You are on page 1of 60

The British Journal of Psychiatry

1–7. doi: 10.1192/bjp.bp.116.187799

Review article

Strategies to prevent death by suicide:


meta-analysis of randomised controlled trials
Natalie B. V. Riblet, Brian Shiner, Yinong Young-Xu and Bradley V. Watts

Background
Few randomised controlled trials (RCTs) have shown of suicide (OR = 0.20, 95% CI 0.09–0.42). Six RCTs (n = 1040)
decreases in suicide. of cognitive–behavioural therapy (CBT) for suicide prevention
and six RCTs of lithium (n = 619) yielded non-significant
Aims findings (OR = 0.34, 95% CI 0.12–1.03 and OR = 0.23, 95%
To identify interventions for preventing suicide. CI 0.05–1.02, respectively).
Method
We searched EMBASE and Medline from inception until Conclusions
31 December 2015. We included RCTs comparing prevention The WHO BIC is a promising suicide prevention strategy.
strategies with control. We pooled odds ratios (ORs) for No other intervention showed a statistically significant effect
suicide using the Peto method. in reducing suicide.

Results Declaration of interest


Among 8647 citations, 72 RCTs and 6 pooled analyses met None.
inclusion criteria. Three RCTs (n = 2028) found that the
World Health Organization (WHO) brief intervention and Copyright and usage
contact (BIC) was associated with significantly lower odds B The Royal College of Psychiatrists 2017.

Suicide is a worldwide health concern. According to the World English language. We required that studies randomly assign
Health Organization (WHO), over 800 000 people die by suicide patients to an intervention aimed at suicide prevention or a
every year.1 The most comprehensive review of available suicide control condition including usual care, placebo or wait-list. We
prevention strategies was published by Mann et al in 2005 and included studies in which patients were aged 18 years or older.
updated by Zalsman et al in 2016.2,3 These reviews concluded that To broaden our search, we included studies if they reported death
restricting access to lethal means prevents suicide, and that by suicide as a primary or secondary outcome. In the event that
clozapine and lithium exert an anti-suicidal effect.2,3 However, death by suicide was a secondary outcome, we required that the
other reviews have drawn different conclusions about the role of primary study aim included the prevention of suicidal ideation
pharmacotherapy, psychotherapy and psychosocial interventions and/or behaviour. We included studies even if there were no
in suicide prevention.4–12 Notably, many of these reviews rely suicide events because this is a more conservative approach and
heavily on observational data to formulate their opinions, and improves the generalisability of our findings.17
their conclusions are largely based on improvements in inter- We included pooled analysis of RCTs in our review. Pooled
mediary outcomes including suicidal ideation and attempts analyses, which use a systematic method to identify suicides in
(rather than death by suicide).4-12 These intermediate outcomes drug trials (e.g. US Food and Drug Administration summary
are susceptible to measurement bias and may not predict basis of approval reports), may offer unique insights into the
suicide.13 The WHO has emphasised the critical need to identify relationship between suicide and psychiatric medications. Because
interventions with proven efficacy for preventing death by it is currently unclear which interventions prevent suicide, we
suicide.1 To address these concerns, we conducted a meta-analysis excluded RCTs that compared two or more active treatments.
of randomised controlled trials (RCTs) comparing the efficacy of Instead, we provided a qualitative summary of these trials in
various interventions versus control to prevent death by suicide online Appendix DS2.
in adults. We believe our review will inform clinical practice and We acknowledge that there are inherent challenges in
illuminate promising areas in need of additional research, by using RCTs to study rare outcomes such as suicide and that
focusing on the hard outcome of death by suicide.14 non-randomised, controlled studies may provide important
information about strategies to prevent suicide. We focused our
Method review on RCTs, however, because this study design is the gold
standard for establishing efficacy, and a sufficient number of RCTs
Review protocol have explored suicide prevention.18
In preparation for our review, we drafted a study protocol
delineating our planned approach for identifying relevant studies.
We did not register our protocol with Prospero, but a copy is Study identification
available in online Appendix DS1. We used the standard We searched EMBASE, Medline, CINAHL, the Cochrane
methodology outlined by Cochrane for our analysis and the Library, and PsycINFO from the inception of each database until
PRISMA guidelines for reporting our methods and findings.15,16 31 December 2015 to identify published articles addressing our
research question. We used exploded MeSH terms and key words
Eligibility criteria to generate the following themes: death by suicide, prevention/
We defined a priori inclusion criteria for this review. We limited control and treatment. We used ‘OR’ to combine prevention/
our review to RCTs and pooled results of RCTs published in the control and treatment. We then used the Boolean term ‘AND’ to

1
Riblet et al

find the intersection between this collective theme and death by domains reported in Mann et al’s review (screening for patients
suicide. We applied a highly sensitive search strategy to identify at high risk, and media reporting guidelines for suicide), because
RCTs in electronic databases. The details of the search strategy we found no related RCTs that met our inclusion criteria.2
are available in online Appendix DS1. To promote homogeneity, we stratified each of the inter-
We attempted to locate additional published and unpublished vention domains based on the targeted population of interest
studies by searching ClinicalTrials.gov from inception to 31 (e.g. patients with schizophrenia). For each domain, we evaluated
December 2015 and reviewing the references of prior review whether the intervention or control was favoured in each study
papers and included articles. When necessary, we contacted and we then determined whether these results were statistically
investigators to determine whether an eligible study met our significant. We then evaluated the effect of combining these
inclusion criteria (online supplemental references35–43). In the studies on the magnitude, direction and statistical significance
event that authors did not respond to our requests, we excluded of the overall summary estimate of the effect size.
these studies (online supplemental references40–43). We used the standard definition of an OR to interpret our
results (i.e. probability of an event occurring versus the probability
that the event will not occur in the intervention versus the
Primary end-point and data abstraction control). We deemed that an OR less than one meant that the
We defined death by suicide using the Centers for Disease Control relative odds of death by suicide was smaller in the intervention
and Prevention (CDC) definition, ‘death caused by self-directed versus the control, and the opposite was true if the OR was greater
injurious behavior with an intent to die as a result of the than one.24 We considered OR to be statistically significant if the
behavior.’19 95% confidence interval did not cross one.
Based on a priori inclusion criteria, two reviewers (N.B.V.R. We used RevMan 5.3 to pool our results.25 We assessed
and B.S.) screened the titles and abstracts of all potentially relevant groupings for heterogeneity using Cochrane’s Q and the I 2
studies. The same reviewers then assessed the full text of the statistic.22 We used a conventional threshold of P50.10 and
remaining studies to make a final determination regarding I 2450% to indicate statistical significance and meaningful
eligibility for review. In the event that it was unclear whether a heterogeneity, respectively.22
full text met all eligibility criteria, a separate reviewer (B.V.W.
independently evaluated these texts for study inclusion.
Discrepancies were resolved through consensus. Confirmatory analysis
Using a piloted, standardised data collection form, two Many concerns have been raised about the validity of available
reviewers (N.B.V.R. and B.S.) extracted data in duplicate from methods for conducting meta-analysis of rare events such as
included studies. We extracted data related to demographics, suicide.26 To address these concerns, we felt that it was appropriate
methods, outcomes and risk of bias. We used the Cochrane to perform a confirmatory analysis. We used a Poisson regression
Risk of Bias Tool to assess study quality.20 In the case of model with random effects and calculated an incidence rate ratio
multiple reports of the same data-sets (online supplemental (IRR) for suicides over person-year for each domain of
references35,36,38,44–72), we selected the study that included the strategies. This approach accounts for differences in exposure time
most comprehensive and up-to-date information (online across studies, addresses any potential heterogeneity between
supplemental references36,60–65,67–72). From pooled analysis, we trials and better accounts for trials with zero events.27,28 The
extracted total number of suicides and person-years of exposure Poisson regression model with random intervention effects has
(number of patients at risk multiplied by the number of years also been used in meta-analysis of rare event data, including
of exposure). To minimise bias, we preferentially selected pooled suicide.27,28 However, because the Poisson regression model is
analyses that limited their analysis to the double-blind period not recommended if there is over-dispersion in the data, we first
and reported person-year exposure. We reviewed pooled analysis performed a boundary likelihood-ratio test, evaluating whether
to ensure that trials were not double counted. Discrepancies were the alpha (the estimate of the dispersion parameter) for domains
resolved through consensus. (and sub-domains) of interest was significantly greater than
zero (defined as P50.05).29 If we encountered significant over-
dispersion, we calculated the IRR using the recommended
Data analysis multilevel mixed-effects negative binomial regression rather than
We evaluated our primary outcome using the Peto method and the Poisson regression.30 Confirmatory analyses were conducted
calculated summary odds ratios (ORs) with 95% confidence using STATA statistical software version 14 (StataCorp).
intervals and P-values.21 More commonly used meta-analysis
methods (e.g. risk ratio) are generally not recommended for the
evaluation of rare outcomes such as suicide.21–23 The Peto method Sensitivity analysis
is a powerful alternative for combining data when event rates are We conducted a sensitivity analysis based on the quality of
below 1%.21–23 We did not apply a continuity correction for trials included studies as judged by the Cochrane Risk of Bias Tool.20
with zero events because this is not recommended with the Peto We evaluated whether the magnitude and direction of the
method.17 summary estimate of each intervention domain changed if we
We formed groups and subgroups of strategies using Mann excluded studies that were judged to be at high risk of bias.
et al’s conceptual framework of suicide prevention–intervention
domains and sub-domains.2 Several interventions included in our
review were not described by Mann et al (i.e. non-cognitive– Assessment of reporting bias
behavioural therapy (CBT), case management, letter/telephone Using STATA, we generated funnel plots for domains that
contact after a suicide attempt, higher-level care interventions, included at least ten studies, and we assessed for publication bias
mood stabilisers, somatic therapies and other classes of using Harbord’s modified test for small-study effects.31 This
antidepressants).2 Therefore, we used consensus among the method uses a modified linear regression analysis to identify
authors to develop additional domains and sub-domains to significant funnel plot asymmetry which would indicate
categorise these interventions. Furthermore, we excluded two publication bias.31 The Harbord method is more powerful in cases

2
Strategies to prevent death by suicide

of few events per trial.31 We considered a P50.05 to suggest


publication bias. 11 866 records identified through 19 additional records
electronic database searching identified through
EMBASE – 3297 other sources
CINAHL – 2171 Reference review – 3
Results Medline – 4641 ClinicalTrials.gov –16
PsycINFO – 1525
As shown in Fig. 1, we identified 8634 records eligible for screening. Cochrane Library – 232

After title and abstract screening, we found that 351 records


remained, and 97 reports (78 studies) met inclusion criteria based 6 6
on full text review (online supplemental references36,60–65,67–137). 8647 unique records from all sources
(duplicates removed)

