2015; Mousavi 2017; Naidoo 2014; O'Connor 2015; O'Connor 2020; Priebe 2012; Sahin 2018; Slee 2008; Tapolaa 2010; Torhorst 1987; Turner 2000; Walton 2020; Wei 2013; Weinberg 2006), followed by self‐reported information supplemented by clinical, hospital, and/or medical records (Bateman 2009; Cedereke 2002; Evans 1999b; Grimholt 2015; Guthrie 2001; Gysin‐Maillart 2016; Hvid 2011; Lin 2020; Mouaffak 2015; Tyrer 2003; Vaiva 2006; Vaiva 2018), or self‐reported information supplemented with information from a collateral informant (e.g. a general practitioner [GP]) (McLeavey 1994; McMain 2017; Van Heeringen 1995; Waterhouse 1990; Welu 1977).
For the remaining 23 trials, information on repetition of SH was obtained from clinical, hospital, and/or medical records (Andreoli 2015; Beautrais 2010; Bennewith 2002; Clarke 2002; Crawford 2010; Evans 1999a; Gibbons 1978; Hatcher 2011; Hatcher 2015; Hatcher 2016; Kapur 2013; Morgan 1993; Morthorst 2012; O'Connor 2017; Owens 2020; Salkovskis 1990; Stewart 2009; Van der Sande 1997), from monitoring systems (Carter 2005; Hawton 1981; Hawton 1987), or from clinical, hospital, and/or medical records supplemented with information from collateral informants (Allard 1992).
In one trial (McAuliffe 2014), mixed methods were used to determine repetition of SH. Specifically, self‐reported information was used at the post‐intervention and six‐month follow‐up assessments, whereas data on hospital representations were used at the 12‐month follow‐up assessment in this trial.
For five trials, information on the source of the data for repetition of SH was not clearly reported (Dubois 1999; Kawanishi 2014; Marasinghe 2012; Torhorst 1988; Wang 2016).
Secondary outcomes
Treatment adherence
Of the 28 trials that reported information on treatment adherence, this was assessed using a variety of methods. This included the proportion of participants who completed the full course of the intervention treatment (Andreoli 2015; Bateman 2009; Bennewith 2002; Brown 2005; Crawford 2010; Evans 1999b; Grimholt 2015; Harned 2014; Husain 2014; Kawanishi 2014; Lin 2020; Linehan 2015; McAuliffe 2014; McLeavey 1994; McMain 2009; McMain 2017; O'Connor 2017; Owens 2020; Priebe 2012; Slee 2008; Sreedaran 2020; Torhorst 1987; Torhorst 1988; Turner 2000; Walton 2020; Van Heeringen 1995), although for some trials, corresponding information on the proportion completing treatment in the comparator arm was not clearly reported. A smaller number of trials reported information on the total number of treatment sessions attended (Allard 1992; Brown 2005; McLeavey 1994; Torhorst 1988; Van der Sande 1997).
Depression
Depression was assessed using the BDI in just under half (47.6%) of the 42 trials that reported information on this outcome (Armitage 2016; Bateman 2009; Brown 2005; Fleischmann 2008; Gibbons 1978; Grimholt 2015; Guthrie 2001; Gysin‐Maillart 2016; Hawton 1987; Husain 2014; Liberman 1981; Linehan 1991; Marasinghe 2012; McAuliffe 2014; McMain 2009; McMain 2017; Salkovskis 1990; Slee 2008; Tapolaa 2010; Turner 2000), followed by the depression subscale of the HADS in six trials (Davidson 2014; Evans 1999b; Hatcher 2011; Hatcher 2015; Hatcher 2016; Tyrer 2003), the Hamilton Rating Scale for Depression (HRSD; Hamilton 1960) in eight trials (Andreoli 2015; Brown 2005; Harned 2014; Lin 2020; Linehan 2006; Linehan 2015; Turner 2000; Wei 2013), the Depression Anxiety Stress Scales (DASS; Lovibond 1995) in two trials (Gratz 2006; Gratz 2014), the Lorr and McNair Mood Scale (LMMS; McNair 1964) in one trial (Hawton 1981), the Montgomery‐Åsberg Depression Rating Scale (MADRS; Montgomery 1979) in one trial (Van der Sande 1997), and the Zung Self‐Rating Depression Scale (ZSRDS; Zung 1965) in one trial (Liberman 1981). In two trials, it was not clear what scale was used to assess depression (Torhorst 1987; Torhorst 1988).
Hopelessness
Hopelessness was assessed using the BHS in the majority (80.9%) of the 21 trials that reported information on this outcome (Brown 2005; Grimholt 2015; Hatcher 2011; Hatcher 2015; Hatcher 2016; Husain 2014; Kawanishi 2014; Lin 2020; Linehan 1991; McAuliffe 2014; McLeavey 1994; Patsiokas 1985; Salkovskis 1990; Stewart 2009; Van der Sande 1997; Waterhouse 1990; Wang 2016), by the future optimism subscale score on the Reasons for Living Inventory (RFL; Osman 1992), which was reverse coded in the present review to indicate a perceived lack of optimism about the future in three trials (Linehan 2015; O'Connor 2015; O'Connor 2020), and by an idiosyncratic scale in one trial (Mousavi 2017).
General functioning
General functioning was assessed using the Global Assessment Scale (GAS; Endicott 1976) in the majority (66.7%) of the six trials that reported information on this outcome (Allard 1992; Andreoli 2015; Bateman 2009; Tyrer 2003), followed by the GAF in the remaining two trials (Cedereke 2002; Sahin 2018).
Social functioning
Social functioning was assessed using the Social Adjustment Scale (SAS; Weissman 1999) in just over half (55.6%) of the nine trials that reported information on this outcome (Bateman 2009; Hawton 1981; Hawton 1987; McMain 2017; Torhorst 1988), the Social Functioning Questionnaire (SFQ; Tyrer 1990) in two trials (Evans 1999b; Tyrer 2003), and the social performance subscale of the Social Behaviour Assessment Schedule (SBAS; Platt 1983) in one trial (Waterhouse 1990). In one further trial, an idiosyncratic scale was used (Gibbons 1978).
Suicidal ideation
Eighteen trials assessed suicidal ideation using the BSSI (Brown 2005; Cedereke 2002; Davidson 2014; Grimholt 2015; Guthrie 2001; Gysin‐Maillart 2016; Hatcher 2011; Hawton 1981; Husain 2014; Lin 2020; Marasinghe 2012; McAuliffe 2014; O'Connor 2015; O'Connor 2020; Patsiokas 1985; Salkovskis 1990; Stewart 2009; Turner 2000), three trials used the Suicide Behaviors Questionnaire (SBQ; Linehan 1981) (Linehan 2006; Linehan 2015; Weinberg 2006), one trial used the Scale for Suicidal Ideators (SSI; Schotte 1982) (Linehan 1991), one the suicidal ideation subscale of the Psychiatric Status Schedule (PSS; Spitzer 1970) (Waterhouse 1990), and one the suicidal ideation subscale of the Suicide Risk Inventory (SRI; Hsu 1997) (Wang 2016). Six trials measured suicidal ideation dichotomously as the proportion with self‐reported suicidal ideation (Andreoli 2015; Hassanian‐Moghaddam 2011; Liberman 1981; Mousavi 2015; Mousavi 2017; Wei 2013).
Suicide
In over half of the 60 trials (60.0%) that recorded information on suicide, the method used to ascertain this outcome was not clearly reported (Andreoli 2015; Bateman 2009; Beautrais 2010; Brown 2005; Clarke 2002; Davidson 2014; Dubois 1999; Gratz 2006; Gratz 2014; Grimholt 2015; Guthrie 2001; Gysin‐Maillart 2016; Harned 2014; Husain 2014; Kapur 2013; Lin 2020; Linehan 1991; Linehan 2006; Linehan 2015; Marasinghe 2012; McLeavey 1994; McMain 2009; McMain 2017; Mousavi 2015; Naidoo 2014; Owens 2020; Priebe 2012; Sahin 2018; Salkovskis 1990; Slee 2008; Sreedaran 2020; Stewart 2009; Tapolaa 2010; Torhorst 1987; Walton 2020; Weinberg 2006). In the remaining trials, a variety of sources were used to ascertain suicide, including: Coroner's records (Allard 1992; Amadéo 2015; Hatcher 2011; Hatcher 2015; Hatcher 2016; Hvid 2011; Tyrer 2003), mortality registers (Carter 2005; Cedereke 2002; Hassanian‐Moghaddam 2011; Kawanishi 2014; Morthorst 2012; Mouaffak 2015; O'Connor 2017; Van Heeringen 1995), mortality statistics supplemented by Coroner’s records (Evans 1999a), hospital records (McAuliffe 2014), medical records (Van der Sande 1997), hospital and/or medical records supplemented by Coroner’s records (Vaiva 2006; Vaiva 2018), or from collateral informant report (Crawford 2010; Fleischmann 2008; Hawton 1987; Wei 2013).
Excluded studies
A total of 157 studies were excluded from this update. The most common reason for exclusion was that not all trial participants had engaged in SH within six months of trial entry (90 studies). Reasons for exclusion for the remaining studies are shown in Figure 1.
Details on the reasons for exclusion for those trials related to psychosocial interventions for SH in adults identified by this update are reported in the Characteristics of excluded studies section.
Ongoing studies
Of the 20 ongoing studies identified in the previous version of this review (Hawton 2016), seven were included in this update (Brown 2005; Lin 2020; O'Connor 2015; O'Connor 2017; O'Connor 2020; Owens 2020; Vaiva 2018). Ten were excluded: for three, trial results were unavailable, for two, trial registration subsequently lapsed, two were suspended due to feasibility difficulties, two subsequently included participants who had not engaged in SH, and one was subsequently published as a qualitative report.
A total of 25 ongoing studies were identified in this update (see Characteristics of ongoing studies section for further information).
Studies awaiting classification
Two potentially eligible trials could not be included in this review (NCT00533117; NCT00834834). Whilst results from these trials have been posted online, we were unable to confirm whether they met our inclusion criteria with the trial authors. Further information is provided in the Characteristics of studies awaiting classification section.
Risk of bias in included studies
Risk of bias was evaluated for the primary outcome repetition of SH at post‐intervention. The results of the 'Risk of bias' assessments can be seen in Figure 2 and Figure 3. Full 'Risk of bias' assessments, including the evidence we used to justify our ratings, are available here: doi.org/10.6084/m9.figshare.14159244.
Summary of 'Risk of bias' assessments
Results of 'Risk of bias' assessments for each study
Bias arising from randomisation process
All trials used random allocation to assign participants to the intervention and comparator arms. Most trials (98.7%) randomised at the individual level. In one trial, cluster randomisation was used (Bennewith 2002). Over half (55.3%) were rated as having a low risk of bias for this domain. Thirty (39.5%) trials were rated as having some concerns for this domain. For most of these, insufficient detail on either generation of the randomisation sequence and/or allocation concealment was reported (Allard 1992; Amadéo 2015; Brown 2005; Cedereke 2002; Dubois 1999; Gratz 2006; Gratz 2014; Kapur 2013; Liberman 1981; Linehan 2006; Marasinghe 2012; McLeavey 1994; Morgan 1993; Mousavi 2015; O'Connor 2020; Patsiokas 1985; Priebe 2012; Sahin 2018; Stewart 2009; Tapolaa 2010; Turner 2000; Van Heeringen 1995; Wang 2016; Wei 2013). For three further trials, baseline differences between the intervention and comparator arms suggested there may have been a problem with the randomisation process (Carter 2005; Guthrie 2001; Hatcher 2011), whilst for two older trials, information on characteristics of the intervention and comparator arms at baseline was not reported (Waterhouse 1990; Welu 1977). Four (5.3%) trials were rated as having high risk of bias for this domain due to insufficient detail on generation of the randomisation sequence and/or allocation concealment coupled with significant differences between the intervention and control arms at baseline on one or more factors, suggesting there may have been a problem with the randomisation process in these trials (Davidson 2014; O'Connor 2015; Torhorst 1987; Torhorst 1988).
Bias due to deviations from intended interventions
Whilst participants and clinical personnel were, typically, not blind to allocation owing to likely differences in treatment intensity between the intervention and comparator arms, most trials (69.9%) were nonetheless rated as being at low risk of bias for this domain as no deviations from the intended intervention were apparent and analyses were conducted on an intention‐to‐treat (ITT) basis, although the statistical method(s) used to undertake these analyses were not always clearly reported. Twenty‐one trials were rated as having some concerns for this domain. For the majority of these, there was insufficient information on the analyses method(s) used (Dubois 1999; Evans 1999b; Fleischmann 2008; Gibbons 1978; Gratz 2006; Hawton 1981; Liberman 1981; Linehan 1991; Linehan 2006; McAuliffe 2014; McLeavey 1994; Mousavi 2015; Mousavi 2017; O'Connor 2015; Patsiokas 1985; Sreedaran 2020; Torhorst 1987; Torhorst 1988; Waterhouse 1990; Welu 1977). For one, some minor departures from the intended intervention occurred as a result of the experimental context. Specifically, some participants moved GP practices. However, as the proportion of participants who moved practices was small and relatively balanced between the intervention and comparator arms in this trial, this was unlikely to have had a substantial impact on the results observed (Bennewith 2002). Two trials were rated as being at high risk of bias for this domain (Carter 2005; Naidoo 2014). In the first, some participants randomised to the control group mistakenly received the intervention treatment and whilst ITT analyses were undertaken, the impact was not balanced between the intervention and comparator arms in this trial (Carter 2005). In the second, the trial authors claimed the intervention was effective in preventing repeat SH even though there were major discrepancies in the data presented in the original trial report that we, as review authors, were unable to clarify through correspondence (Naidoo 2014).
Bias due to missing outcome data
The majority of trials (82.9%) were at low risk of bias for this domain as fewer than 5% of the data were missing at the post‐intervention assessment or, the proportion of missing data was balanced between the intervention and comparator arms at post‐intervention. However, there were some concerns with respect to this domain for four (5.3%) trials. For two of these, no data on repetition of SH were reported (Patsiokas 1985; Sahin 2018). For one, there was evidence of a larger proportion of missing data for the comparator arm as compared to the intervention arm (McAuliffe 2014), whilst, for one, the proportion of missing data was greater for the intervention arm (O'Connor 2015). None of these trials undertook sensitivity analyses to investigate the impact that missing data may have had on the estimate of treatment effectiveness.
Nine (11.8%) trials were rated as being at high risk of bias for this domain. For five of these, there was evidence of a difference in the proportion of missing data between the intervention and comparator arms, no information on likely causes of missingness was reported, and sensitivity analyses were not undertaken to investigate the impact of missing data (Davidson 2014; Evans 1999b; Grimholt 2015; McLeavey 1994; Wei 2013). For one trial, those with a repeat episode of SH during the follow‐up period were excluded from subsequent analyses (Sreedaran 2020); in another, data for over one‐third (38.1%) of the randomised sample were not included in follow‐up assessments (Linehan 1991). For another trial, there were major discrepancies in the data presented in the original trial report that we, as review authors, were unable to clarify through correspondence (Naidoo 2014). Finally, for one multisite and multi‐country trial there was evidence of substantial regional differences in missingness (Fleischmann 2008).
Bias in measurement of the outcome
There were some concerns regarding bias in the measurement of the outcome for just over one‐fifth (21.1%) of the trials included in this review. Typically, this was because repetition of SH was ascertained from self‐reported information alone and participants were either not blind to treatment allocation and/or participant blinding was unlikely to have been possible to achieve given the differences in therapeutic intensity between the intervention and comparator arms (Gratz 2014; Hassanian‐Moghaddam 2011; McAuliffe 2014; Mousavi 2017; O'Connor 2020; Priebe 2012; Slee 2008; Tapolaa 2010; Torhorst 1987; Torhorst 1988; Van Heeringen 1995; Walton 2020; Wei 2013; Weinberg 2006). For one trial, repetition of SH was ascertained from clinical personnel who were not blind to treatment allocation (Waterhouse 1990), whilst in another, repetition of SH was ascertained from unblinded clinical personnel for the majority of those allocated to the intervention arm (Allard 1992).
Ten (13.2%) trials were rated as being at high risk of bias for this domain. For most of these, repetition of SH was ascertained from self‐reported information alone, participant blinding was unlikely to have been possible to achieve given the differences in therapeutic intensity between the intervention and comparator arms, and assessment of the outcome could have been influenced by knowledge of the intervention received (Lin 2020; Linehan 1991; O'Connor 2015; Sahin 2018; Turner 2000; Welu 1977). For one trial, repetition of SH was obtained from self‐report. Given that participants assigned to the intervention and comparator arms in this trial received different treatment modalities (i.e. face‐to‐face versus telephone) and that previous work has shown that participants may be less willing to report sensitive information face‐to‐face than by telephone (Pridemore 2005), ascertainment of the outcome could have differed between groups in this trial (Mousavi 2015). For one further trial, there were major discrepancies for this outcome in the original trial report that we were unable to clarify through correspondence with the trial authors (Naidoo 2014). Finally, for two trials, no information on repetition of SH was reported (Patsiokas 1985; Sreedaran 2020).
Bias in selection of the reported result
Fourteen trials (18.4%) were rated as being at low risk of bias for this domain. Of these, only one clearly reported that data had been analysed in accordance with a prespecified analysis plan that had been finalised before unblinded outcome data and been made available for analysis (Tyrer 2003), whilst for 13 further trials there had been no major departures from the analysis plan as outlined in either a published trial protocol (Linehan 2015) or clinical trials register (Bateman 2009; Gysin‐Maillart 2016; Harned 2014; Hatcher 2011; Hatcher 2015; Hatcher 2016; Husain 2014; McMain 2017; O'Connor 2017; O'Connor 2020; Owens 2020; Vaiva 2018), although it should be noted that one of these trials was retrospectively registered (Bateman 2009).
Most trials (80.3%) were rated as having some concerns for this domain. In the majority of cases, this was because trials were published prior to the International Committee of Medical Journal Editors' (ICMJE) requirement in 2015 that all trials be preregistered in a publicly available clinical trials registry. It was, therefore, difficult to determine whether data had been analysed according to a prespecified plan, although there were no apparent departures from the analyses outlined in the methods section of these trials (Armitage 2016; Amadéo 2015; Andreoli 2015; Mouaffak 2015; Mousavi 2015; Mousavi 2017; Sahin 2018; Walton 2020; Wang 2016). For five trials that were preregistered, this domain was also rated as having some concerns, as the information provided within the clinical trials was not sufficiently detailed to determine whether there had been departures from the proposed analysis plan (Grimholt 2015; Kawanishi 2014; Lin 2020; McMain 2009; O'Connor 2015). For two further trials, there were some concerns for this domain as data on repetition of SH for one or more eligible time point(s) were not reported (Naidoo 2014; Salkovskis 1990); however, in both of these trials, it was unlikely results were selectively reported for favourability. Finally, for two trials, no information on repetition of SH was reported (Patsiokas 1985; Sreedaran 2020).
One trial was rated as being at high risk of bias for this domain as, although repetition of SH was a prespecified outcome, data at post‐intervention and at the 12‐month follow‐up assessment were not reported due to a high proportion of missing outcome data for these time points, although data at longer time points were reported (Sahin 2018). Additionally, we were unable to obtain unpublished data for these time points despite correspondence.
Overall bias
As a consequence, most trials (84.2%) were rated as either having some concerns or being at high risk of bias.
Effects of interventions
See: Summary of findings 1 Comparison 1.1: Individual‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 2 Comparison 1.2: Group‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 3 Comparison 2.1: DBT compared to TAU or another comparator for self‐harm in adults; Summary of findings 4 Comparison 2.2: DBT group‐based skills training compared to TAU or another comparator for self‐harm in adults; Summary of findings 5 Comparison 2.3: DBT individual therapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 6 Comparison 2.4: DBT prolonged exposure protocol compared to TAU or another comparator for self‐harm in adults; Summary of findings 7 Comparison 3: MBT compared to TAU or another comparator for self‐harm in adults; Summary of findings 8 Comparison 4: Emotion‐regulation psychotherapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 9 Comparison 5: Psychodynamic psychotherapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 10 Comparison 6: Case management compared to TAU or another comparator for self‐harm in adults; Summary of findings 11 Comparison 7: Structured GP follow‐up compared to TAU or another comparator for self‐harm in adults; Summary of findings 12 Comparison 9.1: Emergency cards compared to TAU or another comparator for self‐harm in adults; Summary of findings 13 Comparison 9.2: Coping cards compared to TAU or another comparator for self‐harm in adults; Summary of findings 14 Comparison 9.4: Postcards compared to TAU or another comparator for self‐harm in adults; Summary of findings 15 Comparison 9.5: Telephone contact compared to TAU or another comparator for self‐harm in adults; Summary of findings 16 Comparison 9.6: Telephone contact, emergency cards, and letters compared to TAU or another comparator for self‐harm in adults; Summary of findings 17 Comparison 9.7: Telephone‐based psychotherapy compared to TAU or another comparator for self‐harm in adults; Summary of findings 18 Comparison 10: Provision of information and support compared to TAU or another comparator for self‐harm in adults; Summary of findings 19 Comparison 11: Other multimodal interventions compared to TAU or another comparator for self‐harm in adults; Summary of findings 20 Comparison 12.5: General hospital management compared to TAU or another comparator for self‐harm in adults; Summary of findings 21 Comparison 12.8: Long‐term therapy compared to TAU or another comparator for self‐harm in adults
Comparison 1: Cognitive behavioural therapy (CBT)‐based psychotherapy versus TAU or another comparator
Twenty‐one trials assessed the effectiveness of CBT‐based psychotherapy, in which participants in the intervention group were offered some form of specific psychological therapy such as CBT or problem‐solving therapy (PST), for adults (weighted mean age: 31.7 ± 12.1 years; 56.7% female) presenting to services following an episode of SH (Brown 2005, N = 120; Davidson 2014, N = 20; Dubois 1999, N = 102; Evans 1999b, N = 32; Gibbons 1978, N = 400; Guthrie 2001, N = 119; Hatcher 2011, N = 1094; Hawton 1987, N = 80; Husain 2014, N = 221; Lin 2020, N = 147; McAuliffe 2014, N = 433; Mousavi 2017, N = 60; Owens 2020, N = 62; Patsiokas 1985, N = 15; Salkovskis 1990, N = 20; Slee 2008, N = 82; Stewart 2009, N = 32; Tapolaa 2010, N = 16; Tyrer 2003, N = 480; Wei 2013, N = 162; Weinberg 2006, N = 30).
In most trials, therapy was typically very brief (i.e. ≤ 10 sessions) over a relatively brief period (i.e. median: three months; IQR: one to 3.7 months; range: one to 12 months), and was delivered on an individual basis in all but one (McAuliffe 2014).
Comparison 1.1: Individual cognitive behavioural therapy (CBT)‐based psychotherapy versus TAU or another comparator
For one trial (Stewart 2009), there were separate treatment arms for CBT and PST. We therefore combined data from these two conditions, given their similarity. For one further trial (Wei 2013), there were two intervention arms: CBT‐based psychotherapy and telephone contact. We therefore included only data for the CBT‐based psychotherapy arm.
Primary outcome
1.1.1 Repetition of SH
Four trials reported data on the proportion with a repeat episode of SH by the post‐intervention assessment (Mousavi 2017; Stewart 2009; Wei 2013; Weinberg 2006). While there was imprecision in the effect estimate, CBT‐based psychotherapy may reduce repetition of SH by post‐intervention (OR 0.35, 95% CI 0.12 to 1.02; participants = 238; studies = 4; I2 = 0%; Analysis 1.1). The overall risk of bias was high for one trial (Wei 2013) and there were some concerns for the remaining trials. According to GRADE criteria, we judged the evidence to be of low certainty.
By the six‐month follow‐up assessment, on the basis of data from 12 trials, there was evidence of an effect for CBT‐based psychotherapy on repetition of SH (OR 0.52, 95% CI 0.38 to 0.70; N = 1260; k = 12; I² = 2%; Analysis 1.2). For one trial (Owens 2020), data on self‐reported SH were also reported. Using these data did not materially change this result.
There was also evidence of an effect for CBT‐based psychotherapy by the 12‐month follow‐up assessment in nine trials (OR 0.81, 95% CI 0.66 to 0.99; N = 2458; k = 9; I² = 0%; Analysis 1.3). One trial also reported data for the consenting sample (i.e. including only those participants who, following treatment allocation, subsequently consented to participation), rather than all those randomised, as well as data on the proportion of participants self‐reporting an episode of SH rather than those admitted to hospital following an episode of SH, as well as data for the consenting sample (Hatcher 2011). Using either source of data for this trial, however, did not materially affect the overall result. Two trials reported data on repetition of SH by 24 months. An effect for this outcome was found (OR 0.31, 95% CI 0.14 to 0.69; N = 105; k = 2; I² = 0%; Analysis 1.4).
With respect to frequency of SH, data from four trials indicated no effect for CBT‐based psychotherapy on frequency of SH repetition by the post‐intervention assessment (MD ‐0.53, 95% CI ‐1.67 to 0.61; participants = 149; studies = 4; I2 = 62% ; Analysis 1.5). However, there was evidence of an effect for CBT‐based psychotherapy by the six‐month assessment (MD ‐0.71, 95% CI ‐1.32 to ‐0.11; N = 118; k = 4; I² = 0%; Analysis 1.6). One trial reported information on median, rather than mean, number of episodes of SH at six months. However, the authors found that although “[t]he rate of self‐harm episodes was lower in the [experimental] group...[it was not] significantly so” (Evans 1999b, pg.22).
There was evidence of an effect for CBT‐based psychotherapy by the 12‐month assessment in a single trial (mean 1.18, SD 4.22, n = 40 versus mean 4.58; SD 8.37; n = 33; MD ‐3.40, 95% CI ‐6.54 to ‐0.26; N = 73; k = 1; I² = not applicable; Slee 2008). A further trial reported information on median number of episodes of SH at 12 months’ follow‐up, finding that “[t]he median number of self‐harm episodes was two in both [the experimental and TAU] groups” (Tyrer 2003, pg.972).
Three trials reported numeric data on time to SH repetition. There was no evidence of an effect for CBT‐based psychotherapy on time to SH repetition in these three trials (Analysis 1.7). Four trials reported narrative information on time to SH repetition. In one, authors reported "[t]he time till next parasuicide act...showed trends in favour of MACT compared with TAU although [it was not] statistically significant" (Evans 1999b, pg.23), whilst in a second, trial authors reported "[i]n the experimental (treated) group, the mean time to repeat was 9.3 months, while the mean time to repeated attempt for the subjects in the control group was three months" (Salkovskis 1990, pg.874). However, trial authors in a third reported "Kaplan‐Maier survival analysis indicated that MACT effected no significant difference in time to repetition of DSH (log‐rank chi square = .88, df = 15, n.s.)" (Weinberg 2006, pg.487). Finally, in a more recent trial, "time to repetition...showed a more complex pattern, with speedier initial repetition among participants randomised to problem‐solving therapy" (Owens 2020, pg.9).
Secondary outcomes
1.1.2 Treatment adherence
Data on treatment adherence was reported for both the intervention and comparator arms in one trial in which an effect for CBT‐based psychotherapy was found for the proportion of participants who completed all 12 sessions of therapy during the acute phase in addition to the three booster sessions (40/40 versus 33/42; OR 22.97, 95% CI 1.29 to 409.37; N = 82; k = 1; I² = not applicable; Slee 2008).
Five trials reported data on treatment adherence for the intervention arm only (Brown 2005; Evans 1999b; Husain 2014; Lin 2020; Owens 2020). Brown 2005 found that “participants in the cognitive therapy (CT) group participated in a mean (SD) of 8.92 (5.97) CT sessions (range 0‐24). Thirty participants (50%) received ten or more CT sessions” (pg.568). In Evans 1999b, five participants in the intervention group did not have specific sessions of manual‐assisted cognitive‐behaviour therapy (MACT) and received almost all input from the booklet component of CBT alone. Overall, 17 of the 18 participants in the experimental group received the booklets. Husain 2014 found that “more than half of the (intervention) group attended all six sessions (n = 56)” (pg.466). Lin 2020 found that "only around one fourth participants [sic] completed the scheduled intervention" (pg.697). Owens 2020 found that almost half (43.3%) attended all treatment sessions.
1.1.3 Depression
There was evidence of an effect for CBT‐based psychotherapy on depression scores at the post‐intervention assessment (SMD ‐0.28, 95% CI ‐0.54 to ‐0.02; N = 953; k = 4; I² = 71%; Analysis 1.8).
There was no evidence of an effect for CBT‐based psychotherapy on depression scores at the six‐month assessment in eight trials (Analysis 1.9); however, there was evidence of an effect by the 12‐month assessment in seven trials (SMD ‐0.36, 95% CI ‐0.64 to ‐0.07; N = 1130; k = 7; I² = 76%; Analysis 1.10). One trial reported data on median (rather than mean) HDRS scores at six‐ and 12‐months follow‐up (Lin 2020); however, no effect was found for CBT‐based psychotherapy in this trial for this outcome.
Only two trials reported depression scores by the 24‐month follow‐up assessment; however, there was no evidence of an effect for CBT‐based psychotherapy by this time point (Analysis 1.11).
1.1.4 Hopelessness
There was evidence of an effect for CBT‐based psychotherapy on hopelessness scores at the post‐intervention assessment in five trials (MD ‐2.99, 95% CI ‐3.91 to ‐2.07; N = 803; k = 5; I² = 0%; Analysis 1.12). One further trial also reported information on hopelessness scores at post‐intervention (Mousavi 2017); however, as an idiosyncratic scale was used in this trial, data could not be included in this review.
There was also evidence of an effect on hopelessness scores by the six‐month follow‐up assessment in two trials (MD ‐3.14, 95% CI ‐4.78 to ‐1.49; N = 315; k = 2; I² = 0%; Analysis 1.13).
Three trials reported data on hopelessness scores at 12 months, again showing a evidence of an effect for CBT‐based psychological therapy (MD ‐1.89, 95% CI ‐2.97 to ‐0.81; N = 539; k = 3; I² = 16%; Analysis 1.14). One further trial reported data on median (rather than mean) BHS scores at six‐ and 12‐months follow‐up (Lin 2020).
One trial reported data on hopelessness scores at 24 months (Brown 2005). However, there was no evidence of an effect for CBT‐based psychotherapy on hopelessness scores by this time point in this trial (mean 6.07, SD 5.28, n = 45 versus mean 7.24, SD 6.35, n = 40; MD ‐1.17, SD ‐3.67 to 1.33; N = 85; k = 1; I² = not applicable).
1.1.5 General functioning
One trial reported data on general functioning scores by the six‐month assessment. However, whilst there was no evidence of a difference in mean scores on this measure between the CBT‐based psychotherapy and TAU arms in this trial (i.e. 59.4 versus 59.6), insufficient information was reported to enable imputation of missing SDs (Tyrer 2003).
1.1.6 Social functioning
One trial reported data on social functioning scores at post‐intervention and, whilst there was no evidence of a difference in mean scores on this measure between the CBT‐based psychotherapy and TAU arms in this trial (i.e. 2.3 versus 2.3), insufficient information was reported to enable imputation of missing SDs (Hawton 1987).
Three trials reported data on social functioning by the six‐month follow‐up assessment (Evans 1999b; Hawton 1987; Tyrer 2003), however, for two of these trials insufficient information was reported to enable imputation of missing SDs (Hawton 1987; Tyrer 2003). For the remaining trial, there was evidence of an effect for CBT‐based psychological therapy on social functioning scores by this time point (mean 9.8, SD 4.9, n = 18 versus mean 13.1, SD 4.0, n = 12; MD ‐3.30, 95% CI ‐6.50 to ‐0.10; N = 30; k = 1; I² = not applicable; Evans 1999b).
One of these trials also reported data on social functioning scores by the 12‐month follow‐up period. However, whilst those allocated to CBT‐based psychotherapy had lower mean scores (i.e. 1.7 versus 2.1), once again insufficient information was reported to enable imputation of missing SDs (Hawton 1987).
1.1.7 Suicidal ideation
There was evidence of an effect for CBT‐based psychotherapy in five trials at the post‐intervention assessment (SMD ‐0.48, 95% CI ‐0.68 to ‐0.28; N = 718; k = 5; I² = 20%; Analysis 1.15), and by the six‐month follow‐up assessment in four trials (SMD ‐0.38, 95% CI ‐0.60 to ‐0.17; N = 353; k = 4; I² = 0%; Analysis 1.16).
