« Previous | Contents | Next »
Listen
ANNEX E: OVERVIEW OF SYSTEMATIC REVIEWS RETRIEVED
E.1 It did not fall within the remit of the current review to provide an account of prior reviews or meta-analyses. However, we present here a basic overview of the systematic review and meta-analytic material retrieved, because it is comparatively extensive, because people may wish to refer back to earlier reviews, and because it provides some insight into the need for the more 'over-arching' approach adopted in the current review. Tables E1 to E4 below provide brief summary details of each of the reviews and meta-analyses retrieved. In the following discussion,we will not distinguish between meta-analyses and systematic reviews with or without meta-analysis, as the criteria for inclusion were the same for each. To draw in as wide a range of comparatively high quality prior reviews as possible we set the criteria for a 'systematic' approach fairly loosely, requiring only that reviews demonstrate a systematic and replicable search strategy, that the quality of included studies was addressed (but not necessarily that studies were excluded from the review on the basis of poor quality) and that either a narrative or quantitative synthesis of the included data was attempted.
E.2 As noted previously, systematic reviews not referring to suicide, self-harm or suicidal ideation as an explicit outcome are unlikely to have been included. The bulk (49%) of the systematic reviews identified as meeting our criteria addressed pharmaceutical intervention. Whilst a high proportion of primary studies also relate to pharmaceutical intervention, the research bias towards such interventions is more pronounced in systematic reviews. Only five of the reviews identified explored psychotherapeutic/psychosocial interventions and only one addressed educational interventions for the general population (adolescents). The remaining 13 reviews either sought, as with our own review, to identify and evaluate the evidence base for any type of intervention, or focussed on a diverse range of other specific options for intervention. Included within the latter category are reviews relating to the efficacy of suicide prevention centres, 'no-suicide' contracts, Electro-Convulsive Therapy ( ECT), interventions in Accident and Emergency (A&E), contacts with health care and liaison psychiatry.
E.3 Around one third (35%) of the reviews included only material derived from RCTs. A further five reviews included only controlled trials (whether randomised or not). No review reported on outcomes from qualitative research, although nine reviews set their inclusion criteria to retrieve all available studies regardless of design. Nearly half of the reviews (46%) restricted their population of interest to people with a specific psychiatric disorder, primarily depression. Reviews addressing non-pharmaceutical interventions tended to set more open population criteria and were more likely to focus explicitly on people already known to be engaging in suicidal behaviours or ideation.
E.4 In total, 43% of the reviews reported positive outcomes in that they felt adequate evidence existed to conclude that an intervention was effective. The majority of these reviews (11 of 16) related to pharmaceutical interventions. To briefly summarise outcomes, of the 18 pharmaceutical reviews, 11 reported positive outcomes, 5 were equivocal or suggested that further evidence was needed and 2 reported a tendency for suicidal behaviour/ideation to worsen or to show higher incidence where the drug in question was administered. Positive outcomes related to the use of lithium for bipolar, affective and mood disorders; the use of alprazolam for depressed patients; the use of fluvoxamine or paroxetine in depression and fluoxetine in depression and mood disorders and the use of clozapine in patients with schizophrenia or schizoaffective disorders.
E.5 Adverse outcomes (suicidal symptoms increasing or becoming more prevalent during treatment) were reported for the use of levetiracetam in patients with epilepsy (although the authors themselves concluded this related to features of the epilepsy rather than of the drug itself) and naltrexone in opioid dependent individuals. Equivocal outcomes or calls for further evidence were reported for the most recent study of lithium in mood disorders (Burgess et al 2006), for lithium or fluoextine as adjunctive treatments to olanzapine; for SSRIs in the adult population in general and in depressed patients, and pharmacological treatment in general in people with Borderline Personality Disorder.
E.6 The conclusions reached by the authors of all five reviews identified as addressing psychotherapeutic/psychosocial interventions are equivocal at best. However, the use of Cognitive Behavioural Therapy ( CBT) is supported by two reviews as either 'promising' or evidence-based in people already known to have self-harmed or attempted suicide. Other specific interventions cited as promising by one review are Dialectical Behaviour Therapy ( DBT) and green card initiatives (providing a patient with a contact card to arrange readmission). The remaining two reviews in this category flag, respectively, a finding that psychosocial interventions generally may be more effective within higher risk groups and note that there is insufficient evidence available to justify the introduction of any psychosocial programmes for the treatment of depression.