Characteristics of included studies 7 Ongoing studies – 13


Tables 1–3 display the characteristics of included studies, which
spanned five decades. Over 50% were published in the past 10 6
years. Although most trials studied interventions targeted at 8634 records screened
for eligibility
individuals with known risk factors for suicide, some trials
evaluated interventions in a general or primary care population.
8283 records excluded
For example, one trial investigated a central storage facility for 7 by abstract & title review
pesticides in villages practising floriculture in Southern India
(online supplemental reference94). Except for one trial that 6
351 full-text records
used a wait-list control and one trial that used sham (online assessed for eligibility
supplemental references96,136), all trials of non-pharmacological
interventions used a usual care comparison. Pharmacological 254 records excluded by
studies used pill placebo as a control group. full-text review
7 Incorrect study design – 69
Incorrect population – 9
Assessment of quality Incorrect outcome – 156
No usable data – 3
We identified a number of methodological concerns (online Table 6
DS1). All studies were of randomised design, but many trials had 97 reports (78 studies)
small sample sizes, large losses to follow-up, and/or were of short included in quantitative
synthesis
duration. For non-pharmacological studies, it was difficult to
mask patients/personnel, and the usual care arm was not
standardised across studies. Some authors raised concerns about Fig. 1 PRISMA flow diagram.
the fidelity of the intervention. Several authors suggested that
patients in the usual care arm may have benefited from study
assessments, and these benefits may have obscured the true effect intervention included an educational session on suicide
of the intervention. Finally, although many trials used robust prevention followed by regular contact with a trained provider (by
methods to identify suicide, several trials relied on methods that telephone or in person) for up to 18 months (online supplemental
were either unclear or at high risk for bias, which may have references60,73,74). There was no evidence that other complex
resulted in over- or underestimation of the true effect of the psychosocial interventions reduce the risk of suicide (online
intervention. For example, Sun et al used informants (caregivers) Fig. DS1 and online Table DS2).
to report suicide and suicidal behaviour, but raised concerns that
information may have been withheld owing to the stigma of
suicide (online supplemental references65). Effects of psychotherapy
There were 24 unique RCTs (n = 3056) of psychotherapies to
Analysis of heterogeneity address suicidal behaviour (online supplemental references67,68,
95–116
As shown in online Fig. DS1, we encountered little heterogeneity ). In six trials of CBT for suicide prevention, 3 out of 514
in our analysis, except there was modest heterogeneity when we patients in the intervention group and 10 out of 526 patients in
evaluated intensive follow-up interventions (Q = 7.17 (P = 0.05), the control group died by suicide (Fig. 2). The results, however,
I 2 = 48%). This reflects the distinct differences between the types were not statistically significant (OR = 0.34, 95% CI 0.12–1.03,
of interventions tested. Reassuringly, the heterogeneity resolved P = 0.06; IRR = 0.30, 95% CI 0.08–1.11, P = 0.07) (online
when we categorised these interventions into three sub-domains. supplemental references67,95–99). Common features of the
intervention included reviewing a recent suicide attempt, applying
cognitive strategies and learning relapse prevention. There was no
Effects of complex psychosocial interventions evidence that other CBT or non-CBT therapies reduce the risk of
Twenty-nine RCTs (n = 22 135) reported on complex psychosocial suicide (online Fig. DS1 and online Table DS2).
interventions (online supplemental references36,60–65,73–94). In the
three trials of the WHO brief intervention and contact (BIC)
intervention, 3 out of 1041 patients in the intervention group Effects of pharmacotherapy
and 24 out of 987 patients in the control group died by suicide, There were 14 RCTs (n = 2443) reporting on pharmacotherapy
and the difference was significant (OR = 0.20, 95% CI 0.09–0.42, and death by suicide (online supplemental references117–130). After
P50.0001; IRR not calculable owing to insufficient number of accounting for random effects and length of follow-up, there was
studies) (Fig. 2) (online supplemental references60,73–74). The no evidence among the pooled trials that pharmacotherapy reduced
WHO BIC was tested in low- and middle-income countries as part the risk of suicide (OR = 0.21, 95% CI 0.05–0.86; IRR = 0.10, 95%
of the Multisite Intervention Study on Suicidal Behaviours. The CI 0.00–32.27) (online Fig. DS1 and online Table DS2). In 6 trials

3
4
Table 1 Characteristics of included studies: complex psychosocial interventions a
Riblet et al

Years Study Intervention Control Age, years: Female, Setting, Follow-up, months: Attrition %, Rigour in data
covered N N N mean (s.d.) n (%) Europe: n (%) median (IQR) median (IQR) collectionb

Case management – psychosis 1996–2007 4 771 774 33.1 (11.4) 623 (61) 4 (100) 35 (25) 4 (11) B
Intensive follow-up 1992–2015 11 2447 2613 39.4 (15.2) 2868 (56) 5 (46) 12 (3) 16 (19) A
Aftercare strategy 1989 1 68 73 NR NR 1 (100) 12 (0) 4 (0) B
Letter/telephone contact 1976–2011 7 2863 2922 29.4 (12.6) 3436 (62) 4 (57) 13 (30) 13 (18) A
PCMH integration 2005–2010 3 1754 1633 71.0 (7.6) 1597 (67) 0 (0) 24 (27) 6 (4) B
RN education of caregivers 2013 1 51 54 41.0 (16.6) 33 (54) 0 (0) 12 (0) 42 (0) C
Follow-up with GP after SB 2015 1 101 101 37.8 (NR) 111 (55) 1 (100) 6 (0) 13 (0) A
Lethal means restriction 2013 1 4446 3307 NR 3877 (50) 0 (0) 18 19 A
Psychotherapiesc
CBTd 1987–2015 20 1711 1542 30.7 (11.0) 1818 (54) 9 (45) 12 (14) 19 (16) B
Non-CBTd 2001–2016 5 263 278 33.5 (10.0) 222 (41) 4 (80) 18 (12) 23 (10) B

CBT, cognitive–behavioural therapy; GP, general practitioner; IQR, interquartile range; NR, not reported; PCMH, primary care mental health; RN, nurse; SB, suicidal behaviour.
a. Reported results for age, gender and drop-out are limited to those studies for which data were available.
b. Grading of suicide assessment: A: 550% of trials used rigorous methods (e.g. coroner report); B: 550% of trials methods were unclear; C: 550% of trials methods were at risk for bias (e.g. informant).
c. Except for one trial (online CBT module to prevent suicide) which used a wait-list control (van Spijker et al; online supplemental reference96), all trials of non-pharmacological interventions used a usual care or standard care control condition.
d. A third arm of the CBT trial conducted by Tarrier et al (online supplemental reference107) evaluated supportive therapy for psychosis. This arm was included in the analysis of non-CBT therapies.

Table 2 Characteristics of included studies: pharmacological interventionsa

Years Study Intervention Control Age, years: mean Female, Setting, Follow-up, months: Attrition %, Rigour in data
covered N N N (s.d.) n (%) Europe: n (%) median (IQR) median (IQR) collectionb

Randomised placebo-controlled trials of medications


Antidepressants 1982–2013 7 894 891 54.0 (16.0) 966 (57) 3 (43) 3 (3) 27 (16) B
Lithium 1973–2014 6 257 308 42.0 (10.9) 199 (60) 3 (50) 12 (6) 27 (45) B
Omega-3 fatty acids 2007 1 22 27 30.6 (NR) 32 (65) 1 (100) 3 (0) 20 (0) B
Pooled analysis of randomised trials of medications
Antidepressants 2005–2007 3 41 099 14 944 43.6 (2.7) 29 378 (58) n/a PY median: 1367 (2086.7) NR B
Antipsychotics 2003–2013 2 42 203 7042 41 (NR) 15 771 (37) n/a PY median: 3201 (4275.8) NR B
Mood stabilisers 2005 1 2449 1423 NR NR n/a PY median: 730 (0) NR B

IQR, interquartile range; n/a, not available; NR, not reported; PY, person-year exposure.
a. Reported results for age, gender and drop-out are limited to those studies for which data were available.
b. Grading of suicide assessment: A: 550% of trials used rigorous methods (e.g. coroner report); B: 550% of trials methods were unclear; C: 550% of trials methods were at risk for bias (e.g. informant).

Table 3 Characteristics of included studies: higher-level care interventions a


Years Study Intervention Control Age, years: Female, Setting, Follow-up, months: Attrition %, Rigour in data
covered N N N mean (s.d.) n (%) Europe, n (%) median (IQR) median (IQR) collectionb

Partial hospital admission 1999–2008 2 163 87 35.9 (10.7) 122 (49) 2 (100) 15 (3) 28 (15) A/C
Special care unit 1997 1 140 134 36.3 (15.1) 180 (66) 1 (100) 12 (0) NR A
Somatic therapies
ECT 2013 1 28 28 57 (15.8) 28 (50) 1 (100) 12 (0) 0 (0) B
rTMSc 2014 1 20 21 42.5 (15.7) 6 (15) 0 (0) 6 (0) 27 (0) B

ECT, electroconvulsive therapy; IQR, interquartile range; NR = not reported; rTMS, repetitive transcranial magnetic stimulation.
a. Reported results for age, gender and drop-out are limited to those studies for which data were available.
b. Grading of suicide assessment: A: 550% of trials used rigorous methods (e.g. coroner report); B: 550% of trials methods were unclear; C: 550% of trials methods were at risk for bias (e.g. informant)
c. The comparison arm received sham rTMS.
Strategies to prevent death by suicide

Author, year Population Intervention Control


Comprehensive follow-up programmea Events Total Events Total Weight % Odds ratio (95% CI)d
WHO BIC
Amadeo, 2015 Self-harm event(s) 0 100 2 90 6.5 0.12 (0.01–1.94
Fleischmann, 2008 Suicide attempt 2 872 18 827 64.7 0.19 (0.08–0.45)
Mousavi, 2014 Suicide attempt 1 69 4 70 15.9 0.29 (0.05–1.75)
Sub-total 3 1041 24 987 0.20 (0.09–0.42)
Heterogeneity: w2 = 0.33 d.f. = 2 (P = 0.85) I 2 = 0%
Test for overall effect: Z = 4.19 (P50.0001)
Other programmes
Hvid, 2011 (OPAC) Suicide attempt or self harm event(s) 2 69 1 64 9.7 1.82 (0.19–17.88)
Mouaffak, 2015 (OSTA) Suicide attempt 1 152 0 151 3.3 77.34 (0.15–369.95)
Sub-total 3 221 1 215 2.60 (0.36–18.65)
Heterogeneity: w2 = 0.36 d.f. = 1 (P = 0.55) I 2 = 0%
Test for overall effect: Z = 0.95 (P = 0.034)
Total 6 1262 25 1202 100 0.28 (0.14–0.56)
Heterogeneity: w2 = 6.41 d.f. = 4 (P = 0.17) I 2 = 38%
Test for overall effect: Z = 3.57 (P50.017)

Cognitive–behavioural therapy for suicide preventionb


Brown, 2005 Suicide attempt 0 60 1 60 7.8 0.14 (0.00–6.82)
Raj, 2001 Suicide attempt 0 20 1 20 7.8 0.14 (0.00–6.82)
Rudd, 2015 Suicide attempt or SI w/intent 1 76 1 76 15.8 1.00 (0.06–16.14)
Slee, 2008 Self-harm event(s) 0 40 2 42 15.4 0.14 (0.01–2.25)
Tyrer, 2003 Self-harm event(s) 2 202 5 208 53.6 0.43 (0.10–1.92)
Van Spijker, 2014 Mild–moderate SI 0 116 0 120 0.0 No events
Total 3 514 10 526 100 0.34 (0.12–1.03)
Heterogeneity: w2 = 1.50 d.f. = 4 (P = 0.83) I 2 = 0%
Test for overall effect: Z = 1.91 (P = 0.06)