One trial reported information on suicidal ideation scores by 12 months; however, CBT‐based psychotherapy was not associated with an effect by this time point in this trial (mean 3.70, SD 6.70, n = 187 versus mean 4.80, SD 7.40, n = 231; MD ‐1.10, 95% CI ‐2.45 to 0.25; N = 418; k = 1; I² = not applicable; Hatcher 2011). One trial reported data on median (rather than mean) BSSI scores at six‐ and 12‐months follow‐up (Lin 2020); however, no effect was found for CBT‐based psychotherapy in this trial for this outcome.
Two trials reported information on the proportion of participants with suicidal ideation (Mousavi 2017; Wei 2013). At post‐intervention, data from both of these trials indicated no evidence of an effect for CBT‐based psychotherapy. There was also no evidence of an effect by the six‐month or 12‐month follow‐up assessments in one of these trials (Analysis 1.17). However, it is notable that for one of these trials (Wei 2013), results were reported on the basis of those randomised despite the fact that a self‐reported measure was used, and high levels of dropout were observed by the 12‐month follow‐up assessment. As we were unable to confirm these numbers with trial authors, results much be interpreted with caution.
One trial reported data on BSSI scores, however, insufficient information was available to enable imputation of missing SDs (Salkovskis 1990). One further trial reported data on the proportion scoring above zero on the BSSI at the post‐intervention, 6, 12, and 24‐month follow‐up assessments (Brown 2005), however, as this was not based on clinically established cut‐points, these data were unable to be included in this review.
1.1.8 Suicide
Sixteen trials reported data on suicide deaths during follow‐up. There was no evidence of an effect for CBT‐based psychotherapy on suicides by final follow‐up (i.e. up to 24 months) (Analysis 1.18). In one trial, there was one death in the intervention arm that medical staff considered to be a suicide, although the coroner did not record a suicide verdict in this case (Tyrer 2003). Including this death as a suicide did not materially change the overall result.
Subgroup analyses
Two further trials reported data on repetition of SH by repeater status (Hatcher 2011; Lin 2020). There was no evidence of an effect for CBT‐based psychotherapy on repetition of SH by the six‐month follow‐up assessment for either those with a history of repeated SH (7/39 versus 16/49; OR 0.45, 95% CI 0.16 to 1.24; N = 88; k = 1; I² = not applicable; Lin 2020), or those without this history (3/33 versus 7/26; OR 0.27, 95% CI 0.06 to 1.18; N = 59; k = 1; I² = not applicable; Lin 2020) in one trial. However, by the 12‐month follow‐up assessment, data from these two trials indicated that whilst CBT‐based psychotherapy was not associated with an effect on repetition of SH for those without a history of SH prior to trial entry (Analysis 1.19), there was evidence of an effect on this outcome for those with a history of multiple episodes of SH (OR 0.56, 95% CI 0.36 to 0.88; N = 508; k = 1; I² = 0%; Analysis 1.19). However, the test for subgroup differences was not statistically significant for this outcome at this time point (Chi² = 1.97, df = 1, p = 0.16).
With regards to time to SH repetition, data from one of these trials suggested that, whilst there was no evidence of an effect in those without a history of multiple episodes of SH prior to trial entry (HR 1.55, 95% CI 0.98 to 2.48; N = 135; k = 1; I² = not applicable), CBT‐based psychotherapy was associated with an effect on time to SH repetition in those with a history of multiple episodes of SH (HR 0.58, 95% CI 0.36 to 0.94; N = 194; k = 1; I² = not applicable) in this trial (Hatcher 2011).
Sensitivity analyses
One trial used Zelen's design (Hatcher 2011). Omitting this trial did not materially affect results for CBT‐based psychotherapy on repetition of SH by the 12‐month assessment, depression scores at the 12‐month assessment, hopelessness scores by the post‐intervention or 12‐month assessments, or suicidal ideation scores at post‐intervention. However, omitting this trial did cause a reduction in the SMD for CBT‐based psychotherapy on depression scores by the post‐intervention assessment and the estimate was no longer statistically significant (SMD ‐0.19, 95% CI ‐0.40 to 0.02; N = 313; k = 5; I² = 54%).
Two analyses within this comparison were associated with substantial levels of heterogeneity (Analysis 1.10, I² = 76%; Analysis 1.17.1, I² = 85%); however, analyses did not indicate any individual study was associated with excessive influence for either of these outcomes.
Comparison 1.2: Group‐based cognitive behavioural therapy (CBT)‐based psychotherapy versus TAU or another comparator
Primary outcome
1.2.1 Repetition of SH
There was no evidence of an effect for CBT‐based psychotherapy, delivered in a group‐based format, as compared to TAU by the post‐intervention assessment in a single trial (23/171 versus 27/142; OR 0.66, 95% CI 0.36 to 1.21; N = 313; k = 1; I² = not applicable). According to GRADE criteria, we judged the evidence to be of moderate certainty.
There was also no evidence of an effect for group‐based CBT‐based psychotherapy on repetition of SH by the six‐month follow‐up assessment (39/128 versus 26/106; OR 1.35, 95% CI 0.75 to 2.41; N = 234; k = 1; I² = not applicable), or by the 12‐month follow‐up assessment (54/222 versus 50/211; OR 1.03, 95% CI 0.67 to 1.61; N = 433; k = 1; I² = not applicable).
There was also no evidence of an effect for group‐based CBT‐based psychotherapy on frequency of SH repetition by the 12‐month follow‐up assessment in this trial (mean 0.43, SD 1.23, n = 222 versus mean 0.49, SD 1.53, n = 211; MD ‐0.06, 95% CI ‐0.32 to 0.20; N = 433; n = 1; I² = not applicable).
Secondary outcomes
1.2.2 Treatment adherence
Data on treatment adherence was reported for the intervention arm only. The authors found that “almost half of those assigned to [problem‐solving therapy] (103, 46.4%) attended all 6 therapy sessions” (McAuliffe 2014, pg.4).
1.2.3 Depression
There was no evidence of an effect for group‐based CBT‐based psychotherapy on depression by either the post‐intervention (mean 18.20, SD 14.80, n = 171 versus mean 20.60, SD 16.00, n = 142; MD ‐0.16, 95% CI ‐0.38 to 0.07; N = 313; k = 1; I² = not applicable) or the six‐month follow‐up assessment (mean 17.30, SD 15.90, n = 128 versus mean 19.40, SD 17.00, n = 106; MD ‐0.13, 95% CI ‐0.39 to 0.13; N = 234; k = 1; I² = not applicable).
1.2.4 Hopelessness
There was also no evidence of an effect for group‐based CBT‐based psychotherapy on hopelessness scores at the post‐intervention (mean 6.50, SD 6.10, n = 171 versus mean 7.30, SD 6.20, n = 142; MD ‐0.80, 95% CI ‐2.17 to 0.57; N = 313; k = 1; I² = not applicable) or six‐month follow‐up assessment (mean 6.80, SD 6.30, n = 128 versus mean 7.10, SD 6.10, n = 106; MD ‐0.30, 95% CI ‐1.89 to 1.29; N = 234; k = 1; I² = not applicable).
1.2.5 General functioning
No data available.
1.2.6 Social functioning
No data available.
1.2.7 Suicidal ideation
There was no evidence of an effect for group‐based CBT‐based psychotherapy on suicidal ideation by either the post‐intervention (mean 4.30, SD 8.00, n = 171 versus mean 5.80, SD 9.70, n = 142; MD ‐1.50, 95% CI ‐3.50 to 0.50; N = 313; k = 1; I² = not applicable) or the six‐month follow‐up assessment (mean 4.70, SD 8.90, n = 128 versus mean 4.90, SD 8.90, n = 106; MD ‐0.20, 95% CI ‐2.49 to 2.09; N = 234; k = 1; I² = not applicable).
1.2.8 Suicide
There was no evidence of an effect for group‐based CBT‐based psychotherapy on suicide deaths by the final follow‐up assessment (1/222 versus 2/211; OR 0.47, 95% CI 0.04 to 5.25; N = 433; k = 1; I² = not applicable).
Subgroup analyses
Randomisation was stratified by both sex and repeater status in this trial (McAuliffe 2014). However, we were unable to obtain data disaggregated by either factor from the trial authors.
Sensitivity analyses
Not applicable.
Comparison 2: Dialectical behavioural therapy (DBT) versus TAU or another comparator
Ten trials investigated the effectiveness of dialectical behaviour therapy (DBT) as compared to either TAU (Linehan 1991, N = 63; McMain 2009, N = 180; McMain 2017, N = 84; Priebe 2012, N = 80; Sahin 2018, N = 50) or other forms of alternative psychotherapy (Harned 2014, N = 26; Linehan 2006, N = 101; Linehan 2015, N = 99; Turner 2000, N = 24; Walton 2020, N = 166) in adults (weighted mean age: 27.5 ± 11.3 years; 89.7% female) diagnosed with a personality disorder, typically borderline personality disorder, and who were receiving treatment from outpatient mental health services due to recurrent SH.
For one of these trials, there were two different intervention arms: psychodynamic psychotherapy and DBT (Sahin 2018). We therefore included only data for the DBT arm. For a second of these trials, comparison was made between two different forms of DBT: DBT group‐based skills training only (DBT‐S; Linehan 2015a) and DBT individual therapy only (DBT‐I; Linehan 2015b).
Comparison 2.1: Standard DBT
Primary outcome
2.1.1 Repetition of SH
Overall, there was no evidence of an effect for standard DBT compared to either TAU or alternative psychotherapy in terms of the proportion of patients repeating SH by the post‐intervention assessment (OR 0.71, 95% CI 0.32 to 1.55; participants = 502; studies = 6; I2 = 60%; (Analysis 2.1). There was also no difference by comparator (i.e. TAU versus alternative psychotherapy). The overall risk of bias was high for two trials (Linehan 1991; Turner 2000) and there were some concerns for the remaining trials. According to GRADE criteria, we judged the evidence to be of very low certainty.
There was also no evidence of an effect for DBT on repetition of SH by the 12‐month (OR 0.65, 95% CI 0.24 to 1.72; participants = 269; studies = 3; I2 = 47%; Analysis 2.2) assessment. Once again, there was no difference by comparator (i.e. TAU versus alternative psychotherapy).
DBT was associated with an effect for frequency of repeated SH by the post‐intervention assessment when compared to either TAU or alterative psychotherapy (MD ‐5.00, 95% CI ‐8.92 to ‐1.08; N = 659; k = 7; I² = 49%; Analysis 2.3). However, this effect was no longer found by the six‐month follow‐up assessment in one trial (mean 0.32, SD 1.27, n = 42 versus mean 1.14, SD 3.94, n = 42; MD ‐0.82, 95% CI ‐2.07 to 0.43; N = 84; k = 1; I² = not applicable; McMain 2017). Following a 'naturalistic' follow‐up period, data from one further trial indicated that the effectiveness of DBT in terms of frequency of SH was maintained at 24 months in this trial; however, this outcome was only investigated for a proportion of the original participants (61.9%) who the researchers were able to contact at 24 months (Linehan 1991). Results have therefore not been reproduced in this review.
One trial reported information on time to repetition of SH, finding that those "receiving DBT were half as likely to make a suicide attempt (hazard ratio, 2.66; P = .005)" (Linehan 2006, pg.757). However, insufficient information was reported to enable calculation of the accompanying CIs for this outcome for this trial.
Secondary outcomes
2.1.2 Treatment adherence
There was no evidence of an effect for DBT compared with either TAU or alternative psychotherapy for the proportion of participants who completed treatment (Analysis 2.4). There was also no evidence of a difference by comparator. Two further trials did not report information on the proportion completing treatment in the comparator arm (McMain 2017; Priebe 2012), although the comparator in the former trial was a waiting‐list control condition. We were, therefore, unable to incorporate these results in a meta‐analysis.
2.1.3 Depression
There was evidence of an effect for DBT compared to both TAU or alternative psychotherapy on depression scores at post‐intervention (SMD ‐0.37, 95% CI ‐0.58 to ‐0.17; N = 557; k = 6; I² = 23%; Analysis 2.5). There was no evidence of a difference by comparator. One trial also measured depression scores at post‐intervention using the HRSD; however, there was no evidence of an effect for DBT according to this measure in this trial (mean 7.50, SD 5.96, n = 12 versus mean 12.58, SD 3.90, n = 12; MD ‐5.08, 95% CI ‐9.11 to ‐1.05; N = 24; k = 1; I² = not applicable; Turner 2000).
There was no evidence of an effect for DBT on depression scores by the six‐month follow‐up assessment in one trial (mean 27.94, SD 16.08, n = 42 versus mean 29.50, SD 15.71, n = 42; MD ‐1.56, 95% CI ‐8.36 to 5.24; N = 84; k = 1; I² = not applicable; McMain 2017) or by the 12‐month follow‐up assessment in one further trial (mean 12.60, SD 6.80, n = 46 versus mean 14.40, SD 9.10, n = 35; MD ‐1.80, 95% CI ‐5.40 to 1.80; N = 81; k = 1; I² = not applicable; Linehan 2006).
2.1.4 Hopelessness
We obtained data on hopelessness by correspondence for one trial (Linehan 1991); however, there was no evidence of an effect for DBT as compared to TAU by the 24‐month follow‐up assessment in this trial (mean 10.86, SD 6.04, n = 7 versus mean 10.69, SD 6.18, n = 11; MD 0.17, 95% CI ‐5.61 to 5.95; N = 18; k = 1; I² = not applicable).
2.1.5 General functioning
No data available.
2.1.6 Social functioning
One trial reported data on social functioning scores at post‐intervention and by the six‐month follow‐up assessment (McMain 2017). Whilst there was evidence of an effect for DBT on social functioning scores at post‐intervention in this trial (mean 2.50, SD 0.56, n = 42 versus mean 2.88, SD 0.59, n = 42; MD ‐0.38, 95% CI ‐0.63 to ‐0.13; N = 84; k = 1; I² = not applicable), this effect was no longer apparent by the six‐month follow‐up assessment (mean 2.60, SD 0.70, n = 42 versus mean 2.87, SD 0.65, n = 42; MD ‐0.27, 95% CI ‐0.56 to 0.02; N = 84; k = 1; I² = not applicable).
2.1.7 Suicidal ideation
There was no evidence of an effect for DBT when compared either to TAU or to alternative psychotherapy on suicidal ideation scores at the post‐intervention assessment in three trials (Analysis 2.6), or by the 12‐month assessment in one of these trials (mean 24.10, SD 19.80, n = 46 versus mean 31.92, SD 26.80, n = 35; MD ‐7.80, 95% CI ‐18.38 to 2.74; N = 81; k = 1; I² = not applicable; Linehan 2006). There was also no evidence of a difference by comparator.
2.1.8 Suicide
There was no evidence of an effect for DBT as compared to either TAU or alternative psychotherapy on suicide deaths at either post‐intervention (Analysis 2.7), the 12‐month (Analysis 2.8), or 24‐month follow‐up period (Analysis 2.9).
Subgroup analyses
Randomisation was stratified by sex in one trial (Walton 2020). Whilst data obtained by correspondence from the trial authors indicated that, compared with alternative psychotherapy, DBT was not associated with an effect on repetition of SH by the post‐intervention assessment for females (21/38 versus 33/50; OR 0.64, 95% CI 0.27 to 1.52; N = 88; k = 1; I² = not applicable), DBT may be associated with a greater proportion of males engaging in a repeat episode of SH at this time point (12/13 versus 6/13; OR 14.00, 95% CI 1.39 to 141.49; N = 26; k = 1; I² = not applicable).
With regards to frequency of SH repetition, data obtained by correspondence for this trial indicated that, whilst DBT was associated with an effect at post‐intervention for females (mean 5.08, SD 11.13, n = 38 versus mean 17.66, SD 33.10, n = 50; MD ‐12.58, 95% CI ‐22.41 to ‐2.75; N = 88; k = 1; I² = not applicable), there was no evidence of an effect for males (mean 5.77, SD 5.34, n = 13 versus mean 4.93, SD 7.24, n = 13; MD 0.84, 95% CI ‐4.05 to 5.73; N = 26; k = 1; I² = not applicable).
There was no evidence of an effect for DBT as compared to alternative psychotherapy on the proportion of males (9/19 versus 11/18; OR 0.57, 95% CI 0.15 to 2.12; N = 37; k = 1; I² = not applicable) or females (35/62 versus 37/63; OR 0.91, 95% CI 0.45 to 1.85; N = 125; k = 1; I² = not applicable) who completed the full course of treatment on this trial.
Finally, whilst DBT was associated with an effect on depression scores as compared to alternative psychotherapy at the post‐intervention assessment for males (mean 13.54, SD 12.82, n = 13 versus mean 23.70, SD 12.30, n = 12; MD ‐10.16, 95% CI ‐20.01 to ‐0.31; N = 25; k = 1; I² = not applicable), there was no evidence of an effect on depression scores for females in this trial (mean 15.34, SD 14.01, n = 37 versus mean 21.88, SD 19.00, n = 48; MD ‐6.54, 95% CI ‐13.56 to 0.48; N = 85; I² = not applicable).
Sensitivity analyses
Not applicable.
Comparison 2.2: DBT group‐based skills training
For one trial (i.e. Linehan 2015), there were two different intervention arms: DBT group‐based skills training only (DBT‐S; Linehan 2015a) and DBT individual therapy only (DBT‐I; Linehan 2015b). For a detailed description of the therapeutic content of the two different intervention arms included in this trial, see the Characteristics of included studies section.
Here we included results for DBT group‐based skills training only as compared with standard DBT in adult women (mean age: 30.4 ± 7.5 years) diagnosed with borderline personality disorder referred for outpatient treatment for repeated SH (Linehan 2015a, N = 66).
Primary outcome
2.2.1 Repetition of SH
In this trial, data on repetition of SH were reported separately for episodes of attempted suicide and NSSI. As we were unable to obtain data on combined SH from the trial authors despite correspondence, we have reproduced the results for attempted suicide and NSSI separately.
There was no evidence of an effect for DBT group‐based skills training only as compared with standard DBT on suicide reattempts (9/33 versus 12/33; OR 0.66, 95% CI 0.23 to 1.86; N = 66; k = 1; I² = not applicable) or NSSI (18/33 versus 19/33; OR 0.88, 95% CI 0.33 to 2.34; N = 66; k = 1; I² = not applicable) at post‐intervention. According to GRADE criteria, we judged the evidence to be of moderate certainty.
There was also no evidence of an effect for DBT group‐based skills training only on suicide reattempts (6/33 versus 2/33; OR 3.44, 95% CI 0.64 to 18.51; N = 66; k = 1; I² = not applicable) or NSSI (15/33 versus 15/33; OR 1.00, 95% CI 0.38 to 2.64; N = 66; k = 1; I² = not applicable) by the 12‐month assessment in this trial.
There was no evidence of an effect for DBT group‐based skills training only on frequency of suicide reattempts (mean 2.60, SD 2.90, n = 33 versus mean 3.40, SD 4.60, n = 33; MD ‐0.80, 95% CI ‐2.66 to 1.06; N = 66; k = 1; I² = not applicable) or episodes of NSSI (mean 9.90, SD 19.70, n = 33 versus mean 10.20, SD 16.30, n = 33; MD ‐0.30, 95% CI ‐9.02 to 8.42; N = 66; k = 1; I² = not applicable) at post‐intervention. Nor was there any evidence of an effect for this intervention on frequency of suicide reattempts (mean 1.50, SD 0.60, n = 33 versus mean 2.00, SD 1.40, n = 33; MD ‐0.50, 95% CI ‐1.02 to 0.02; N = 66; k = 1; I² = not applicable) or episodes of NSSI (mean 9.40, SD 20.40, n = 33 versus mean 7.90, SD 8.50, n = 33; MD 1.50, 95% CI ‐6.04 to 9.04; N = 66; k = 1; I² = not applicable) by the 12‐month follow‐up assessment in this trial.
Finally, with regards to time to SH repetition, "[s]urvival analysis...indicated no difference between groups in the time to the first suicide attempt (χ² = 1.4 [P = .50])" (Linehan 2015, pg.4).
Secondary outcomes
2.2.2 Treatment adherence
There was no evidence of an effect for DBT group‐based skills training, as compared with standard DBT, on the proportion of participants who completed treatment by the post‐intervention assessment in this trial (20/33 versus 25/33; OR 0.49, 95% CI 0.17 to 1.42; N = 66; k = 1; I² = not applicable; Linehan 2015a).
There was also no evidence of an effect for DBT group‐based skills training, as compared with standard DBT, on the number of group therapy sessions attended (mean 4.20, SD 0.11, n = 33 versus mean 4.20, SD 0.12, n = 33; MD 0.00, 95% CI ‐0.06 to 0.06; N = 66; k = 1; I² = not applicable; Linehan 2015a). Information on the number of individual therapy sessions attended could not be estimated as individuals assigned to the group‐based skills training arm in this trial did not receive any individual therapy sessions.
2.2.3 Depression
There was no evidence of a significant treatment effect for DBT group‐based skills training, as compared with standard DBT, on depression scores at either the post‐intervention (mean 10.40, SD 6.40, n = 33 versus mean 12.30, SD 8.00, n = 33; MD ‐1.90, 95% CI ‐5.40 to 1.60; N = 66; k = 1; I² = not applicable), or 12‐month assessment (mean 11.90, SD 8.80, n = 33 versus mean 15.20, 8.60, n = 33; MD ‐3.30, 95% CI ‐7.50 to 0.90; N = 66; k = 1; I² = not applicable; Linehan 2015a).
2.2.4 Hopelessness
There was also no evidence of an effect for DBT group‐based skills training, as compared with standard DBT, on hopelessness at either the post‐intervention (mean ‐155.30, SD 41.40, n = 33 versus mean ‐168.70, SD 45.3, n = 33; MD 13.40, 95% CI ‐7.54 to 34.34; N = 66; k = 1; I² = not applicable) or 12‐month follow‐up assessment (mean ‐152.20, SD 43.20, n = 33 versus mean ‐159.90, SD 40.20, n = 33; MD 7.40, 95% CI ‐12.73 to 27.53; N = 66; k = 1; I² = not applicable; Linehan 2015a) in this trial.
2.2.5 General functioning
No data available.
2.2.6 Social functioning
No data available.
2.2.7 Suicidal ideation
There was no evidence of an effect for DBT group‐based skills training, as compared with standard DBT, on suicidal ideation scores at either the post‐intervention (mean 27.50, SD 19.10, n = 33 versus mean 32.00, SD 21.60, n = 33; MD ‐4.50, 95% CI ‐14.34 to 5.34; N = 66; k = 1; I² = not applicable) or 12‐month follow‐up assessment (mean 21.20, SD 19.20, n = 33 versus mean 28.90, SD 16.60, n = 33; MD ‐7.80, 95% CI ‐16.46 to 0.86; N = 66; k = 1; I² = not applicable).
2.2.8 Suicide
There was one suicide death in the standard DBT arm by the 24‐month follow up assessment in this trial; however, there was no evidence of an effect for DBT group‐based skills training, as compared with standard DBT, on this outcome by this time point (0/33 versus 1/33; OR 0.32, 95% CI 0.01 to 8.23; N = 66; k = 1; I² = not applicable).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 2.3: DBT individual therapy
For one trial (i.e. Linehan 2015), there were two different intervention arms: DBT group‐based skills training only (DBT‐S; Linehan 2015a) and DBT individual therapy only (DBT‐I; Linehan 2015b). For a detailed description of the therapeutic content of the two different intervention arms included in this trial, see the Characteristics of included studies section.
Here we included results for DBT individual therapy only as compared with standard DBT in adult women (mean age: 30.6 ± 7.9 years) diagnosed with borderline personality disorder referred for outpatient treatment for repeated SH (Linehan 2015b, N = 66).
Primary outcome
2.3.1 Repetition of SH
As before, in this trial, data on repetition of SH were reported separately for episodes of attempted suicide and NSSI. As we were unable to obtain data on combined SH from the trial authors despite correspondence, we have reproduced the results for attempted suicide and NSSI separately.
There was no evidence of an effect for DBT individual therapy only as compared with standard DBT on suicide reattempts (15/33 versus 12/33; OR 1.46, 95% CI 0.54 to 3.91; N = 66; k = 1; I² = not applicable) or NSSI (21/33 versus 19/33; OR 1.29, 95% CI 0.48 to 3.47; N = 66; k = 1; I² = not applicable) at post‐intervention. According to the GRADE criteria, we judged the evidence to be of moderate certainty.
There was also no evidence of an effect for DBT individual therapy only on suicide reattempts (7/33 versus 2/33; OR 4.17, 95% CI 0.80 to 21.85; N = 66; k = 1; I² = not applicable) or NSSI (13/33 versus 15/33; OR 0.78, 95% CI 0.29 to 2.07; N = 66; k = 1; I² = not applicable) by the 12‐month assessment in this trial.
There was no evidence of an effect for DBT individual therapy only on frequency of suicide reattempts (mean 2.90, SD 3.00, n = 33 versus mean 3.40, SD 4.60, n = 33; MD ‐0.50, 95% CI ‐2.37 to 1.37; N = 66; k = 1; I² = not applicable) or episodes of NSSI (mean 20.60, SD 33.10, n = 33 versus mean 10.20, SD 16.30, n = 33; MD 10.40, 95% CI ‐2.19 to 22.99; N = 66; k = 1; I² = not applicable) at post‐intervention. Nor was there any evidence of an effect for this intervention on frequency of suicide reattempts (mean 3.60, SD 3.20, n = 33 versus mean 2.00, SD 1.40, n = 33; MD 1.60, 95% CI 0.41 to 2.79; N = 66; k = 1; I² = not applicable) or episodes of NSSI (mean 16.00, SD 32.60, n = 33 versus mean 7.90, SD 8.50, n = 33; MD 8.10, 95% CI ‐3.39 to 19.59; N = 66; k = 1; I² = not applicable) by the 12‐month follow‐up assessment in this trial.
Finally, with regards to time to SH repetition, "[s]urvival analysis...indicated no difference between groups in the time to the first suicide attempt (χ² = 1.4 [P = .50])" (Linehan 2015, pg.4).
Secondary outcomes
2.3.2 Treatment adherence
There was evidence of an effect for DBT individual therapy, as compared with standard DBT, on the proportion of participants who completed treatment by the post‐intervention assessment in this trial (17/33 versus 25/33; OR 0.34, 95% CI 0.12 to 0.97; N = 66; k = 1; I² = not applicable; Linehan 2015b).
There was no evidence of an effect for DBT individual therapy, as compared with standard DBT, on the number of individual therapy sessions attended (mean 4.16, SD 0.18, n = 33 versus mean 4.20, SD 0.18, n = 33; MD ‐0.04, 95% CI ‐0.13 to 0.05; N = 33; k = 1; I² = not applicable; Linehan 2015b). Information on the number of group therapy sessions attended could not be estimated as participants assigned to the individual therapy arm in this trial did not receive any group‐based therapy sessions.
2.3.3 Depression
There was evidence of an increase in depression scores at post‐intervention for those receiving DBT individual therapy (mean 18.20, SD 7.90, n = 33 versus mean 12.30, SD 8.00, n = 33; MD 5.90, 95% CI 2.06 to 9.74; N = 66; k = 1; I² = not applicable). By the 12‐month follow‐up assessment, however, there was no longer evidence of an effect on depression scores for DBT individual therapy, as compared with standard DBT in this trial (mean 13.90, SD 9.60, n = 33 versus mean 15.20, 8.60, n = 33; MD ‐1.30, 95% CI ‐5.70 to 3.10; N = 66; k = 1; I² = not applicable).
2.3.4 Hopelessness
There was no evidence of an effect for DBT individual therapy, as compared with standard DBT, on hopelessness at either the post‐intervention (mean ‐177.90, SD 53.10, n = 33 versus mean ‐168.70, SD 45.30, n = 33; MD ‐9.20, 95% CI ‐33.01 to 14.61; N = 66; k = 1; I² = not applicable), or 12‐month follow‐up assessment (mean ‐178.40, SD 53.70, n = 33 versus mean ‐159.90, SD 40.20, n = 33; MD ‐18.50, 95% CI ‐41.39 to 4.39; N = 66; k = 1; I² = not applicable) in this trial.
2.3.5 General functioning
No data available.
2.3.6 Social functioning
No data available.
2.3.7 Suicidal ideation
There was also no evidence of an effect for DBT individual therapy, as compared with standard DBT, on suicidal ideation scores at either the post‐intervention (mean 30.30, SD 27.50, n = 33 versus mean 32.00, SD 21.60, n = 33; MD ‐1.70, 95% CI ‐13.63 to 10.23; N = 66; k = 1; I² = not applicable) or 12‐month follow‐up assessment (mean 25.50, SD 20.80, n = 33 versus mean 28.90, SD 16.60, n = 33; MD ‐3.40, 95% CI ‐12.48 to 5.68; N = 66; k = 1; I² = not applicable).
2.3.8 Suicide
There was one suicide death in the standard DBT arm by the 24‐month follow‐up assessment in this trial; however, there was no evidence of an effect for DBT individual therapy only on this outcome by this time point (0/33 versus 1/33; OR 0.32, 95% CI 0.01 to 8.23; N = 66; k = 1; I² = not applicable).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 2.4: DBT prolonged‐exposure protocol
The effectiveness of two forms of DBT were compared over a three‐month period in one small trial of adult (mean age: 32.6 ± 12.0 years) women diagnosed with comorbid borderline personality disorder and post‐traumatic stress disorder and who were referred to outpatient clinical services due to recurrent SH (Harned 2014, N = 26). In this trial, participants assigned to the intervention arm received, in addition to the standard DBT protocol, additional weekly therapy sessions involving in vivo and imaginal exposure to previously traumatic experiences (i.e. DBT prolonged‐exposure protocol).
Primary outcome
2.4.1 Repetition of SH
Data obtained by correspondence indicated there was no evidence of an effect for the DBT prolonged‐exposure protocol on repetition of SH by the post‐intervention assessment (3/12 versus 2/6; OR 0.67, 95% CI 0.08 to 5.68; N = 18; k = 1; I² = not applicable). According to GRADE criteria, we judged the evidence to be of moderate certainty.
There was also no evidence of an effect for the DBT prolonged‐exposure protocol on repetition of SH by the six‐month follow‐up assessment (3/12 versus 2/6; OR 0.67, 95% CI 0.08 to 5.68; N = 18; k = 1; I² = not applicable).
Data on frequency of SH, obtained following correspondence with trial authors, also suggested no effect for the DBT prolonged‐exposure protocol by either the post‐intervention (mean 0.92, SD 1.97, n = 12 versus mean 1.17, SD 2.40, n = 6; MD ‐0.25, 95% CI ‐2.47 to 1.97; N = 18; k = 1; I² = not applicable) or six‐month follow‐up assessments (mean 0.67, SD 1.50, n = 12 versus mean 0.33, SD 0.52, n = 6; MD 0.34, 95% CI ‐0.61 to 1.29; N = 18; k = 1; I² = not applicable).
Secondary outcomes
2.4.2 Treatment adherence
There was no evidence of an effect for the DBT prolonged‐exposure protocol on the proportion of participants who completed the full one‐year course of treatment (10/17 versus 5/9; OR 1.14, 95% CI 0.22 to 5.84; N = 26; k = 1; I² = not applicable). According to the trial authors, however, "...one therapist...was not adherent to DBT and had a 100% dropout rate" (Harned 2014, p.12). Excluding the four participants treated by this therapist did not, however, materially affect this result.
2.4.3 Depression
There was no evidence of an effect for the DBT prolonged‐exposure protocol for depression scores at either the post‐intervention (mean 11.80, SD 8.00, n = 12 versus mean 15.50, SD 6.50, n = 6; MD ‐3.70, 95% CI ‐10.59 to 3.19; N = 18; k = 1; I² = not applicable) or six‐month follow‐up (mean 12.50, SD 8.20, n = 12 versus mean 16.80, SD 3.40, n = 6; MD ‐4.30, 95% CI ‐9.68 to 1.08; N = 18; k = 1; I² = not applicable) assessments.
2.4.4 Hopelessness
No data available.
2.4.5 General functioning
No data available.
2.4.6 Social functioning
No data available.
2.4.7 Suicidal ideation
No data available.