E.7 The one review we identified addressing general population educational initiatives (Ploeg et al 1996) concluded that there is insufficient evidence currently available to support curriculum-based initiatives for adolescents. Within the rather broader category of reviews addressing 'any' or 'other' interventions, there are also few positive messages. Only two reviews report unequivocally positive outcomes. These support the preventive benefits of physician education in recognising and treating depression; restriction of access to means and, in a generic review of all available interventions, CBT and interpersonal therapy. In addition, one review (Gunnell et al 2005b) provides evidence-based recommendations that strategies to reduce suicide by hanging should focus on controlled environments, the emergency management of 'near hanging' and on suicide prevention in general.
E.8 The remaining reviews either conclude that there is no convincing evidence base for any type of intervention, or cite a lack of evidence and need for more research in relation to specific options for prevention. Such options include suicide prevention centres, no-suicide contracts, prevention programmes in general hospital and A&E settings, ECT, contact with clinicians and liaison psychiatry. In the case of liaison psychiatry the review concludes more strongly (Rudd et al 2005) that many areas of liaison psychiatry are not evidence-based.
E.9 The above reviews provide useful outcomes which, in the main, do not differ from our own conclusions in respect of the specific evidence bases to which they refer. However, they are all comparatively narrow in focus in respect either of study design, population, or the number of studies included. One exception to this is Gunnell et al 2005a, which, although restricting study design to RCTs, otherwise goes to considerable lengths to identify all available relevant evidence. However, even in the latter case, the justifiable focus on a single intervention limits the usefulness of the review for practitioners and policy makers, who need to know not just whether any specific intervention option is effective, but also which approaches are, on balance, the best available.
E.10 A key advantage of broad approaches such as that used for the current review is that direct comparisons can be drawn between interventions, populations, settings and also types of evidence. This option for head-to-head comparison of outcomes modified by the key parameters which are of interest to practitioners and policy makers is a substantial advantage of an 'holistic' approach to the review process. Analogously to the advantage of collecting standardised data on two interventions in a single trial, the advantages of collecting evidence across a range of interventions and intervention parameters following a common paradigm for data collection at the same point in time, provides considerable insight into the current state of the evidence base which could not be drawn from a series of independent reviews on distinct interventions conducted in different ways and at different points in time.
Table E.1 Summary of Systematic Reviews Retrieved : Pharmaceutical Interventions
Study ID | Intervention | Design | Population | N of Studies Included | Main Conclusion(s) |
|---|
Pharmaceutical | | | | | |
|---|
Baldessarini et al 2003 | Lithium | No specific design, reviews pharmaceutical studies whether or not lithium is evaluated | Bipolar Disorder | 34 | Major reductions in suicide attempts with lithium maintenance |
|---|
Binks et al 2006 | Pharmacological | RCTs | Borderline Personality disorder ( BPD) | 10 | Current research provides inadequate evidence |
|---|
Burgess et al 2006 | Lithium | RCTs | Mood Disorders | 9 | No definitive evidence for whether or not lithium has an anti-suicidal effect |
|---|
Cipriania et al 2005 | Lithium | RCTs | Mood disorders | 32 | Data from 7 trials suggests patients receiving lithium less likely to die by suicide; composite measure of suicide plus self-harm also lower |
|---|
Cramer et al 2003 | Levetiracetam ( LEV) | Placebo controlled trials | Adults with epilepsy, cognitive disorders or anxiety disorder | Not stated | Suicidal symptoms were significantly more common amongst patients with epilepsy treated with LEV than amongst similarly treated patients with cognitive disorders or anxiety disorders |
|---|
Digiusto et al 2004 | Pharmacotherapy | Prospective Longitudinal follow-up of post-trial data | Opioid dependent individuals | 12 | Clinicians should alert addicts taking naltrexone of the possible risks of heroin overdose |
|---|
Filteau et al 1993 | Selective Serotonin Re-uptake Inhibitors ( SSRIs) | Double blind clinical trials | Depressed patients | 11 | Significant rapid and effective lessening of suicidal ideation during SSRI treatment |
|---|