Treatment with lithiumc


Bauer, 2000 MDD in remission 0 14 1 15 14.4 0.14 (0.00–7.31)
Girlanda, 2014 Refractory MDD & self-harm 1 29 0 25 14.3 76.44 (0.13–327.93)
Khan, 2011 Depressive disorder & SI 0 40 0 40 0.0 No events
Lauterbach, 2008 Depressive disorder & suicide attempt 0 84 3 83 42.6 0.13 (0.01–1.27)
Prien, 1973a Bipolar & unipolar depression 0 45 1 39 14.3 0.12 (0.00–5.91)
Prien, 1973b Bipolar depression 0 101 1 104 14.4 0.14 (0.00–7.02)

Total 1 313 6 306 100 0.23 (0.05–1.02)


Heterogeneity: w2 = 3.23 d.f. = 4 (P = 0.52) I 2 = 0%
Test for overall effect: Z = 1.94 (P = 0.05)
0.0 0.50 1.0 50.0 100.0

Fig. 2 Forest plots of the odds of suicide with three different targeted interventions to prevent suicide v. control condition.
MDD, major depressive disorder; OPAC, outreach, problem solving, adherence and continuity; SI, suicidal ideation; w2, Cochrane’s Q; WHO BIC, World Health Organization brief
intervention and contact programme.
a. Programmes included: WHO BIC (educational intervention plus telephone or face-to-face contact with providers trained in suicide prevention), OPAC (a nurse specialising in
suicide prevention was assigned to follow the patient throughout the course of the intervention), and OSTA (regular telephone and letter contact with patient plus interprofessional
collaboration). Study duration ranged from 12 to 18 months.
b. Patients received between 5 and 12 sessions of the therapy intervention.
c. The study duration ranged from 1 month to 24 months.
d. The odds ratio has a skewed distribution. Although the lower end of the odds ratio is bounded by zero (an odds ratio cannot be negative), the upper end can reach infinity
(online supplemental reference139).

of lithium, 1 out of 313 patients in the intervention group and references70–72,131–133), and there was insufficient granularity to
6 out of 306 patients in the control group died by suicide. The compare specific pharmacotherapy agents.
results, however, were not statistically significant (OR = 0.23, 95%
CI 0.05–1.02, P = 0.05; IRR = 0.14, 95% CI 0.00–9.41, P40.1) Other interventions
(online Fig. 2 and online Table DS2) (online supplemental
We found no evidence that higher-level care interventions such as
references117–122). Several trials tested lithium in patients with
partial hospital admission (2 trials, n = 432; OR = 0.36, 95% CI 0.07–
depressive symptoms and suicidal behaviour (Fig. 2). With the
1.86, P40.10) (online supplemental references69,134,135) or somatic
exception of one trial that evaluated low-dose lithium (300 mg/
therapies such as electroconvulsive therapy (2 trials, n = 92;
day) (online supplemental references120), trials were designed to
OR = 0.14, 95% CI 0.00–6.82, P40.10) (online supplemental refer-
reach a target therapeutic lithium level (online supplemental
ences136,137) reduce the risk of suicide (online Fig. DS1). There were
references117–119,121,122). We identified no placebo-controlled trials
too few studies available to calculate an IRR for these interventions.
of clozapine for suicide prevention.
A large amount of data were included in pooled analyses
(6 trials, 23 016 person-years) to evaluate pharmacotherapy. The Sensitivity analysis
overall summary estimate yielded non-significant results We performed a sensitivity analysis in which we excluded trials
(IRR = 1.15, 95% CI 0.56–2.39, P40.10) (online supplemental that had one or more high risk for bias based on the Cochrane

5
Riblet et al

Risk of Bias Tool (online supplemental references63–65,67,69,71,72,75–82, may reflect our inclusion of two additional RCTs and our focus
84,86,88–91,93,94,96,97,99,101,104–108,110–114,116,118,119,134,137)
(online Table on death by suicide rather than on intermediary outcomes.2,3
DS1). This did not change our results in a substantial way, except Recently, Inagaki et al also reported that intensive follow-up
for our analysis of lithium trials. The results of the summary significantly reduced suicides.7 These authors, however, did not
estimate for lithium became statistically significant after removing a specifically evaluate the WHO BIC programme.7
more recent study (online supplemental references118) with several Akin to our results, a Cochrane review concluded that there is
methodological limitations (5 trials; OR = 0.13, 95% CI 0.03–0.66, an insufficient number of high-quality trials available to draw any
P = 0.01, test of heterogeneity P = 1.00 (Q = 0.01, I 2 = 0%)). firm conclusions about the role of medications in patients who
self-harm.5 Although Zalsman et al reported that antidepressants
Reporting bias reduced the risk of suicide in adults,3 we did not replicate this
We did not find any evidence to suggest publication bias among finding. The majority of RCTs of antidepressants included in
complex psychosocial interventions, psychotherapy, intensive our review reported no suicides. In addition, unlike previous
follow-up strategies or CBT (Harbord’s modified test for small-study reviews,2,3 we did not find that lithium significantly reduced
effects; P = 0.20, P = 0.47, P = 0.57 and P = 0.71, respectively). We suicide. This may be explained by our inclusion of a recent RCT
were unable to formally assess for publication bias among the with negative findings (online supplemental reference118).
remaining domains and sub-domains (as domains included Furthermore, although Mann et al 2 and Zalsman et al 3 concluded
fewer than ten studies or most of the studies in the domain that clozapine has an anti-suicidal effect, we did not replicate this
reported no events). finding. This difference may reflect our decision to limit our
review to placebo-controlled trials. We located no placebo-
controlled trials of clozapine for suicide prevention. We also
Discussion focused our review on suicide deaths, rather than on intermediary
outcomes. Furthermore, prior reviews have relied heavily on the
Summary of main results
results of the InterSePT study to support clozapine’s anti-suicidal
The amount of research on suicide prevention, and specifically the effect (online supplemental reference138).2 We excluded the
number of RCTs targeting suicide, has increased substantially over InterSePT study because clozapine was compared with olanzapine
the past decade. We located 56 RCTs of suicide prevention (online supplemental reference138) rather than placebo. It is
strategies in adults that have been published since Mann et al’s notable, however, that the InterSePT study reported a higher
review.2 This suggests that more research using RCT methodology number of suicide deaths in the clozapine arm versus olanzapine,
is being done in the area of suicide prevention. Although most although the results were not significant (online supplemental
interventions did not lead to a significant reduction in suicide, reference138). Others have also raised concerns about the lack of
we did find that the WHO’s BIC intervention was associated with strong evidence to support clozapine’s anti-suicidal effect.32,33
significantly lower odds of death by suicide. Although trials of Although many reviews stress the role of restricting access to
lithium and CBT for suicide prevention showed fewer deaths by lethal methods in suicide prevention,2,3 we were unable to
suicide among the intervention groups than the controls, we were systematically study this type of intervention because only one
unable to draw any definitive conclusions, as the confidence study met our inclusion criteria. Restricting access to lethal means
interval for the summary estimates spanned no difference. Trials such as gun control is not easily tested under randomised
had several limitations. Most lithium trials had small sample sizes conditions, although a plethora of observational data have
(585 patients) and trials of CBT for suicide prevention were demonstrated that restricting access to lethal means can prevent
generally of short duration (median 10.5 months). It is also worth suicide.2,3 Finally, Zalsman et al concluded that other strategies
noting that the WHO’s BIC intervention may not be generalisable such as screening programmes and media education required
to high-income countries, and that lithium was only studied in more testing.3 We were unable to assess these strategies because
patients with unipolar and bipolar depression. no studies met our inclusion criteria.

Comparison with other studies


Meaning and implications of the review
As with a meta-analysis of findings from the most recent
The WHO BIC intervention was associated with significantly
Cochrane review of psychosocial interventions for self-harm, we
lower odds of suicide, but it will be important to test this strategy
found that CBT-based therapies were associated with lower odds
in other populations. Although the summary estimates for CBT
of suicide, but the results were not significant.8 Because these
for suicide prevention and lithium showed fewer deaths by suicide
authors did not perform a subgroup analysis, we cannot compare
among the intervention group than the control group, the results
our results for CBT for suicide prevention.8 Other reviews, how-
were not statistically significant. Therefore, we believe our results
ever, have suggested that CBT for suicide prevention may prevent
should be interpreted with caution.34 Our findings do not suggest
suicidal behaviour in high-risk populations.2,4 Unlike prior
that lithium or CBT for suicide prevention cause harm, but they
reviews, we found no evidence that problem-solving therapy or
also do not provide clear evidence of effectiveness. Our findings
dialectical behaviour therapy reduce the risk of suicide.2,3 Our
suggest the need for further study of these interventions.
differing conclusions may reflect our focus on suicide deaths,
rather than on intermediary outcomes.
Consistent with others, we found that most psychosocial Strengths and weaknesses of the review
interventions were ineffective,2,3 but we differed from Milner We performed a rigorous search of the literature and did not
et al 6 in that we found no evidence to support letter/telephone- exclude studies based on quality or relevance. This has been a
driven interventions. We believe the divergence in our results may criticism of the Zalsman review.35 Although our focus on death
reflect our decision to evaluate postcard/telephone interventions by suicide can be viewed as a strength,14,35 it is also a potential
and intensive follow-up strategies separately.6 Furthermore, unlike limitation.13 RCTs offer the best evidence, but there are inherent
Mann et al 2 and Zalsman et al,3 we felt that there was strong limitations in developing and testing targeted interventions to
evidence that the WHO BIC programme was associated with address rare events such as suicide in the context of an RCT.13,18
significantly lower odds of suicide. Our divergent conclusions Many studies were at risk for bias in their assessment of suicide,

6
Strategies to prevent death by suicide

and this may have obscured the true effect. Furthermore, for many 9 Hawton K, Taylor Salisbury TL, Arensman E, Gunnell D, Townsend E,
van Heeringen K, et al. Interventions for self-harm in children and
interventions there were insufficient data (i.e. limited number of adolescents (review). Cochrane Database Syst Rev 2015; 12: CD012013.
trials and/or small sample sizes) to draw any definitive conclusions 10 Harrod CS, Goss CW, Stallones L, Diguiseppi C. Interventions for primary
about their efficacy. In addition, owing to the small sample sizes, prevention of suicide in university and other post-secondary educational
we had poor precision around the summary estimate of the effect settings (review). Cochrane Database Syst Rev 2016; 10: CD009439.
size, further limiting our evaluation of these interventions. Since 11 Wei Y, Kutcher S, LeBlanc JC. Hot idea or hot air: a systematic review of
our analysis did not include patient-level data, we were unable evidence for two widely marketed youth suicide prevention programs and
recommendations for implementation. J Can Acad Child Adolesc Psychiatry
to explore potential moderators or mediators of the efficacy of 2015; 24: 5–16.
suicide prevention interventions. Peto ORs can also yield biased 12 Kutcher S, Wei Y, Behzadi P. School- and community-based youth suicide
results when there are substantial differences in study arm sizes. prevention interventions: hot idea, hot air, or sham? Can J Psychiatry 2016;
Reassuringly, the comparator arms of included RCTs were well pii: 0706743716659245 (Epub ahead of print).
balanced. We did not calculate Peto ORs for pooled analysis 13 Glenn CR, Nock MK. Improving the short-term prediction of suicidal behavior.
Am J Prev Med 2014; 47: S176–80.
because of large differences in study arms. Finally, although many
14 Perlis RH. Hard outcomes: clinical trials to reduce suicide. Am J Psychiatry
novel approaches to suicide prevention (e.g. ketamine) are
2011; 168: 1009–11.
emerging, no studies of these met our full inclusion criteria.
15 Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of
Overall, our review suggests that the WHO BIC intervention is Interventions Version 5. 1.0. The Cochrane Collaboration, 2011.
associated with significantly lower odds of suicide. Although trials 16 Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for
of CBT for suicide prevention and lithium showed fewer deaths by systematic reviews and meta-analyses: the PRISMA statement. Ann Intern
suicide among the intervention groups than the controls, the Med 2009; 151: 264–9.