2.4.8 Suicide
There was no evidence of an effect for the DBT prolonged‐exposure protocol on death by suicide by the follow‐up assessment in this trial (0/17 versus 1/9; OR 0.16, 95% CI 0.01 to 4.41, N = 26; k = 1; I² = not applicable).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 3: Mentalisation‐based therapy (MBT) versus TAU or another comparator
A single trial investigated the effectiveness of mentalisation‐based therapy (MBT) in adults (mean age: 31.1 ± 7.7 years; 79.9% female) diagnosed with BPD referred to a specialist personality disorder treatment service following a suicide attempt or episode of life‐threatening SH (Bateman 2009, N = 134).
Primary outcome
3.1 Repetition of SH
MBT may reduce repetition of SH by the conclusion of the 18‐month treatment period based on data obtained by correspondence (18/71 versus 31/63; OR 0.35, 95% CI 0.17 to 0.73; N = 134; k = 1; I² = not applicable). According to GRADE criteria, we judged the evidence to be of high certainty.
There was also evidence of an effect for MBT on frequency of SH episodes by the post‐intervention assessment according to data obtained by correspondence (mean 0.38, SD 0.38, n = 71 versus mean 1.66, SD 2.87, n = 63; MD ‐1.28, 95% CI ‐2.01 to ‐0.55; N = 134; k = 1; I² = not applicable).
Secondary outcomes
3.2 Treatment adherence
There was no evidence of an effect for MBT on the proportion of participants who completed the full course of treatment (52/71 versus 47/63; OR 0.93, 95% CI 0.43 to 2.02; N = 134; k = 1; I² = not applicable).
3.3 Depression
MBT was associated with an effect for depression at the post‐intervention assessment (mean 14.80, SD 8.55, n = 71 versus mean 18.68, SD 8.76, n = 63, MD ‐3.88, 95% CI ‐6.82 to ‐0.94; N = 134; k = 1; I² = not applicable).
3.4 Hopelessness
No data available.
3.5 General functioning
There was evidence of an effect for MBT on general functioning at the post‐intervention assessment (mean 60.9, SD 15.8, n = 71 versus mean 53.2, SD 11.7, n = 63; MD 7.70, 95% CI 3.03 to 12.37; N = 134; k = 1; I² = not applicable).
3.6 Social functioning
There was also evidence of an effect for MBT on social functioning at the post‐intervention assessment in this trial (mean 1.76, SD 0.50, n = 71 versus mean 2.17, SD 0.64, n = 63; MD ‐0.41, 95% CI ‐0.61 to ‐0.21; N = 134; k = 1; I² = not applicable).
3.7 Suicidal ideation
No data available.
3.8 Suicide
There were no suicides in either treatment arm by the time of the post‐intervention assessment.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 4: Emotion‐regulation psychotherapy versus TAU or another comparator
Two trials investigated the effectiveness of emotion‐regulation psychotherapy, delivered in a group‐based format, in adult women (weighted mean age: 33.2 ± 11.2 years) diagnosed with BPD referred for outpatient treatment as a result of recurrent SH (Gratz 2006, N = 22; Gratz 2014, N = 61).
Primary outcome
4.1 Repetition of SH
Group‐based emotion‐regulation psychotherapy may reduce repetition of SH by the post‐intervention assessment based on evidence from two trials (OR 0.34, 95% CI 0.13 to 0.88; N = 83; k = 2; I² = 0%; Analysis 3.1). According to GRADE criteria, we judged the evidence to be of moderate certainty.
There was no evidence of an effect for group‐based emotion‐regulation psychotherapy on frequency of repetition of SH by the post‐intervention assessment (Analysis 3.2).
Secondary outcomes
4.2 Treatment adherence
No data available.
4.3 Depression
There was evidence of an effect for group‐based emotion‐regulation therapy on depression scores at the post‐intervention assessment (MD ‐9.59, 95% CI ‐13.43 to ‐5.75; N = 83; k = 2; I² = 0%; Analysis 3.3).
4.4 Hopelessness
No data available.
4.5 General functioning
No data available.
4.6 Social functioning
No data available.
4.7 Suicidal ideation
No data available.
4.8 Suicide
There were no suicides in either the intervention or comparator arms in either trial by the post‐intervention assessment.
Subgroup analyses
Randomisation was stratified by repeater status in one trial (Gratz 2014). However, we were unable to obtain disaggregated data from the trial authors despite correspondence.
Sensitivity analyses
Not applicable.
Comparison 5: Psychodynamic psychotherapy versus TAU or another comparator
Two trials investigated psychodynamic psychotherapeutic approaches as compared to TAU in adults (weighted mean age: 31.7 ± 9.8 years; 88.9% female) who either presented to EDs following an episode of SH or who were receiving outpatient treatment for frequent repetition of SH (Andreoli 2015, N = 170; Sahin 2018, N = 71).
For the first trial (Andreoli 2015), there were two separate intervention arms: psychodynamic psychotherapy delivered either by certified psychotherapists or trained nurses. We therefore combined data from these two conditions given their similarity. For the second trial, however, there were two different intervention arms: psychodynamic psychotherapy and DBT (Sahin 2018). We therefore included here only data for the psychodynamic psychotherapy arm.
Primary outcome
5.1 Repetition of SH
There was no evidence of an effect for psychodynamic psychotherapy by the post‐intervention assessment in one of these trials (9/140 versus 4/30; OR 0.45, 95% CI 0.13 to 1.56; N = 170; k = 1; I² = not applicable; Andreoli 2015). According to GRADE criteria, we judged the evidence to be of moderate certainty.
This trial also reported information on time to repetition of SH, finding that "participants who received either form of AP...had increased survival to suicidal crisis relapse compared to patients assigned to TAU (AP‐P vs. TAU log‐rank test: Mantel χ² = 7.63, df = 1, p = .006; AP‐N vs. TAU log‐rank test: Mantel χ² = 9.87, df = 1, p = .002)" (Andreoli 2015, pg.11).
Secondary outcomes
5.2 Treatment adherence
There was evidence of an effect for psychodynamic psychotherapy on the proportion of participants who completed treatment by the post‐intervention assessment in one trial (131/140 versus 19/30; OR 8.43, 95% CI 3.09 to 22.99; N = 170; k = 1; I² = not applicable; Andreoli 2015).
5.3 Depression
There was evidence of an effect for psychodynamic psychotherapy on depression scores at post‐intervention in one trial (mean 9.90, SD 5.70, n = 140 versus mean 13.40, SD 6.40, n = 30; MD ‐3.50, 95% CI ‐5.98 to ‐1.02; N = 170; k = 1; I² = not applicable; Andreoli 2015).
5.4 Hopelessness
No data available.
5.5 General functioning
There was also evidence of an effect for psychodynamic psychotherapy on general functioning scores by the post‐intervention assessment in one trial (mean 62.90, SD 6.20, n = 140 versus mean 57.6, SD 7.60, n = 30; MD 5.30, 95% CI 2.39 to 8.21; N = 170; k = 1; I² = not applicable; Andreoli 2015).
5.6 Social functioning
No data available.
5.7 Suicidal ideation
One trial assessed suicidal ideation as the frequency of episodes necessitating additional intensive psychiatric care (Andreoli 2015). There was evidence of an effect for this intervention on frequency of episodes of suicidal ideation scores at post‐intervention in this trial (mean 0.30, SD 0.60, n = 140 versus mean 1.00, SD 1.10, n = 30; MD ‐0.70, 95% CI ‐1.11 to ‐0.29; N = 170; k = 1; I² = not applicable; Andreoli 2015).
5.8 Suicide
One participant died by suicide in the TAU arm in one trial by the post‐intervention assessment (0/140 versus 1/30; OR 0.07, 95% CI 0.00 to 1.76; N = 170; k = 1; I² = not applicable; Andreoli 2015).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 6: Case management versus TAU or another comparator
Five trials investigated the provision of case management for the prevention of SH compared either to TAU (Clarke 2002, N = 467; Hvid 2011, N = 133; Van Heeringen 1995, N = 516) or enhanced usual care (EUC; Kawanishi 2014, N = 914; Morthorst 2012, N = 243) in adults (weighted mean age 37.9 ± 13.7 years; 56.2% female) admitted to hospital following an episode of SH. Although the intervention in three of these trials included aspects of PST, this was not the primary or only element of the case management strategy (Hvid 2011; Kawanishi 2014; Morthorst 2012). We therefore thought these trials were sufficiently similar to the others to justify pooling within a meta‐analysis.
Primary outcome
6.1 Repetition of SH
There was no evidence of an effect for case management on repetition of SH by the post‐intervention assessment in four of these trials, nor was there any evidence of a difference by comparator (i.e. TAU versus EUC) (OR 0.78, 95% CI 0.47 to 1.30; participants = 1608; studies = 4; I2 = 54%; Analysis 4.1). Supplementing hospital‐recorded episodes of SH with self‐reported data for Morthorst 2012 did not materially affect this result. Multiple readmissions for SH were, however, significantly more common in the case management group than in the TAU group in one trial (9/220 versus 2/247; OR 5.23, 95% CI 1.12 to 24.45; N = 467; k = 1; I² = not applicable; Clarke 2002). According to GRADE criteria, we judged the evidence to be of low certainty.
There was also no evidence of an effect for case management as compared with TAU on repetition of SH by the 12‐month follow‐up assessment in one further trial (21/196 versus 34/195; OR 0.57, 95% CI 0.32 to 1.02; N = 391; k = 1; I2 = not applicable; Van Heeringen 1995).
Two trials provided information on time to SH repetition. In one, trial authors reported that "[t]he log‐rank test shows a significant difference in favour of fewer events in the intervention arm, log‐rank test 4.16 (P = 0.0414)" (Hvid 2011, pg.295). However, in the second, trial authors reported "the survival curve for the assertive case management group was not significantly different from that for the control group (log‐rank P = 0.258, Wilcoxon P = 0·103)" Kawanishi 2014, pg.197).
Secondary outcomes
6.2 Treatment adherence
The authors of one trial reported that "11 participants in the assertive case management group did not receive the intervention" (Kawanishi 2014, pg.197). However, as corresponding numbers for the EUC group were not reported, we were unable to analyse the effect of assertive case management on treatment adherence for this trial.
Case management was associated with an effect on the proportion of participants who completed treatment as compared with TAU by the 12‐month assessment in one trial (129/252 versus 102/256; OR 1.58, 95% CI 1.11 to 2.25; N = 508; k = 1; I2 = not applicable; Van Heeringen 1995).
6.3 Depression
No data available.
6.4 Hopelessness
Although the BHS was administered to participants in one trial (Kawanishi 2014), no data on this outcome was reported. Correspondence with trial authors for the previous version of this review indicated they were analysing these data to present in a future report. However, data on this outcome has not been reported in any of the secondary publications of this trial identified by this update of the review.
6.5 General functioning
No data available.
6.6 Social functioning
No data available.
6.7 Suicidal ideation
No data available.
6.8 Suicide
There was no evidence of an effect for case management on suicide by the post‐intervention assessment, nor was there any evidence of a significant difference by comparator for this outcome (Analysis 4.2). There was also no evidence of an effect for case management on suicide by the 12‐month follow‐up assessment in one further trial (6/196 versus 7/195; OR 0.85, 95% CI 0.28 to 2.57; N = 391; k = 1; I² = not applicable; Van Heeringen 1995).
Subgroup analyses
One trial stratified randomisation by sex (Hvid 2011). Although there was no evidence of an effect in males in this trial (4/20 versus 4/18; OR 0.88, 95% CI 0.18 to 4.17; N = 38; k = 1; I² = not applicable), case management was associated with an effect on repetition of SH in females (2/49 versus 10/46; OR 0.15, 95% CI 0.03 to 0.74; N = 95; k = 1; I² = not applicable).
A second trial stratified randomisation by repeater status (Morthorst 2012). However, we were unable to obtain disaggregated data from the trial authors despite correspondence.
Sensitivity analyses
Not applicable.
Comparison 7: Structured general practitioner (GP) follow‐up versus TAU or another comparator
A single trial investigated the effectiveness of structured follow‐up by the participants' general practitioner (GP) compared with TAU over a six‐month period in adults (mean age: 38.2 years, SD not reported; 74.5% female) admitted to the acute ward of a general hospital following an episode of self‐poisoning (Grimholt 2015, N = 202).
Primary outcome
7.1 Repetition of SH
There was no evidence of an effect for structured GP follow‐up on the proportion repeating SH by the post‐intervention assessment according either to hospital records (8/60 versus 11/83; OR 1.01, 95% CI 0.38 to 2.68; N = 143; k = 1; I² = not applicable) or emergency medical records (9/54 versus 5/69; OR 2.56, 95% CI 0.80 to 8.15; N = 123; k = 1; I² = not applicable). According to GRADE criteria, we judged the evidence to be of low certainty.
The trial authors also reported data on self‐reported repeat episodes of self‐poisoning, self‐cutting, and other methods of SH. Whilst a greater proportion of those assigned to structured GP follow‐up reported repeat episodes of self‐poisoning by the post‐intervention assessment (15/38 versus 9/57; OR 3.48, 95% CI 1.33 to 9.12; N = 95; k = 1; I² = not applicable), there was no evidence of an effect for this intervention on self‐reported episodes of self‐cutting (10/40 versus 9/58; OR 1.81, 95% CI 0.66 to 4.98; N = 98; k = 1; I² = not applicable) or other methods of SH (5/18 versus 9/17; OR 0.34, 95% CI 0.08 to 1.39; N = 35; k = 1; I² = not applicable) by this time point.
Secondary outcomes
7.2 Treatment adherence
There was no evidence of an effect for structured GP follow‐up on the proportion of participants who completed treatment by the post‐intervention assessment in this trial (47/62 versus 64/87; OR 1.13, 95% CI 0.53 to 2.39; N = 149; k = 1; I² = not applicable).
7.3 Depression
Although those allocated to the intervention arm in this trial had slightly higher depression scores by the post‐intervention assessment, there was no evidence of an effect (mean 24.90, SD 10.10, n = 18 versus mean 18.80, SD 12.10, n = 22; MD 6.10, 95% CI ‐0.78 to 12.98; N = 40; k = 1; I² = not applicable); however, there were substantial missing data for this outcome (109/149; 73.2%).
7.4 Hopelessness
There was also no evidence of an effect for structured GP follow‐up on hopelessness scores by the post‐intervention assessment in this trial (mean 11.70; SD 5.60, n = 18 versus mean 8.70, SD 6.70, n = 26; MD 3.00, 95% CI ‐0.65 to 6.65, N = 44; k = 1; I² = not applicable). There were, however, substantial missing data for this outcome (105/149; 70.5%).
7.5 General functioning
No data available.
7.6 Social functioning
No data available.
7.7 Suicidal ideation
There was no evidence of an effect for structured GP follow‐up on suicidal ideation scores by the post‐intervention assessment (mean 19.80, SD 5.30, n = 6 versus mean 17.10, SD 7.20, n = 7; MD 2.70, 95% CI ‐4.11 to 9.51; N = 13; k = 1; I² = not applicable). Again, there were substantial missing data for this outcome (136/149; 91.3%).
7.8 Suicide
There was one suicide death in each trial arm by the post‐intervention assessment. However, there was no evidence of an effect for structured GP follow‐up by the post‐intervention assessment for this outcome (1/62 versus 1/87; OR 1.41, 95% CI 0.09 to 22.98; N = 149; k = 1; I² = not applicable).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 8: Brief emergency department‐based interventions versus TAU or another comparator
Five trials investigated the effectiveness of a one‐off brief (i.e. between 30 to 57 minutes' duration) intervention conducted in the emergency department (ED) in adults (weighted mean age 35.4 ± 13.5 years; 61.2% female) presenting to EDs following an episode of SH (Armitage 2016, N = 226; O'Connor 2015, N = 30; O'Connor 2017, N = 518; O'Connor 2020, N = 48).
Comparison 8.1: Brief Collaborative Assessment and Management of Suicidality (CAMS)‐based intervention
Two trials investigated the effectiveness of a brief intervention informed by therapeutic principles of both the Collaborative Assessment and Management of Suicidality (CAMS) model, in which a collaborative rapport is developed together with the participant, and the functional analysis model of DBT in which validation and guided discovery is used to understand the participant's motivations for SH, in adults (weighted mean age: 42.1 ± 8.7 years; 43.6% female) presenting to the ED following an episode of SH (O'Connor 2015, N = 30; O'Connor 2020, N = 48).
Primary outcome
8.1.1 Repetition of SH
There was no evidence of an effect for brief CAMS and DBT‐based interventions on repetition of SH by the 12‐month follow‐up assessment in one of these trials (1/13 versus 2/10; OR 0.33, 95% CI 0.03 to 4.32; N = 23; k = 1; I² = not applicable; O'Connor 2020).
One trial also reported data on frequency of repeat SH by the 12‐month follow‐up assessment which could be estimated from the data reported. Once again, however, there was no evidence of an effect for a brief ED‐based intervention on frequency of repeat SH in this trial (mean 0.08, SD 0.28, n = 13 versus mean 1.30, SD 3.47, n = 10; MD ‐1.22, 95% CI ‐3.38 to 0.94; N = 23; k = 1; I² = not applicable; O'Connor 2020).
Secondary outcomes
8.1.2 Treatment adherence
No data available.
8.1.3 Depression
No data available.
8.1.4 Hopelessness
Data on hopelessness were available for both of these trials; however, the trial authors did not report information on SDs or other information to enable imputation of SDs.
By the six‐month follow‐up assessment, authors of one of these trials reported "[s]ignificant improvements across time on Reasons for Living improvements were also observed for the [intervention] group compared to the usual care group..." (O'Connor 2015, pg.431). However, data estimated from graphics presented in the trial report suggested any such difference was likely to be modest (estimated mean ‐138.00, n = 11 versus estimated mean ‐136.50, n = 13).
By the 12‐month follow‐up assessment, authors of the second of these trials reported there were "no statistically significant differences between groups on reported reasons for living...at any time point" (O'Connor 2020, pg.116).
8.1.5 General functioning
No data available.
8.1.6 Social functioning
No data available.
8.1.7 Suicidal ideation
Data on suicidal ideation scores were available for both of these trials; however, the trial authors did not report information on SDs or other information to enable imputation of SDs.
By the six‐month assessment, authors of the first trial reported "[a]lthough a trend was observed for greater improvement of suicidal ideation for the TAU group..." (O'Connor 2015, pg.431), this was not significant. Indeed, data estimated from graphics presented in the trial report suggested that suicidal ideation scores were similar between the intervention and control arms by this time point (estimated mean 7.00, n = 11 versus estimated mean 6.00, n = 13).
For the second trial (O'Connor 2020), data estimated from graphics presented in the trial report indicated there was no meaningful difference in suicidal ideation scores by the 12‐month follow‐up assessment in this trial (estimated mean 5.50, n = 13 versus estimated mean 5.30, n = 10).
8.1.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 8.2: Brief guided Integrated Motivational‐Volitional‐focused intervention
Two trials investigated a guided intervention, based on the Integrated Motivational‐Volitional help‐sheet model, in which participants are encouraged to identify situations in which they may be triggered to engage in SH and link these with alternative responses that may help them to avoid SH (Armitage 2016, N = 151; O'Connor 2017, N = 518).
For one of these trials (Armitage 2016), there were two separate intervention arms: a guided implementation‐focused intervention based on the volitional help‐sheet model (Armitage 2016a), and a self‐directed implementation intention‐focused intervention (Armitage 2016b). Given that the model used in Armitage 2016a was based in a large part on O'Connor 2017, and involved authors of this latter trial in the design of the intervention, we grouped these two trials within a single analysis. For a detailed description of the therapeutic content of the two different intervention arms included in this trial, see the Characteristics of included studies section.
Primary outcome
8.2.1 Repetition of SH
There was no evidence of an effect for this intervention on repetition of SH by the six‐month follow‐up assessment in one of these trials (72/259 versus 72/259; OR 1.00, 95% CI 0.68 to 1.47; N = 518; k = 1; I² = not applicable; O'Connor 2017).
One trial reported data on frequency of repeat SH episodes by the six‐month follow‐up assessment; however, there was no evidence of an effect for this intervention in this trial (mean 0.68, SD 2.53, n = 259 versus mean 0.85, SD 2.78, n = 259; MD ‐0.17, 95% CI ‐0.63 to 0.29; N = 518; k = 1; I² = not applicable; O'Connor 2017). For the second of these trials (Armitage 2016a), data on frequency of SH could not be disaggregated from data on frequency of suicidal ideation and therefore could not be included in the review.
There was also no evidence of an effect for this intervention on time to SH repetition in one of these trials (HR 0.80, 95% CI 0.55 to 1.18, N = 518; k = 1; I² = not applicable; O'Connor 2017).
Secondary outcomes
8.2.2 Treatment adherence
Data from one trial indicated that almost all (95.7%) of those assigned to the intervention arm completed the intervention in the hospital. However, as corresponding numbers were not reported for the comparator arm in this trial, we could not calculate treatment effect sizes for this outcome (O'Connor 2017).
8.2.3 Depression
One trial reported data on depression scores by the six‐month follow‐up assessment (Armitage 2016a); however, there was no evidence of an effect for this intervention (mean 16.65, SD 5.92, n = 75 versus mean 17.78, SD 6.51, n = 73; MD ‐1.13, 95% CI ‐3.14 to 0.88; N = 148; k = 1; I² = not applicable).
8.2.4 Hopelessness
No data available.
8.2.5 General functioning
No data available.
8.2.6 Social functioning
No data available.
8.2.7 Suicidal ideation
No data available.
8.2.8 Suicide
There was no evidence of an effect for this intervention on suicide deaths by the six‐month follow‐up in one trial (1/259 versus 2/259; OR 0.50, 95% CI 0.04 to 5.53; N = 518; k = 1; I² = not applicable; O'Connor 2017).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 8.3: Brief self‐guided Integrated Motivational‐Volitional‐focused intervention
The second intervention arm in Armitage 2016 investigated the effectiveness of a self‐guided intervention, based in the Integrated Motivational‐Volitional help‐sheet model, as compared to TAU in adults (mean age: 29.3 ± 11.9 years; 69.9% female) presenting to the emergency department following an episode of SH (Armitage 2016b, N = 151).
Primary outcome
8.3.1 Repetition of SH
For this trial, data on frequency of SH could not be disaggregated from data on frequency of suicidal ideation and therefore could not be included in the review.
Secondary outcomes
8.3.2 Treatment adherence
No data available.
8.3.3 Depression
No effect for this intervention approach was found for depression scores by the six‐month follow‐up assessment in this trial (mean 16.20, SD 6.56, n = 78 versus mean 17.78, SD 6.51, n = 73; MD ‐1.58, 95% CI ‐3.67 to 0.51; N = 151; k = 1; I² = not applicable; Armitage 2016b).
8.3.4 Hopelessness
No data available.
8.3.5 General functioning
No data available.
8.3.6 Social functioning
No data available.
8.3.7 Suicidal ideation
No data available.
8.3.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 8.4: Brief alcohol‐focused intervention
One trial investigated the effectiveness of a brief ED‐based intervention for alcohol misuse on repetition of SH over a six‐month follow‐up period in adults (mean age: 37.2 ± 12.0 years; 50.0% female) who were misusing alcohol and were admitted to the ED following an episode of SH (Crawford 2010, N = 103).
Primary outcome
8.4.1 Repetition of SH
There was no evidence of an effect for brief alcohol‐focused ED‐based interventions on repetition of SH by the six‐month follow‐up assessment in this trial (7/52 versus 11/51; OR 0.57, 95% CI 0.20 to 1.60; k = 1; I² = not applicable; Crawford 2010).
Secondary outcomes
8.4.2 Treatment adherence
The trial authors reported that only around half (47.1%) of those randomised to the intervention arm attended the brief alcohol treatment session (Crawford 2010, pg.1826). However, as corresponding numbers were not available for those allocated to the control arm, who did not receive an invitation to brief alcohol treatment sessions, we could not calculate treatment effect sizes for this outcome.
8.4.3 Depression
No data available.
8.4.4 Hopelessness
No data available.
8.4.5 General functioning
No data available.
8.4.6 Social functioning
No data available.
8.4.7 Suicidal ideation
No data available.
8.4.8 Suicide
Correspondence with trial authors confirmed that no participants died by suicide in either arm over the course of the six‐month follow‐up period. However, the authors warned that, as they were unable to track participants via their National Health Service (NHS) identity numbers, they were unable to confirm numbers of suicides from national mortality data. Thus, there may have been suicide deaths amongst those participants whom the authors were unable to contact by the six‐month follow‐up assessment in this trial.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 9: Remote contact interventions versus TAU or another comparator
A number of trials investigated the effectiveness of a variety of remote contact interventions, including: emergency cards (Evans 1999a; Morgan 1993), coping cards (Wang 2016), general practitioners' (GP) letters (Bennewith 2002), postcards (Beautrais 2010; Carter 2005; Hassanian‐Moghaddam 2011; Kapur 2013), telephone contact (Cedereke 2002; Mousavi 2015; Vaiva 2006), telephone contact combined with emergency cards and letters (Mouaffak 2015; Vaiva 2018), and telephone‐based psychotherapy (Marasinghe 2012; Sreedaran 2020; Wei 2013).
Comparison 9.1: Emergency cards
Two trials investigated the effectiveness of providing an emergency contact card ('green card') providing 24‐hour access to emergency advice from a psychiatrist in addition to TAU in adults (weighted mean age: 32.6 ± 11.7 years; 55.4% female) admitted to general hospitals following an episode of SH, most commonly self‐poisoning (Evans 1999a, N = 827; Morgan 1993, N = 212). Note, repetition of SH data for Evans 1999a was reported in a secondary publication (Evans 2005).
Primary outcome
9.1.1 Repetition of SH
There was no evidence of an effect for emergency cards on repetition of SH by the post‐intervention assessment in these two trials (OR 0.82, 95% CI 0.31 to 2.14; participants = 1039; studies = 2; I2 = 68%; Analysis 5.1). According to GRADE criteria, we judged the evidence to be of low certainty.
There was also no evidence of an effect for emergency cards by the 12‐month follow‐up assessment in the one trial for which data were available (Analysis 5.2).
One trial reported data on frequency of repetition of SH as the proportion of participants who had no episodes at follow‐up, the proportion with a single episode at follow‐up, and the proportion with two or more repeat episodes of SH by the 12‐month follow‐up assessment (Evans 1999a). There was no difference between groups in the number of participants who had none (347/417 versus 351/410; OR 0.83, 95% CI 0.57 to 1.21; N = 827; k = 1; I² = not applicable), one (46/417 versus 32/410; OR 1.46, 95% CI 0.91 to 2.35; N = 827; k = 1; I² = not applicable), or two or more (24/417 versus 27/410; OR 0.87, 95% CI 0.49 to 1.53; N = 827; k = 1; I² = not applicable) further episodes of SH.
One trial also reported information on time to SH repetition. The authors found that whilst "[t]he median time to first repeat among those given a green card was 33 days (range 1‐180); in the control group it was 40 days (range 2‐156). This difference is not significant (log rank test: chi‐square 0.82, P = 0.36" (Evans 1999a, pg.25).
Secondary outcomes
9.1.2 Treatment adherence
No data available.
9.1.3 Depression
No data available.
9.1.4 Hopelessness
No data available.
9.1.5 General functioning
No data available.
9.1.6 Social functioning
No data available.
9.1.7 Suicidal ideation
No data available.
9.1.8 Suicide
Data on suicides were reported in only one trial (Evans 1999a). There was no evidence of an effect for emergency cards on suicide by the post‐intervention assessment in this trial (2/417 versus 1/410; OR 1.97, 95% CI 0.18 to 21.82; N = 827; k = 1; I² = not applicable).
Subgroup analyses
Data on repetition of SH was available by repeater status (i.e. those without a history of multiple episodes of SH versus those with a history of multiple episodes of SH) in one trial (Evans 1999a). There was no evidence of an effect for emergency cards on repetition of SH in those without a history of multiple episodes of SH; however, emergency cards were associated with an increased risk of repetition of SH in those with a history of multiple episodes of SH by the 12‐month follow‐up assessment in this trial (Analysis 5.2).
For those without a history of multiple episodes of SH, there was no difference between groups in the number of participants who had none (203/221 versus 181/206; OR 1.56, 95% CI 0.82 to 2.95; N = 427; k = 1; I² = not applicable), one (13/221 versus 16/206; OR 0.74, 95% CI 0.35 to 1.58; N = 427; k = 1; I² = not applicable), or two or more (5/221 versus 9/206; OR 0.51, 95% CI 0.17 to 1.54; N = 427; k = 1; I² = not applicable) repeat episodes of SH (Evans 1999a).
For those with a history of multiple episodes of SH, however, receipt of an emergency card was associated with a reduction in the number of participants with no further episodes of SH (142/194 versus 167/200; OR 0.54, 95%CI 0.33 to 0.88; N = 394; k = 1; I² = not applicable) coupled with an increase in the number of participants with one repeat episode of SH (33/194 versus 15/200; OR 2.53, 95% CI 1.33 to 4.82; N = 394; k = 1; I² = not applicable). There was no difference between the intervention and comparator groups with respect to the number of participants with or two or more subsequent episodes of SH for those with a history of multiple episodes of SH, however (19/194 versus 18/200; OR 1.10, 95% CI 0.56 to 2.16; N = 394; k = 1; I² = not applicable) (Evans 1999a).
Sensitivity analyses
Not applicable.
Comparison 9.2: Coping cards
A single, small trial investigated the effectiveness of coping cards, in which participants were encouraged to note down alternative activities they could engage in when feeling suicidal, resources they found helpful when seeking help, as well as the numbers of a 24‐hour crisis line and local medical services, in addition to TAU as compared to TAU alone in adults (mean age: 37.9±11.1 years; 73.4% female) referred to a specialist suicide prevention service following a suicide attempt (Wang 2016, N = 64).
Primary outcome
9.2.1 Repetition of SH
There was no evidence of an effect for coping cards on the proportion of participants re‐attempting suicide by the post‐intervention assessment (0/32 versus 5/32; OR 0.08, 95% CI 0.00 to 1.45; N = 64; k = 1; I² = not applicable; Wang 2016). According to GRADE criteria, we judged the evidence to be of moderate certainty.
Regarding time to repetition, whilst numeric data were not provided the authors report "[r]esults indicated that participants in the coping card group had a significantly longer time to re‐attempt than those in the TAU group...[at the] 3‐month" (i.e., post‐intervention) period (Wang 2016, p.112).
Secondary outcomes
9.2.2 Treatment adherence
No data available.
9.2.3 Depression
The authors indicate they measured depression; however, no data on this outcome were reported in the trial report.
9.2.4 Hopelessness
There was no evidence of an effect for coping cards on hopelessness scores at post‐intervention (mean 6.81, SD 2.98, n = 32 versus mean 8.38, SD 3.96, n = 32; MD ‐1.57, 95% CI ‐3.29 to 0.15; N = 64; k = 1; I² = not applicable; Wang 2016).
9.2.5 General functioning
No data available.
9.2.6 Social functioning
No data available.
9.2.7 Suicidal ideation
Suicidal ideation was assessed in this trial; however, an idiosyncratic six‐point scale was used. Data on this outcome were therefore unable to be included in this review.
9.2.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 9.3: GP letters
A single cRCT compared the effectiveness of a letter from participants' general practitioners following discharge from hospital offering an appointment and advice compared to TAU over a 12‐month period in adults (mean age: 32.5 ± 13.2 years; 59.0% female) presenting to hospital following an episode of SH (Bennewith 2002, clusters = 98, N = 1932).
As the trial authors were unable to provide values for either the inter‐cluster correlation coefficient or the design effect, and further, there was no similar cluster RCT of this intervention approach from which these values could be approximated, we were unable to statistically account for the effects of clustering. Results presented in this section may therefore overestimate the effectiveness of this intervention.
Primary outcome
9.3.1 Repetition of SH
There was no evidence of an effect for a letter from participants' GPs on repetition of SH by the 12‐month follow‐up assessment in this trial (211/964 versus 189/968; OR 1.15, 95% CI 0.93 to 1.44; N = 1932; k = 1; I² = not applicable).
There was also no evidence of an effect for GP letters on time to SH repetition in this trial (HR, 1.15, 95% CI 0.94 to 1.42; N = 1932; k = 1; I² = not applicable).