Gunnell et al 2005 | SSRIs | RCTs | Adults | 477 (including trials not specifically addressing suicide or self-harm but reporting on adverse events which include these) | Increased risks of suicide and self-harm caused by SSRIs cannot be ruled out, but the risks need to be weighed against the effectiveness of SSRIs in treating depression |
|---|
Jonas & Hearron 1996 | Alprazolam | Controlled trials | Depressed patients | 22 | There was no significant difference between alprazolam and other active comparators in either increasing or decreasing suicidal ideation; alprazolam is superior to placebo in reducing suicidal ideation |
|---|
Khan et al 2003 | Anti-depressants including SSRIs | FDARCTs | Depressed patients | Unclear, but includes trials not directly addressing the issue of suicide or self-harm but reporting these as adverse events | There is no support for either an overall difference in suicide risk between anti-depressant and placebo treated subjects or for a difference between SSRIs and other types of anti-depressant |
|---|
Letizia et al 1996 | Fluvoxamine | RCTs | Major depressive disorder | 19 | Treatment with fluvoxamine is associated with a significantly greater improvement in suicidal ideation than placebo |
|---|
Montgomery et al 1995 | Paroxetine | Controlled and open trials | Major depression | Unclear, but includes studies not directly focussed on but reporting suicidal ideation as an adverse event | Fewer instances of suicidal ideation emerge in paroxetine treated patients compared with placebo |
|---|
Szanto et al 2003 | Paroxetine or nortriptyline | 'federally funded treatment studies' | Depression in the elderly | 3 | Suicidal ideation resolves rapidly with pharmaceutical treatment but resolution of thoughts about death is more gradual |
|---|
Tollefson et al 1993 | fluoxetine | RCTs | Patients with mood disorders vs patients with non-mood disorders | 63, including trails which did not address suicidality directly but report adverse incident data | Substantial suicidal ideation emerged less frequently with fluoxetine than with placebo in patients with mood disorders. Too few occurrences of suicidality in patients with non-mood disorders were noted to draw comparisons |
|---|
Tondo et al 2001 | lithium | Any design providing suicide rates during lithium maintenance therapy | Affective disorders | 22 | Suicide risk was consistently lower during long-term treatment with lithium in all identified studies |
|---|
Vieta et al 2004 | Olanzapine with adjunctive lithium or fluoxetine | Short-term blinded trials with open-label extensions | Rapid-cycling vs non-rapid cycling bipolar disorder | 2 | Adjunctive use of lithium or fluoxetine was not associated with suicide attempts, no significant differences were noted between rapid-cycling and non-rapid cycling bipolar disorder |
|---|
Wagstaff & Perry 2003 | clozapine | All available designs | Patients with schizophrenia or schizo-affective disorder | 9 (1 RCT, 3 prospective studies, 5 retrospective studies) | In the one RCT available, (Inter SEPT trial) clozapine had a greater preventive effect on suicidality at patients at high risk of suicidality than olanzapine |
|---|
Wernicke et al 1997 | Fluoxetine combined with adjunctive centrally-acting medication | RCTs | Depression | 25, including studies not directly addressing suicide or self-harm but including adverse incident data | Fluoxetine is associated with a significantly superior reduction in suicidal acts and ideation than placebo, independently of concomitant medication. Fluoxetine is superior to TCAs in patients not taking concomitant medication in respect of suicidal acts and ideation |
|---|
Table E.2 Summary of Systematic Reviews Retrieved: Psychotherapeutics/Psychosocial interventions
Study ID | Intervention | Design | Population | N of Studies Included | Main Conclusion(s) |
|---|
Psychotherapeutic/Psychosocial | | | | | |
|---|
Hepp et al 2004 | Psychological & psychosocial interventions | RCTs | People who had attempted suicide or engaged in deliberate self-harm | 25 | Minimal interventions (e.g. green card initiatives) and psychodynamic interventions (e.g. CBT and DBT) show promise but more research is needed to provide adequate evidence |
|---|
Linehan 1997 | Psychosocial and behavioural interventions | RCTs and studies assigning participants using an alternating sequential design | Any reported population | 20 (of which, 18 RCTs) | Psychosocial interventions appear to be most effective with the more high-risk individuals |
|---|
Macgowan 2004 | Psychosocial treatments | All available designs | Adolescents | 10 | A number of treatments are cited as promising, but the authors conclude that current evidence of efficacy is weak and research designs are poor |
|---|
Merry et al 2004 | Psychological and/or psycho-educational interventions | RCTs | Children & adolescents | 13 | There is insufficient evidence to warrant the introduction of depression prevention programmes to reduce suicide attempts and completed suicide |
|---|
Vandersande et al 1997 | Psychosocial intervention | RCTs | Suicide attempters | 15 | Currently there is evidence only to support CBT approaches in preventing repeated suicide attempts |