differences were not statistically significant. Available studies also 17 Friedrich JO, Adhikari NK, Beyene J. Inclusion of zero total event trials in
meta-analyses maintains analytic consistency and incorporates all available
have several limitations that may threaten their internal and data. BMC Med Res Methodol 2007; 7: 5.
external validity. More research is needed to evaluate the efficacy 18 Rosen L, Manor O, Engelhard D, Zucker D. In defense of the randomized
of these and other interventions in a range of settings. controlled trial for health promotion research. Am J Public Health 2006; 96:
1181–6.
Natalie B. V. Riblet, MD, MPH, Brian Shiner, MD, MPH, Veterans Affairs Medical
19 Centers for Disease Control and Prevention. Definitions: self-directed
Center, Vermont, Geisel School of Medicine at Dartmouth, Hanover, New Hampshire,
and The Dartmouth Institute for Health Policy and Clinical Practice, Lebanon, New violence. CDC, 2015 (http://www.cdc.gov/violenceprevention/suicide/
Hampshire, USA; Yinong Young-Xu, DSc, MS, Bradley V. Watts, MD, MPH, definitions.html).
Veterans Affairs Medical Center, Vermont, and Geisel School of Medicine at 20 Higgins JPT, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The
Dartmouth, Hanover, New Hampshire, USA Cochrane Collaboration’s tool for assessing risk of bias in randomised trials.
BMJ 2011; 43: d5928.
Correspondence: Natalie Riblet, MD, MPH, Veterans Affairs Medical Center,
215 North Main Street, White River Junction, VT 05009, USA. Email: 21 Higgins JPT, Green S (eds). Chapter 16.9.5. Validity of methods of meta
Natalie.Riblet@dartmouth.edu analysis for rare events. In Cochrane Handbook for Systematic Reviews of
Interventions Version 5. 1. 0. The Cochrane Collaboration, 2011 (http://
First received 14 May 2016, final revision 30 Jan 2017, accepted 4 Mar 2017
handbook.cochrane.org/).
22 Deeks JJ, Higgins JPT, Altman DG. Chapter 9. Analysing data and undertaking
meta-analyses. In Cochrane Handbook for Systematic Reviews of
Funding Interventions Version 5. 1. 0 (eds JPT Higgins, S Green). The Cochrane
Collaboration, 2011 (http://handbook.cochrane.org/).
This work was supported by the VA National Center for Patient Safety Center of Inquiry
23 Lane PW. Meta-analysis of incidence of rare events. Stat Methods Med Res
Program (PSCI-WRJ-SHINER). B.S. is the recipient of a VA Health Services Research and
Development Career Development Award (CDA11-263). 2013; 22: 117-32.
24 Szumilas M. Explaining odds ratios. J Can Acad Child Adolesc Psychiatry
Acknowledgements 2010; 19: 227–9.
25 The Cochrane Collaboration. Review Manager (RevMan) Version 5.3. Nordic
We thank Dr Gregory McHugo, Research Professor of Community and Family Medicine, and Cochrane Centre, 2014.
of Psychiatry, of The Dartmouth Institute at the Geisel School of Medicine at Dartmouth, for 26 Shuster JJ, Walker MA. Low-event-rate meta-analyses of clinical trials:
providing us with editing suggestions. implementing good practices. Stat Med 2016; 35: 2467–78.
27 Cai T, Parast L, Ryan L. Meta-analysis for rare events. Stat Med 2010; 29:
References 2078–89.
28 Spittal MJ, Pirkis J, Gurrin LC. Meta-analysis of incidence rate data in the
1 World Health Organization. Preventing Suicide: A Global Imperative. WHO, 2014. presence of zero events. BMC Med Res Methodol 2015; 15: 1–16.
2 Mann JJ, Apter A, Bertolote J, Beautrais A, Currier D, Haas A, et al. Suicide 29 Hilbe JM. Negative Binomial Regression (2nd edn). Cambridge University
prevention strategies: a systematic review. JAMA 2005; 294: 2064–74. Press, 2011.
3 Zalsman G, Hawton K, Wasserman C, Van Heeringen K, Arensman E, 30 Multilevel mixed-effects negative binomial regression. STATA, 2013 (http://
Sarchiapone M, et al. Suicide prevention strategies revisited: 10-year www.stata.com/manuals13/memenbreg.pdf).
systematic review. Lancet Psychiatry 2016; 3: 646–59.
31 Harbord RM, Egger M, Sterne JA. A modified test for small-study effects in
4 Brown GK, Jager-Hyman S. Evidence-based psychotherapies for suicide meta-analysis of controlled trials with binary endpoints. Stat Med 2006; 25:
prevention. Am J Prev Med 2014; 47: S186–94. 3443–57.
5 Hawton K, Witt KG, Taylor Salisbury TL, Arensman E, Gunnell D, Hazell P, et al. 32 Baethge C. Keine hochwertige evidence fuer verhinderung von suiziden
Pharmacological interventions for self-harm in adults. Cochrane Database durch antidepressiva und Clozapin [No high level evidence for suicide
Syst Rev 2015; 7: CD011777. prevention with antidepressants or clozapine] Nervenarzt 2017; 88: 309–10.
6 Milner AJ, Carter G, Pirkis J, Robinson J, Spittal MJ. Letters, green cards, 33 Lobos CA, Komossa K, Rummel-Kluge C, Hunger H, Schmid F, Schwarz S,
telephone calls and postcards: systematic and meta-analytic review of brief et al. Clozapine versus other atypical antipsychotics for schizophrenia.
contact interventions for reducing self-harm, suicide attempts and suicide. Cochrane Database Syst Rev 2010; 11: CD006633.
Br J Psychiatry 2015; 206: 184–90.
34 Wood J, Freemantle N, King M, Nazareth I. Trap of trends to statistical
7 Inagaki M, Kawashima Y, Kawanishi C, Yonemoto N, Sugimoto T, Furuno T, significance: likelihood of near significant P value becoming more significant
et al. Interventions to prevent repeat suicidal behavior in patients admitted with extra data. BMJ 2014; 348: g2215.
to an emergency department for a suicide attempt: a meta-analysis. J Affect 35 Kutcher S, Wei Y. The vexing challenge of suicide prevention: a research
Disord 2015; 175: 66–78. informed perspective on a recent systematic review. National Elf Service
8 Hawton K, Witt KG, Taylor Salisbury TL, Arensman E, Gunnell D, Hazell P, et al. 2016; 12 August (http://www.nationalelfservice.net/mental-health/suicide/
Psychosocial interventions following self-harm in adults: a systematic review vexing-challenge-suicide-prevention-research-informed-perspective-recent-
and meta-analysis. Lancet Psychiatry 2016; 3: 740–50. systematic-review/).

7
Data supplement to:

Strategies to prevent death by suicide: a meta-analysis of randomised


controlled trials
Natalie Riblet, Brian Shiner, Yinong Young-Xu and Bradley Watts

British Journal of Psychiatry doi: 10.1192/bjp.bp.116.187799


Supplemental references

35. Hassanian-Moghaddam H, Sarjami S, Kolahi AA, Carter GL. Postcards in persia: randomised

controlled trial to reduce suicidal behaviours 12 months after hospital-treated self-poisoning. Br J

Psychiatry 2011;198:309-16.

36. Hassanian-Moghaddam H, Sarjami S, Kolahi, AA, Lewin T, Carter G. Postcards in Persia: a twelve to

twenty-four month follow-up of a randomized controlled trial for hospital-treated deliberate self-

poisoning. Archives of Suicide Research 2015; doi: 10.1080/13811118.2015.1004473.

37. Crawford MJ, Csipke E, Brown A, Reid S, Nilsen K, Redhead J, et al. The effect of referral for brief

intervention for alcohol misuse on repetition of deliberate self-harm: an exploratory randomized

controlled trial. Psychol Med 2010;40:1821-8.

38. Nordentoft M, Jeppesen P, Abel M, Kassow P, Petersen L, Thorup A, et al. OPUS study: suicidal

behaviour, suicidal ideation and hopelessness among patients with first-episode psychosis. One-year

follow-up of a randomised controlled trial. Br J Psychiatry Suppl 2002;43:s98-s106.

39. Almeida OP, Pirkis J, Kerse N, Sim M, Flicker L, Snowdon J, et al. A randomized trial to reduce the

prevalence of depression and self-harm behavior in older primary care patients. Ann Fam Med 2012

10: 347-56.

40. Wei S, Liu L, Bi B, Li H, Hou J, Tan S, et al. An intervention and follow-up study following a suicide

attempt in the emergency departments of four general hospitals in Shenyang, China. Crisis

2013;34:107-15.

41. Yip PSF, Law CK, Fu KW, Law YW, Wong PWC, Xu Y. Restricting the means of suicide by

charcoal burning. Br J Psychiatry 2010;196:241-2.

42. Hatcher S, Sharon C, House A, Collins N, Collings S, Pillai A. The ACCESS study: Zelen

randomised controlled trial of a package of care for people presenting to hospital after self-harm. Br J

Psychiatry 2015;206:229-36.

43. Hatcher S, Sharon C, Parag V, Collins N. Problem-solving therapy for people who present to hospital

with self-harm: Zelen randomised controlled trial. Br J Psychiatry 2011;199:310-6.

1
44. McMain SF, Links PS, Gnam WH, Guimond T, Cardish RJ, Korman L, et al. A randomized trial of

dialectical behavior therapy versus general psychiatric management for borderline personality

disorder. Am J Psychiatry 2009;166:1365-74.

45. Tyrer P, Tom B, Byford S, Schmidt U, Jones V, Davidson K, et al. Differential effects of manual

assisted cognitive behavior therapy in the treatment of recurrent deliberate self-harm and personality

disturbance: the POPMACT study. J Pers Disord 2004;18:102-16.

46. Tyrer P, Jones V, Thompson S, Catalan J, Schmidt U, Davidson K, et al. Service variation in baseline

variables and prediction of risk in a randomised controlled trial of psychological treatment in repeated

parasuicide: the popmact study. Int J Soc Psychiatry 2003;49:58-69.

47. Vijayakumar L, Umamaheswari C, Sultana Z, Ali S, Devaraj P, Kesavan K. Intervention for suicide

attempters: a randomized controlled study. Indian J Psychiatry 2011;53:244-8.

48. Hassanzadeh M, Khajeddin N, Nojomi M, Fleischmann A, Eshrati T. Brief intervention and contact

after deliberate self-harm: an iranian randomized controlled trial. Iran J Psychiatry Behav Sci

2010;4(2):5-12.

49. Motto J. Suicide prevention for high-risk persons who refuse treatment. Suicide Life Threat Behav

1976;6:223-30.

50. Moller H. Efficacy of different strategies of aftercare for patients who have attempted suicide. J R Soc

Med 1989;82:643-7.

51. Gallo JJ, Hwang S, Joo JH, Bogner HR, Morales KH, Bruce ML, et al. Multimorbidity, Depression,

and Mortality in Primary Care: Randomized Clinical Trial of an Evidence-Based Depression Care

Management Program on Mortality Risk. J Gen Intern Med 2016; 31: 380-6.