Secondary outcomes
9.3.2 Treatment adherence
There was no effect for the number of participants with a least one contact with treatment services by the 12‐month follow‐up assessment (351/599 versus 387/681; OR 1.08, 95% CI 0.86 to 1.34; N = 1280; k = 1; I² = not applicable).
9.3.3 Depression
No data available.
9.3.4 Hopelessness
No data available.
9.3.5 General functioning
No data available.
9.3.6 Social functioning
No data available.
9.3.7 Suicidal ideation
No data available.
9.3.8 Suicide
No data available.
Subgroup analyses
Data on repetition of SH was available by sex and repeater status (i.e. those without a history of multiple episodes of SH versus those with a history of multiple episodes of SH). A post hoc analysis by sex suggested that whilst there was no effect for males (82/383 versus 84/413; OR 1.07, 95% CI 0.76 to 1.50; N = 796; k = 1; I² = not applicable), a GP letter was associated with an effect on repetition of SH for females (30/581 versus 105/555; OR 0.23, 95% CI 0.15 to 0.36; N = 1136; k = 1; I² = not applicable).
The trial authors also reported data on repetition of SH by repeater status at trial entry and concluded that "[t]he odds ratio for the effect of the intervention in patients with a history of self‐harm was 0.57 (0.33 to 0.98), indicating a beneficial effect, and in those with no history was 1.32 (1.02 to 1.70), indicating a harmful effect" (Bennewith 2002, pg.1258). However, as the raw data on which these subgroup results were based were not reported, we were unable to reproduce these results in this review.
Sensitivity analyses
Not applicable.
Comparison 9.4: Postcards
Four trials assessed the effectiveness of sending postcards on a regular basis over a 12‐month period to adults (weighted mean age: 27.1 ± 8.9 years; 63.0% female) presenting to hospital following an episode of SH (Beautrais 2010, N = 327; Hassanian‐Moghaddam 2011, N = 2113; Kapur 2013, N = 66), including one that used Zelen's design and reported data only for the randomised sample (Carter 2005, N = 772).
Primary outcome
9.4.1 Repetition of SH
Overall, there was no evidence of an effect for postcards on the proportion of patients repeating SH by the post‐intervention assessment (OR 0.87, 95% CI 0.62 to 1.23; participants = 3277; studies = 4; I2 = 51%; Analysis 6.1). According to GRADE criteria, we judged the evidence to be of very low certainty.
There was also no evidence of an effect for postcards on repetition of SH by the 12‐month follow‐up assessment in two of these trials (Analysis 6.2).
With respect to frequency of SH, we obtained data on the mean number of repeat SH episodes by correspondence for three of the four trials of postcards (Carter 2005; Hassanian‐Moghaddam 2011; Kapur 2013). Overall, there was no evidence of an effect for postcards on frequency of repeated SH by the post‐intervention assessment in these three trials (Analysis 6.3). For the one remaining trial (Beautrais 2010), the available data indicated a reduced mean number of SH episodes for the intervention group at post‐intervention (0.57 versus 0.78); however, SDs were not presented and insufficient information was reported to enable imputation of SDs. There was similarly no evidence of an effect for postcards either by the 12‐month follow‐up assessment in two trials (Analysis 6.4), or by the 24‐month follow‐up assessment in one trial (Analysis 6.5).
Secondary outcomes
9.4.2 Treatment adherence
No data available.
9.4.3 Depression
No data available.
9.4.4 Hopelessness
No data available.
9.4.5 General functioning
No data available.
9.4.6 Social functioning
No data available.
9.4.7 Suicidal ideation
One trial reported information on suicidal ideation at both the post‐intervention assessment and 12‐ to 24‐months' follow‐up (Hassanian‐Moghaddam 2011). There was an effect for postcards on the proportion of participants reporting suicidal ideation at the post‐intervention assessment (302/1043 versus 446/1070; OR 0.57, 95% CI 0.48 to 0.68; N = 2113; k = 1; I² = not applicable). Data reported by the same trial authors in a subsequent follow‐up publication suggested that this effect was maintained by the 12‐ to 24‐month follow‐up assessment (465/997 versus 588/1004; OR 0.62, 95% CI 0.52 to 0.74; N = 2001; k = 1; I² = not applicable; Hassanian‐Moghaddam 2017).
9.4.8 Suicide
There was no evidence of an effect for postcards on suicide by the post‐intervention assessment in these four trials (Analysis 6.6).
Data on suicides by the 12‐month follow‐up assessment were reported in one trial; however, no effect was found by this time point either (2/378 versus 5/394; OR 0.41, 95% CI 0.08 to 2.15; N = 772; k = 1; I² = not applicable; Carter 2005).
Subgroup analyses
Randomisation was stratified by sex in one of these trials (Carter 2005); however, there was no evidence of an effect for postcards in either sex at either the post‐intervention (Analysis 6.1) or the the 12‐month follow‐up assessment (Analysis 6.2).
Through correspondence, we were also able to obtain data on frequency of SH repetition by the post‐intervention assessment by both sex and repeater status for three trials (Carter 2005; Hassanian‐Moghaddam 2011; Kapur 2013). There was no evidence of an effect for postcards either by sex (test for subgroup differences: Chi² = 0.07, df = 1, P = 0.79; Analysis 6.3) or by repeater status (i.e. first SH episode versus repeat SH episode) (test for subgroup differences: Chi² = 0.59, df = 1, P = 0.44; Analysis 6.3) in these trials. There was also no evidence of an effect for postcards either by sex (test for subgroup differences: Chi² = 1.25, df = 1, P = 0.26; Analysis 6.4) or by repeater status (test for subgroup differences: Chi² = 0.60, df = 1, P = 0.44; Analysis 6.4) by the 12‐month follow‐up assessment in two trials, or by the 24‐month follow‐up period in one trial (sex: test for subgroup differences: Chi² = 0.57, df = 1, P = 0.45; repeater status: test for subgroup differences: Chi² = 0.35, df = 1, P = 0.55; Analysis 6.5).
Sensitivity analyses
One trial used Zelen's design (Carter 2005). Omitting this trial did not materially affect results for postcards on repetition of SH by the post‐intervention assessment. However, by the 12‐month assessment, excluding this trial caused the result for repetition of SH to become statistically significant (OR 0.67, 95% CI 0.52 to 0.86; N = 2113; k = 1; I² = not applicable). Additionally, excluding this trial suggested a potentially harmful effect of postcards on suicides by the post‐intervention assessment (OR 3.74, 95% CI 1.04 to 13.51; N = 2692; k = 3; I² = 0%).
Four analyses within this comparison were associated with substantial levels of heterogeneity (Analysis 6.4.1, I² = 84%; Analysis 6.4.1, I² = 84%; Analysis 6.4.3, I² = 76%; Analysis 6.4.4, I² = 81%); however, analyses did not indicate any individual study was associated with excessive influence for either of these outcomes.
Comparison 9.5: Telephone contact
Three trials investigated the effectiveness of telephone contact in adults (weighted mean age: 36.5 ± 13.5 years; 72.1% female) presenting to emergency departments following a suicide attempt (Cedereke 2002, N = 216; Mousavi 2015, N = 55; Vaiva 2006, N = 605).
Primary outcome
9.5.1 Repetition of SH
There was no evidence of an effect for telephone contact on repetition of SH by the post‐intervention in one trial (1/29 versus 2/26; OR 0.43, 95% CI 0.04 to 5.02; N = 55; k = 1; I² = not applicable; Mousavi 2015). According to GRADE criteria, we judged the evidence to be of low certainty.
There was also no evidence of an effect for telephone contact on repetition of SH by the 12‐month assessment in a second trial (14/83 versus 15/89; OR 1.00, 95% CI 0.45 to 2.23; N = 172; k = 1; I² = not applicable; Cedereke 2002), or by the 24‐month follow‐up assessment in a third trial (44/293 versus 59/312; OR 0.76, 95% CI 0.49 to 1.16; N = 605; k = 1; I² = not applicable; Vaiva 2006).
With respect to frequency of repetition of SH, there was no evidence of an effect for telephone contact by the post‐intervention assessment in one trial (mean 0.04, SD 0.20, n = 25 versus mean 0.18, SD 0.66, n = 22; MD ‐0.14, 95% CI ‐0.43 to 0.15; N = 47; k = 1; I² = not applicable; Mousavi 2015). The mean number of episodes of SH was similar between arms in both Cedereke 2002 by the 12‐month assessment (0.31 versus 0.30) and in Vaiva 2006 by the 24‐month assessment (0.15 versus 0.19). However, as trial authors provided insufficient information to enable imputation of SDs, we were unable to calculate the mean difference in the number of repeat episodes of SH between arms in these two trials.
Secondary outcomes
9.5.2 Treatment adherence
No data available.
9.5.3 Depression
No data available.
9.5.4 Hopelessness
No data available.
9.5.5 General functioning
Whilst data on general functioning were available for one of these trials by the 12‐month follow‐up assessment (mean 61.4, SD 20.4 versus mean 58.6, 20.2), it was unclear how many of the 163 participants with data on this outcome had been allocated to the intervention and control arms (Cedereke 2002). We were therefore unable to include these data in our review. The trial authors, however, reported that "[a]t 12 months there had been significant improvements within both the intervention and control groups in global functioning (GAF)..." (Cedereke 2002, pg.87).
9.5.6 Social functioning
No data available.
9.5.7 Suicidal ideation
Telephone contact was not associated with an effect on suicidal ideation scores by the 12‐month follow‐up assessment in one trial (mean 5.80, SD 7.80, N = 5 versus mean 4.00, SD 6.20, N = 8; MD 1.80, 95% CI ‐6.27 to 9.87; N = 13; k = 1; I² = not applicable; Cedereke 2002).
One further trial reported information on the proportion of participants reporting suicidal ideation; however, once again telephone contact was not associated with an effect by the post‐intervention assessment in this trial (2/25 versus 1/22; OR 1.83, 95% CI 0.15 to 21.64; N = 47; k = 1; I² = not applicable; Mousavi 2015).
9.5.8 Suicide
There was no evidence of an effect for telephone contact on suicides either by the post‐intervention (0/25 versus 1/22; OR 0.28, 95% CI 0.01 to 7.26; N = 47; k = 1; I² = not applicable; Mousavi 2015), 12‐month (1/107 versus 1/109; OR 1.02, 95% CI 0.06 to 16.50; N = 216; k = 1; Cedereke 2002) or 24‐month (1/293 versus 2/312; OR 0.52, 95% CI 0.05 to 5.89; N = 605; Vaiva 2006) follow‐up assessments.
Subgroup analyses
Whilst one trial stratified randomisation, this trial compared those who engaged in three or fewer suicide attempts over the three years prior to trial entry to those who had engaged in four or more episodes over this time frame, rather than those engaging in a first episode versus a repeat episode (Vaiva 2006). We were therefore unable to include these data in our review.
Sensitivity analyses
Not applicable.
Comparison 9.6: Telephone contact combined with emergency cards and letters
Two relatively large trials assessed the effectiveness of telephone contact combined with emergency cards and letters (for those unable to be contacted via telephone) compared to TAU in adults (weighted mean age: 38.4 ± 13.2 years; 64.5% female) presenting to the emergency department following a suicide attempt (Mouaffak 2015, N = 302; Vaiva 2018, N = 987).
Primary outcome
9.6.1 Repetition of SH
There was no evidence of an effect for telephone contact combined with emergency cards and letters on repetition of SH by the post‐intervention assessment in one of these trials (22/152 versus 21/151; OR 1.05, 95% CI 0.55 to 2.00; N = 303; k = 1; I² = not applicable; Mouaffak 2015). According to GRADE criteria, we judged the evidence to be of moderate certainty.
For the second of these trials, information on SH repetition was combined with that for suicide deaths (Vaiva 2018). Despite correspondence, we were unable to obtain disaggregated data for these outcomes. We were therefore unable to include these data in our analyses.
There was also no evidence of an effect for this intervention on frequency of repeated SH by the post‐intervention assessment in one of these trials (mean 0.20, SD 0.58, n = 152 versus mean 0.23, SD 0.84, n = 150; MD ‐0.03, 95% CI ‐0.19 to 0.13; N = 302; k = 1; I² = not applicable; Mouaffak 2015).
Although neither trial provided numeric data, both reported there was no effect for telephone contact combined with emergency cards and letters on time to SH repetition: "[t]he intervention had no significant impact on...the duration of the period that patients were free of attempts (Vaiva 2018, pg.5), and, there was no "significant differences between the two groups" (Mouaffak 2015, pg.916).
Secondary outcomes
9.6.2 Treatment adherence
No data available.
9.6.3 Depression
No data available.
9.6.4 Hopelessness
No data available.
9.6.5 General functioning
No data available.
9.6.6 Social functioning
No data available.
9.6.7 Suicidal ideation
No data available.
9.6.8 Suicide
There was no evidence of an effect for telephone contact combined with emergency cards and letters by the post‐intervention assessment in one trial (1/152 versus 0/150; OR 2.98, 95% CI 0.12 to 73.74; N = 302; k = 1; I² = not applicable; Mouaffak 2015), or by the 13‐month follow‐up assessment in the second trial (3/493 versus 8/494; OR 0.37, 95% CI 0.10 to 1.41; N = 987; k = 1; I² = not applicable; Vaiva 2018).
Subgroup analyses
Randomisation was stratified by history of SH prior to trial entry in one of these trials (Mouaffak 2015); however, there was no evidence of an effect for telephone contact combined with emergency cards and letters on repetition of SH by the post‐intervention assessment either in those with a history of SH (18/79 versus 17/72; OR 0.95, 95% CI 0.45 to 2.03; N = 151; k = 1; I² = not applicable) or in those without this history at trial entry (4/77 versus 4/74; OR 0.96, 95% CI 0.23 to 3.98; N = 151; k = 1; I² = not applicable).
Sensitivity analyses
Not applicable.
Comparison 9.7: Telephone‐based psychotherapy
Three trials from developing countries (i.e. China, India, and Sri Lanka) assessed the effectiveness of psychotherapy, incorporating principles from CBT‐based psychotherapy, third‐wave techniques, and social support delivered by telephone over either a one‐month (Sreedaran 2020, N = 28), three‐month (Wei 2013, N = 157), or six‐month (Marasinghe 2012, N = 68) period in adults (weighted mean age: 32.6 ± 14.0; 69.6% female) presenting to services following an episode of SH.
As this latter trial used a cross‐over design, we reported only data from the post‐intervention assessment (i.e. prior to cross‐over) in this review. Additionally, for one of these trials (i.e. Wei 2013), there were two intervention arms: CBT‐based psychotherapy and telephone‐based psychotherapy. We therefore included here only data for the telephone contact arm, and have split the comparator arm data following the advice in Higgins 2011.
Primary outcome
9.7.1 Repetition of SH
For one trial, correspondence with trial authors indicated there were "no suicide attempts in both groups in those who did not drop out [at post‐intervention]...We were unable to obtain information from those who dropped out with respect to suicide attempts. We did have one suicide attempt in the [intervention] group, but this was before all interventions were completed and so this was excluded from final analysis" (Sreedaran 2020, personal communication).
For the remaining two trials, there was no evidence of an effect for telephone‐based psychotherapy on repetition of SH by the post‐intervention in two of these trials (OR 0.36, 95% CI 0.01 to 8.94; participants = 185; studies = 2; I2 = 0%; Analysis 7.1). According to GRADE criteria, we judged the evidence to be of low certainty.
There was also no evidence of an effect for telephone‐based psychotherapy on repetition of SH by the six‐month (1/41 versus 4/40; OR 0.23, 95% CI 0.02 to 2.11; N = 81; k = 1; I² = not applicable; Wei 2013), or 12‐month follow‐up assessments in one of these trials (1/36 versus 5/27; OR 0.13, 95% CI 0.01 to 1.15; N = 63; k = 1; I² = not applicable; Wei 2013).
Secondary outcomes
9.7.2 Treatment adherence
There was no effect for mobile telephone‐based psychotherapy on the proportion of participants who completed treatment by the post‐intervention assessment in one of these trials (8/13 versus 7/12; OR 1.14, 95% CI 0.23 to 5.67; N = 25; k = 1; I² = not applicable; Sreedaran 2020).
9.7.3 Depression
There was also no evidence of an effect for mobile telephone‐based psychotherapy on depression scores by the post‐intervention assessment in two of these trials (Analysis 7.2), or by the six‐month (mean 6.01, SD 8.87, n = 41 versus mean 5.85, SD 8.16, n = 40; MD 0.16, 95% CI ‐3.55 to 3.87; N = 81; k = 1; I² = not applicable) or 12‐month (mean 5.73, SD 8.71, n = 36 versus mean 5.84, SD 8.23, n = 27; MD ‐0.11, 95% CI ‐4.32 to 4.10; N = 63; k = 1; I² = not applicable) follow‐up assessments in one of these trials (Wei 2013).
9.7.4 Hopelessness
No data available.
9.7.5 General functioning
No data available.
9.7.6 Social functioning
No data available.
9.7.7 Suicidal ideation
There was evidence of an effect for mobile telephone‐based psychotherapy for suicidal ideation scores at the post‐intervention assessment in one trial (mean 3.60, SD 1.60, n = 34 versus mean 7.30, SD 5.50, n = 34; MD ‐3.70, 95% CI ‐5.63 to ‐1.77; N = 68; k = 1; I² = not applicable; Marasinghe 2012).
One trial reported information on the proportion of participants reporting suicidal ideation (Wei 2013); however, there was no evidence of an effect for telephone‐based psychotherapy on the proportion reporting suicidal ideation at either the post‐intervention (37/80 versus 32/77; OR 1.21, 95% CI 0.64 to 2.27; N = 157; k = 1; I² = not applicable), six‐month (26/80 versus 24/77; OR 1.06, 95% CI 0.54 to 2.08; N = 157; k = 1; I² = not applicable), or 12‐month (24/80 versus 25/77; OR 0.89, 95% CI 0.45 to 1.75; N = 157; k = 1; I² = not applicable) follow‐up assessments in this trial. However, it is notable that, for this trial, results were reported on the basis of those randomised despite the fact that a self‐reported measure was used, and high levels of dropouts were observed by the 12‐month follow‐up assessment. As we were unable to confirm these numbers with trial authors, results much be interpreted with caution.
9.7.8 Suicide
Data obtained by correspondence indicated that there was no evidence of an effect for telephone‐based psychotherapy on suicide by the post‐intervention assessment (Analysis 7.3).
Subgroup analyses
Telephone‐based psychotherapy was associated with an effect for depression scores at the post‐intervention assessment in males (mean 5.90, SD 2.40, n = 17 versus mean 13.30, SD 6.10, n = 17; MD ‐7.40, 95% CI ‐10.52 to ‐4.28; N = 34; k = 1; I² = not applicable), but not in females (mean 8.10, SD 6.30, n = 17 versus mean 11.60, SD 6.50, n = 17; MD ‐3.50, 95% CI ‐7.80 to 0.80; N = 34; k = 1; I² = not applicable) in one trial (Marasinghe 2012).
However, for suicidal ideation scores, whilst there was no evidence of an effect for telephone‐based psychotherapy at the post‐intervention assessment in males in this trial (mean 3.50, SD 1.80, n = 17 versus mean 6.20, SD 5.50, n = 17; MD ‐2.70, 95% CI ‐5.45 to 0.05; N = 34; k = 1; I² = not applicable), there was, however, an effect for females (mean 3.80, SD 1.40, n = 17 versus mean 8.90, SD 6.20, n = 17; MD ‐5.10, 95% CI ‐8.12 to ‐2.08; N = 34; k = 1; I² = not applicable) (Marasinghe 2012).
Sensitivity analyses
One analysis within this comparison was associated with substantial levels of heterogeneity (Analysis 7.2, I² = 90%); however, analyses did not indicate any individual study was associated with excessive influence for this outcome.
Comparison 10: Provision of information and support versus TAU or another comparator
Two multicentre and multi‐country studies investigated the effectiveness of providing a one‐off hospital‐based information session combined with regular home visits and/or telephone contact in addition to TAU over an 18‐month period in adults (weighted mean age: 26.7 years, SD not reported; 62.6% female) presenting to the emergency department following an episode of SH: the START Study (Amadéo 2015, N = 190) and the SUPRE‐MISS Study (Fleischmann 2008, N = 1699). In one further trial, participants nominated a support person in addition to receiving information and support using the original SUPRE‐MISS model (Naidoo 2014, N = 688).
Primary outcome
10.1 Repetition of SH
There was no evidence of an effect for information and support on repetition of SH by the 18‐month post‐intervention assessment in two trials (OR 1.09, 95% CI 0.79 to 1.50; participants = 1853; studies = 2; I2 = 0%; Analysis 8.1). Whilst a third trial also reported information on repetition of SH by the post‐intervention assessment, there were major discrepancies for this outcome that we were unable to clarify with trial authors (Naidoo 2014). According to GRADE criteria, we judged the evidence to be of very low certainty.
As randomisation was conducted locally for each country included in the SUPRE‐MISS study, we were also able to include results for each of the five countries separately. Although there was no difference between groups for the individual sites in Campinas, Brazil (21/71 versus 10/64; OR 2.27, 95% CI 0.97 to 5.28; N = 135; k = 1; I² = not applicable), Colombo, Sri Lanka (3/130 versus 5/121; OR 0.55, 95% CI 0.13 to 2.34; N = 251; k = 1; I² = not applicable), Karaj, Iran (33/303 versus 28/298; OR 1.18, 95% CI 0.69 to 2.00; N = 601; k = 1; I² = not applicable), and Yuncheng, China (1/58 versus 0/38; OR 2.01, 95% CI 0.08 to 50.60; N = 96; k = 1; I² = not applicable), fewer participants in the intervention group had repeated SH by the 18‐month period at the Chennai, India site (8/301 versus 17/260; OR 0.39, 95% CI 0.17 to 0.92; N = 561; k = 1; I² = not applicable).
One secondary publication for the SUPRE‐MISS study also reported data on frequency of SH for one site (i.e. Karaj, Iran) by the six‐month follow‐up assessment (Hassanzadeh 2010). At this site, there was evidence of a increase in frequency of SH repetition in the intervention and support arm relative to the TAU arm (mean 1.63, SD 1.19, n = 319 versus mean 1.17, SD 0.38, n = 310; MD 0.46, 95% CI 0.32 to 0.60; N = 629; k = 1; I² = not applicable).
Secondary outcomes
10.2 Treatment adherence
No data available.
10.3 Depression
Correspondence with trial authors revealed that information on depression was recorded at one site only: Yuncheng, China (Xu 2012). There was evidence of an effect for information and support on depression scores at this site by the post‐intervention assessment (mean 0.64, SD 0.29, n = 57 versus mean 0.52, SD 0.30, n = 54; MD 0.12, 95% CI 0.01 to 0.23; N = 111; k = 1; I² = not applicable).
10.4 Hopelessness
No data available.
10.5 General functioning
No data available.
10.6 Social functioning
No data available.
10.7 Suicidal ideation
No data available.
10.8 Suicide
There was evidence of an effect for information and support on suicide deaths by the 18‐month post‐intervention assessment (OR 0.12, 95% CI 0.03 to 0.44; N = 1889; k = 2; I² = 0%; Analysis 8.2).
For the SUPRE‐MISS study, data on suicide deaths were also available for three sites in related publications (i.e. Chennai, India [Vijayakumar 2011], Karaj, Iran [Hassanzadeh 2010], and Yuncheng, China [Xu 2012]). There was evidence of an effect for information and support on suicide deaths by the 18‐month assessment at the Chennai, India site (1/302 versus 9/320; OR 0.11, 95% CI 0.01 to 0.91; N = 622; k = 1; I² = not applicable), but not at the Karaj, Iran (2/319 versus 2/310; OR 0.97, 95% CI 0.14 to 6.94; N = 629; k = 1; I² = not applicable), or the Yuncheng, China (0/57 versus 2/54; OR 0.18, 95% CI 0.01 to 3.89; N = 111; k = 1; I² = not applicable) sites. Notably, the number of completed suicides in the intervention group reported for these three subsamples was greater than the number reported for the overall SUPRE‐MISS cohort in the primary trial (i.e. Fleischmann 2008). We were unable to confirm the correct number of completed suicides in the intervention group with the authors. Including the one additional suicide for the intervention group identified from the three subsample publications with the data reported in the primary study reference, however, did not materially affect the result obtained for the overall SUPRE‐MISS cohort.
Subgroup analyses
For the SUPRE‐MISS study, data on repetition of SH were also available for males and females separately (Fleischmann 2008). Overall, across all five sites in this trial, there was no evidence of an effect on repetition of SH by the post‐intervention assessment in either males (30/349 versus 27/340; OR 1.09, 95% CI 0.63 to 1.88; N = 689; k = 1; I² = not applicable) or females (36/514 versus 33/460; OR 0.97, 95% CI 0.60 to 1.59; N = 974; k = 1; I² = not applicable). Breaking these results down by sex for the individual sites involved in the SUPRE‐MISS study revealed no effect for information and support on repetition of SH by the 18‐month assessment for either sex at any of the five study sites: Campinas, Brazil (males: 4/21 versus 3/25; OR 1.73, 95% CI 0.34 to 8.76; N = 46; k = 1; I² = not applicable; females: 17/50 versus 7/39; OR 2.35, 95% CI 0.86 to 6.44; N = 89; k = 1; I² = not applicable), Chennai, India (males: 5/148 versus 7/125; OR 0.59, 95% CI 0.18 to 1.91; N = 273; ; k = 1; I² = not applicable; females: 3/153 versus 10/153; OR 0.29, 95% CI 0.08 to 1.06; N = 306; k = 1; I² = not applicable), Colombo, Sri Lanka (males: 1/54 versus 3/53; OR 0.31, 95% CI 0.03 to 3.12; N = 107; k = 1; I² = not applicable; females: 2/76 versus 2/68; OR 0.89, 95% CI 0.12 to 6.51; N = 144; k = 1; I² = not applicable), Karaj, Iran (males 19/109 versus 14/118; OR 1.57, 95% CI 0.74 to 3.31; N = 227; k = 1; I² = not applicable; females: 14/194 versus 14/180; OR 0.92, 95% CI 0.43 to 1.99; N = 374; k = 1; I² = not applicable), Yuncheng, China (males: 1/17 versus 0/19; OR 3.55, 95% CI 0.14 to 93.01; N = 36; k = 1; I² = not applicable; females 0/41 versus 0/38; OR not estimable, 95% CI not estimable; N = 79; k = 1; I² = not applicable).
For the START study, data on repetition of SH were reported separately for those with a history of SH prior to trial entry; however, there was no evidence of an effect for information and support on repetition of SH by the post‐intervention assessment in this group (7/14 versus 9/17; OR 0.89, 95% CI 0.22 to 3.66; N = 31; k = 1; I² = not applicable; Amadéo 2015).
Sensitivity analyses
Not applicable.
Comparison 11: Other multimodal interventions versus TAU or another comparator
Three Zelen RCTs investigated the effectiveness of a package of interventions, including CBT‐based psychotherapy, remote contact interventions (e.g. telephone contact, letters, and/or postcards), and GP vouchers in adults (weighted mean age: 35.5 ± 12.5; 54.7% female) admitted to emergency departments following an episode of SH (Gysin‐Maillart 2016, N = 120; Hatcher 2015, N = 1474; Hatcher 2016, N = 365). For one of these trials, the treatment package was culturally adapted for those who identified as of Māori ethnicity (Hatcher 2016).
Primary outcome
11.1 Repetition of SH
There was no evidence of an effect for this package of interventions at post‐intervention in terms of repetition of SH (Analysis 9.1). Using data for the consenting sample (i.e. including only those participants who, following treatment allocation, subsequently consented to participation), rather than all those randomised, for Hatcher 2015 and Hatcher 2016, did not materially affect these results. According to GRADE criteria, we judged the evidence to be of very low certainty.
One trial reported data on frequency of SH at post‐intervention. There was evidence of an effect for a multimodel package of interventions on frequency of SH in this trial (mean 0.08, SD 0.28, n = 60 versus mean 0.82, SD 1.89, n = 50; MD ‐0.74, 95% CI ‐1.27 to ‐0.21; N = 110; k = 1; I² = not applicable; Gysin‐Maillart 2016).
With regard to time to SH repetition, there was no evidence of a significant treatment effect for this package of interventions in two trials (Analysis 9.2). Information on time to first SH presentation was also reported in Hatcher 2015; however, "...there was no significant difference between the two groups in time to first re‐presentation (log‐rank test P = 0.6564)" (Hatcher 2015, pg.232). As hazard ratios (HRs) and their accompanying 95% CIs were not reported in this trial, we were unable to include this result in our review.
Secondary outcomes
11.2 Treatment adherence
No data available.
11.3 Depression
There was no effect for this package of interventions on depression scores at the post‐intervention assessment (Analysis 9.3).
11.4 Hopelessness
There was also no apparent effect for this package of interventions on hopelessness scores at the post‐intervention assessment (Analysis 9.4).
However, the trial authors for one of these trials noted that "whilst there was a greater change in hopelessness scores at...12 months [i.e. the post‐intervention assessment] in the intervention group, the control group had a significantly lower baseline score [;]...because of the significant differences in baseline scores and missing follow up data we...used a mixed linear model to estimate the differences in scores at [the post‐intervention assessment]". Using this model, the trial authors found "there was a decrease in [h]opelessness scores in the treatment group compared to the usual care group but this was statistically non‐significant" (Hatcher 2016, pg.889).
11.5 General functioning
No data available.
11.6 Social functioning
No data available.
11.7 Suicidal ideation
One trial reported information on suicidal ideation scores (Gysin‐Maillart 2016); however, there was no evidence of an effect for this package of interventions on suicidal ideation scores by the post‐intervention assessment in this trial (mean 4.80, SD 7.88, n = 35 versus mean 4.62, SD 6.48, n = 34; MD 0.18, 95% CI ‐3.22 to 3.58; N = 69; k = 1; I² = not applicable).
11.8 Suicide
There was no effect for this intervention package on suicides by the post‐intervention assessment in any trial (Analysis 9.5). For one of these trials (i.e. Hatcher 2015), one death in the intervention arm and one in the control arm were due to uncertain causes and had yet to be investigated by the Coroner. However, assuming these deaths to be attributable to suicide did not materially affect this result.
Subgroup analyses
Randomisation was stratified by history of SH ( i.e. first SH episode versus repeat SH episode) in two of these trials (Hatcher 2015; Hatcher 2016); however, there was no difference in repetition of SH at post‐intervention in either trial for either those without a history of multiple episodes of SH prior to the index episode (i.e. first SH episode) or for those with a history of multiple episodes of SH (i.e. repeat SH episode) (Analysis 9.1).
Sensitivity analyses
As all trials included in this comparison used Zelen's design, we were unable to undertake sensitivity analyses to investigate the impact use of this design may have had on the results observed.
Two analyses within this comparison were associated with substantial levels of heterogeneity (Analysis 9.1.1, I² = 81%; Analysis 9.2, I² = 89%); however, analyses did not indicate any individual study was associated with excessive influence for any of these outcomes.
Comparison 12: Other mixed interventions versus TAU or another comparator
A number of single trials investigated the effectiveness of other mixed types of interventions compared with TAU or another comparator, including: continuity of care by the same therapist (Torhorst 1987), interpersonal problem‐solving therapy (McLeavey 1994), behaviour therapy (Liberman 1981), intensive in‐ and out‐patient treatment (Van der Sande 1997), general hospital admission (Waterhouse 1990), intensive outpatient treatment (Allard 1992; Welu 1977), home‐based psychotherapy and telephone contact (Hawton 1981), and long‐term therapy (Torhorst 1988).
Comparison 12.1: Continuity of care by the same therapist
One trial investigated the effectiveness of continuing aftercare with the same therapist (defined as continued therapeutic contact with the original hospital therapist in an outpatient setting) versus changing to a different therapist (defined as receiving therapy in a specialised suicide prevention centre which involved changing both therapist and institution) over a 12‐month follow‐up period in adults (mean ± SD age not reported; 69.5% female) admitted to hospital following an episode of self‐poisoning (Torhorst 1987, N = 141).
Primary outcome
12.1.1 Repetition of SH
There was no evidence of an effect for this intervention on repetition of SH by the 12‐month follow‐up assessment in this trial (3/70 versus 9/66; OR 0.28, 95% CI 0.07 to 1.10; N = 136; k = 1; I² = not applicable; Torhorst 1987).