|---|
Table E.3 Summary of Systematic Reviews Retrieved: Educational Interventions
Study ID | Intervention | Design | Population | N of Studies Included | Main Conclusion(s) |
|---|
Educational/Training | | | | | |
|---|
Ploeg et al 1996 | Curriculum-based prevention programmes | Prospective studies with a control group or before/after evaluation | Adolescents | 11 | There is currently insufficient evidence to support curriculum-based prevention programmes. The evidence suggest there may be both beneficial and harmful effects on attitudes related to suicide |
|---|
Table E.4 Summary of Systematic Reviews Retrieved: 'Any' or 'Other' Interventions
Study ID | Intervention | Design | Population | N of Studies Included | Main Conclusion(s) |
|---|
Other/'Any initiative' | | | | | |
|---|
Comtois 2002 | Any intervention for parasuicide | Experimental and quasi-experimental designs | Parasuicidal individuals | Not stated | Empirically supported treatments are rarely used as part of usual care; standard treatments including hospitalization of expensive |
|---|
Dew et al 1987 | Suicide Prevention Centres | Any available design | Unspecified | 7 | There is no evidence for the efficacy of suicide Prevention Centres in reducing the suicide rate |
|---|
Gould et al 2003 | Any intervention | Any design | Young people | Unclear | Whilst several interventions are promising, none have proven efficacy |
|---|
Gunnell et al 2005b | Interventions to reduce suicide by hanging | All available designs | Any reported population | unclear | Strategies to reduce suicide by hanging should focus on controlled environments , the emergency management of 'near hanging' and on the primary prevention of suicide in general |
|---|
Hawton et al 1998 | Psychological and pharmacological treatments | RCTs | Patients who had deliberately harmed themselves | 20 | There remains considerable uncertainty about which forms of psychosocial and physical treatments of patients who harm themselves are most effective. |
|---|
Kelly & Knudson 2000 | 'no suicide' contracts | All available designs | Any reported population | 0 | No evidence exists to evaluate the efficacy of no-suicide contracts |
|---|
Links & Hoffman 2005 (this updates an earlier review by Gunnell & Frankel 1994) | Prevention programmes in psychiatric wards or units within general hospitals | All available designs | Psychiatric patients | 34 | A number of programme and policy recommendations are made but these are based on the potential of current treatments to reduce suicide risk, firm evidence for efficacy is not presented. |
|---|
Mann et al 2005 | Any prevention initiative | Systematic reviews & meta-analyses, RCTs, cohort studies, ecological or population based studies | Any reported population | Systematic reviews & meta-analyses (10), RCTs (18), cohort studies (24) ,ecological or population based studies 41) | Physician education in depression recognition and treatment and restricting access to lethal methods reduce suicide other interventions require more evidence of efficacy |
|---|
O'Leary et al 2001 | ECT and anti-depressant availability | Follow-up studies | Affective disorder | 75 | The availability of ECT and anti-depressants may have contributed to decline in suicide during follow-up |
|---|
Pirkis & Burgess 1998 | Contacts with health care | All available designs | People known to have committed suicide (Retrospective analysis) | 24 | Contact with clinicians may help prevent suicide but more evidence is needed, in particular to identify which risk groups this applies to |
|---|
Repper 1999 | Interventions in A&E | Unclear, authors refer to RCTs only in tabulated data but imply a wider retrieval of study design in the text | Persons presenting to A&E | Unclear, authors refer only to 7 RCTs in tabulated data, to 8 studies in the abstract and to 'all UK studies' in the text | There is inadequate information regarding the targeting of clients at risk of suicide, no specific intervention strategy in A&E has proven efficacy |
|---|
Ruddy & House 2005 | Liaison psychiatry | Systematic reviews | Any persons in contact with liaison psychiatry | 64, including studies not directly addressing suicide or self-harm but recording relevant outcomes | Many areas of liaison psychiatry are not based on high quality evidence, more research is needed |
|---|
Ryan 2005 | Any treatment for depression | Unclear, the authors do not cite any exclusion criteria based on design, but text and tables refer only to controlled trials | Children and adolescents with depression | Unclear | Cognitive behavioural therapy and interpersonal therapy are better than treatment as usual ; several anti-depressants are more efficacious than placebo, there is a correlation between treatment with SSRIs and a decrease in completed suicide, however comparing all anti-depressants as a single group the association is with an increase in suicide |
|---|
« Previous | Contents | Next »