52. Gallo JJ, Morales KH, Bogner HR, Raue PJ, Zee J, Bruce ML, et al. Long term effect of depression

care management on mortality in older adults: follow-up of cluster randomized clinical trial in

primary care. BMJ 2013;346:f2570.

53. Bateman A, Fonagy P. Effectiveness of partial hospitalization in the treatment of borderline

personality disorder: a randomized controlled trial. Am J Psychiatry 1999;156:1563-9.

2
54. Khan A, Khan S, Kolts R, Brown WA. Suicide rates in clinical trials of SSRIs, other antidepressants,

and placebo: analysis of FDA reports. Am J Psychiatry 2003; 160: 790-2.

55. Khan A, Khan SR, Leventhal R, Brown WA. Symptom reduction and suicide risk among patients

treated with placebo in antipsychotic clinical trials: an analysis of the Food and Drug Administration

Database. Am J Psychiatry 2001; 158: 1449-54.

56. Khan A, Khan SR, Leventhal RM, Brown WA. Symptom reduction and suicide risk in patients

treated with placebo in antidepressant clinical trials: a replication analysis of the Food and Drug

Administration Database. Int J Neuropsychopharmacol 2001; 4: 113-8.

57. Khan A, Warner HA, Brown WA. Symptom reduction and suicide risk in patients treated with

placebo in antidepressant clinical trials. Arch Gen Psychiatry 2000; 57: 311-7.

58. Khan A, Kolts RL, Brodhead AE, Krishnan R, Brown WA. Suicide risk analysis among patients

assigned to psychotropics and placebo. Psychopharmacol Bull 2006; 39: 6-14.

59. Torhorst A, Moller HJ, Burk F, Kurz A, Wachtler C, Lauter H. The psychiatric management of

parasuicide patients: a controlled clinical study comparing different strategies of outpatient treatment.

Crisis 1987;8:53-61.

60. Fleischmann A, Bertolote JM, Wasserman D, De Leo D, Bolhari J, Botega NJ, et al. Effectiveness of

brief intervention and contact for suicide attempters: a randomized controlled trial in five countries.

Bull World Health Organ 2008;86:703-9.

61. Bertelsen M, Jeppesen P, Petersen L, Thorup A, Ohlenschlaeger J, Quach P, et al. Suicidal behaviour

and mortality in first-episode psychosis: the OPUS trial. Br J Psychiatry 2007;191(Suppl S1):s140-s6.

62. Motto J, Bostrom A. A randomized controlled trial of postcrisis suicide prevention. Psychiatr Serv

2001;52:828-33.

63. Raue PJ, Morales KH, Post EP, Bogner HR, Have TT, Bruce ML. The wish to die and 5-year

mortality in elderly primary care patients. Am J Geriatr Psychiatry 2010;18:341-50.

64. Moller H. Attempted suicide: efficacy of different aftercare strategies. Int Clin Psychopharmacol

1992;6(Suppl 6):58-69.

3
65. Sun F-K, Chiang C-Y, Yu P-J, Lin C-H. A suicide education programme for nurses to educate the

family caregivers of suicidal individuals: a longitudinal study. Nurse Education Today 2013; 10:

1192-1200.

66. Sun F-K, Chiang C-Y, Lin Y-H, Chen T-B. Short-term effects of a suicide education intervention for

family caregivers of people who are suicidal. J Clin Nurs 2013; 10: 91 – 102.

67. Tyrer P, Thompson S, Schmidt U, Jones V, Knapp M, Davidson K, et al. Randomized controlled trial

of brief cognitive behaviour therapy versus treatment as usual in recurrent deliberate self-harm: the

POPMACT study. Psychol Med 2003;33:969-76.

68. McMain SF, Guimond T, Streiner DL, Cardish RJ, Links PS. Dialectical behavior therapy compared

with general psychiatric management for borderline personality disorder: clinical outcomes and

functioning over a 2-year follow-up. Am J Psychiatry 2012;169:650-61.

69. Bateman A, Fonagy P. 8-year follow-up of patients treated for borderline personality disorder:

mentalization-based treatment versus treatment as usual. Am J Psychiatry 2008;165:631-8.

70. Hammad T, Laughren T, Racoosin J. Suicide rates in short-term randomized controlled trials of

newer antidepressants. J Clin Psychopharmacol 2006;26:203-7.

71. Khan A, Schwartz K. Suicide risk and symptom reduction in patients assigned to placebo in

duloxetine and escitalopram clinical trials: analysis of the FDA summary basis of approval reports.

Ann Clin Psychiatry 2007;19:31-6.

72. Khan A, Faucett J, Morrison S, Brown W. Comparative mortality risk in adult patients with

schizophrenia, depression, bipolar disorder, anxiety disorders, and attention-deficit/hyperactivity

disorder participating in psychopharmacology clinical trials. JAMA Psychiatry 2013;70:1091-9.

73. Amadeo S, Rereao M, Malogne A, Favro P, Nguyen NL, Jehel L, et al. Testing brief intervention and

phone contact among subjects with suicidal behavior: A randomized controlled trial in French

Polynesia in the frames of the World Health Organization/Suicide Trends in At-Risk Territories

study. Ment Illn 2015; 7: 48-53.

4
74. Mousavi SG, Zohreh R, Maracy MR, Ebrahimi A, Sharbafchi MR. The efficacy of telephonic follow

up in prevention of suicidal reattempt in patients with suicide attempt history. Adv Biomed Res

2014;3:198.

75. Hvid M, Vanborg K, Sorensen HJ, Nielsen IK, Stenborg JM, Wang AG. Preventing repetition of

attempted suicide--II. The Amager Project, a randomized controlled trial. Nord J Psychiatry

2011;65:292-8.

76. Mouaffak F, Marchand A, Castaigne E, Arnoux A, Hardy P. OSTA program: a French follow up

intervention program for suicide prevention. Psychiatry Res 2015;230:913-8.

77. Kawanishi C, Aruga T, Ishizuka N, Yonemoto N, Otsuka K, Kamijo Y, et al. Assertive case

management versus enhanced usual care for people with mental health problems who had attempted

suicide and were admitted to hospital emergency departments in Japan (ACTION-J): a multicentre,

randomised controlled trial. Lancet Psychiatry 2014;1:193-201.

78. Allard R, Marshall M, Plante MC. Intensive follow-up does not decrease the risk of repeat suicide

attempts. Suicide Life Threat Behav 1992;22(3):303-14.

79. van Heeringen C, Jannes S, Buylaert W, Henderick H, De Bacquer D, van Remoortel J. The

management of non-compliance with referral to out-patient after-care among attempted suicide

patients: a controlled intervention study. Psychol Med 1995;25:963-70.

80. De Leo D, Heller T. Intensive case management in suicide attempters following discharge from

inpatient psychiatric care. Aust J Prim Health 2007;13:49-58.

81. Clarke T, Baker P, Watts CJ, Williams K, Feldman RA, Sherr L. Self-harm in adults: a randomised

controlled trial of nurse-led case management versus routine care only. J Ment Health 2002;11:167-

76.

82. Morthorst B, Krogh J, Erlangsen A, Alberdi F, Nordentoft M. Effect of assertive outreach after

suicide attempt in the AID (assertive intervention for deliberate self harm) trial: randomised

controlled trial. BMJ 2012;345:e4972.

5
83. Connolly J, Marks I, Lawrence R, McNamee G, Muijen M. Observations from community care for

serious mental illness during a controlled study. Psychiatr Bull 1996;20:3-7.

84. Walsh E, Harvey K, White I, Higgitt A, Fraser J, Murray R. Suicidal behaviour in psychosis:

prevalence and predictors from a randomised controlled trial of case management. Br J Psychiatry

2001;178:255-60.

85. Dekker J, Wijdenes W, Koning YA, Gardien R, Hermandes-Willenborg L, Nusselder H. Assertive

community treatment in Amsterdam. Community Ment Health J 2002;38:425-34.

86. Morgan HG, Jones EM, Owen JH. Secondary prevention of non-fatal deliberate self-harm: the green

card study. Br J Psychiatry 1993;163:111-2.

87. Evans MO, Morgan HG, Hayward A, Gunnell DJ. Crisis telephone consultation for deliberate self-

harm patients: effects on repetition. Br J Psychiatry 1999;175:23-7.

88. Cedereke M, Monti K, Ojehagen A. Telephone contact with patients in the year after a suicide

attempt: does it affect treatment attendance and outcome? A randomised controlled study. Eur

Psychiatry 2002;17:82-91.

89. Carter GL, Clover K, Whyte IM, Dawson AH, D'Este C. Postcards from the EDge: 5-year outcomes

of a randomised controlled trial for hospital-treated self-poisoning. Br J Psychiatry 2013;202:372-80.

90. Vaiva G, Ducrocq F, Meyer P, Mathieu D, Philippe A, Libersa C, et al. Effect of telephone contact on

further suicide attempts in patients discharged from an emergency department: randomised controlled

study. BMJ 2006;332:1241-5.

91. Grimholt TK, Jacobsen D, Haavet OR, Sandvik L, Jorgensen T, Norheim AB, et al. Effect of

Systematic Follow-Up by General Practitioners after Deliberate Self-Poisoning: A Randomised

Controlled Trial. PLoS One 2015; 10: e0143934.

92. Schulberg HC, Lee PW, Bruce ML, Raue PJ, Lefever JJ, Williams JW, et al. Suicidal ideation and

risk levels among primary care patients with uncomplicated depression. Ann Fam Med 2005;3:523-8.

93. Unutzer J, Tang L, Oishi S, Katon W, Williams Jr JW, Hunkeler E, et al. Reducing suicidal ideation

in depressed older primary care patients. J Am Geriatr Soc 2006; 10: 1550-6.

6
94. Vijayakumar L, Jeyaseelan L, Kumar S, Mohanraj R, Devika S, Manikandan S. A central storage

facility to reduce pesticide suicides--a feasibility study from India. BMC Public Health 2013;13:850.

95. Brown GK, Ten HT, Henriques GR, Xie SX, Hollander JE, Beck AT. Cognitive therapy for the

prevention of suicide attempts: a randomized controlled trial. JAMA 2005;294:563-70.

96. van Spijker BA, van Straten A, Kerkhof AJ. Effectiveness of online self-help for suicidal thoughts:

results of a randomised controlled trial. PLoS ONE 2014;9:e90118.

97. Slee N, Garnefski N, van der Leeden R, Arensman E, Spinhoven P. Cognitive-behavioural

intervention for self-harm: randomised controlled trial. Br J Psychiatry 2008;192:202-11.

98. Rudd MD, Bryan CJ, Wertenberger EG, Peterson AL, Young-McCaughan S, Mintz J, et al. Brief

cognitive-behavioral therapy effects on post-treatment suicide attempts in a military sample: results of

a randomized clinical trial with 2-year follow-up. Am J Psychiatry 2015; 172: 441-9.

99. Raj MA, Kumaraiah V, Bhide AV. Cognitive-behavioural intervention in deliberate self-harm. Acta

Psychiatr Scand 2001;104:340-5.

100. Linehan MM, Armstrong HE, Suarez A, Allmon D, Heard HL. Cognitive-behavioral treatment of

chronically parasuicidal borderline patients. Arch Gen Psychiatry 1991;48:1060-4.

101. Linehan MM, Comtois KA, Murray AM, Brown MZ, Gallop RJ, Heard HL, et al. Two-year

randomized controlled trial and follow-up of dialectical behavior therapy vs therapy by experts for

suicidal behaviors and borderline personality disorder.[Erratum appears in Arch Gen Psychiatry 2007;

64:1401]. Arch Gen Psychiatry 2006;63:757-66.