Secondary outcomes
12.1.2 Treatment adherence
There was evidence of an effect for this intervention on the proportion of participants who completed treatment compared to alternative psychotherapy (49/68 versus 36/73; OR 2.65, 95% CI 1.32 to 5.34; N = 141; k = 1; I² = not applicable; Torhorst 1987).
12.1.3 Depression
Depression scores did not differ between groups at the 12‐month follow‐up assessment in this trial (mean 6.20, SD 6.90, n = 65 versus mean 7.60, SD 9.20, n = 62; MD ‐1.40, 95% CI ‐4.24 to 1.44; N = 127; k = 1; I² = not applicable; Torhorst 1987).
12.1.4 Hopelessness
No data available.
12.1.5 General functioning
No data available.
12.1.6 Social functioning
No data available.
12.1.7 Suicidal ideation
No data available.
12.1.8 Suicide
There was no evidence of an effect for this intervention on the number of participants who died by suicide by the 12‐month follow‐up period (2/70 versus 3/66; OR 0.62, 95% CI 0.10 to 3.82; N = 136; k = 1; I² = not applicable; Torhorst 1987).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.2: Interpersonal problem‐solving therapy
One small trial compared the effectiveness of interpersonal problem‐solving skills training (IPSST) with alternative psychotherapy (i.e. brief problem‐oriented therapy) in adults (mean age: 23.9 ± 7.2 years; 74.3% female) admitted to accident and emergency facilities following an episode of self‐poisoning (McLeavey 1994, N = 39).
Primary outcome
12.2.1 Repetition of SH
There was no evidence of an effect for repetition of SH, defined as a 'self‐poisoning act', by the 12‐month follow‐up period (2/17 versus 4/16; OR 0.40, 95% CI 0.06 to 2.57; N = 33; k = 1; I² = not applicable) in this trial.
Secondary outcomes
12.2.2 Treatment adherence
There was no evidence of an effect on the proportion of participants who completed the full course of treatment in this trial (2/19 versus 3/20; OR = 0.67, 95% CI 0.10 to 4.51; N = 39; k = 1; I² = not applicable).
There was, however, evidence of an effect for interpersonal problem‐solving therapy in terms of the number of treatment sessions attended (mean 5.30, SD 0.48, n = 17 versus mean 4.20, SD 1.32, n = 16; MD 1.10, 95% CI 0.41 to 1.79; N = 33; k = 1; I² = not applicable).
12.2.3 Depression
No data available.
12.2.4 Hopelessness
There was no evidence of an effect for hopelessness by the six‐month follow‐up assessment in this trial (mean 6.12, SD 4.61, n = 19 versus mean 4.35, SD 4.39, n = 20; MD 1.77, 95% CI ‐1.06 to 4.60; N = 39; k = 1; I² = not applicable).
12.2.5 General functioning
No data available.
12.2.6 Social functioning
No data available.
12.2.7 Suicidal ideation
No data available.
12.2.8 Suicide
There were no suicide deaths in either the intervention or comparator arm by the 12‐month follow‐up period.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.3: Behaviour therapy
One small trial compared the effectiveness of behaviour therapy versus alternative psychotherapy (i.e. insight‐oriented therapy) in adults (mean age: 29.7 ± 8.8 years; 66.6% female) referred for inpatient treatment following a suicide attempt (Liberman 1981, N = 24).
Primary outcome
12.3.1 Repetition of SH
There was no evidence of an effect with regards to the proportion of participants repeating SH by the 24‐month follow‐up period (2/12 versus 3/12; OR 0.60, 95% CI 0.08 to 4.45; N = 24; k = 1; I² = not applicable).
Secondary outcomes
12.3.2 Treatment adherence
No data available.
12.3.3 Depression
Depression was measured using both the BDI and ZSRDS in one trial. There was evidence of an effect for behaviour therapy at the post‐intervention assessment according to both measures (BDI: mean 4.00, SD 4.00, n = 12 versus mean 14.00, SD 12.00, n = 12; MD ‐10.00, 95% CI ‐17.16 to ‐2.84; N = 24; k = 1; I² = not applicable; ZSRDS: mean 32.00, SD 8.00, n = 12 versus mean 43.00, SD 14.00, n = 12; MD ‐11.00, 95% CI ‐20.12 to ‐1.88; N = 24; k = 1; I² = not applicable).
By the six‐month follow‐up assessment, although there was no effect for behaviour therapy on depression according to the ZSRDS (mean 34.00, SD 8.00, n = 12 versus mean 41.00, SD 13.00, n = 12; MD ‐7.00, 95% CI ‐15.64 to 1.64; N = 24; k = 1; I² = not applicable), BDI scores did show an effect (mean 4.00, SD 6.00, n = 12 versus mean 13.00, SD 11.00, n = 12; MD ‐9.00, 95% CI ‐16.09 to ‐1.91; N = 24; k = 1; I² = not applicable).
12.3.4 Hopelessness
No data available.
12.3.5 General functioning
No data available.
12.3.6 Social functioning
No data available.
12.3.7 Suicidal ideation
There was no evidence of an effect for behaviour therapy on the proportion of participants reporting suicidal ideation by the 24‐month follow‐up assessment (5/12 versus 9/12; OR 0.24, 95% CI 0.04 to 1.36; N = 24; k = 1; I² = not applicable).
12.3.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.4: Intensive in‐ and outpatient treatment
One trial compared the effectiveness of brief psychiatric inpatient admission followed by regular outpatient appointments and 24‐hour access to the psychiatric unit with TAU over a 12‐month follow‐up period in adults (mean age: 36.3 ± 15.1 years; 57.7% female) admitted to a general hospital following a suicide attempt (Van der Sande 1997, N = 274).
Primary outcome
12.4.1 Repetition of SH
There was no evidence of an effect for intensive in‐ and outpatient treatment on repetition of SH by the 12‐month follow‐up period (24/140 versus 20/134; OR 1.18, 95% CI 0.62 to 2.25; N = 274; k = 1; I² = not applicable).
There was no evidence of an effect for intensive in‐ and outpatient treatment on frequency of repetition of SH (mean 0.23, SD 0.57, n = 140 versus mean 0.23, SD 0.81, n = 134; MD 0.00, 95% CI ‐0.17 to 0.17, N = 274; k = 1; I² = not applicable).
There was also no evidence of an effect for intensive in‐ and outpatient treatment on time to SH repetition in this trial (HR 1.24, 95% CI 0.68 to 2.27; N = 274; k = 1; I² = not applicable).
Secondary outcomes
12.4.2 Treatment adherence
There was no effect for intensive in‐ and outpatient treatment in the total number of treatment sessions attended (mean 14.30, SD 24.20, n = 140 versus mean 11.40, SD 27.70, n = 134; MD 2.90, 95% CI ‐3.27 to 9.07; N = 274; k = 1; I² = not applicable).
12.4.3 Depression
There was also no effect for intensive in‐ and outpatient treatment on depression scores by the 12‐month follow‐up assessment (mean 30.80, SD 15.90, n = 94 versus mean 35.80, SD 16.20, n = 50; MD ‐5.00, 95% CI ‐10.52 to 0.52; N = 144; k = 1; I² = not applicable).
12.4.4 Hopelessness
There was no effect for intensive in‐ and outpatient treatment on hopelessness scores by the 12‐month follow‐up assessment in this trial (mean 6.10, SD 5.00, n = 94 versus mean 7.50, SD 5.90, n = 50; MD ‐1.40, 95% CI ‐3.32 to 0.52; N = 144; k = 1; I² = not applicable).
12.4.5 General functioning
No data available.
12.4.6 Social functioning
No data available.
12.4.7 Suicidal ideation
No data available.
12.4.8 Suicide
There was also no evidence of an effect for suicide by the 12‐month follow‐up assessment in this trial (1/140 versus 2/134; OR 0.47, 95% CI 0.04 to 5.30; N = 274; k = 1; I² = not applicable).
Subgroup analyses
Whilst randomisation was stratified, data disaggregated by sex were not reported.
Sensitivity analyses
Not applicable.
Comparison 12.5: General hospital admission
One trial investigated the effectiveness of admission to a general hospital versus non‐admission over a four‐month follow‐up period in adults (mean age: 30.0 years, SD not reported; 62.3% female) presenting to an emergency room following an episode of self‐poisoning and who had no immediate medical or psychiatric treatment needs (Waterhouse 1990, N = 77). In this trial, admission was described as consisting of little more than admission to an inpatient bed. The trial authors did not attempt to influence referral to psychiatric or other treatment services. The median length of admission for those allocated to the experimental group was 17 hours.
Primary outcome
12.5.1 Repetition of SH
There was no evidence of an effect for hospital admission on repetition of SH at the post‐intervention assessment (2/38 versus 2/39; OR 1.03, 95% CI 0.14 to 7.69; N = 77; k = 1; I² = not applicable). According to the GRADE criteria, we judged the evidence to be of moderate certainty.
There was also no evidence of an effect for general hospital management on repetition of SH by the four‐month follow‐up assessment (3/38 versus 4/39; OR 0.75, 95%CI 0.16 to 3.60; N = 77; k = 1; I² = not applicable) in this trial.
Secondary outcomes
12.5.2 Treatment adherence
No data available.
12.5.3 Depression
No data available.
12.5.4 Hopelessness
The authors reported there was no difference in hopelessness scores at the post‐intervention assessment (mean 10.29, SD 5.68 versus mean 10.21, SD 4.97); however, as they did not provide the numbers of patients in each group, we were unable to calculate the MD and its associated 95% CI. No data on hopelessness scores by the follow‐up assessment were reported.
12.5.5 General functioning
No data available.
12.5.6 Social functioning
There was no evidence of an effect for hospital admission on social functioning by the four‐month assessment in this trial (mean 0.41, SD 0.44, n = 16 versus mean 0.51, SD 0.40, n = 20; MD ‐0.10, 95% CI ‐0.38 to 0.18; N = 36; k = 1; I² = not applicable).
12.5.7 Suicidal ideation
There was no evidence of an effect for hospital admission on suicidal ideation scores by the four‐month follow‐up assessment (mean 0.22, SD 0.85, n = 27 versus mean 0.04, SD 0.20, n = 25; MD 0.18, 95% CI ‐0.15 to 0.51; N = 52; k = 1; I² = not applicable).
12.5.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.6: Intensive outpatient treatment
Two trials compared the effectiveness of intensive outreach interventions with standard outpatient care in adults (weighted mean age: 29.0 years, SD not reported; 55.3% female) admitted to emergency departments following a suicide attempt (Allard 1992, N = 150; Welu 1977, N = 119). The first compared an intensive intervention, involving psychiatrists and a social worker, a schedule of visits including at least one home visit, therapy provided where needed, reminders (telephone or written), and home visits with treatment by regular personnel in the same hospital over a 12‐month period (Allard 1992). In the second, a specialist, intensive outreach programme in which a community mental health team contacted participants immediately after discharge and arranged home visits and weekly or bi‐weekly contact with therapists alongside routine psychiatric consultation was compared with TAU.
Primary outcome
12.6.1 Repetition of SH
There was no evidence of an effect for intensive outpatient treatment on repetition of SH by either the four‐month (3/62 versus 9/57; OR 0.27, 95% CI 0.07 to 1.06; N = 119; k = 1; I² = not applicable; Welu 1977) or 24‐month follow‐up assessment (22/63 versus 19/63; OR 1.24, 95% CI 0.59 to 2.62; N = 126; k = 1; I² = not applicable; Allard 1992).
In the one trial that included information on frequency of repeated SH by the 24‐month follow‐up assessment, the authors reported that "the experimental subjects did not make fewer attempts than the comparison subjects" (Allard 1992, pg.310).
Secondary outcomes
12.6.2 Treatment adherence
Data on treatment adherence were only available for one of these trials (Allard 1992); however, as insufficient information was reported to enable imputation of missing SDs in this trial, we were unable to calculate the mean difference in the number of treatment sessions attended. Nevertheless, the trial authors reported that "[t]he mean numbers of encounters with psychiatrists were 12.35 versus 1.54 (P < 0.001) in the first year and 2.11 versus 0.64 (P = 0.071) in the second year" (Allard 1992, pg.311).
12.6.3 Depression
No data available.
12.6.4 Hopelessness
No data available.
12.6.5 General functioning
For one of these trials, the authors found "[t]here was no difference...in scores on the...GAS" (Allard 1992, pg.311) by the 24‐month follow‐up period. However, numerical data for this outcome were not reported.
12.6.6 Social functioning
No data available.
12.6.7 Suicidal ideation
No data available.
12.6.8 Suicide
One trial reported data on suicide deaths (Allard 1992). There was no evidence of an effect for the intensive outpatient intervention on suicides by the 24‐month follow‐up assessment in this trial (3/76 versus 1/74; OR 3.00, 95% CI 0.30 to 29.52; N = 150; k = 1; I² = not applicable).
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.7: Home‐based psychotherapy and telephone contact
One trial investigated the effectiveness of problem‐oriented counselling delivered in two different ways, namely as a flexibly‐timed home‐based therapy, combined with open access via telephone services to the general psychiatric service, versus treatment in weekly outpatient clinics, in adults (mean age: 25.2 ± 8.2 years; 69.8% female) referred to the psychiatric department of a general hospital following admission for self‐poisoning (irrespective of suicidal intent) (Hawton 1981, N = 96).
Primary outcome
12.7.1 Repetition of SH
There was no evidence of an effect for home‐based psychotherapy and telephone contact on repetition of SH by the 12‐month follow‐up assessment in this trial (5/48 versus 7/48; OR 0.68, 95% CI 0.20 to 2.32; N = 96; k = 1; I² = not applicable).
Secondary outcomes
12.7.2 Treatment adherence
There was, however, evidence of an effect for home‐based psychotherapy and telephone contact on the proportion of participants who completed treatment over the course of the 12‐month follow‐up period (40/48 versus 20/48; OR 7.00, 95% CI 2.70 to 18.13; N = 96; k = 1; I² = not applicable).
12.7.3 Depression
Although this trial included data on depression, the trial authors modified the scale used. We were therefore unable to include data on this outcome in this review.
12.7.4 Hopelessness
No data available.
12.7.5 General functioning
No data available.
12.7.6 Social functioning
Although this trial included data on social functioning, the trial authors modified the scale used. We were therefore unable to include data on this outcome in this review.
12.7.7 Suicidal ideation
Data obtained by correspondence suggested there was no effect for suicidal ideation at either the post‐treatment assessment (Mann‐Whitney U = 984, P = 0.29) or six‐month follow‐up assessment (Mann‐Whitney U = 726, P = 0.14). As only median, rather than mean, scores were available for this outcome, we were unable to reproduce the mean difference in suicidal ideation scores between the experimental and control groups in this review.
12.7.8 Suicide
No data available.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Comparison 12.8: Long‐term therapy
One trial investigated the effectiveness of long‐term (one session per month over 12 months) versus short‐term (one session per week over 12 weeks) outpatient psychotherapy on repetition of SH over a 12‐month follow‐up period in adults (mean age: not reported, SD not reported; 100% female) admitted to hospital following repeated episodes of self‐poisoning (Torhorst 1988, N = 80). The content of therapy was not specified.
Primary outcome
12.8.1 Repetition of SH
There was no evidence of an effect for long‐term therapy on repetition of SH by the post‐intervention assessment in this trial (9/40 versus 9/40; OR 1.00, 95% CI 0.35 to 2.86; N = 80; k = 1; I² = not applicable). According to GRADE criteria, we judged the evidence to be of low certainty.
Secondary outcomes
12.8.2 Treatment adherence
The trial authors did not provide numerical data on treatment adherence, although they stated, "[a]ttendance at the first session was about equal for both groups (about 60%)" (Torhorst 1988, pg.420). However, they further stated that "participation of the 12‐month (long‐term therapy) group dropped drastically by the second session to under 40%, while the participation of the patients in the 3‐month (intensive short‐term therapy) program remained higher" (Torhorst 1988, pg.420). However, it was unclear whether this difference was statistically significant.
Overall, treatment adherence also appears to have been very low in both groups as the "average number of sessions was 3.9 (out of a possible 12 sessions) in the three‐month group and 2.6 (out of a possible 12 sessions) for the 12 month group" (Torhorst 1988, pg.420). Again, however, it was unclear whether this difference was statistically significant as insufficient information was reported to enable imputation of missing SDs in this trial.
12.8.3 Depression
Although numerical data on depression scores was not reported, means estimated by the review authors from a graph in the original report suggested there was little difference in depression scores between those allocated to long‐term therapy and those allocated to short‐term therapy by the 12‐month follow‐up assessment (estimated means 9.3 versus 6.7). However, insufficient information was reported to enable imputation of missing SDs for this outcome.
The trial authors, however, stated that "self‐evaluated depressivity...improved considerably more for the patients of the three‐month program than for those of the 12‐month program" (Torhorst 1988, pg.421). This improvement was described by the trial authors as statistically significant.
12.8.4 Hopelessness
No data provided.
12.8.5 General functioning
No data provided.
12.8.6 Social functioning
Whilst "[p]atients in the 3‐month group evaluated themselves according to the Social Adjustment Scale (SAS) significantly more positively...than did the patients of the 12‐month group" (Torhorst 1988, pg.421), numerical data for this outcome were not reported.
12.8.7 Suicidal ideation
No data provided.
12.8.8 Suicide
No data provided.
Subgroup analyses
No included trial stratified randomisation by sex or repeater status.
Sensitivity analyses
Not applicable.
Discussion
This review included 76 trials, 21 (27.6%) of which have been completed since the previous version of this review (Hawton 2016). Previously, we commented that the promising results for CBT‐based psychological therapy and DBT warranted further investigation of these approaches to understand which patients benefit from these types of interventions. There were only a few, generally small, trials of most other forms of psychosocial therapy, providing little evidence of beneficial effects; however, these could not be ruled out.
Summary of main results
The trials included in this review investigated the effectiveness of various forms of psychosocial interventions for SH in adults.
CBT‐based psychotherapy
Individual‐based CBT‐based psychotherapy
Whilst 20 trials investigated the effectiveness of individual‐based cognitive‐behavioural therapy (CBT)‐based psychotherapy, only four (20.0%) reported data on the primary outcome of this review (i.e., repetition of SH at post‐intervention). Although there was imprecision in the effect estimate, on the basis of data from four of these trials, individual‐based CBT‐based psychotherapy may reduce repetition of SH as compared to TAU or another comparator by the end of the intervention. At longer follow‐up time points, there was more robust evidence of effect for this intervention. However, there was no apparent effect for CBT‐based psychotherapy on the frequency of SH, or time to SH repetition.
Few trials reported information on secondary outcomes. CBT‐based psychotherapy was associated with a relatively consistent but small beneficial effect on depression, hopelessness, and suicidal ideation scores as compared with TAU or another comparator over time. There was no apparent effect for this intervention on suicide, however.
Group‐based CBT‐based psychotherapy
On the basis of data from a single trial in which CBT‐based psychotherapy was delivered in a group‐based format, there is probably little to no effect of group‐based CBT‐based psychotherapy on repetition of SH as compared to TAU by the end of the intervention. There was also no evidence of effect for this intervention approach at longer follow‐up assessment points.
There was also no apparent effect of this intervention approach on any of the secondary outcome measures assessed by this review.
Dialectical behaviour therapy (DBT)
Standard DBT
On the basis of data from six trials, the evidence remains uncertain as to whether standard DBT has any effect on absolute repetition of SH by the post‐intervention assessment compared to either TAU or alternative psychotherapy in terms of the proportion of participants engaging in repeat SH by the post‐intervention or 12‐month follow‐up assessments. However, standard DBT may be associated with an effect for frequency of repeated SH by the post‐intervention assessment.
Whilst there was an apparent effect for standard DBT on depression and suicidal ideation scores compared with TAU or another comparator at post‐intervention, these effects were generally no longer apparent at longer follow‐up time points; however, few trials investigated these secondary outcomes. There was no apparent effect for standard DBT on suicide.
DBT group‐based skills training
Based on one trial in which two variants of DBT were compared to standard DBT, there is probably little to no effect of DBT group‐based skills training on absolute or frequency of suicide reattempts or NSSI (Linehan 2015a). For this trial, we were unable to obtain data on combined SH from the trial authors despite correspondence. There was also no apparent effect for DBT group‐based skills training only on any of the secondary outcomes investigated in this trial.
DBT individual therapy
There is also probably little to no effect for DBT individual skills training only on absolute repetition or frequency of suicide reattempts or NSSI as compared to standard DBT, based on findings of the single trial which included this comparison (Linehan 2015b). Again, however, we were unable to obtain data on combined SH from the trial authors despite correspondence. There was also no apparent effect for DBT individual skills training only on any of the secondary outcomes investigated in this trial.
DBT prolonged‐exposure protocol
On the basis of data from a single trial, there is probably little to no effect of an experimental form of DBT in which participants were given significantly longer cognitive exposure to stressful events coupled with the standard DBT protocol on repetition of SH (Harned 2014). There was also no apparent effect for the DBT prolonged exposure intervention on frequency of SH repetition, treatment adherence, depression scores, or suicide.
Mentalisation‐based therapy (MBT)
On the basis of data from a single trial, mentalisation‐based therapy (MBT) for adults referred to a specialist personality disorder treatment service reduces both absolute repetition of SH and frequency of SH by the post‐intervention assessment as compared with TAU (Bateman 2009).
There was also an apparent effect for MBT on depression scores, general functioning scores, and social functioning scores by this time point. However, there was no apparent effect for this intervention on treatment adherence. No participants died by suicide in either trial arm by the post‐intervention assessment.
Emotion regulation psychotherapy
On the basis of data from two trials conducted by the same research group, emotion‐regulation psychotherapy for women diagnosed with borderline personality disorder provided in a group‐based setting probably reduces absolute repetition of SH by the post‐intervention assessment (Gratz 2006; Gratz 2014). There was no apparent effect for this intervention on frequency of SH repetition. However, correspondence with authors suggested that this treatment did not require participants to abstain from SH. Instead, participants were encouraged to work on resisting urges to engage in SH and, when SH occurred, to learn responses to it.
There was also evidence of an apparent effect for group‐based emotion regulation psychotherapy on depression scores by the post‐intervention assessment. No participants died by suicide in either trial arm by the post‐intervention assessment.
Psychodynamic psychotherapy
Whilst two trials investigated the effectiveness of psychodynamic psychotherapeutic approaches as compared to TAU (Andreoli 2015; Sahin 2018), we were only able to include data from the first of these trials despite attempts to obtain data for the second trial by correspondence. Therefore, based on data from one trial, there is probably little to no effect of psychodynamic psychotherapy on repetition of SH by the post‐intervention assessment. However, there was an apparent effect of this intervention on time to SH repetition.
There was also evidence of an apparent effect for this intervention on treatment adherence, depression scores, general functioning scores, and frequency of suicidal ideation episodes necessitating additional intensive psychiatric care by the post‐intervention assessment. There was no apparent effect on suicide deaths. However, these results were all based on a single trial.
Case management
Five trials investigated the effectiveness of case management as compared to either TAU (Clarke 2002; Hvid 2011; Van Heeringen 1995) or enhanced usual care (Kawanishi 2014; Morthorst 2012). On the basis of data from four of these trials, case management may have little to no effect on repetition of SH by the post‐intervention assessment. The impact of this intervention over longer term follow‐up is uncertain as only one of these trials investigated outcomes beyond the post‐intervention assessment time point. There were inconsistent findings for case management on time to SH repetition in two trials.
Findings from a single trial indicated that case management may be associated with an effect on the proportion of participants who completed treatment, which was based on attendance for outpatient clinic appointments, the patient sample having been identified before the trial as having failed to attend such appointments (Van Heeringen 1995). However, few studies reported data on other secondary outcomes. For one trial, although data on hopelessness scores at post‐intervention were collected, data on this outcome have not been reported in any secondary publications of this trial to date (Kawanishi 2014), indicating that publication bias may have been present for this outcome. There was no apparent effect for case management on suicide.
Structured general practitioner (GP) follow‐up
Based on results of a single trial of structured follow‐up by the participants' general practitioner (GP) compared with TAU in adults admitted to the acute ward of a general hospital following an episode of self‐poisoning (Grimholt 2015), this intervention may have little to no effect on repetition of SH by the post‐intervention assessment.
There was also no apparent effect for structured GP follow‐up on treatment adherence, depression scores, hopelessness scores, and suicidal ideation scores, although missing data for these secondary outcomes was substantial (i.e. 70.5% to 91.3%). Therefore, the evidence for these secondary outcomes is uncertain. There was also no apparent effect for this intervention on suicide.
Brief emergency department‐based interventions
Five trials investigated one‐off brief interventions delivered in the emergency department. Two were based, in a large part, on the Collaborative Assessment and Management of Suicidality (CAMS) approach (Jobes 2012), two were based on the Integrated Motivational‐Volitional theoretical model of suicidal behaviour (O'Connor 2011), and one focused on addressing alcohol misuse.
Brief Collaborative Assessment and Management of Suicidality (CAMS)‐based intervention
On the basis of data from a single trial, there is likely no effect of a CAMS‐based approach on repetition of SH by the 12‐month follow‐up assessment (O'Connor 2020). There was also no apparent effect for this intervention on frequency of SH repetition by this time point based on data estimated from graphics presented in the report of this trial.
Limited data were available on any of the secondary outcomes assessed in this review. Although trial authors reported that there were improvements in hopelessness and suicidal ideation scores by the six‐month assessment in the original trial of this intervention (O'Connor 2015), this effect was no longer apparent by the 12‐month follow‐up assessment in a subsequent trial of this intervention (O'Connor 2020).
Brief guided Integrated Motivational‐Volitional‐focused intervention
On the basis of data from a single trial, there is likely no effect of the Integrated Motivational‐Volitional model either on the proportion of participants who engaged in a repeat episode of SH, the frequency of SH repetition, or in the proportion dying by suicide, as compared with TAU (O'Connor 2017).
There was some evidence of an effect for this intervention on treatment adherence in this trial, however, there was also no indication that this approach led to a reduction in depression scores in a second trial (Armitage 2016a). There was also no evidence of an apparent effect for this intervention on suicide deaths.
Brief self‐guided Integrated Motivational‐Volitional‐focused intervention
The second intervention arm in Armitage 2016 investigated the effectiveness of a self‐guided version of the Integrated Motivational‐Volitional model as compared to TAU (i.e. Armitage 2016b). However, in this trial, data on repetition of non‐fatal SH could not be disaggregated from data on suicidal ideation, despite correspondence. Therefore, the evidence for these secondary outcomes is uncertain. There was no apparent effect for this intervention approach on depression scores by the post‐intervention assessment, however.
Brief alcohol‐focused intervention
Based on a single trial of a brief intervention for alcohol misuse in SH patients, there is likely no effect of this intervention on repetition of SH by the six‐month follow‐up assessment (Crawford 2010). However, only around half (47.1%) of those randomised to the brief alcohol‐focused intervention attended the treatment session.
No suicide deaths were observed amongst the group of participants for whom information was available at the six‐month follow‐up. However, the trial authors warned that, as they were unable to track participants via their National Health Service (NHS) identity numbers, they were unable to confirm numbers of suicides from national mortality data. Thus, there may have been suicide deaths amongst those participants whom the authors were unable to contact by the six‐month follow‐up assessment in this trial. The original report did, however, identify a non‐significant trend towards reduced alcohol consumption per drinking day in those allocated to the intervention arm.
Remote contact interventions
A number of trials investigated the effectiveness of a variety of remote contact interventions, including: emergency cards (Evans 1999a; Morgan 1993), coping cards (Wang 2016), general practitioners' (GP) letters (Bennewith 2002), postcards (Beautrais 2010; Carter 2005; Hassanian‐Moghaddam 2011; Kapur 2013), telephone contact (Cedereke 2002; Mousavi 2015; Vaiva 2006), telephone contact combined with emergency cards and letters (Mouaffak 2015; Vaiva 2018), and telephone‐based psychotherapy (Marasinghe 2012; Sreedaran 2020; Wei 2013).
Emergency cards
On the basis of data from two trials, emergency cards may have little to no effect on repetition of SH by the post‐intervention assessment (Evans 1999a; Morgan 1993). In the original report for one of these trials, however, a post hoc subgroup analysis indicated that receipt of the emergency card may be associated with an increased risk of repetition of SH in those with a history of multiple episodes of SH prior to the index episode (Evans 1999a).
There was also no apparent effect for this intervention on suicide deaths on the basis of data reported in one of these trials (Evans 1999a). No data on other secondary outcomes were reported in either of these two trials.
Coping cards
A single trial investigated the effectiveness of a coping card intervention, in which participants were encouraged to note down alternative activities they could engage in when feeling suicidal, resources they found helpful when seeking help, as well as the numbers of a 24‐hour crisis line and local medical services (Wang 2016). On the basis of data from this trial, there is probably little to no effect of coping cards on repetition of SH by the post‐intervention assessment. There may be an apparent effect for this intervention on time to SH repetition.
There was no apparent effect of coping cards on hopelessness scores at post‐intervention. However, whilst the authors measured depression and suicidal ideation scores at post‐intervention, we were unable to obtain data on these outcomes despite correspondence. Therefore, the evidence for these secondary outcomes is uncertain.
General practitioner (GP) letters
There is likely no effect of general practitioners (GP) letters sent to participants following their discharge from hospital after SH on repetition of SH by the six‐month follow‐up assessment in a single cluster‐RCT (Bennewith 2002). There was also no apparent effect for this intervention on time to SH repetition. These results were based on a single cluster‐RCT, which may have overestimated the effectiveness of this intervention. No data on any of the secondary outcomes were reported, however.
Postcards
Four trials investigated the effectiveness of postcards sent on a regular basis over a 12‐month period as compared to TAU (Beautrais 2010; Carter 2005; Hassanian‐Moghaddam 2011; Kapur 2013). On the basis of data from these four trials, the evidence remains uncertain as to whether postcards have any effect on repetition of SH by the post‐intervention assessment. The single largest trial of this intervention, however, did find fewer participants repeating SH in the experimental group (Hassanian‐Moghaddam 2011). This result is notable because the comparator used in this trial would have consisted of little more than discharge because of the paucity of psychiatric services in Iran as compared to Australia, New Zealand, and the UK, where these trials were conducted and which have well‐developed services. This raises the possibility that this type of intervention may be more effective in such settings. Additionally, the postcards used in this trial included religious and philosophical messages in addition to providing general support, which may also explain their apparent efficacy in reducing SH and suicidal behaviour in this setting.
There was no apparent effect of postcards on frequency of SH repetition by this time point. It should be noted that the positive effect on frequency of repetition of SH as reported in one of these trials was, according to the trial's author, largely accounted for by difference in repetition in a small subsample (fewer than 3% of the total sample) of women with a history of three or more episodes of SH prior to trial entry (Carter 2005). There was also no apparent effect of this intervention on suicide deaths.
Telephone contact
On the basis of data from one trial at each time point, telephone contact may have little to no effect on repetition of SH by either the post‐intervention (Mousavi 2015), 12‐month (Cedereke 2002), or 24‐month (Vaiva 2006) follow‐up assessments. There was also no evidence of an effect for this intervention on frequency of SH repetition. There was also no apparent effect for telephone contact on suicide, or on any of the other secondary outcomes.
Telephone contact combined with emergency cards and letters
Two trials investigated the effect of combining telephone contact with emergency cards and letters for those unable to be contacted by telephone (Mouaffak 2015; Vaiva 2006). On the basis of data from the first of these trials, telephone contact combined with emergency cards and letters probably has little to no effect on repetition of SH by the post‐intervention assessment. However, data from the second trial of this intervention approach was excluded as information on non‐fatal SH repetition was combined with that for suicide deaths and we were unable to obtain disaggregated data from the trial authors in time for publication. There was also no apparent effect for this intervention on frequency of SH repetition, time to SH repetition, or on suicide. Neither trial investigated other secondary outcomes, including depression, general or social functioning, or suicidal ideation.
Telephone‐based psychotherapy
Three trials investigated the effectiveness of psychotherapy delivered via telephone (Marasinghe 2012; Sreedaran 2020; Wei 2013). In all three, the intervention approach incorporated principles of CBT‐based psychotherapy, third‐wave techniques, and social support. On the basis of data from two of these trials, telephone‐based psychotherapy may have little to no effect on repetition of SH by the post‐intervention assessment, or by the six‐ or 12‐month follow‐up assessments based on results in one of these trials. There was no apparent effect for telephone‐based psychotherapy on treatment adherence, depression, suicidal ideation, or suicide.