102. Davidson KM, Tyrer P, Norrie J, Palmer SJ, Tyrer H. Cognitive therapy v. usual treatment for

borderline personality disorder: prospective 6-year follow-up. Br J Psychiatry 2010;197:456-62.

103. Davidson KM, Brown TM, Vairi J, Kirk J, Richardson J. Manual-assisted cognitive therapy for

self-harm in personality disorder and substance misuse: a feasibility trial. Psychiatr Bull 2014;38:108-

11.

7
104. Blum N, St John D, Pfohl B, Stuart S, McCormick B, Allen J, et al. Systems training for

emotional predictability and problem solving (STEPPS) for outpatients with borderline personality

disorder: a randomized controlled trial and 1-year follow-up. Am J Psychiat 2008; 165: 468-78.

105. Power PJ, Bell RJ, Mills R, Herrman-Doig T, Davern M, Henry L, et al. Suicide prevention in

first episode psychosis: the development of a randomised controlled trial of cognitive therapy for

acutely suicidal patients with early psychosis. Aust N Z J Psychiatry 2003;37:414-20.

106. Grawe RW, Falloon IR, Widen JH, Skogvoll E. Two years of continued early treatment for

recent-onset schizophrenia: a randomised controlled study. Acta Psychiatr Scand 2006;114:328-36.

107. Tarrier N, Haddock G, Lewis S, Drake R, Gregg L, SoCrates Trial Group. Suicide behaviour over

18 months in recent onset schizophrenic patients: the effects of CBT. Schizophr Res 2006;83:15-27.

108. Morley KC, Sitharthan G, Haber PS, Tucker P, Sitharthan T. The efficacy of an opportunistic

cognitive behavioral intervention package (OCB) on substance use and comorbid suicide risk: a

multisite randomized controlled trial. J Consult Clin Psychol 2014;82:130-40.

109. Husain N, Afsar S, Ara J, Fayyaz H, Rahman RU, Tomenson B, et al. Brief psychological

intervention after self-harm: randomised controlled trial from Pakistan. Br J Psychiatry

2014;204:462-70.

110. Rudd MD, Rajab MH, Orman DT, Joiner T, Stulman DA, Dixon W. Effectiveness of an

outpatient intervention targeting suicidal young adults: preliminary results. J Consult Clin Psychol

1996;64:179-90.

111. Hawton K, McKeown S, Day A, Martin P, O'Connor M, Yule J. Evaluation of out-patient

counselling compared with general practitioner care following overdoses. Psychol Med 1987;17:751-

61.

112. McAuliffe C, McLeavey BC, Fitzgerald T, Corcoran P, Carroll B, Ryan L, et al. Group problem-

solving skills training for self-harm: randomised controlled trial. Br J Psychiatry 2014; 204: 383-90.

8
113. Gregory RJ, Chlebowski S, Kang D, Remen AL, Soderberg MG, Stepkovitch J, et al. A

controlled trial of psychodynamic psychotherapy for co-occurring borderline personality disorder and

alcohol use disorder. Psychother 2008; 45: 28-41.

114. Winter D, Sireling L, Riley T, Metcalfe C, Quaite A, Bhandari S. A controlled trial of personal

construct psychotherapy for deliberate self-harm. Psychol Psychother 2007;80:23-37.

115. Guthrie E, Kapur N, Mackway-Jones K, Chew-Graham C, Moorey J, Mendel E, et al.

Randomised controlled trial of brief psychological intervention after deliberate self poisoning. BMJ

2001;323:1-5.

116. Gysin-Maillart AG, Schwab S, Soravia L, Megert M, Michel K. A novel brief therapy for patients

who attempt suicide: a 24-months follow-up randomized controlled study of the attempted suicide

short intervention program (ASSIP). PLoS ONE 2016:1-21.

117. Bauer M, Bschor T, Kunz D, Berghoefer A, Stroehle A, Mueller-Oerlinghausen B. Double-blind,

placebo-controlled trial of the use of lithium to augment antidepressant medication in continuation

treatment of unipolar major depression. Am J Psychiatry 2000;157:1429-35.

118. Girlanda F, Cipriani A, Agrimi E, Appino M, Barichello A, Beneduce R, et al. Effectiveness of

lithium in subjects with treatment-resistant depression and suicide risk: results and lessons of an

underpowered randomised clinical trial. BMC Research Notes 2014;7:1-8.

119. Lauterbach E, Felber W, Muller-Oerlinghausen B, Ahrens B, Bronisch T, Meyer T, et al.

Adjunctive lithium treatment in the prevention of suicidal behaviour in depressive disorders: a

randomised, placebo-controlled, 1-year trial. Acta Psychiatr Scand 2008;118:469-79.

120. Khan A, Khan SRF, Hobus J, Faucett J, Mehra V, Giller EL, et al. Differential pattern of response

in mood symptoms and suicide risk measures in severely ill depressed patients assigned to citalopram

with placebo or citalopram combined with lithium: role of lithium levels. J Psychiatr Res

2011;45:1489-96.

121. (a) Prien RF, Klett CJ, Caffey EM. Lithium carbonate and imipramine in prevention of affective

episodes. Arch Gen Psychiatry 1973;29:420-5.

9
122. (b) Prien RF, Point P, Caffey EM, Klett CJ. Prophylactic efficacy of lithium carbonate in manic-

depressive illness. Arch Gen Psychiatry 1973;28:337-41.

123. Nelson JC, Delucchi K, Schneider LS. Suicidal thinking and behavior during treatment with

sertraline in late-life depression. Am J Geriatr Psychiatry 2007; 15: 573-80.

124. Zisook S, Kasckow J, Golshan S, Fellows I, Solorzano E, Lehman D, et al. Citalopram

augmentation for subsyndromal symptoms of depression in middle-aged and older outpatients with

schizophrenia and schizoaffective disorder: a randomized controlled trial. J Clin Psychiatry 2009; 70:

562-71.

125. Rosenthal JZ, Boyer P, Vialet C, Hwang E, Tourian KA. Efficacy and safety of desvenlafaxine 50

mg/d for prevention of relapse in major depressive disorder:a randomized controlled trial. J Clin

Psychiatry 2013;74(2):158-66.

126. Phillips KA, Kelly MM. Suicidality in a placebo-controlled fluoxetine study of body dysmorphic

disorder. Int Clin Psychopharmacol 2009;24(1):26-8.

127. Hirsch SR, Walsh C, Draper R. Parasuicide: a review of treatment interventions. J Affect Disord

1982;4:299-311.

128. Montgomery SA, Roy D, Montgomery DB. The prevention of recurrent suicidal acts. Br J Clin

Pharmac 1983;15:183S-8S.

129. Verkes RJ, Van der Mast RC, Hengeveld MW, Tuyl JP, Zwinderman AH, Van Kempen GM.

Reduction by paroxetine of suicidal behavior in patients with repeated suicide attempts but not major

depression. Am J Psychiatry 1998;155(4):543-7.

130. Hallahan B, Hibbeln JR, Davis JM, Garland MR. Omega-3 fatty acid supplementation in patients

with recurrent self-harm: single-centre double-blind randomised controlled trial. Br J Psychiatry

2007; 190: 118-122.

131. Storosum JG, Wohlfarth T, Gispen-de Wied CC, Linszen DH, Gersons BPR, van Zwieten BJ, et

al. Suicide risk in placebo-controlled trials of treatment for acute manic episode and prevention of

manic-depressive episode. Am J Psychiatry 2005;162:799-802.

10
132. Storosum JG, Van Zwieten BJ, Wohlfarth T, de Haan L, Khan A, Van den Brink W. Suicide risk

in placebo vs active treatment in placebo-controlled trials for schizophrenia. Arch Gen Psychiatry

2003;60:365-8.

133. Aursnes I, Tvete IF, Gaasemyr J, Natvig B. Suicide attempts in clinical trials with paroxetine

randomised against placebo. BMC Medicine 2005;3:1-5.

134. Jones G, Gavrilovic J, McCabe R, Becktas C, Priebe S. Treating suicidal patients in an acute

psychiatric day hospital: a challenge to assumptions about risk and overnight care. J Ment Health

2008;17(4):375-87.

135. van der Sande R, van Rooijen L, Buskens E, Allart E, Hawton K, van der Graaf Y, et al. Intensive

in-patient and community intervention versus routine care after attempted suicide: a randomised

controlled intervention study. Br J Psychiatry 1997;171:35-41.

136. George MS, Raman R, Benedek DM, Pelic CG, Grammer GG, Stokes KT, et al. A two-site pilot

randomized 3 day trial of high dose left prefrontal repetitive transcranial magnetic stimulation (rTMS)

for suicidal inpatients. Brain Stimul 2014; 7: 421-31.

137. Nordenskjold A, Von Knorring L, Lyung T, Carlborg A, Brus O, Engstrom I. Continuation

electroconvulsive therapy with pharmacotherapy versus pharmacotherapy alone for prevention of

relapse of depression: a randomized controlled trial. J ECT 2013; 29: 86-92.

138. Meltzer HY, Alphs L, Green AI, Altamura AC, Anand R, Bertoldi A, et al. Clozapine treatment

for suicidality in schizophrenia: international suicide prevention trial (InterSePT). Arch Gen

Psychiatry 2003;60:82-91.

139. Bland JM, Altman DG. Statistics notes: the odds ratio. BMJ 2000; 320: 1468.

11
Table DS1 Methodological quality and results of trials evaluating suicide prevention strategiesa
Random Blinding of Blinding of Incomplete
sequence Allocation participants outcome outcome Selective Other
Author, Year OR (95%CI) generation Concealment & personnel assessment data Reporting Biases
Cognitive Behavioral Therapies
Blum, 2008 No events Low Unclear High Unclear Unclear Low Unclear

Brown, 2005 0.14 (0.00 – 6.82) Low Unclear Unclear Unclear Low Low Low

Davidson, 2010 0.77 (0.05 – 12.80) Low Low Unclear Low Low Low Unclear

Davidson, 2014 3.91 (0.05 – 328.91) Low Low Unclear Unclear Low Low Unclear

Grawe, 2006 No events Low Low High Unclear Low Unclear Unclear

Hawton, 1987 8.41 (0.17 – 426. 74) Unclear Unclear Unclear High Unclear Unclear Unclear

Husain, 2014 1.10 (0.15 – 7.93) Low Low Unclear Unclear Low Low Unclear

Linehan, 1991 6.80 (0.13 – 343.88) Unclear Unclear Unclear Unclear Unclear Unclear Unclear

Linehan, 2006 No events Low Unclear High Unclear Unclear Unclear Unclear

McAuliffe, 2014 0.49 (0.05 – 4.70) Low Low High Unclear Unclear Unclear High

McMain, 2012 No events Low Low Unclear Low Low Unclear Unclear

Morley, 2014 No events Unclear Low Unclear Unclear Unclear Unclear High

Power, 2003 1.00 (0.06 – 16.55) Unclear Unclear Unclear Low Low Unclear High

Raj, 2001 0.14 (0.00 – 6.82) High Unclear Unclear Unclear Unclear Unclear Unclear