Provision of information and support
On the basis of data from two multicentre and multi‐country trials, one of the START model (Amadéo 2015) and a second of the SUPRE‐MISS model (Fleischmann 2008), in which participants received a hospital‐based information service combined with regular home support, telephone support or both, the evidence remains uncertain as to whether this intervention has any effect on repetition of SH by the post‐intervention assessment. We had to exclude one further trial of this approach from our analyses as there were major discrepancies in the data reported which we were unable to clarify with the trial authors despite correspondence (Naidoo 2014).
There was an apparent effect for this intervention approach on suicide deaths. However, we have noted that there is a discrepancy between findings for the overall cohort for one of these trials (i.e. Fleischmann 2008) and those reported for the individual sites (i.e. Hassanzadeh 2010; Vijayakumar 2011; Xu 2012), in that the number of suicides reported for the overall cohort for the intervention arm was fewer than that reported in the three local site reports (i.e. two versus three respectively). Data on secondary outcomes were only reported for the individual trial sites, rather than for the overall cohort for these trials.
Other multimodal interventions
Three Zelen RCTs investigated the effectiveness of a package of interventions, including CBT‐based psychotherapy, remote contact interventions (e.g. telephone contact, letters, and/or postcards), and GP vouchers in adults admitted to emergency departments following an episode of SH (Gysin‐Maillart 2016; Hatcher 2015; Hatcher 2016). For one of these trials, the treatment package was culturally adapted for those who identified as being of Māori ethnicity (Hatcher 2016). The evidence remains uncertain as to whether this package of interventions has any effect on repetition of SH by the post‐intervention assessment, or on time to SH repetition. There was an apparent effect on frequency of SH repetition in one of these trials (Gysin‐Maillart 2016).
There was no evidence of an apparent effect for this package of interventions on depression, hopelessness, or suicidal ideation scores by the post‐intervention assessment. There was also no evidence of an apparent effect for this package of interventions on suicide by the post‐intervention assessment.
Other mixed interventions
Continuity of care by the same therapist
Based on a single trial, there is likely no effect of continuity of care by the same therapist compared to receiving therapeutic contact from a different therapist on repetition of SH by the 12‐month assessment (Torhorst 1987). There was an apparent effect for this intervention on treatment adherence, but no such effect on either depression scores or suicide.
Interpersonal problem‐solving therapy
Based on a single trial, there is likely no effect of interpersonal problem‐solving skills training (IPSST) compared to alternative psychotherapy (i.e. brief problem‐oriented psychotherapy) on repetition of SH, hopelessness scores, and suicide by the 12‐month follow‐up assessment (McLeavey 1994). There was also no overall effect for this intervention on treatment adherence, although those assigned to the IPSST group did attend a greater number of treatment sessions.
Behaviour therapy
Based on a single trial, there is likely no effect of behaviour therapy compared to alternative psychotherapy (i.e. insight‐oriented therapy) on repetition of SH by the 24‐month follow‐up assessment, although there was an apparent effect for behaviour therapy on depression scores as compared to alternative psychotherapy, at least in the short term (i.e. post‐intervention assessment) (Liberman 1981). Behaviour therapy was also associated with mixed findings with respect to suicidal ideation scores in this trial.
Intensive in‐ and outpatient treatment
Based on a single trial, there is likely no effect of brief inpatient psychiatric admission admission followed by regular outpatient appointments plus 24‐hour access to a treatment service (i.e. intensive in‐ and outpatient treatment) compared to TAU on repetition of SH, frequency of SH repetition, or on time to SH repetition by the 12‐month follow‐up assessment (Van der Sande 1997). There was also no apparent effect for this intervention on treatment adherence, depression scores, hopelessness scores, or suicide.
General hospital admission
Based on a single trial involving patients who would normally have been admitted to a general hospital following SH, there is probably little to no effect of general hospital admission compared with discharge from the hospital on repetition of SH by either the post‐intervention or the four‐month follow‐up assessment (Waterhouse 1990). There was also no apparent effect for this intervention on social functioning or suicidal ideation scores by either time point. However, as trial entry was limited to low‐risk participants, only around 15% of presenting patients were eligible for inclusion in this trial.
Intensive outpatient treatment
Based on two trials that compared a combination of intensive therapies, including psychotherapy, behaviour therapy, and family therapy versus standard outpatient care, there is likely no effect of intensive outpatient treatment on repetition of SH by either the four‐month (Welu 1977) or 24‐month (Allard 1992) follow‐up assessments. There is likely no effect of this intervention on treatment adherence, general functioning, or suicide deaths; although secondary outcomes were only investigated in one of these two trials (Allard 1992).
Home‐based psychotherapy and telephone contact
Based on a single trial, there is likely no effect of home‐based psychotherapy on repetition of SH by the 12‐month follow‐up assessment (Hawton 1981). Whilst there was an apparent effect for this intervention on treatment adherence, there was no such effect on suicidal ideation scores. As modified scales were used to assess depression and social functioning, however, we were unable to include data for these secondary outcomes in our review.
Long‐term therapy
Based on a single trial, long‐term therapy (the therapeutic content and focus of this approach was not specified) may have little to no effect on repetition of SH by the post‐intervention assessment compared with short‐term intensive therapy (Torhorst 1988). Overall treatment adherence was fairly low in both groups, however. Estimates of scores from graphs also suggests there was no apparent effect for long‐term therapy on depression scores.
Overall completeness and applicability of evidence
Completeness of evidence
Where it was unclear whether a trial satisfied our inclusion criteria, we contacted corresponding trial authors for clarification. We also contacted corresponding authors where data were either not clearly reported, or where we required data reported in a different format to allow for their inclusion in a meta‐analysis. However, despite engaging in over 200 emails with corresponding authors we were not always able to obtain all relevant data. This was due to a combination of non‐response to our enquiries and to authors being unable to access relevant data, often due to moving on to later positions or as a result of working from home due to COVID‐19 pandemic quarantine orders. This is a common problem in meta‐analyses (Selph 2014).
There are some concerns with regards to completeness of evidence for two large multicentre and multi‐country trials of information and support. For the START Study, although the trial was originally scheduled to be conducted in 13 countries (De Leo 2013), data from only one of these sites (i.e. French Polynesia; Amadéo 2015) has been published to date. Additionally, while the SUPRE‐MISS trial was scheduled to be conducted in 18 sites (Fleischmann 2002), data from only five of these were reported in Fleischmann 2008 and a related publication (Bertolote 2010). Secondary data from three sites (i.e. China [Xu 2012], Iran [Hassanzadeh 2010], and India [Vijayakumar 2011]) have also been published. However, data for the remaining sites involved in this trial have not been published to date.
Presence of publication bias could only be formally evaluated for three meta‐analyses for CBT‐based psychotherapy with respect to repetition of SH at six months (Figure 4) and 12 months (Figure 5), and for suicide deaths at final follow‐up (Figure 6). In all cases, some funnel plot asymmetry was apparent and seemed to affect the right side of the plot suggesting that there may be unpublished trials in which the intervention was found to be ineffective. However, as the remaining meta‐analyses in this review included fewer than 10 trials, we were unable to investigate publication bias for these analyses.
Funnel plot: CBT‐based psychotherapy on repetition of SH at six‐months.
Funnel plot: CBT‐based psychotherapy on repetition of SH at 12‐months.
Funnel plot: CBT‐based psychotherapy on suicide deaths by final follow‐up.
Whilst all but two of the included trials reported information on repetition of SH, publication bias may have been more common for the secondary outcomes assessed by this review. However, formal testing of publication bias was not possible due to the small number of trials. Therefore, we cannot rule out the possibility that publication bias may have affected the results of the other analyses in this review. This is a problem that commonly affects clinical data (Easterbrook 1991).
Applicability of evidence
The majority of participants in these trials were female, reflecting the typical pattern in hospital‐presenting populations (Hawton 2008). Only four trials stratified randomisation by sex (Hvid 2011; McAuliffe 2014; Van der Sande 1997; Walton 2020) whilst data were analysed separately by sex in six further trials (Bennewith 2002; Carter 2005; Fleischmann 2008; Hassanian‐Moghaddam 2011; Kapur 2013; Marasinghe 2012). Trials where sex was reported largely indicated benefits for females, but not males, in line with a previous review (Krysinska 2016). However, as these trials generally included fewer males, it may be that subgroup results for males were underpowered. Given that there are some differences in the motives for SH in males as compared to females (Claes 2007), further work on the treatment needs and preferences of males who engage in SH, as well as their experiences of clinical services, and how these may differ from females who engage in SH, is warranted. Further, a binary approach to gender was evident in the majority of trials. There was therefore poor attention to, and reporting of, outcomes for those with a non‐binary gender identity. There is good evidence that there are higher risks of self‐harm in those who are gender diverse (Newcomb 2020).
The majority of trials included either patients who have all engaged in intentional drug overdoses or self‐poisoning, or samples where the majority had, again reflecting the typical pattern observed in patients who present to general hospitals following SH (Hawton 2007). However, there are other important patient subgroups, such as those who engage in self‐cutting, who may have different treatment needs (Hawton 2004). None of the trials included in this review specifically focused on these patients; although it should be noted that method switching is common in those who engage in repeat episodes of SH (Witt 2019). In some trials there was a lack of information relating to the method of SH. Seventeen (22.4%) trials focused on those with a history of repeated SH, which is a particular issue in this clinical population, given the association of individuals with a history of repeat episodes having a greater risk of suicide (Zahl 2004). However, only 14 investigated impacts of psychosocial interventions for those with an initial episode of SH versus those engaging in repeated SH (Amadéo 2015; Bennewith 2002; Carter 2005; Evans 1999a; Gratz 2014; Hatcher 2011; Hatcher 2015; Hatcher 2016; Hassanian‐Moghaddam 2011; Kapur 2013; Lin 2020; McAuliffe 2014; Morthorst 2012; Mouaffak 2015). Outcomes were generally not stratified by recruitment setting (e.g. emergency department, community, and/or GP‐based referrals). This could be a focus of future trials and reviews. Secondary outcomes were also infrequently assessed which limits interpretation of possible mechanisms of effect.
Compared to previous versions of this review (Hawton 2016), there is now greater representation of trials from low‐to‐middle‐income countries, including: Iran (Hassanian‐Moghaddam 2011; Mousavi 2015; Mousavi 2017), Taiwan (Lin 2020; Wang 2016), China (Wei 2013), India (Sreedaran 2020), Malaysia (Armitage 2016), Pakistan (Husain 2014), South Africa (Naidoo 2014), and Sri Lanka (Marasinghe 2012). Additionally, two multicentre and multi‐country trials were conducted in a number of countries, including several low‐to‐middle‐income countries (e.g. Brazil, China, India, and Iran (Fleischmann 2008), and French Polynesia (Amadéo 2015)).
This review focused exclusively on those who engaged in SH. As a result, we have excluded trials in which participants were diagnosed with conditions such as borderline personality disorder but where SH was not required for trial entry. We also excluded trials in which participants engaged in repetitive self‐injurious behaviour in the context of an intellectual disability or developmental disorder (e.g. an autism spectrum disorder). Readers interested in the use of psychosocial interventions for these patient groups are instead referred to the relevant reviews (Cristea 2017; Oliver 2010).
Quality of the evidence
Certainty of evidence, as assessed using the GRADE approach, was generally moderate to very low suggesting that further research is likely to have an important impact on our confidence in the estimates of treatment effectiveness, and may in fact change the estimates. This is particularly likely to affect results for those interventions that so far have only been assessed in single trials.
Additionally, using the Cochrane 'Risk of bias' tool version 2.0 (Sterne 2019), for almost all (85.5%) trials included in this review there were some concerns or a high risk of bias in relation to at least one aspect of trial design, with weaknesses most commonly observed with respect to selection of the reported result and measurement of the outcome.
For most trials, insufficient information was reported to determine whether data were analysed in accordance with a prespecified plan. In 2015, the International Committee of Medical Journal Editors (ICMJE) recommended all clinical trials should be preregistered in a public trials registry (Witt 2020a). Whilst around half (54.5%) of the 22 trials published subsequent to this recommendation were preregistered, in some cases there was insufficient detail provided within the clinical trial register to determine how key outcome(s) were defined. This made it difficult to determine whether there had been any substantive changes to the proposed analysis plan and, if so, the reasons for any such departures. Future trials should provide sufficient detail within the clinical trial register to determine how key outcome(s) are defined and measured to aid in the determination as to whether there has been any substantive changes to the proposed analysis plan, and if so, the reasons for any such departures.
There were also some concerns relating to bias in the measurement of the outcome. This was typically because repetition of SH was based on self‐reported information. Given that around two‐thirds of SH recorded in medical and clinical records are not reported by participants, prevalence estimates derived from self‐reported information alone may underestimate the true rate of SH (Mitchell 2016). By supplementing data on self‐reported SH with information from clinical or medical records, future trials could compare results based on self‐reported information with those obtained from objective sources to investigate what impact, if any, this bias may have had on the estimate of treatment effectiveness.
Additionally, participants and clinical personnel were, typically, not blind to allocation owing to likely differences in treatment intensity between the intervention and control arms (Witt 2020a). Indeed, due to safety considerations, unblinding may be unavoidable in these types of trials. However, given that repetition of SH was based on self‐report in a number of the included trials (Amadéo 2015; Armitage 2016; Brown 2005; Davidson 2014; Fleischmann 2008; Gratz 2006; Gratz 2014; Harned 2014; Hassanian‐Moghaddam 2011; Husain 2014; Liberman 1981; Linehan 1991; Linehan 2006; Linehan 2015; McMain 2009; Mousavi 2015; Mousavi 2017; Naidoo 2014; O'Connor 2015; O'Connor 2020; Priebe 2012; Sahin 2018; Slee 2008; Tapolaa 2010; Torhorst 1987; Turner 2000; Walton 2020; Wei 2013; Weinberg 2006), this may introduce potential bias.
Lastly, the trials included in this review were, in general, relatively small to detect differences in proportions of patients who engage in a repeat episode of SH, although it is acknowledged that some of the trials were feasibility studies (Davidson 2014; Evans 1999b; Gibbons 1978; Harned 2014; Kapur 2013; O'Connor 2015; O'Connor 2020; Owens 2020; Sreedaran 2020). Whilst sample sizes have increased over time, most trials in this field are still underpowered. We have previously calculated that trials in this field may need to recruit up to a minimum of 1862 participants per arm to detect an effect for repetition of SH with 80% power at the conventional alpha level (Witt 2020a). Future trials should therefore supply a priori power calculations to justify their sample size.
Potential biases in the review process
We are confident that we have identified all relevant psychosocial interventions for SH in adults. However, we cannot rule out the possibility that some relevant outcome data may be missing from this review. Although data on repetition of SH were available for all but two of the included trials, limited data were available on secondary outcomes. Nevertheless, by using the random‐effects model in all analyses, our results possess greater generalisability than if we had used the fixed‐effect model (Erez 1996).
It is worth noting that the categorisation and grouping of interventions in this review may, in some cases, obscure the nature of the discrete interventions within a particular comparison. For example, a decision was made to group brief problem solving therapy (PST) with CBT into a single comparison we have referred to as 'CBT‐based psychotherapy'. As PST is usually a component of CBT, we have included trials where PST was the key therapeutic modality. It is important that those making decisions about the interventions examined in this review attend to the particular therapeutic components used in each trial included in any particular comparison (as we have detailed in the Characteristics of included studies table).
Agreements and disagreements with other studies or reviews
This review is an update of the 2016 Cochrane Review on psychosocial interventions for SH in adults (Hawton 2016). The previous review included 55 trials of seven different approaches, finding that CBT‐based psychological therapy may result in fewer individuals repeating SH whilst DBT may lead to a reduction in frequency of repeated SH. This update concurs with its previous iteration. There appears to be a probable beneficial effect for CBT‐based psychotherapy on repetition of SH as compared to TAU or another comparator immediately after the end of therapy, and this effect appears to be maintained over time. There also appears to be limited positive findings for DBT in terms of reduced frequency of SH.
We identified 18 further reviews that included a focus on psychosocial interventions for SH in adults that have been completed since the previous version of this review was published. Six were systematic reviews (D'Anci 2019; Hanratty 2019; Inagaki 2019; Krysinska 2017; Milner 2015; Tighe 2018), 10 included meta‐analysis (Bornheimer 2020; Briggs 2019; Calati 2016; DeCou 2019; Gøtzsche 2017; Hetrick 2016; Meerwijk 2016; Milner 2015; Padmanathan 2020; Riblet 2017), and two were narrative reviews (Ghanbari 2015; Pirkis 2020). The majority of these reviews included studies where participants had mixed clinical presentations (i.e. clinically significant levels of suicidal ideation and/or SH), and not all were limited to RCTs.
Of these reviews, those that have focused on specific intervention approaches, including acceptance and commitment therapy (ACT)‐based approaches (Tighe 2018), brief contact interventions (Ghanbari 2015; Milner 2015; Milner 2016), CBT (Gøtzsche 2017), the Collaborative Assessment and Management of Suicidality (CAMS) model (Hanratty 2019), DBT (DeCou 2019), psychoanalytic and psychodynamically‐oriented therapy (Briggs 2019), and provision of information and support (Pirkis 2020), have generally reported evidence of benefit for the approach examined; although two reviews of brief contact interventions (Milner 2015; Milner 2016) and one each of the ACT (Tighe 2018) and CAMS model (Hanratty 2019) indicated there was limited evidence for these approaches on repetition of SH, in line with findings from our review.
Of these reviews that have adopted a broader focus, whilst they generally indicated that delivering some form of psychotherapy is likely to be more effective than nothing (e.g. Bornheimer 2020; Calati 2016; Hetrick 2016; Inagaki 2019; Meerwijk 2016; Riblet 2017), these reviews have tended to statistically pool results from very different interventions together and so the results are largely meaningless for clinical practice as they provide little insight into which approach may be most beneficial for particular clinically relevant subgroups of patients.
Future reviews should undertake network meta‐analysis to investigate which component(s) of these typically multi‐component intervention approaches are most effective. Individual participant data meta‐analyses would also assist with the identification of clinically relevant subgroups of patients who may benefit from certain more intensive forms of intervention.
Summary of 'Risk of bias' assessments
Results of 'Risk of bias' assessments for each study
Funnel plot: CBT‐based psychotherapy on repetition of SH at six‐months.
Funnel plot: CBT‐based psychotherapy on repetition of SH at 12‐months.
Funnel plot: CBT‐based psychotherapy on suicide deaths by final follow‐up.
Comparison 1: CBT‐based psychotherapy, Outcome 1: Repetition of SH at post‐intervention
Comparison 1: CBT‐based psychotherapy, Outcome 2: Repetition of SH at six months
Comparison 1: CBT‐based psychotherapy, Outcome 3: Repetition of SH at 12 months
Comparison 1: CBT‐based psychotherapy, Outcome 4: Repetition of SH at 24 months
Comparison 1: CBT‐based psychotherapy, Outcome 5: Frequency of SH at post‐intervention
Comparison 1: CBT‐based psychotherapy, Outcome 6: Frequency of SH at six months
Comparison 1: CBT‐based psychotherapy, Outcome 7: Time to SH repetition
Comparison 1: CBT‐based psychotherapy, Outcome 8: Depression scores at post‐intervention
Comparison 1: CBT‐based psychotherapy, Outcome 9: Depression scores at six months
Comparison 1: CBT‐based psychotherapy, Outcome 10: Depression scores at 12 months
Comparison 1: CBT‐based psychotherapy, Outcome 11: Depression scores at 24 months
Comparison 1: CBT‐based psychotherapy, Outcome 12: Hopelessness scores at post‐intervention
Comparison 1: CBT‐based psychotherapy, Outcome 13: Hopelessness scores at six months
Comparison 1: CBT‐based psychotherapy, Outcome 14: Hopelessness scores at 12 months
Comparison 1: CBT‐based psychotherapy, Outcome 15: Suicidal ideation scores at post‐intervention
Comparison 1: CBT‐based psychotherapy, Outcome 16: Suicidal ideation scores at six months
Comparison 1: CBT‐based psychotherapy, Outcome 17: Proportion reporting suicidal ideation
Comparison 1: CBT‐based psychotherapy, Outcome 18: Suicide deaths by final follow‐up
Comparison 1: CBT‐based psychotherapy, Outcome 19: Repetition of SH at 12 months (by repeater status)
Comparison 2: DBT, Outcome 1: Repetition of SH by post‐intervention
Comparison 2: DBT, Outcome 2: Repetition of SH by 12 months
Comparison 2: DBT, Outcome 3: Frequency of SH by post‐intervention
Comparison 2: DBT, Outcome 4: Treatment adherence: Proportion completing treatment
Comparison 2: DBT, Outcome 5: Depression scores by post‐intervention
Comparison 2: DBT, Outcome 6: Suicidal ideation scores by post‐intervention
Comparison 2: DBT, Outcome 7: Suicide deaths by post‐intervention
Comparison 2: DBT, Outcome 8: Suicide deaths by 12 months
Comparison 2: DBT, Outcome 9: Suicide deaths by 24 months
Comparison 3: Group‐based emotion‐regulation psychotherapy, Outcome 1: Repetition of SH at post‐intervention
Comparison 3: Group‐based emotion‐regulation psychotherapy, Outcome 2: Frequency of repetition of SH at post‐intervention
Comparison 3: Group‐based emotion‐regulation psychotherapy, Outcome 3: Depression scores at post‐intervention
Comparison 4: Case management, Outcome 1: Repetition of SH at post‐intervention
Comparison 4: Case management, Outcome 2: Suicide deaths at post‐intervention
Comparison 5: Remote contact interventions: Emergency cards, Outcome 1: Repetition of SH at post‐intervention
Comparison 5: Remote contact interventions: Emergency cards, Outcome 2: Repetition of SH at 12 months
Comparison 6: Remote contact interventions: Postcards, Outcome 1: Repetition of SH at post‐intervention
Comparison 6: Remote contact interventions: Postcards, Outcome 2: Repetition of SH at 12 months
Comparison 6: Remote contact interventions: Postcards, Outcome 3: Frequency of SH at post‐intervention
Comparison 6: Remote contact interventions: Postcards, Outcome 4: Frequency of SH at 12 months
Comparison 6: Remote contact interventions: Postcards, Outcome 5: Frequency of SH at 24 months
Comparison 6: Remote contact interventions: Postcards, Outcome 6: Suicide deaths at post‐intervention
Comparison 7: Remote contact interventions: Telephone‐based psychotherapy, Outcome 1: Repetition of SH at post‐intervention
Comparison 7: Remote contact interventions: Telephone‐based psychotherapy, Outcome 2: Depression scores at post‐intervention
Comparison 7: Remote contact interventions: Telephone‐based psychotherapy, Outcome 3: Suicide deaths at post‐intervention
Comparison 8: Provision of information and support, Outcome 1: Repetition of SH at post‐intervention
Comparison 8: Provision of information and support, Outcome 2: Suicide deaths at post‐intervention
Comparison 9: Other multimodal interventions, Outcome 1: Repetition of SH at post‐intervention
Comparison 9: Other multimodal interventions, Outcome 2: Time to SH repetition
Comparison 9: Other multimodal interventions, Outcome 3: Depression scores at post‐intervention
Comparison 9: Other multimodal interventions, Outcome 4: Hopelessness scores at post‐intervention
Comparison 9: Other multimodal interventions, Outcome 5: Suicide deaths at post‐intervention
| Comparison 1.1: Individual‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with CBT‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.35 | 238 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of case management on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 200 per 1000 | 80 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 1.2: Group‐based CBT‐based psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with CBT‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.66 | 313 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of group‐based CBT‐based psychotherapy on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 190 per 1000 | 134 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.1: DBT compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT | |||||
| Repetition of SH by post‐intervention | Study population | OR 0.71 | 502 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate of DBT on repetition of SH at post‐intervention. The true effect is likely to be substantially different from the estimate of effect. | |
| 682 per 1000 | 604 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by two levels as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for two or more of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.2: DBT group‐based skills training compared to TAU or alternative psychotherapy for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT group‐based skills training | |||||
| Repetition of attempted suicide at post‐intervention | Study population | OR 0.66 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT group‐based skills training on repetition of attempted suicide at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 364 per 1000 | 274 per 1000 | |||||
| Repetition of NSSI at post‐intervention | Study population | OR 0.88 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT group‐based skills training on repetition of NSSI at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 576 per 1000 | 544 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.3: DBT individual therapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT individual therapy | |||||
| Repetition of attempted suicide at post‐intervention | Study population | OR 1.46 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT individual therapy on repetition of attempted suicide at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 364 per 1000 | 455 per 1000 | |||||
| Repetition of NSSI at post‐intervention | Study population | OR 1.29 | 66 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT individual therapy on repetition of NSSI at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 576 per 1000 | 636 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 2.4: DBT prolonged exposure protocol compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with DBT prolonged exposure protocol | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.67 | 18 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of DBT prolonged exposure protocol on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 333 per 1000 | 251 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 3: MBT compared TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with MBT | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.35 | 134 | ⊕⊕⊕⊕ | We are very confident that the true effect for MBT on repetition of SH at post‐intervention lies close to that of the estimate of the effect. | |
| 492 per 1000 | 253 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Comparison 4: Emotion‐regulation psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Group‐based emotion‐regulation psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.34 | 83 | ⊕⊕⊕⊝ | We are moderately confident that the true effect of group‐based emotion‐regulation psychotherapy on repetition of SH at post‐intervention lies close to that of the estimate of the effect. | |
| 775 per 1000 | 539 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgradedthis domain by one level as one study was suggestive of benefit, whilst the second trial of this intervention was not. | ||||||
| Psychodynamic psychotherapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Psychodynamic psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.45 | 170 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of psychodynamic psychotherapy on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 133 per 1000 | 65 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Case management compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Case management | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.78 | 1608 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of case management on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 114 per 1000 | 91 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Structured general practitioner (GP) follow‐up compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Structured general practitioner (GP) follow‐up | |||||
| Repetition of SH at post‐intervention (hospital records) | Study population | OR 1.01 | 143 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of structured general practitioner (GP) follow‐up on repetition of SH (according to hospital records) at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 133 per 1000 | 134 per 1000 | |||||
| Repetition of SH at post‐intervention (emergency records) | Study population | OR 2.56 | 123 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of structured general practitioner (GP) follow‐up on repetition of SH (according to emergency records) at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 72 per 1000 | 167 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.1: Emergency cards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Emergency cards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.82 | 1039 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of emergency cards on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 171 per 1000 | 145 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.2: Coping cards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Coping cards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.08 | 64 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of coping cards on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 156 per 1000 | 15 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.4: Postcards compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Postcards | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.87 | 3277 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for postcards on repetition of SH. The true effect is likely to be substantially different from the estimate of effect. | |
| 132 per 1000 | 117 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the I2 value indicated substantial levels of heterogeneity. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.5: Telephone contact compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Telephone contact | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.43 | 55 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of telephone contact on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 77 per 1000 | 35 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.6: Telephone contact, emergency cards, and letters compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Telephone contact, emergency cards, and letters | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.05 | 303 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of telephone contact, emergency cards, and letters on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 139 per 1000 | 145 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 9.7: Telephone‐based psychotherapy compared to TAU or alternative psychotherapy for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Remote contact interventions: Telephone‐based psychotherapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.36 | 185 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of telephone‐based psychotherapy on repetition of SH at post‐intervention is limited. The true effect may be substantially different from the estimate of the effect. | |
| 11 per 1000 | 4 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 10: Provision of information and support compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Provision of information and support | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.09 | 1853 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for information and support on repetition of SH by the post‐intervention assessment. The true effect is likely to be substantially different from the estimate of effect. | |
| 90 per 1000 | 97 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Publication bias is suspected as data from some centres have not been published. 2 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 3 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 11: Other multimodal interventions compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Other multimodal interventions | |||||
| Repetition of SH at post‐intervention | Study population | OR 0.61 | 1937 | ⊕⊝⊝⊝ | We have very little confidence in the effect estimate for other multimodal interventions on repetition of SH at post‐intervention. The true effect is likely to be substantially different from the estimate of effect. | |
| 252 per 1000 | 189 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by two levels as the I2 value indicated considerable levels of heterogeneity. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 12.5: General hospital management compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with General hospital management | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.03 | 77 | ⊕⊕⊕⊝ | We are moderately confident in the effect estimate of general hospital management on repetition of SH at post‐intervention. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | |
| 51 per 1000 | 53 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Comparison 12.8: Long‐term therapy compared to TAU or another comparator for self‐harm in adults | ||||||
| Patient or population: self‐harm in adults Intervention: Long term therapy | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with comparator | Risk with Long term therapy | |||||
| Repetition of SH at post‐intervention | Study population | OR 1.00 | 80 | ⊕⊕⊝⊝ | Our confidence in the effect estimate of long‐term therapy on repetition of SH at post‐intervention is limited.The true effect may be substantially different from the estimate of the effect. | |
| 225 per 1000 | 225 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 We downgraded this domain by one level as we rated any of the sources of risk of bias (as described in Assessment of risk of bias in included studies) at high risk for one of the studies included in the pooled estimate. 2 We downgraded this domain by one level as the 95% CI for the pooled effect included the null value. | ||||||
| Reference | Method | |||
| Self‐poisoning n (%) | Self‐injury n (%) | Combined self‐poisoning and self‐injury n (%) | Unspecified n (%) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 250 (76.7) | 64 (19.6) | ‐ | 15 (4.6) | |
| 7,733 (89.7) | 158 (8.2) | ‐ | 41 (2.1) | |
| 70 (58.3) | 33 (27.5) | ‐ | 17 (14.2) | |
| 772 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 442 (94.6) | 25 (5.3) | 8 (1.70) | ‐ | |
| 74 (71.8) | 25 (24.3) | ‐ | 4 (3.9) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 808 (97.7) | ‐ | ‐ | 19 (2.3) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 400 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 149 (100) | ‐ | ‐ | ‐ | |
| 119 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | 26 (100) | ‐ | ‐ | |
| 2,300 (100) | ‐ | ‐ | ‐ | |
| 471 (85.3) | 81 (14.7) | ‐ | ‐ | |
| 532 (77.8) | 125 (18.3) | 27 (3.9) | ‐ | |
| 115 (68.9) | 41 (24.5) | 11 (6.6) | ‐ | |
| 96 (100) | ‐ | ‐ | ‐ | |
| 80 (100) | ‐ | ‐ | ‐ | |
| 217 (98.2) | 4 (1.8) | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 707 (77.3) | 332 (36.3) | ‐ | 42 (4.6) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 161 (37.2) | 57 (13.2) | ‐ | 4 (0.9) | |
| 39 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 207 (97.6) | ‐ | ‐ | 5 (2.4) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 55 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 504 (97.3) | ‐ | ‐ | 14 (2.7) | |
| ‐ | ‐ | ‐ | ‐ | |
| 40 (64.6) | 12 (19.3) | 10 (16.1) | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 73 (89.0) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 141 (100) | ‐ | ‐ | ‐ | |
| 80 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 605 (100) | ‐ | ‐ | ‐ | |
| 926 (93.8) | ‐ | ‐ | ‐ | |
| 232 (84.7) | ‐ | ‐ | 42 (15.3) | |
| 463 (89.7) | ‐ | ‐ | 53 (10.3) | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| 77 (100) | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | |
| ‐ | 120 (100) | ‐ | ‐ | |
| 1 Percentages greater than 100% as participants could have used multiple methods. 2 Methods of self‐harm for the remaining 221 (48.7%) participants were not reported. 3 The majority of participants engaged in 'non‐violent' methods (n = 258; 85.2%), followed by 'violent' methods (n = 45; 14.8%). The methods considered to be 'non‐violent' or 'violent' in this trial were not reported, however. 4 The majority of participants engaged in 'low lethality' methods (n = 57; 89.1%), followed by 'high lethality' methods, including hanging, charcoal burning, or jumping from a height (n = 7; 10.9%). The methods considered to be 'low lethality' in this trial were not reported, however. | ||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Repetition of SH at post‐intervention Show forest plot | 4 | 238 | Odds Ratio (M‐H, Random, 95% CI) | 0.35 [0.12, 1.02] |
| 1.2 Repetition of SH at six months Show forest plot | 12 | 1260 | Odds Ratio (M‐H, Random, 95% CI) | 0.52 [0.38, 0.70] |