Rudd, 1996 No events Unclear Unclear Unclear Unclear High Unclear Unclear

Rudd, 2015 1.00 (0.06 – 16.14) Low Unclear Unclear Low Low Unclear Unclear

Slee, 2008 0.14 ( 0.01 – 2.25) Low Low Unclear Unclear Unclear Low High
CBT: 6.84 (0.14 – 345.58)
Tarrier, 2006 Low Low Unclear High Unclear Unclear Low
Sup: 6.61 (0.41 – 107.35)
Tyrer, 2003 0.43 ( 0.10 – 1.92) Low Unclear High Unclear Low Low Unclear

Van Spijker, 2014 No events Low Low High High Low Low High

OR = Odds Ratio; 95%CI = 95 percent confidence interval; CBT = Cognitive Behavioral Therapy; Sup = Supportive Therapy
a. Based on the Cochrane Risk of Bias Tool; Low = low risk of bias; High = high risk of bias; Unclear= Insufficient evidence to judge risk of bias

Table DS1 Methodological quality and results of trials evaluating suicide prevention strategiesa (continued)
Random Blinding of Blinding of
sequence Allocation participants & outcome Incomplete Selective Other
Author, Year OR (95% CI) generation Concealment personnel assessment outcome data Reporting Biases
Non-Cognitive Behavioral Therapies
Gregory, 2008 0.14 (0.00 – 6.82) Low Unclear High Unclear Low Unclear High

Guthrie, 2001 No events Low Unclear Unclear Unclear Low Unclear Unclear
Gysin-Maillart,
1.00 (0.06 – 16.18) Low Low High Unclear Low Low Unclear
2016
Winter, 2007 0.83 (0.08 – 8.95) High High Unclear Unclear Unclear Unclear Unclear
Complex Psychosocial Interventions
Allard, 1992 2.79 (0.38 – 20.26) Unclear Low High Low Low Unclear High

Amadeo, 2015 0.12 (0.01 – 1.94) Low Unclear Unclear Low Low Low Unclear

Bertelsen, 2007 0.59 (0.13 – 2.67) Unclear Unclear Unclear Low Low Unclear Unclear

Carter, 2005 0.87 (0.26 – 2.85) Low Low High Low Low Unclear High

Cedereke, 2002 1.07 (0.07 – 17.32) Unclear Low High Low Low Unclear High

Clarke, 2002 1.12 (0.07 – 18.10) Low Unclear Unclear Unclear Low Unclear High

Connolly, 1996 1.42 (0.31 – 6.40) Unclear Unclear Unclear Unclear Unclear Unclear Unclear
Dekker, 2002 7.71 (0.48 – 125.04) Unclear Unclear Unclear Unclear Low Unclear Unclear

De Leo, 2007 No events Low Low High Unclear High Unclear High

Evans, 1999 1.92 (0.20 – 18.49) Unclear Low Unclear Low Low Unclear Unclear

OR = Odds Ratio; 95%CI = 95 percent confidence interval


a. Based on the Cochrane Risk of Bias Tool; Low = low risk of bias; High = high risk of bias; Unclear= Insufficient evidence to judge risk of bias

Table DS1 Methodological quality and results of trials evaluating suicide prevention strategiesa (continued)
Random Blinding of Blinding of
sequence Allocation participants & outcome Incomplete Selective Other
OR (95%CI) generation Concealment personnel assessment outcome data Reporting Biases
Complex Psychosocial Interventions (continued)
Fleischmann,
0.19 (0.08 – 0.45) Low Low Unclear Unclear Unclear Low Unclear
2008
Grimholt, 2015 1.41 (0.08 – 23.64) Low Low High Low Low Low High
Hassanian-
Moghaddam, 1.95 (0.63 – 6.07) Low Low Unclear Low Low Unclear Low
2015
Hvid, 2011 1.82 (0.19 – 17.88) Low Low Unclear Low Low Low High

Kawanishi, 2014 0.88 (0.52 – 1.51) Low Low High Low Low Low Unclear

Moller, 1992 1.61 (0.27 – 9.52) High Unclear Unclear Unclear Low Unclear Unclear

Morgan, 1993 No events Unclear Low Unclear High Low Unclear Unclear

Morthorst, 2012 7.21 (0.14 – 363.52) Low Low Unclear High Low Low High

Motto, 2001 1.13 (0.64 – 1.99) Unclear Unclear Unclear Low Low Unclear Unclear

Mouaffak, 2015 7.34 (0.15 – 369.95) Unclear Low High Low Low Unclear Unclear
Mousavi, 2014 0.29 (0.05 – 1.75) Unclear Unclear Unclear Unclear Unclear Low Unclear

Raue, 2010 6.50 (0.13 – 330.64) Low Unclear High Low Low Low Unclear

Schulberg, 2005 No events Low Low Unclear Unclear Low Unclear Unclear

Sun, 2012 1.06 (0.15 – 7.76) Unclear Low Unclear High High Unclear High

Unutzer, 2006 No events Low Low High Low Unclear Low Unclear

Vaiva, 2006 0.70 (0.07 – 6.99) Low Low High Low Low Unclear High
Van Heeringen,
0.85 (0.28 – 2.56) Low Unclear Unclear Low Low Unclear High
1995
Vijayakumar,
0.61 (0.08 – 4.58) Unclear Unclear Unclear Low High Unclear Unclear
2013
Walsh, 2001 1.23 (0.33 – 4.59) Unclear Low Unclear High Low Unclear High

OR = Odds Ratio; 95%CI = 95 percent confidence interval


a. Based on the Cochrane Risk of Bias Tool: Low = low risk of bias; High = high risk of bias; Unclear= Insufficient evidence to judge risk of bias

Table DS1 Methodological quality and results of studies evaluating strategies to prevent suicidea (continued)
Random Blinding of Blinding of
sequence Allocation participants & outcome Incomplete Selective Other
OR (95%CI) generation Concealment personnel assessment outcome data Reporting Biases
Higher Level Care Interventions
Bateman, 2008 0.12 (0.00 – 5.89) Unclear Unclear High High Unclear Unclear Unclear

Jones, 2008 0.40 (0.02 – 8.21) High Low High Low Low Unclear High
Van Der Sande,
0.49 (0.05 – 4.74) Low Low Unclear Low Low Unclear Unclear
1997
Somatic Therapies

George, 2014 No events Unclear Unclear Low Low Unclear Unclear Unclear
Nordenskjold,
0.14 (0.00 – 6.82) Low Low Unclear Unclear Low Unclear High
2013
Randomized Controlled Trials of Medications

Antidepressants

Hirsch, 1982 No events Unclear Unclear Unclear Unclear Unclear Unclear Unclear
Montgomery,
No events Unclear Unclear Low Unclear Unclear Unclear Unclear
1983
Nelson, 2007 No events Unclear Unclear Low Low Low Unclear Unclear

Phillips, 2009 No events Low Low Low Unclear Unclear Unclear Unclear

Rosenthal, 2013 No events Low Low Low Unclear Low Low Unclear

Verkes, 1998 0.13 (0.00 – 6.67) Unclear Unclear Low Unclear Low Low Low

Zisook, 2011 No events Unclear Unclear Low Low Low Unclear Unclear

Mood Stabilizers

Bauer, 2000 0.14 (0.00 – 7.31) Unclear Unclear Low Unclear Low Unclear Unclear

Girlanda, 2014 6.44 (0.13 – 327.93) Low Low Unclear Unclear High Low High

Khan, 2011 No events Low Low Low Unclear Low Low Low
Lauterbach,
0.13 (0.01 – 1.27) Low Unclear Low Unclear Unclear Low High
2008

OR = Odds Ratio; 95%CI = 95 percent confidence interval


a. Based on the Cochrane Risk of Bias Tool: Low = low risk of bias; High = high risk of bias; Unclear= Insufficient evidence to judge risk of bias
Table DS1 Methodological quality and results of studies evaluating strategies to prevent suicidea (continued)
Random Blinding of Blinding of
sequence Allocation participants & outcome Incomplete Selective Other
OR (95% CI) generation Concealment personnel assessment outcome data Reporting Biases
Mood Stabilizers (continued)
Prien, 1973a 0.12 (0.00 – 5.91) Unclear Unclear Unclear Low Unclear Unclear Unclear

Prien, 1973b 0.14 (0.00 – 7.02) Unclear Unclear Unclear Low Unclear Unclear Unclear

Nutritional Supplement
Hallahan,
No events Low Low Low Unclear Low Unclear Unclear
2007
Pooled Analysis of Randomized Controlled Trials of Medicationsb

Antidepressants

Aursnes, 2005 n/a Low Low Low Low Low Unclear Low

Hammad, 2006 n/a Low Low Low Low Low Unclear Low

Khan, 2007 n/a Low Low Low Low Low Unclear High

Antipsychotics

Khan, 2013 n/a Low Low Low Low Low Unclear High

Storosum, 2003 n/a Low Low Low Low Low Unclear Low

Mood Stabilizers

Storosum, 2005 n/a Low Low Low Low Low Unclear Low

OR = Odds Ratio; 95%CI = 95 percent confidence interval; N/A = not applicable

a. Based on the Cochrane Risk of Bias Tool; Low = low risk of bias; High = high risk of bias; Unclear= Insufficient evidence to judge risk of bias
b. Peto odds ratios could not be calculated for the results of individual pooled analysis of randomized controlled trials due to large imbalances in treatment arms.
Table DS2 Incidence rate ratios of death by suicide for various suicide prevention strategiesa

Intervention domain N IRR 95% Confidence Interval

Complex Psychosocial Interventions 29 0.93 0.65 – 1.33

Intensive follow-up programs 11 0.71 0.34 – 1.48

Comprehensive follow-up programs 5 0.30 0.07 – 1.22

Case management after suicidal behavior 4 0.93 0.56 – 1.54

Case management for psychosis 4 1.16 0.50 – 2.55

Letter/Phone Contact 7 1.16 0.75 – 1.79

Psychotherapyb 24 0.79 0.41 – 1.53

Cognitive Behavioral Therapies 20 0.72 0.34 – 1.53

Cognitive Behavioral Therapy-Suicide Prevention 6 0.30 0.08 – 1.11


Problem Solving Therapy 4 1.00 0.25 – 4.04

Non-Cognitive Behavioral Therapies 5 1.10 0.27 – 4.38

Medications (Randomized Trials) 14 0.10 0.00 – 32.27

Lithium 6 0.14 0.00 – 9.41

Medications (Pooled Analysis) 6 1.15 0.56 – 2.39

IRR = incidence rate ratio; N = number of studies


a. There were too few studies to calculate IRR for the following domains: means restriction; cognitive behavioral therapies (except for cognitive behavioral therapy designed to address
suicidal behavior and problem solving therapy); non-cognitive behavioral therapies; aftercare; other types of intensive follow-up programs, primary mental health integrated care; nursing-
led caregiver education; higher-level care; somatic therapies, and medications by type (except for lithium).
b. A third arm of the CBT trial conducted by Tarrier et al (online supplemental references107) evaluated supportive therapy for psychosis. This arm was included in the analysis of non-CBT
therapies.
Fig. DS1
Systematic Review and Meta-Analysis Protocol

Appendix DS1

Title: Strategies to prevent death by suicide: a meta-analysis of randomized controlled trials

Authors: Riblet NB, Shiner B, Young-Xu Y, Watts BV

Introduction:

Suicide is a significant public health concern in the United States. The Centers for Disease