| 1.3 Repetition of SH at 12 months Show forest plot | 9 | 2458 | Odds Ratio (M‐H, Random, 95% CI) | 0.81 [0.66, 0.99] |
| 1.4 Repetition of SH at 24 months Show forest plot | 2 | 105 | Odds Ratio (M‐H, Random, 95% CI) | 0.31 [0.14, 0.69] |
| 1.5 Frequency of SH at post‐intervention Show forest plot | 4 | 149 | Mean Difference (IV, Random, 95% CI) | ‐0.53 [‐1.67, 0.61] |
| 1.6 Frequency of SH at six months Show forest plot | 4 | 118 | Mean Difference (IV, Random, 95% CI) | ‐0.71 [‐1.32, ‐0.11] |
| 1.7 Time to SH repetition Show forest plot | 3 | Hazard Ratio (IV, Random, 95% CI) | 0.81 [0.61, 1.08] | |
| 1.8 Depression scores at post‐intervention Show forest plot | 5 | 953 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.28 [‐0.54, ‐0.02] |
| 1.9 Depression scores at six months Show forest plot | 8 | 934 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.24 [‐0.50, 0.01] |
| 1.10 Depression scores at 12 months Show forest plot | 7 | 1130 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.36 [‐0.64, ‐0.07] |
| 1.11 Depression scores at 24 months Show forest plot | 2 | 225 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.48, 0.05] |
| 1.12 Hopelessness scores at post‐intervention Show forest plot | 5 | 803 | Mean Difference (IV, Random, 95% CI) | ‐2.99 [‐3.91, ‐2.07] |
| 1.13 Hopelessness scores at six months Show forest plot | 2 | 315 | Mean Difference (IV, Random, 95% CI) | ‐3.14 [‐4.78, ‐1.49] |
| 1.14 Hopelessness scores at 12 months Show forest plot | 3 | 539 | Mean Difference (IV, Random, 95% CI) | ‐1.89 [‐2.97, ‐0.81] |
| 1.15 Suicidal ideation scores at post‐intervention Show forest plot | 5 | 718 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.48 [‐0.68, ‐0.28] |
| 1.16 Suicidal ideation scores at six months Show forest plot | 4 | 353 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.38 [‐0.60, ‐0.17] |
| 1.17 Proportion reporting suicidal ideation Show forest plot | 2 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.17.1 Post‐intervention | 2 | 219 | Odds Ratio (M‐H, Random, 95% CI) | 0.44 [0.08, 2.26] |
| 1.17.2 Six months | 1 | 159 | Odds Ratio (M‐H, Random, 95% CI) | 1.27 [0.66, 2.46] |
| 1.17.3 12 months | 1 | 159 | Odds Ratio (M‐H, Random, 95% CI) | 1.20 [0.62, 2.31] |
| 1.18 Suicide deaths by final follow‐up Show forest plot | 16 | 2130 | Odds Ratio (M‐H, Random, 95% CI) | 0.79 [0.34, 1.80] |
| 1.19 Repetition of SH at 12 months (by repeater status) Show forest plot | 2 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.19.1 History of prior SH | 2 | 508 | Odds Ratio (M‐H, Random, 95% CI) | 0.56 [0.36, 0.88] |
| 1.19.2 No history of prior SH | 2 | 733 | Odds Ratio (M‐H, Random, 95% CI) | 1.10 [0.48, 2.52] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Repetition of SH by post‐intervention Show forest plot | 6 | 502 | Odds Ratio (M‐H, Random, 95% CI) | 0.71 [0.32, 1.55] |
| 2.1.1 Comparator: TAU | 3 | 267 | Odds Ratio (M‐H, Random, 95% CI) | 0.59 [0.16, 2.15] |
| 2.1.2 Comparator: Alternative psychotherapy | 3 | 235 | Odds Ratio (M‐H, Random, 95% CI) | 0.67 [0.16, 2.78] |
| 2.2 Repetition of SH by 12 months Show forest plot | 3 | 269 | Odds Ratio (M‐H, Random, 95% CI) | 0.65 [0.24, 1.72] |
| 2.2.1 Comparator: TAU | 2 | 172 | Odds Ratio (M‐H, Random, 95% CI) | 0.36 [0.05, 2.47] |
| 2.2.2 Comparator: Alternative psychotherapy | 1 | 97 | Odds Ratio (M‐H, Random, 95% CI) | 1.18 [0.35, 3.95] |
| 2.3 Frequency of SH by post‐intervention Show forest plot | 7 | 659 | Mean Difference (IV, Random, 95% CI) | ‐5.00 [‐8.92, ‐1.08] |
| 2.3.1 Comparator: TAU | 4 | 376 | Mean Difference (IV, Random, 95% CI) | ‐10.09 [‐21.57, 1.40] |
| 2.3.2 Comparator: Alternative psychotherapy | 3 | 283 | Mean Difference (IV, Random, 95% CI) | ‐5.00 [‐7.91, ‐2.09] |
| 2.4 Treatment adherence: Proportion completing treatment Show forest plot | 4 | 467 | Odds Ratio (M‐H, Random, 95% CI) | 1.40 [0.73, 2.67] |
| 2.4.1 Comparator: TAU | 1 | 180 | Odds Ratio (M‐H, Random, 95% CI) | 0.95 [0.52, 1.74] |
| 2.4.2 Comparator: Alternative psychotherapy | 3 | 287 | Odds Ratio (M‐H, Random, 95% CI) | 1.79 [0.65, 4.89] |
| 2.5 Depression scores by post‐intervention Show forest plot | 6 | 557 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.37 [‐0.58, ‐0.17] |
| 2.5.1 Comparator: TAU | 3 | 282 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.27 [‐0.50, ‐0.03] |
| 2.5.2 Comparator: Alternative psychotherapy | 3 | 275 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.50 [‐0.85, ‐0.14] |
| 2.6 Suicidal ideation scores by post‐intervention Show forest plot | 3 | 194 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.31 [‐0.64, 0.02] |
| 2.6.1 Comparator: TAU | 1 | 81 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.34 [‐0.78, 0.10] |
| 2.6.2 Comparator: Alternative psychotherapy | 2 | 113 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.40 [‐1.11, 0.31] |
| 2.7 Suicide deaths by post‐intervention Show forest plot | 5 | 563 | Odds Ratio (M‐H, Random, 95% CI) | 3.00 [0.12, 76.49] |
| 2.7.1 Comparator: TAU | 4 | 401 | Odds Ratio (M‐H, Random, 95% CI) | 3.00 [0.12, 76.49] |
| 2.7.2 Comparator: Alternative psychotherapy | 1 | 162 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.8 Suicide deaths by 12 months Show forest plot | 2 | 254 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.8.1 Comparator: TAU | 2 | 254 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.8.2 Comparator: Alternative psychotherapy | 0 | 0 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.9 Suicide deaths by 24 months Show forest plot | 2 | 254 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.9.1 Comparator: TAU | 2 | 254 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| 2.9.2 Comparator: Alternative psychotherapy | 0 | 0 | Odds Ratio (M‐H, Random, 95% CI) | Not estimable |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 Repetition of SH at post‐intervention Show forest plot | 2 | 83 | Odds Ratio (M‐H, Random, 95% CI) | 0.34 [0.13, 0.88] |
| 3.2 Frequency of repetition of SH at post‐intervention Show forest plot | 2 | 83 | Mean Difference (IV, Random, 95% CI) | ‐12.76 [‐34.92, 9.40] |
| 3.3 Depression scores at post‐intervention Show forest plot | 2 | 83 | Mean Difference (IV, Random, 95% CI) | ‐9.59 [‐13.43, ‐5.75] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 Repetition of SH at post‐intervention Show forest plot | 4 | 1608 | Odds Ratio (M‐H, Random, 95% CI) | 0.78 [0.47, 1.30] |
| 4.1.1 TAU (whole sample) | 2 | 600 | Odds Ratio (M‐H, Random, 95% CI) | 0.59 [0.25, 1.40] |
| 4.1.2 Enhanced usual care (whole sample) | 2 | 1008 | Odds Ratio (M‐H, Random, 95% CI) | 0.99 [0.42, 2.31] |
| 4.2 Suicide deaths at post‐intervention Show forest plot | 4 | 1757 | Odds Ratio (M‐H, Random, 95% CI) | 0.95 [0.57, 1.57] |
| 4.2.1 TAU | 2 | 600 | Odds Ratio (M‐H, Random, 95% CI) | 1.50 [0.24, 9.35] |
| 4.2.2 Enhanced usual care | 2 | 1157 | Odds Ratio (M‐H, Random, 95% CI) | 0.91 [0.54, 1.55] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 Repetition of SH at post‐intervention Show forest plot | 2 | 1039 | Odds Ratio (M‐H, Random, 95% CI) | 0.82 [0.31, 2.14] |
| 5.2 Repetition of SH at 12 months Show forest plot | 1 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 5.2.1 Emergency cards (whole sample) | 1 | 827 | Odds Ratio (M‐H, Random, 95% CI) | 1.19 [0.85, 1.67] |
| 5.2.2 Emergency cards (history of prior SH) | 1 | 394 | Odds Ratio (M‐H, Random, 95% CI) | 1.85 [1.14, 3.03] |
| 5.2.3 Emergency cards (no history of prior SH) | 1 | 427 | Odds Ratio (M‐H, Random, 95% CI) | 0.64 [0.34, 1.22] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 6.1 Repetition of SH at post‐intervention Show forest plot | 4 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 6.1.1 Postcards (whole sample) | 4 | 3277 | Odds Ratio (M‐H, Random, 95% CI) | 0.87 [0.62, 1.23] |
| 6.1.2 Postcards (males) | 1 | 247 | Odds Ratio (M‐H, Random, 95% CI) | 0.86 [0.42, 1.75] |
| 6.1.3 Postcards (females) | 1 | 524 | Odds Ratio (M‐H, Random, 95% CI) | 0.89 [0.56, 1.41] |
| 6.2 Repetition of SH at 12 months Show forest plot | 2 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 6.2.1 Postcards (whole sample) | 2 | 2885 | Odds Ratio (M‐H, Random, 95% CI) | 0.76 [0.57, 1.02] |
| 6.2.2 Postcards (males) | 1 | 247 | Odds Ratio (M‐H, Random, 95% CI) | 0.95 [0.50, 1.84] |
| 6.2.3 Postcards (females) | 1 | 524 | Odds Ratio (M‐H, Random, 95% CI) | 1.19 [0.78, 1.80] |
| 6.3 Frequency of SH at post‐intervention Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.3.1 Postcards (whole sample) | 3 | 1097 | Mean Difference (IV, Random, 95% CI) | ‐0.07 [‐0.32, 0.18] |
| 6.3.2 Postcards (males) | 3 | 401 | Mean Difference (IV, Random, 95% CI) | ‐0.01 [‐0.13, 0.12] |
| 6.3.3 Postcards (females) | 3 | 695 | Mean Difference (IV, Random, 95% CI) | ‐0.04 [‐0.29, 0.20] |
| 6.3.4 Postcards (history of prior SH) | 3 | 339 | Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.68, 0.51] |
| 6.3.5 Postcards (no history of prior SH) | 3 | 758 | Mean Difference (IV, Random, 95% CI) | 0.23 [‐0.32, 0.77] |
| 6.4 Frequency of SH at 12 months Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.4.1 Postcards (whole sample) | 2 | 984 | Mean Difference (IV, Random, 95% CI) | ‐0.19 [‐0.58, 0.20] |
| 6.4.2 Postcards (males) | 2 | 336 | Mean Difference (IV, Random, 95% CI) | 0.03 [‐0.11, 0.16] |
| 6.4.3 Postcards (females) | 2 | 647 | Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.62, 0.18] |
| 6.4.4 Postcards (history of prior SH) | 2 | 296 | Mean Difference (IV, Random, 95% CI) | ‐0.64 [‐2.07, 0.80] |
| 6.4.5 Postcards (no history of prior SH) | 2 | 688 | Mean Difference (IV, Random, 95% CI) | ‐0.07 [‐0.22, 0.09] |
| 6.5 Frequency of SH at 24 months Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.5.1 Postcards (whole sample) | 1 | 472 | Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.16, 0.10] |
| 6.5.2 Postcards (males) | 1 | 220 | Mean Difference (IV, Random, 95% CI) | 0.04 [‐0.21, 0.29] |
| 6.5.3 Postcards (females) | 1 | 252 | Mean Difference (IV, Random, 95% CI) | ‐0.07 [‐0.21, 0.07] |
| 6.5.4 Postcards (history of prior SH) | 1 | 338 | Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.31, 0.13] |
| 6.5.5 Postcards (no history of prior SH) | 1 | 134 | Mean Difference (IV, Random, 95% CI) | ‐0.01 [‐0.16, 0.14] |
| 6.6 Suicide deaths at post‐intervention Show forest plot | 4 | 3464 | Odds Ratio (M‐H, Random, 95% CI) | 1.86 [0.61, 5.72] |
| 6.6.1 Postcards (whole sample) | 4 | 3464 | Odds Ratio (M‐H, Random, 95% CI) | 1.86 [0.61, 5.72] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 7.1 Repetition of SH at post‐intervention Show forest plot | 2 | 185 | Odds Ratio (M‐H, Random, 95% CI) | 0.36 [0.01, 8.94] |
| 7.2 Depression scores at post‐intervention Show forest plot | 2 | 185 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.40 [‐1.39, 0.59] |
| 7.3 Suicide deaths at post‐intervention Show forest plot | 3 | 240 | Odds Ratio (M‐H, Random, 95% CI) | 3.09 [0.12, 78.55] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 8.1 Repetition of SH at post‐intervention Show forest plot | 2 | 1853 | Odds Ratio (M‐H, Random, 95% CI) | 1.09 [0.79, 1.50] |
| 8.2 Suicide deaths at post‐intervention Show forest plot | 3 | 2577 | Odds Ratio (M‐H, Random, 95% CI) | 0.15 [0.05, 0.48] |
| 8.2.1 Provision of information and support | 2 | 1889 | Odds Ratio (M‐H, Random, 95% CI) | 0.12 [0.03, 0.44] |
| 8.2.2 Provision of information and support with peer support | 1 | 688 | Odds Ratio (M‐H, Random, 95% CI) | 0.33 [0.03, 3.20] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 9.1 Repetition of SH at post‐intervention Show forest plot | 3 | Odds Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 9.1.1 Whole sample | 3 | 1937 | Odds Ratio (M‐H, Random, 95% CI) | 0.69 [0.37, 1.30] |
| 9.1.2 First SH episode | 2 | 901 | Odds Ratio (M‐H, Random, 95% CI) | 1.02 [0.70, 1.47] |
| 9.1.3 Repeat SH episode | 2 | 938 | Odds Ratio (M‐H, Random, 95% CI) | 0.92 [0.52, 1.62] |
| 9.2 Time to SH repetition Show forest plot | 2 | Hazard Ratio (IV, Random, 95% CI) | 0.39 [0.09, 1.76] | |
| 9.3 Depression scores at post‐intervention Show forest plot | 3 | 651 | Mean Difference (IV, Random, 95% CI) | ‐0.52 [‐2.08, 1.05] |
| 9.4 Hopelessness scores at post‐intervention Show forest plot | 2 | 556 | Mean Difference (IV, Random, 95% CI) | ‐0.29 [‐1.27, 0.69] |
| 9.5 Suicide deaths at post‐intervention Show forest plot | 3 | 971 | Odds Ratio (M‐H, Random, 95% CI) | 0.63 [0.13, 3.06] |
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Mousavi 2017 |  | |  | | | | ||||||
| "Eligible subjects were allocated to two groups (intervention and control), with simple randomization by a third party physician using tables of random numbers" (p.2). No information on allocation concealment reported. "There is no difference between the two groups in terms of demographic and clinical characteristics" (p.3). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. There was no information as to whether intention‐to‐treat analyses had been conducted. Subsequent analyses appear to be based on those participants with available information suggesting modified intention to treat analyses were used. | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| Data on repetition of self‐harm was obtained via self‐report. However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). As both the intervention and comparator arms received face‐to‐face therapy in this trial, there was no significant difference in treatment modality. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| Data on repetition of self‐harm was obtained via self‐report. All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Stewart 2009 | | | | | | | ||||||
| Correspondence with trial authors clarified that "[n]ames of treatment groups were drawn from a container and participants were allocated to a treatment group". However, no information on allocation concealment reported. Referring to Table 1 (p.542), only data on certain clinical characteristics (e.g., hopelessness, satisfaction with treatment, social problem solving, and suicidal ideation scores at baseline were reported. Whilst there were no significant differences between the intervention and comparator arms on these variables, there were no data reported on any demographic factors at baseline. | ||||||||||||
| Correspondence with authors clarified that the “treatment condition was offered to the client via a phone call”, suggesting that participants would have known to which arm they had been allocated. Correspondence with trial authors further clarified that the therapist running the research was aware of which treatment condition the participant was being offered. There were no apparent deviations from the intended intervention. Correspondence with authors clarified that 10 participants dropped out of the comparator arm, 12 dropped out of the CBT arm, and 11 dropped out of the PST arm. Data were analysed using the modified intention to treat principle. | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "…repeated attempt data from hospital chart audits that recorded re‐presentation to hospital for suicide attempts"(p.541). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Correspondence with authors clarified that the "therapist collected outcome data via self‐report measures and chart audits." However, neither clinical personnel were not blind to treatment allocation. Repetition of SH as determined by clinical and hospital records represents an observer‐reported outcome. | ||||||||||||
| "Measures were administered...directly following treatment (for the PST and CBT groups), and at two months follow‐up (for the TAU group)" (p.542). All eligible reported results for repetition of SH correspond to all intended outcome. "…repeated attempt data from hospital chart audits that recorded re‐presentation to hospital for suicide attempts" (p.541). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Wei 2013 | | |  | | | | ||||||
| "[P]articipants...were randomly assigned...using a computerized randomization program" (p. 109). However, no information on allocation concealment reported. "The three groups did not significantly differ on age, education, minority race, marriage, sex, occupation, and mental disorder diagnosis" (p.110). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "All analyses were conducted using the intent‐to‐treat (ITT) principle..." (p.110). | ||||||||||||
| Data were available for 72.0% of those randomised. Sensitivity analyses were not undertaken. "...the cumulative dropout rate [at post‐intervention] was 32.9% (n = 27) for the cognitive therapy group, 30.0% (n = 24) for the telephone intervention group, and 20.8% (n = 16) for the control group; the proportion of participants with missed assessment of three groups was not significantly different" (p.110). No other information on causes of missingness reported. | ||||||||||||
| "...research assessors followed up patients by telephone using measures of a detailed structured questionnaire...[to assess] suicidal history of patients" (p.110). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If clinical personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "At the 3 months, 6 months, and 12 months post baseline assessments, the research assessors followed up patients..." (p.110). All eligible reported results for repetition of SH correspond to all intended outcome. "...research assessors followed up patients by telephone using measures of a detailed structured questionnaire...[to assess] suicidal history of patients" (p.110). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as some concerns for risk of bias as missingness in the outcome could have depended on its true value and analyses did not correct for bias, nor were sensitivity analyses undertaken to investigate the potential effect of missing data. | ||||||||||||
| Weinberg 2006 | | | | | | | ||||||
| "Subjects were randomly assigned..." (p. 485). Correspondence with trial authors clarified that "subjects were asked to choose between 2 similar envelopes..." However, detail on the randomisaton method was not reported. Additionally, it was not clear whether these envelopes were sequentially numbered. There were no significant differences between the intervention and comparator arms for any of the demographic characteristics and clinical variables measured (Table 1, p.486). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. Although "[a]ll [intervention] participants completed 6 sessions of MACT. Two [comparator] group participants were not available for the post‐treatment assessments" (p. 485). Nevertheles, "[a]ll participants were interviewed at the 6 months follow up" (p.485), suggesting that intention‐to‐treat analyses were undertaken. | ||||||||||||
| Data were available for 100% of those randomised (by the six‐month assessment). | ||||||||||||
| Repetition of self‐harm was ascertained from the "Parasuicide History Interview‐PHI " (p.486). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If clinical personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "The follow‐up assessments by [sic] performed after completion of the [intervention] (e.g, 6‐8 weeks after baseline assessments) and at a similar time point in the control group" (p.486). All eligible reported results for repetition of SH correspond to all intended outcome. Repetition of self‐harm was ascertained from the "Parasuicide History Interview‐PHI " (p.486). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 2.1.1 Comparator: TAU | ||||||||||||
| Linehan 1991 | | | | | | | ||||||
| "...randomized” (p. 1060). Correspondence with authors clarified that they used a computer programme to generate the random sequence. Additionally, correspondence with authors further clarified that allocation had been concealed. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. There was no information as to whether intention‐to‐treat analyses had been conducted. Subsequent analyses appear to be based on those participants with available information suggesting modified intention to treat analyses were used. | ||||||||||||
| "Ten subjects dropped out of the study during the pretreatment assessment…seven subjects were dropped following pretreatment assessment for refusal or inability to meet study conditions…Two subjects...were dropped from all analyses [due to treatment non‐adherence]" (p.1061). Therefore, data were available for 69.8% of those randomised. Sensitivity analyses were not undertaken. Data were available for 68.7% of those randomised to the intervention arm, and 71.0% of those randomised to the comparator arm. Additionally, data for 24/63 (38.1%) participants were deliberately not included in the 24‐month follow‐up assessment. | ||||||||||||
| "The PHI obtained information about all parasuicide behavior" (p.1061). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. As the intervention was found to be significant, this is possible. | ||||||||||||
| "The treatment lasted 1 year, with assessment every 4 months" (p.1060). All eligible reported results for repetition of SH correspond to all intended outcome. "The PHI obtained information about all parasuicide behavior" (p.1061). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as unblinded personnel were acting as outcome assessors which may have influenced assessment of the outcome. As the intervention was found to be significant this is possible, particularly given that no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| McMain 2009 | | | | | | | ||||||
| "Eligible participants were randomly assigned to treatment arms using a pre‐generated block randomization scheme developed and held by the statistician who prepared 45 sealed envelopes, each containing the group allocations in random order for four participants" (p. 1366). "...there were no between‐group differences on baseline demographic characteristics, diagnostic variables, or suicidal and nonsuicidal self‐injurious behaviors" (p.1371). | ||||||||||||
| This was a single‐blind study and only "[a]ssessors...were...blind to treatment assignment" (p. 1366). Participants were therefore not blind to treatment allocation. Additionally, "[t]he study coordinator...was not blind to treatment assignment" (p. 1366). There were no apparent deviations from the intended intervention. "All results were analysed using an intent‐to‐treat analysis..." (p.1370). | ||||||||||||
| Data were available for 82.2% of those randomised. "...participants with missing data were still included in the model. Each model was reanalyzed using preselected covariates that are known or theorized to affect the outcome" (p.1371). | ||||||||||||
| "The primary outcome measures were frequency and severity of suicidal and nonsuicidal self‐injurious behavior episodes, as assessed by the Suicide Attempt Self‐Injury Interview" (p.1370). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Participants were not blind to treatment allocation. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| Referring back to the trial protocol (NCT00154154), no specific information was reported. "The primary outcome measures, assessed at baseline and every 4 months over the treatment period, were frequency and severity of suicidal and nonsuicidal self‐harm episodes" (p.1365). All eligible reported results for repetition of SH correspond to all intended outcome measurements. "The primary outcome measures were frequency and severity of suicidal and nonsuicidal self‐injurious behavior episodes, as assessed by the Suicide Attempt Self‐Injury Interview..." (p.1370). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Priebe 2012 | | | | | | | ||||||
| "Randomisation was computer generated with a 1:1 allocation...using 6 blocks of 12 randomly permuted treatment allocation sequences, with a final block of 8" (p. 358). However, no information on allocation concealment reported. There were no significant differences between the intervention and comparator arms for any of the demographic characteristics and clinical variables measured (Table 1, p.360). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "[M]issing covariate values were...[estimated using] maximum likelihood estimation [to] ensure...unbiased parameter estimates. We...[also] conducted a sensitivity analysis with last observation carried forward" (p.358). | ||||||||||||
| Data were available for 93.7% of those randomised. "[M]issing covariate values were...[estimated using] maximum likelihood estimation [to] ensure...unbiased parameter estimates. We...[also] conducted a sensitivity analysis with last observation carried forward" (p.358). Moreover, for the primary outcome, "[t]he sensitivity analysis with last observation carried forward showed a very similar result..." (p.360). | ||||||||||||
| "The primary outcome was self‐harm...recorded in an interview on a structured form" (p.358). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "The primary outcome was self‐harm...[t]his was assessed...at each 2‐month follow‐up thereafter" (p.358). All eligible reported results for repetition of SH correspond to all intended outcome. "The primary outcome was self‐harm...recorded in an interview on a structured form" (p.358). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Subgroup 2.1.2 Comparator: Alternative psychotherapy | ||||||||||||
| Linehan 2006 | | | | | | | ||||||
| "Computerized adaptive minimization randomization procedure" (p. 758). No information on allocation concealment reported. "The treatment groups did not significantly differ on demographic characteristics, diagnoses, or pretreatment number of suicide attempts or nonsuicidal self‐injuries" (p.760). | ||||||||||||
| "Initial assessments were done before informing subjects of treatment assignment" (p. 758), suggesting that participants were made aware of treatment assignment during the trial. Additionally, given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. There was no information as to whether intention‐to‐treat analyses had been conducted. Subsequent analyses appear to be based on those participants with available information suggesting modified intention to treat analyses were used. Moreover, "we examined the effects of differential missing data and treatment dropout on each of our major outcome variables and found no evidence that the findings were biased by these differences" (p. 760). | ||||||||||||
| Data were available for 96% of those randomised. | ||||||||||||
| "The Suicide Attempt Self‐Injury Interview measured...suicide intent, and medical severity of each suicide attempt and nonsuicidal self‐injury" (p.758). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "Initial assessments were done before informing subjects of treatment assignment" (p. 758), suggesting that participants were made aware of treatment assignment during the trial. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "Our primary method for analyzing repeated‐measures data (pretreatment through the 24‐month assessment)..." (p.760). All eligible reported results for repetition of SH correspond to all intended outcome. "The Suicide Attempt Self‐Injury Interview measured...suicide intent, and medical severity of each suicide attempt and nonsuicidal self‐injury" (p.758). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias no specific information on allocation concealment was reported. Additionally, no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Turner 2000 | | | | | | | ||||||
| "Participants [were] randomly assigned..." (p. 414). However, without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. Additionally, no information on allocation concealment reported. "To determine if the random assignment procedure worked, we examined the pretreatment values of the dependent variables for each...outcome...there were no significant differences between the groups" (pp.416‐7). However, information on any differences in demographic characteristics at baseline was not reported. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "All 24 patients participated in the 6 month and 12 month assessments and composed the intention‐to‐treat sample for the analyses" (p. 414). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "Patients...maintained daily logs of suicidal urges and attempts" (p.415). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. As the intervention was found to be significant, this is possible, and given the magnitude of the effect size, this is likely. | ||||||||||||
| "Outcome assessment was conducted at pretreatment, 6 months, and 12 months" (p.415). All eligible reported results for repetition of SH correspond to all intended outcome. "Patients...maintained daily logs of suicidal urges and attempts" (p.415). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as unblinded personnel were acting as outcome assessors. They may have influenced assessment of the outcome. As the intervention was found to be significant, this is possible, and given the magnitude of the effect size, this is likely. | ||||||||||||
| Walton 2020 | | | | | | | ||||||
| "A stratified randomisation procedure was used to maximise the likelihood of comparable distributions across groups for gender and antidepressant use. A computerised formula with blocked randomisation (blocks of 4) was used by an independent research manager of the health service. A sealed, opaque envelope containing randomisation status was given to participants at the end of their assessment" (p.3). "The two groups showed no imbalances at baseline after randomisation with regard to sociodemographic and clinical characteristics" (p.7). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. Referring back to the trial protocol (ACTRN12612001187831), there was no apparent deviation from the intended intervention. "Intent‐to‐treat analyses..." (p.7). All patients who were randomised and attended a minimum of one treatment session after the baseline assessment were included in the analysis. | ||||||||||||
| Data were available for 74.7% of those randomised. "To take into account missing data and maximise usage of all available data, major analyses comprised a series of generalised linear models (and, where appropriate, GEE analyses), together with some supplementary sensitivity analyses (e.g. restricted to participants completing posttreatment assessments)" (p.7). | ||||||||||||
| "Combined outcome of any episode of suicidal and non‐suicidal self‐injury [using] the SASI‐Count" (p.4). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, given that data obtained by correspondence indicated the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "This study utilised an adapted time frame of behaviour occurring within the past 7 months" (p.4). All eligible reported results for repetition of SH correspond to all intended outcome. "Combined outcome of any episode of suicidal and non‐suicidal self‐injury [using] the SASI‐Count" (p.4). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Gratz 2006 | | | | | | | ||||||
| "Random assignment" (p. 30). However, without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. Additionally, no information on allocation concealment reported. "A series of...were conducted on demographic and clinical characteristic variables to determine equivalence across conditions. Results indicate no significant between‐group differences on any of these variables" (p.30). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. There was no information as to whether intention‐to‐treat analyses had been conducted. "Two participants dropped out of the study (one from each condition)" (p.27). Subsequent analyses appear to be based on those participants with available information suggesting modified intention to treat analyses were used. | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The following self‐report measures...were used to assess outcome: Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "...self‐report measures [were] administered before and after treatment, were used to assess outcome" (p.29). All eligible reported results for repetition of SH correspond to all intended outcome. "The following self‐report measures...were used to assess outcome: | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information on randomisation or allocation concealment was reported. Additionally, no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Gratz 2014 | | | | | | | ||||||
| "Participants...were matched on four prognostic variables [emotion dysregulation, number of lifetime incidents of SH, global assessment of functioning (GAF) scores, and age] and randomly assigned...using a stratified randomization procedure" (p. 2100). However, without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. Additionally, no information on allocation concealment reported. There were no significant differences between the intervention and comparator arms for any of the demographic characteristics and clinical variables measured (Table 1, p.2101). Furthermore, the trial authors undertook a series of tests "...on pre‐treatment scores on outcome measures to determine equivalence across conditions. The results indicate no significant between‐group differences...with one exception: the [intervention group] had significantly lower scores...on the DERS lack of clarity subscale...(p <0.05)" (p.2104). However, this may have been compatible with chance. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "We adopted a Bayesian approach...using the Markov chain Monte Carlo routines...This approach implements a multiple imputation strategy to handle missing data...enabling an analysis of the intent‐to‐treat (ITT) sample" (p. 2104). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The following measures were administered...to assess treatment efficacy...The DSHI [and] The Self‐Harm Inventory..." (pp.2100‐2). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. As the intervention was found to be significant, this is possible, but given the magnitude of the effect size, this is unlikely. | ||||||||||||
| "...measures were administered pre and post‐treatment or ‐waitlist to assess treatment efficacy, and at 3 and 9 months post‐treatment" (p.2100). All eligible reported results for repetition of SH correspond to all intended outcome. "The following measures were administered...to assess treatment efficacy...The DSHI [and] The Self‐Harm Inventory..." (pp.2100‐2). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information on randomisation or allocation concealment was reported. Additionally, no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 4.1.1 TAU (whole sample) | ||||||||||||
| Clarke 2002 | | | | | | | ||||||
| "Randomisation was conducted using random numbered lists, stratified for sex and admitting hospital..." (p.169). Additionally, "[t]he random number lists were constructed independently of the research team and they did not have sight of them....The researchers were required to telephone an administrator with possible candidates’ details and were then informed of the treatment group" (p. 169). There were no significant differences between the intervention and comparator arms for any of the demographic characteristics and clinical variables measured (Table 1, p.172). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Data analysis proceeded on an intention to treat basis..." (p.170). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| Repetition of self‐harm was ascertained from readmission to hospital. However, only repeat episodes of self‐harm resulting in presentation to the study hospital were counted. Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. In this trial, researchers also acted as case managers. Repetition of SH as determined by clinical and hospital records represents an observer‐reported outcome. | ||||||||||||
| "The main outcome measure was readmission to A&E as a result of self‐harm within 12 months" (p.186). All eligible reported results for repetition of SH correspond to all intended outcome. "The main outcome measure was readmission to A&E as a result of self‐harm..." (p.186). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Hvid 2011 | | | | | | | ||||||
| "We applied a stratified randomization procedure...this stratified randomization procedure...created eight categories and randomization was performed for each independently" (p. 293). Without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. However, "[r]andomisation was performed by an independent office" (p. 293). There were "no significant differences with respect to age, sex, method, type (violent/non‐violent), alcohol influence, psychiatric assessment, earlier psychiatric contact, earlier suicide attempt and reason (problem type)" (p.294). | ||||||||||||
| "The patient...knew who was a case and who was a control" (p. 293). Additionally, "...intervention staff knew who was a case and who was a control" (p. 293). There were no apparent deviations from the intended intervention. "Outcomes were measured by an intent‐to‐treat design in which all patients were followed until the end of the trial, irrespective of whether the patient was still receiving or complying with the assigned treatment" (p. 294). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "...information about suicidal behaviour...was obtained by our team through interview, public registrations, health records and staff" (p.293). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "[I]ndependent assessors (three psychiatrists) who reviewed all incidents did not have this information" (p. 294) on participants' treatment allocation. | ||||||||||||
| "Outcome data were obtained...in a 1‐year period, including 6 months after the completion of the 6‐month intervention period" (p.294). All eligible reported results for repetition of SH correspond to all intended outcome. "...information about suicidal behaviour...was obtained by our team through interview, public registrations, health records and staff" (p.293). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Subgroup 4.1.2 Enhanced usual care (whole sample) | ||||||||||||
| Kawanishi 2014 | | | | | | | ||||||
| "Participants were randomly assigned (1:1) by and Internet‐based system...to either the intervention group (assertive case management) or the control group (enhanced usual care). Assignment was by the minimisation method,with four factors: participating hospital, sex, age...and history of previous suicide attempts before the current episode..." (p. 194). Additionally, "[p]articipants were randomly assigned...by an Internet‐based system operated by a central, independent data centre" (p. 194). "Baseline characteristics were well balanced between the groups" (p.197). | ||||||||||||
| "...patients and case managers who provided the interventions were not [blind to treatment allocation]" (p. 194). There were no apparent deviations from the intended intervention. "Analyses were done in accordance with the intention‐to‐treat principle" (p.196). The sample was biased towards more compliant patients who were willing and able to attend a psychoeducation session seminar at the commencement of treatment and were able to attend hospital regularly for face‐to‐face interviews and case management sessions. | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The primary outcome measure was the incidence of first recurrent suicidal behaviour" (p.196). However, even after referring to the trial protocol (NCT00736918) it is unclear how repetition of self‐harm was ascertained. It is likely that repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "Outcome assessors were masked to group assignment...The assessors did not know the participants’ assigned groups, the status of implementation of the intervention or information about events obtained by other on‐site staff" (p. 194). | ||||||||||||
| Referring back to the trial protocol (NCT00736918 and UMIN‐CTR C000000444), no specific information was reported. Referring back to the trial protocol (NCT00736918 and UMIN‐CTR C000000444), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (NCT00736918 and UMIN‐CTR C000000444), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Morthorst 2012 | | | | | | | ||||||
| "Computer randomisation was done by an independent research assistant stratified by whether the patient had previously attempted suicide (first attempt v previous attempt), previous psychiatric contacts or hospitalisations (none v previous contacts), and alcohol consumption at the time of suicide attempt (none v alcohol consumption)...The randomisation procedure ensured adequate...allocation concealment" (p.3). "The only differences between participants at baseline were that patients in the intervention group were more frequently being treated with antidepressants at inclusion [OR 1.99, 95% CI 1.18 to 3.55] and more frequently used narcotics as method of suicide attempt" (p.3). However, data on the proportion using narcotics at the index attempt were not reported making it difficult to determine whether this difference could have been compatible with chance. | ||||||||||||
| "[P]articipants were immediately informed of the outcome [i.e., allocation]" (p. 3). Additionally, "[o]wing to the nature of the study design, the intervention staff were not blinded" (p. 3) There were no apparent deviations from the intended intervention. "All participants were included in the analysis regardless of subsequent adherence to treatment, according to the intention to treat principle" (p.3). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The outcomes of interest were repeated suicide attempt...We obtained information on these events through ICD‐10 diagnoses (recorded as part of standardised hospital registration procedure) and review of medical records. Additionally, we collected self reported data on events...through telephone interviews with patients" (p.3). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "An external medical evaluation committee conducted a blinded outcome assessment using medical records" (p. 3). However, trial authors later state "[t]he researcher conducting the analyses on self‐reported outcomes was...not blinded" (p.3). Repetition of SH as determined by clinical and hospital records represents an observer‐reported outcome. | ||||||||||||
| "Main outcome...[was r]epeated suicide attempt and death by suicide...at 1‐year follow‐up" (p.1). All eligible reported results for repetition of SH correspond to all intended outcome. "The outcomes of interest were repeated suicide attempt...We obtained information on these events through ICD‐10 diagnoses (recorded as part of standardised hospital registration procedure) and review of medical records. Additionally, we collected self reported data on events...through telephone interviews with patients" (p.3). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Evans 1999a | | | | | | | ||||||
| "Randomised on a 1:1 basis...using the sealed envelope technique, ensuring that it was impossible to tell from feeling or looking at the envelopes whether they contained a green card or a ’dummy card’ (which was not given out)" (p. 23). However, without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. "These [demographic and clinical] characteristics were simllar in the two groups, although there were slighily more males in the control group [OR 0.80, 95% CI 0.61 to 1.06]" (p.25). This may be compatible with chance. | ||||||||||||
| "Those randomised to receive a green card were offered the card immediately after the psychiatric assessment" (p.23). Given this, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "...all analyses were conducted on an intention‐to‐treat basis" (p. 24). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "All subiects' repeat hospital attendanccs for DSH...were monitored (blind to their study group) by means of a computerised case register based on routine accident and emergency and admission data for each of the three general hospitals" (p.24). Whilst only repeat episodes of self‐harm resulting in presentation to the study hospitals were counted, the authors do note that "...other accident and emergency departments were at least 13 miles away" (p.24). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "All subjects’ repeat hospital attendances for SH...were monitored (blind to their study group)..." (p. 24). | ||||||||||||
| "All subjects' repeat hospital attendanccs for DSH within six months of randomisation were monitored..." (p.24). All eligible reported results for repetition of SH correspond to all intended outcome. "All subiects' repeat hospital attendanccs for DSH...were monitored (blind to their study group) by means of a computerised case register..." (p.24). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as there was a higher incidence of personality disorders as well as a higher number of subjects with multiple suicide attempts at baseline amongst those allocated to the intervention arm, suggesting there may have been a problem with the randomisation process. | ||||||||||||
| Morgan 1993 | | | | | | | ||||||
| "Allocation to experimental or control group was carried out by random selection from a supply of closed envelopes, half of which contained the green card" (p.111). However, it is unclear whether these were opaque. Additionally, no information on allocation concealment reported. "The groups were similar in age...ethnic origin, marital, and occupational status. In both groups, depressive disorder was the most common diagnostic category" (p.111). | ||||||||||||
| "[Patients] receive[ed] the green card" (p. 111). Additionally, "GPs were also sent copies of the green card" (p. 111), suggesting that both participants and clinical personnel were aware of treatment allocation. There were no apparent deviations from the intended intervention. "Data concerning outcome were obtained for all patients included in the study" (p. 111). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "...follow up data were obtained by review of medical and psychiatric notes, as well as OP [outpatient] records of all patients" (p.111). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. No specific information on whether outcome assessors were aware of participants' treatment allocation was reported. Repetition of SH as determined by clinical and hospital records represents an observer‐reported outcome. | ||||||||||||
| "One year later, follow up data were obtained..." (p.111). All eligible reported results for repetition of SH correspond to all intended outcome. "...follow up data were obtained by review of medical and psychiatric notes, as well as OP [outpatient] records of all patients" (p.111). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 6.1.1 Postcards (whole sample) | ||||||||||||
| Beautrais 2010 | | | | | | | ||||||
| “Participants were randomised 1:1...using predetermined computer‐generated random numbers...using a uniform distribution to generate a sequence of random numbers between 0 and 1. Numbers of 0.5 or above were classified as the intervention group; numbers below 0.5 were classified as the control group" (p. 56). Additionally, "[r]andomisation was performed...by research staff who were not involved in the recruitment or clinical care of participants" (p.56). "There were no significant differences between the intervention and control groups in age, gender, marital status, method of self‐harm, length of hospital stay or history of attendance for self‐harm in the 12 months prior to the index presentation" (p.57) However, whilst the trial authors state "...there was a significant difference between the groups in the number of prior attendances for self‐harm in the previous 12 months...with the number of prior attendances being lower in the intervention than in the control group" (p.57), this may have been compatible with chance. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. However, "[p]articipants’ randomisation status was not conveyed to clinical...staff" (p.56). There were no apparent deviations from the intended intervention. "...results of the trial were analysed using the intention‐to‐treat design" (p. 56). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "Re‐presentations for self‐harm were assessed by monitoring two sources of re‐presentation information. First, psychiatric emergency service records were checked...to identify attendances by study participants. Second, participants’ hospital medical records were reviewed..." (p.56). However, only repeat episodes of self‐harm resulting in presentation to the study hospital were counted. Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "Participants’ randomisation status was not conveyed to...data‐collection staff" (p. 56). | ||||||||||||
| "Outcome measures were the proportion of participants re‐presenting with self‐harm and the number of re‐presentations for self‐harm in the 12 months following the initial presentation" (p.55). All eligible reported results for repetition of SH correspond to all intended outcome. "Re‐presentations for self‐harm were assessed by monitoring two sources of re‐presentation information. First, psychiatric emergency service records were checked...Second, participants’ hospital medical records were reviewed..." (p.56). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Carter 2005 | | | | | | | ||||||
| "Randomisation was by database (HanDBase, version 2.0, DDH Softwards, FL, USA) on a personal digital assistant (Palm III, Palm, CA, USA) that was populated with a pre‐generated randomisation schedule (in blocks of 10)..." (p. 2). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. Referring to Table 1 (p.3), there was a lower proportion of females assigned to the intervention as compared to the comparator arm (OR 0.57, 95% CI 0.42 to 0.77). This is unlikely to have been compatible with chance. However, there were no other significant differences between the intervention and comparator arms for any of the other demographic characteristics and clinical variables measured (Table 1, p.3). | ||||||||||||
| "We used a randomised consent design, using the single consent version (Zelen’s design). This design is a variation on the standard randomised controlled experimental design, where participants are randomised...before consent is sought" (p.2). Additionally, "...the secretary responsible for managing the mailing database and postcards [was] not blind to allocation status" (p. 4). As clinical personnel knew whether or not a postcard had been sent, they could not have been blinded to allocation status. "Twenty participants in the control group received the intervention due to clerical errors but were included in the control group for the intention to treat analyses" (p.2). The proportion of participants crossing over was modest (6.6%). However, the proportion of participants crossing over was not balanced (i.e., 0% in the intervention arm compared with 6.6% in the comparator arm). "We assessed the outcomes by an intention to treat analysis on the basis of allocation" (p. 2). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "...research staff remained blinded to allocation" (p. 2). | ||||||||||||
| "Main outcome measures [were the p]roportion of patients with one or more repeat episodes of deliberate self poisoning and the number of repeat episodes for deliberate self poisoning per person in 12 months" (p.1). All eligible reported results for repetition of SH correspond to all intended outcome. "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as there deviations from the intended intervention that arose because of the trial context that may have affected the outcome. Specifically, a number of participants assigned to the control group received the intervention due to clerical errors. | ||||||||||||
| Hassanian‐Moghaddam 2011 | | | | | | | ||||||
| "Block randomisation (blocks of 100) was undertaken using a random digit table” (p. 310). Furthermore, "[t]o maintain masking to allocation, randomisation was not revealed to the recruiting toxicologist until all information was entered and eligibility determined" (p.310). The trial authors note that although "...this older form of randomisation is potentially liable to interference...no imbalances at baseline suggest that the randomisation was likely to have been successful" (p.314). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. However, "[o]ther staff were masked to allocation status during hospital treatment” (p.310). There were no apparent deviations from the intended intervention. "All outcomes were analysed on randomisation status at baseline for 12‐month follow‐up" (p. 311). | ||||||||||||
| Data were available for 91.9% of those randomised. Sensitivity analyses were not undertaken. Data were available for 90.7% of those randomised to the intervention arm, and 93.0% of those randomised to the comparator arm. Additionally, the trial authors report "[there were no significant differences for gender, age, employment status or previous suicide attempt...[and] a small, significant difference for marital status...for missing versus study group respectively" (p.313). | ||||||||||||
| "Outcomes were determined by direct questions, for example ‘Did you have any self‐mutilation (self‐cutting) during the study period?'" (p.310). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If unblinded personnel were acting as outcome assessors, they may have influenced assessment of the outcome. As the intervention was found to be significant, this is possible, but given the magnitude of the effect size, this is unlikely. | ||||||||||||
| "All outcomes were analysed on randomisation status at baseline for 12‐month follow‐up" (p.311). All eligible reported results for repetition of SH correspond to all intended outcome. "Outcomes were determined by direct questions..." (p.310). | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Kapur 2013 | | | | | | | ||||||
| "Randomisation was conducted via a remote Internet‐based service (www.sealedenvelope.com)" (p. 73). However, no information on allocation concealment reported. Information on any differences in demographic characteristics and clinical variables between the intervention and control groups at baseline was not reported. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Primary analysis was on an intention‐to‐treat basis" (p. 73). | ||||||||||||
| Data were available for 98.5% of those randomised. | ||||||||||||
| "We investigated the proportion of patients with at least one repeat episode of self‐harm resulting in hospital attendance...identified from the hospital information systems" (p.73). However, only repeat episodes of self‐harm resulting in presentation to the study hospital were counted. Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "All outcome data were collected by researchers masked to allocation status" (p.73). | ||||||||||||
| We investigated the proportion of patients with at least one repeat episode of self‐harm resulting in hospital attendance within 12 months" (p.73). All eligible reported results for repetition of SH correspond to all intended outcome. "We investigated the proportion of patients with at least one repeat episode of self‐harm resulting in hospital attendance...identified from the hospital information systems" (p.73). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Subgroup 6.1.2 Postcards (males) | ||||||||||||
| Carter 2005 | | | | | | | ||||||
| "Randomisation was by database (HanDBase, version 2.0, DDH Softwards, FL, USA) on a personal digital assistant (Palm III, Palm, CA, USA) that was populated with a pre‐generated randomisation schedule (in blocks of 10)..." (p. 2). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. Referring to Table 1 (p.3), there was a lower proportion of females assigned to the intervention as compared to the comparator arm (OR 0.57, 95% CI 0.42 to 0.77). This is unlikely to have been compatible with chance. However, there were no other significant differences between the intervention and comparator arms for any of the other demographic characteristics and clinical variables measured (Table 1, p.3). | ||||||||||||
| "We used a randomised consent design, using the single consent version (Zelen’s design). This design is a variation on the standard randomised controlled experimental design, where participants are randomised...before consent is sought" (p.2). Additionally, "...the secretary responsible for managing the mailing database and postcards [was] not blind to allocation status" (p. 4). As clinical personnel knew whether or not a postcard had been sent, they could not have been blinded to allocation status. "Twenty participants in the control group received the intervention due to clerical errors but were included in the control group for the intention to treat analyses" (p.2). The proportion of participants crossing over was modest (6.6%). However, the proportion of participants crossing over was not balanced (i.e., 0% in the intervention arm compared with 6.6% in the comparator arm). "We assessed the outcomes by an intention to treat analysis on the basis of allocation" (p. 2). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "...research staff remained blinded to allocation" (p. 2). | ||||||||||||
| "Main outcome measures [were the p]roportion of patients with one or more repeat episodes of deliberate self poisoning and the number of repeat episodes for deliberate self poisoning per person in 12 months" (p.1). All eligible reported results for repetition of SH correspond to all intended outcome. "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as there deviations from the intended intervention that arose because of the trial context that may have affected the outcome. Specifically, a number of participants assigned to the control group received the intervention due to clerical errors. | ||||||||||||
| Subgroup 6.1.3 Postcards (females) | ||||||||||||
| Carter 2005 | | | | | | | ||||||
| "Randomisation was by database (HanDBase, version 2.0, DDH Softwards, FL, USA) on a personal digital assistant (Palm III, Palm, CA, USA) that was populated with a pre‐generated randomisation schedule (in blocks of 10)..." (p. 2). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. Referring to Table 1 (p.3), there was a lower proportion of females assigned to the intervention as compared to the comparator arm (OR 0.57, 95% CI 0.42 to 0.77). This is unlikely to have been compatible with chance. However, there were no other significant differences between the intervention and comparator arms for any of the other demographic characteristics and clinical variables measured (Table 1, p.3). | ||||||||||||
| "We used a randomised consent design, using the single consent version (Zelen’s design). This design is a variation on the standard randomised controlled experimental design, where participants are randomised...before consent is sought" (p.2). Additionally, "...the secretary responsible for managing the mailing database and postcards [was] not blind to allocation status" (p. 4). As clinical personnel knew whether or not a postcard had been sent, they could not have been blinded to allocation status. "Twenty participants in the control group received the intervention due to clerical errors but were included in the control group for the intention to treat analyses" (p.2). The proportion of participants crossing over was modest (6.6%). However, the proportion of participants crossing over was not balanced (i.e., 0% in the intervention arm compared with 6.6% in the comparator arm). "We assessed the outcomes by an intention to treat analysis on the basis of allocation" (p. 2). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "...research staff remained blinded to allocation" (p. 2). | ||||||||||||
| "Main outcome measures [were the p]roportion of patients with one or more repeat episodes of deliberate self poisoning and the number of repeat episodes for deliberate self poisoning per person in 12 months" (p.1). All eligible reported results for repetition of SH correspond to all intended outcome. "We extracted the descriptive variables of the sample from the toxicology service’s database, which were derived from the standardised clinical assessment of patients with deliberate self poisoning" (p.1). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as there deviations from the intended intervention that arose because of the trial context that may have affected the outcome. Specifically, a number of participants assigned to the control group received the intervention due to clerical errors. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Marasinghe 2012 | | | | | | | ||||||
| "The participants were randomly allocated" (p. 152). However, without further specific information on the method used it is difficult to ascertain whether the randomisation sequence was adequate. Additionally, no information on allocation concealment reported. "There were no significant differences between the groups at baseline" (p.152). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Intention to treat analyses..." (p.152) | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| No specific information on how repeat self‐harm was ascertained was reported. It is likely that repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The assessor was blind to the treatment” (p. 152). | ||||||||||||
| "At six months, an independent assessor repeated the baseline measures with the participants...[a]t the end of 12 months another independent assessor assessed the outcomes" (p.152). All eligible reported results for repetition of SH correspond to all intended outcome. No specific information on how repeat self‐harm was ascertained was reported. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias no specific information on randomistion or allocation concealment was reported. Additionally, no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Wei 2013 | | | | | | | ||||||
| "[P]articipants...were randomly assigned...using a computerized randomization program" (p. 109). However, no information on allocation concealment reported. "The three groups did not significantly differ on age, education, minority race, marriage, sex, occupation, and mental disorder diagnosis" (p.110). | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "All analyses were conducted using the intent‐to‐treat (ITT) principle..." (p.110). | ||||||||||||
| Data were available for 72.0% of those randomised. Sensitivity analyses were not undertaken. "...the cumulative dropout rate [at post‐intervention] was 32.9% (n = 27) for the cognitive therapy group, 30.0% (n = 24) for the telephone intervention group, and 20.8% (n = 16) for the control group; the proportion of participants with missed assessment of three groups was not significantly different" (p.110). No other information on causes of missingness reported. | ||||||||||||
| "...research assessors followed up patients by telephone using measures of a detailed structured questionnaire...[to assess] suicidal history of patients" (p.110). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant blinding could have been convincingly achieved. If clinical personnel were acting as outcome assessors, they may have influenced assessment of the outcome. However, as the intervention was not found to be significant, this is unlikely. | ||||||||||||
| "At the 3 months, 6 months, and 12 months postbaseline assessments, the research assessors followed up patients..." (p.110). All eligible reported results for repetition of SH correspond to all intended outcome. "...research assessors followed up patients by telephone using measures of a detailed structured questionnaire...[to assess] suicidal history of patients" (p.110). All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as some concerns for risk of bias as missingness in the outcome could have depended on its true value and analyses did not correct for bias, nor were sensitivity analyses undertaken to investigate the potential effect of missing data. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Amadéo 2015 | | | | | | | ||||||
| "A sequence based on a random procedure (block randomization) was used to assign all enrolled subjects" (p.49). However, detail on the randomisaton method was not reported. Additionally, no information on allocation concealment reported. "Over half of subjects in the [intervention] group had a diagnosis of mood disorder. A large proportion of cases (63% in the [intervention], 72% in [control] group) had previous history of suicidal behaviors before the index episode" (p.49). These differences may have been compatible with chance. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant or clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. Analyses were made "...using intention to treat approach" (p.52). | ||||||||||||
| Data were available for 95.0% of those randomised. | ||||||||||||
| Repetition of SH was ascertained from "several self‐report scales" (p.49). "Additional information from the hospital emergency department was obtained because participants may fail to report new episodes of non‐fatal suicidal behavior" (p.49). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. No specific information on whether outcome assessors were aware of participants' treatment allocation was reported. Repetition of SH as determined by clinical and hospital records represents an observer‐reported outcome. | ||||||||||||
| "This randomized controlled trial used number of suicides and repeated non‐fatal suicidal behavior (NFSB), as primary outcome measures [at 18 months]" (p.49). Repetition of self‐harm was ascertained from participant self‐report. All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at some concerns for risk of bias as no specific information on randomisation or allocation concealment was reported. Additionally, no specific information was reported as to whether the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis. | ||||||||||||
| Fleischmann 2008 | | | | | | | ||||||
| "An allocation sequence based on a random‐number table was used to randomly assign all enrolled subjects....the allocation sequence was maintained in a separate location to prevent...bias" (p. 704). "No differences in the sociodemographic variables and items related to the current attempt between the TAU and [intervention] groups were found among the subjects analysed" (p.705). | ||||||||||||
| "The subjects were blinded as to their assignment" (p. 704). No information was reported on clinical personnel blinding, however, given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. There was no information as to whether intention‐to‐treat analyses had been conducted. Subsequent analyses appear to be based on those participants with available information suggesting modified intention to treat analyses were used. | ||||||||||||
| Data were available for 89.1% of those randomised (by the 18‐month assessment). Sensitivity analyses were not undertaken. Data were available for 93.6% of those randomised to the intervention arm, and 84.6% of those randomised to the comparator arm. However, the authors do note that "[n]o subjects at all were lost during follow‐up in Yuncheng; 3%, 4%, 11%, and 15% were lost at the final follow‐up in Campinas, Karaj, Colombo, and Chennai respectively" (p.705), suggesting there may have also been regional differences. | ||||||||||||
| "The questionnaire for the comprehensive assessment of all suicide attempters enrolled was commonly applied across all sites, translated into the local language of each site, adapted to take into account cultural specificities, and pilot‐tested to assess face and content validity. It was largely based on the European Parasuicide Study Interview Schedule (EPSIS), which had been applied in the WHO/EURO Multicentre Study on Suicidal Behaviour" (p.705). However, previous work has demonstrated that self‐harm prevalence estimates derived from self‐report may be underestimated, and supplementing prevalence estimates with medical or clinical record information is advisable (Mars 2016). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The subjects were blinded as to their assignment" (p. 704). | ||||||||||||
| "The primary study outcome measurement was...at 18‐month follow‐up" (p.703). All eligible reported results for repetition of SH correspond to all intended outcome. Repetition of self‐harm was ascertained from participant self‐report. All eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as high risk of bias as missingness in the outcome likely depended on its true value and analyses did not correct for bias, nor were sensitivity analyses undertaken to investigate the potential effect of missing data. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 9.1.1 Whole sample | ||||||||||||
| Gysin‐Maillart 2016 | | | | | | | ||||||
| "Simple randomization was carried out using shuffled unmarked sealed envelopes. Similar to a modified Zelen design...participants were contacted following randomization and informed that they had been allocated to one of the two groups...Written informed consent was obtained from participants of both groups after detailed description of the study" (p.3). Referring to Table 1 (p.9), those assigned to the intervention arm were less likely to have engaged in multiple episodes of SH (i.e., ≥1) (OR 0.58, 95% CI 0.28 to 1.20) but had attended a greater number of outpatient treatment sessions in the six months preceding trial entry (i.e., 8 versus 2.5, p<0.01). This is may have been compatible with chance. | ||||||||||||
| Given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that participant and clinical personnel blinding could have been convincingly achieved. Referring back to the trial protocol (NCT02505373), there was no apparent deviation from the intended intervention. "Analysis was conducted on the intention‐to‐treat (ITT) population to ensure unbiased comparisons between the two groups and to address concerns regarding withdrawal and noncompliance. To address the issue of missing data, multiple imputations by chained equations were implemented" (p.6). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The questions related to suicidal behavior and self‐harm were as follows: (1) 'Did you in the past 6 months try to take your own life?'; (2) 'How often did you in the past 6 months have thoughts about taking your own life?'; and (3) 'How often did you in the past 6 months physically harm yourself (e.g., cutting, burning, hitting)?'...we [also] searched hospital records and contacted the responsible general practitioners and therapists to check for suicide attempts..." (p.6). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. This was an open label trial. Self‐reported repetition of self‐harm was supplemented with clinical and hospital records which represents an observer‐reported outcome. | ||||||||||||
| Referring back to the trial protocol (NCT02505373), there were no apparent departures from the analysis plan. Referring back to the trial protocol (NCT02505373), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (NCT02505373), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at low risk of bias as all domains were rated as at low risk of bias. | ||||||||||||
| Hatcher 2015 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomisation system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors between the study groups" (p.231). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "There were no significant differences between those allocated to the two groups who provided consent to be in the study" (p.232). | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Analysis of the primary outcome was in everyone who was randomised..." (p.232). "Analysis of the primary outcome was in everyone who was randomised... The secondary outcomes were analysed only in those people who had consented to be in the study” | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The primary outcome measure was re‐presentation to any hospital in New Zealand for self‐harm...This was assessed by interrogating district health board electronic systems and Ministry Of Health Information Directorate data" (p.230). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were masked to treatment allocation" (p.231). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000641291), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at low risk of bias as all domains were rated as at low risk of bias. | ||||||||||||
| Hatcher 2016 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomization system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors..." (p.888). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "Comparing participants who consented there were no significant differences between those randomised to the intervention and control groups except for the Beck Hopelessness Scores which were significantly lower in the controls [MD ‐3.00, 95% CI ‐5.19 to ‐1.41]" (p.888). This is unlikely to have been compatible with chance. | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, "the introduction to the study differed depending on which arm of the trial [the participant was] randomised to" (p.3, unpublished manuscript). Given the difference in therapeutic intensity between the intervention and comparator arms, it is also unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "The intention to treat analyses consisted of all patients that were randomized..." (p.892). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| Repetition of self‐harm was ascertained from "a review of DHB records; and by using the National Minimum Dataset from the Ministry of Health Information Directory to record hospital contacts...[and t]he National Minimum Dataset [which] contains routinely collected information on all hospital discharges in New Zealand linked to a patient’s individual National Health Index number" (pp.887‐8). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were blind to treatment allocation" (p.888). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000952246), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as some concerns for risk of bias as there was a higher severity of hopelessness at baseline amongst those allocated to the comparator arm, suggesting there may have been a problem with the randomisation process. | ||||||||||||
| Subgroup 9.1.2 First SH episode | ||||||||||||
| Hatcher 2015 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomisation system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors between the study groups" (p.231). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "There were no significant differences between those allocated to the two groups who provided consent to be in the study" (p.232). | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Analysis of the primary outcome was in everyone who was randomised..." (p.232). "Analysis of the primary outcome was in everyone who was randomised... The secondary outcomes were analysed only in those people who had consented to be in the study” | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The primary outcome measure was re‐presentation to any hospital in New Zealand for self‐harm...This was assessed by interrogating district health board electronic systems and Ministry Of Health Information Directorate data" (p.230). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were masked to treatment allocation" (p.231). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000641291), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at low risk of bias as all domains were rated as at low risk of bias. | ||||||||||||
| Hatcher 2016 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomization system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors..." (p.888). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "Comparing participants who consented there were no significant differences between those randomised to the intervention and control groups except for the Beck Hopelessness Scores which were significantly lower in the controls [MD ‐3.00, 95% CI ‐5.19 to ‐1.41]" (p.888). This is unlikely to have been compatible with chance. | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, "the introduction to the study differed depending on which arm of the trial [the participant was] randomised to" (p.3, unpublished manuscript). Given the difference in therapeutic intensity between the intervention and comparator arms, it is also unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "The intention to treat analyses consisted of all patients that were randomized..." (p.892). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| Repetition of self‐harm was ascertained from "a review of DHB records; and by using the National Minimum Dataset from the Ministry of Health Information Directory to record hospital contacts...[and t]he National Minimum Dataset [which] contains routinely collected information on all hospital discharges in New Zealand linked to a patient’s individual National Health Index number" (pp.887‐8). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were blind to treatment allocation" (p.888). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000952246), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as some concerns for risk of bias as there was a higher severity of hopelessness at baseline amongst those allocated to the comparator arm, suggesting there may have been a problem with the randomisation process. | ||||||||||||
| Subgroup 9.1.3 Repeat SH episode | ||||||||||||
| Hatcher 2015 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomisation system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors between the study groups" (p.231). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "There were no significant differences between those allocated to the two groups who provided consent to be in the study" (p.232). | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, given the difference in therapeutic intensity between the intervention and comparator arms, it is unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "Analysis of the primary outcome was in everyone who was randomised..." (p.232). "Analysis of the primary outcome was in everyone who was randomised... The secondary outcomes were analysed only in those people who had consented to be in the study” | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| "The primary outcome measure was re‐presentation to any hospital in New Zealand for self‐harm...This was assessed by interrogating district health board electronic systems and Ministry Of Health Information Directorate data" (p.230). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were masked to treatment allocation" (p.231). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000641291), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000641291), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as at low risk of bias as all domains were rated as at low risk of bias. | ||||||||||||
| Hatcher 2016 | | | | | | | ||||||
| "All eligible participants were allocated randomly to the intervention or usual care groups using a central computerised randomization system at the Clinical Trials Research Unit (subsequently the National Institute for Health Innovation)...Stratified minimisation randomization was used to ensure a balance in key prognostic factors..." (p.888). This trial implemented Zelen’s post‐consent method. Therefore, following allocation participants were given the option to change treatment arms. "Comparing participants who consented there were no significant differences between those randomised to the intervention and control groups except for the Beck Hopelessness Scores which were significantly lower in the controls [MD ‐3.00, 95% CI ‐5.19 to ‐1.41]" (p.888). This is unlikely to have been compatible with chance. | ||||||||||||
| As this trial implemented Zelen’s post‐consent method, participants were given the option to change treatment arms following allocation.Therefore, participants cannot have been blind to treatment allocation. Additionally, "the introduction to the study differed depending on which arm of the trial [the participant was] randomised to" (p.3, unpublished manuscript). Given the difference in therapeutic intensity between the intervention and comparator arms, it is also unlikely that clinical personnel blinding could have been convincingly achieved. There were no apparent deviations from the intended intervention. "The intention to treat analyses consisted of all patients that were randomized..." (p.892). | ||||||||||||
| Data were available for 100% of those randomised. | ||||||||||||
| Repetition of self‐harm was ascertained from "a review of DHB records; and by using the National Minimum Dataset from the Ministry of Health Information Directory to record hospital contacts...[and t]he National Minimum Dataset [which] contains routinely collected information on all hospital discharges in New Zealand linked to a patient’s individual National Health Index number" (pp.887‐8). Repetition of self‐harm was ascertained in the same way between the intervention and comparator arms in this trial. "The research assistants were blind to treatment allocation" (p.888). | ||||||||||||
| Referring back to the trial protocol (ACTRN12609000952246), there were no apparent departures from the analysis plan. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended outcome measurements. Referring back to the trial protocol (ACTRN12609000952246), all eligible reported results for repetition of SH correspond to all intended analyses of the data. | ||||||||||||
| Overall, this trial was rated as some concerns for risk of bias as there was a higher severity of hopelessness at baseline amongst those allocated to the comparator arm, suggesting there may have been a problem with the randomisation process. | ||||||||||||