Control estimates that over 30,000 people die by suicide each year in the U.S.1 Suicide is also the

tenth leading cause of death among people of all ages1 and the third leading cause of death among

adolescents ages 15 to 24 years old.1 Suicidal behavior is associated with significant costs to

society with an estimated annual loss of 34.6 billion dollars due to healthcare costs and lost

earnings.2 Caring for a loved one who is suicidal can also cause considerable emotional toll on

family members and loved ones.3

Although many strategies have been proposed to prevent suicide, the efficacy of these various

interventions remains unclear.4-5 For example, antidepressants may have a protective effect in

mood disorders but these findings are largely based on studies which are underpowered and of

short duration.5 Similarly, it is unclear whether antipsychotics or mood stabilizers have an anti-

suicidal effect.5 The role of behavioral interventions and more comprehensive suicide prevention

programs in reducing suicide also remains unclear. Safer packaging of medications has shown

some promise but these studies may be limited by confounding and inadequate duration.6

In 2005, Mann et al published a comprehensive systematic review of all available suicide

prevention strategies.7 These authors concluded that there was evidence to support the following

interventions including physician education in depression recognition, gatekeeper education and

lethal means restriction.7 The authors, however, determined that the efficacy of individual

components of multifaceted strategies for suicide prevention remains unclear.7

Page 1 of 23
Systematic Review and Meta-Analysis Protocol

Since Mann et al’s review, there have been several additional studies of suicide prevention

interventions. These studies continue to produce mixed results. For example, a randomized

controlled trial of a worldwide disseminated gatekeeper training called ASIST found that there

was no significant improvement in gatekeeper knowledge about suicide prevention with this

training.8 On the other hand, a randomized controlled trial of a brief psychological intervention

after self harm found that the therapeutic intervention was associated with a significant (and

sustained) reduction in symptoms of suicidal ideation.9

Our current review provides an update to the previous comprehensive systematic review

completed by Mann et al in 2005.7 We have used the conceptual framework previously developed

by Mann et al to understand and to evaluate existing strategies for suicide prevention.7 Our aim is

to provide a complete and all-inclusive summary of all of the available evidence for interventions

to prevent suicide. We believe that this information will help inform clinical decisions about

suicide prevention and illuminate areas in need of additional research.

Research Question: Among adults 18 years and older (P), which interventions that are designed

to prevent death by suicide (or suicidal behavior or ideation) (I) have greater efficacy than usual

care condition, placebo or waitlist (C) at preventing death by suicide (O)?

STUDY INCLUSION CRITERIA

Inclusion Criteria Table & Outcomes of Interest:

Category Inclusion Criteria Justification/Explanation Order

Study Design: Randomized While RCTs are not ideal for studying a rare 1

Controlled Trials outcome such as death by suicide, we will

(RCTs), pooled limit our review to this study design because

analysis of RCTs RCTs are the highest quality study design

and the gold standard for evaluating the

efficacy of an intervention.10-11 We will also

Page 2 of 23
Systematic Review and Meta-Analysis Protocol

include the results of pooled analysis of

RCTs of pharmacological interventions to

better understand the role of medications in

preventing death by suicide. We have chosen

this approach because death by suicide is a

rare outcome and individual drug trials are

unlikely to be powered to evaluate for this

particular outcome but pooled analysis using

a systematic approach to identify death by

suicide in drug trials may offer unique

insights.

Population Any adult age 18 Since the risk for death by suicide spans all 2

and older types of adult populations and covers a wide

range of risk factors (e.g. known mental

health problems, access to lethal means,

Veteran status), we plan to broaden our

population to include interventions targeted

at reducing risk for suicide in all adults over

the age of 18. 11 In order to be comprehensive

and include as many studies as possible, we

will include studies if the total population (or

the large majority) was 18 years and older.

We have decided to exclude children and

adolescents from our review since this

population may have uniquely different

Page 3 of 23
Systematic Review and Meta-Analysis Protocol

responses to potential therapeutic

interventions to prevent death by suicide.

Therefore, a comprehensive review of

interventions to prevent death by suicide in

the child/adolescent population would be

warranted but is beyond the scope of the

current review.

Intervention: Interventions We will use a framework based on the work 3

targeted at presented by Mann et al’s in their systematic

preventing death by review of the literature.7 These interventions

suicide (or suicidal fall into the following major categories:7

behavior or 1) Education and Awareness Programs

ideation) in a 2) Screening for individuals at high risk

population 3) Pharmacotherapy

4) Psychotherapy

5) Follow-up Care for Suicide Attempts

6) Restriction of Access to Lethal Means

7) Media Reporting Guidelines for suicide

If necessary, we will create additional

domains and sub-domains (or remove

domains) based on consensus if we identify

new interventions which were not covered

in Mann et al’s review and do not readily fit

into the domains developed in their review

or we identify that there are certain domains

Page 4 of 23
Systematic Review and Meta-Analysis Protocol

which were not studied using a randomized

design methodology.7

Comparison Usual care or Since the primary concern in the literature is 4

placebo condition whether any suicide prevention strategies are

(placebo or sham) effective, we will include as our comparator

or waitlist arm usual care condition or placebo or

waitlist. Usual care may include any

intervention that would be considered

standard of care based on current clinical

practice and would be unethical to withhold

from a patient. This may include typical

supportive care interventions such as routine

visits with a physician and provider

education as well as allowing patients to seek

out psychotherapy or continue taking

pharmacological agents. Similarly, we will

allow waitlist comparison. We will also

allow the comparison arm to include placebo

or sham as this would be relevant (and

expected) in the case of pharmacotherapy

trials or somatic therapies such as

electroconvulsive therapy and transcranial

magnetic stimulation. Since there is limited

knowledge about the efficacy of any

available interventions to prevent death by

Page 5 of 23
Systematic Review and Meta-Analysis Protocol

suicide, we have not included active

treatment as a comparison. The aim of this

review is to identify which interventions are

superior to standard treatment. Our review

will not address issues of equivalency or

non-inferiority. We will, however, briefly

summarize in a qualitative fashion the results

of trials that evaluate active treatment as

comparison.

Outcomes: Death by suicide We will require that included studies report 5

on death by suicide. While studies are more

likely to report on intermediary outcomes

such as suicidal behavior or suicidal ideation,

these outcomes are known to be more

susceptible to measurement bias.12

Furthermore, it is clear that from a societal

perspective the prevention of death by

suicide is the most relevant clinical

outcome.1-2 To broaden our search and be as

inclusive as possible, we will include studies

that report death by suicide as a primary

outcome or a secondary outcome. In the

event that death suicide was a secondary

outcome, we will require that the primary

aim of the report included the prevention of

Page 6 of 23
Systematic Review and Meta-Analysis Protocol

suicidal ideation and/or suicidal behavior.

We have selected this method to help

minimize biases due to the fact that these

studies will be far less likely to be powered

to evaluate for death by suicide. In addition,

we will exclude a study if they do not clearly

state whether or not death by suicide

occurred in one or both arms (i.e. we will not

assume that no mention of death by suicide

means no death by suicide occurred).

However, to be as inclusive as possible, we

will include studies even if the methods that

they used to assess for death by suicide were

at higher risk for bias (e.g. informants, chart

review). Finally, given that death by suicide

is a rare outcome, we will include studies

even if they report that no events (death by

suicide) occurred in either arm.

Outcomes of Interest:

Primary Outcome:

Odds of death by suicide

Justification: This is the most relevant question with regards to whether suicide

prevention strategies are effective.1-2

Search Strategy:

Page 7 of 23
Systematic Review and Meta-Analysis Protocol

Databases:

We used the following databases for our review including Medline (via Ovid), the

Cochrane Library, PsycINFO, Excerpta Medica Database (EMBASE) and The

Cumulative Index to Nursing and Allied Health Literature (CINAHL).

Search terms:

We used exploded MeSH terms and key words to generate the following themes: suicide,

prevention and control and treatment. We then used “OR” to combine the theme

prevention and control and treatment and the Boolean term “AND” to find their

intersection with the theme suicide (see search strategy below). We applied this approach

during our search of the Medline database and modified our approach as necessary to

search the Cochrane Library, PsycINFO, EMBASE, CINAHL and clinicaltrials.gov.

Limits:

We applied a randomized controlled trial limit in our search of the Cochrane Library.

Special Strategies:

We applied Cochrane’s recommended highly sensitive search strategy for identifying

randomized controlled trials in our Medline, EMBASE and PsycINFO search because our

original search yielded an unfeasible number of studies to search manually.13-16 Similarly,

we applied a sensitive search strategy for identifying randomized controlled trials in

CINAHL recommended by the Scottish Intercollegiate Guidelines Network.17

Results of Search Strategy:

Using the above search terms, we identified 11,866 potentially eligible studies through

our electronic database searching. Details of the structure and findings of our various

search strategies are provided below (see search strategy).

Additional Search methods:

In order to identify relevant published and unpublished studies which may have been

missed in our preliminary search, we searched clinicaltrials.org and manually reviewed

Page 8 of 23
Systematic Review and Meta-Analysis Protocol

references of relevant articles. The results of our clinicaltrials.gov search and reference

review are provided below (see search strategy).

Prior Reviews:

There have been several prior reviews of this topic both in the general population as well as in

specific subpopulations such as Veterans.4-7 The most recent comprehensive systematic review

that has been published, however, was the review undertaken by Mann et al in 2005.7 We feel

that updating the work of Mann et al is important since there have been several additional

randomized controlled trials which have been published since 2005 which may help shed

additional light on this topic.8-9

Results of Preliminary Search

We have provided an example of an abstract that we located during our initial search which meets

all of our inclusion criteria (see preliminary abstracts).

Protocol Amendments:

Since our protocol was written, we have made one amendment. Please see the protocol

amendments section below for further details.

Page 9 of 23
Systematic Review and Meta-Analysis Protocol

Search Terms Table

Theme MeSH Terms Key Words

Suicide Suicidal Ideation, Suicide, Attempted, Suicide Suicide

Prevention and Health education, Health promotion, mass Health education, Health
control screening, risk assessment, school health promotion, Mass screening,
services, student health services, firearms, gas risk assessment, school
poisoning, pesticides, drug overdose, mass health services, student
media, harm reduction, emergency services, health services, firearms, gas
hospital, poisoning, pesticide, drug
overdose, mass media, harm
reduction, emergency service
hospital, means restriction,
restricted access, media,
suicide prevention

pc.fs (free floating


subheading of prevention &
control)

Treatment antidepressive agents/pd [pharmacology], serotonin uptake inhibitors,


serotonin uptake inhibitors, antipsychotic ECT, TCA, SSRI, MAOI,
agents/pd [pharmacology], psychotropic CBT, rTMS, ketamine, VNS,
drugs/pd [pharmacology], antimanic vagus nerve stimulator,
agents/pd [pharmacology], electroconvulsive TENS
therapy, monoamine oxidase inhibitors/pd
[pharmacology], psychotherapy, cognitive
therapy, transcranial magnetic stimulation,
ketamine, vagus nerve stimulation,
transcutaneous electric nerve stimulation

Search Strategy

Page 10 of 23
Systematic Review and Meta-Analysis Protocol

Database: Ovid MEDLINE(R) inception (1945) to Present (December 31, 2015)


The search strategy was created with assistance from a Librarian at the Dartmouth Hitchcock
Medical Library

Number Search terms

1 Suicide.mp or exp Suicide/ or exp Suicide, Attempted/

2 Suicidal ideation.mp or exp Suicidal Ideation/

3 1 or 2

4 Health education.mp or exp Health Education/

5 Health promotion.mp or exp Health Promotion/


6 Mass screening.mp or exp Mass Screening/
7 Risk assessment.mp or exp Risk Assessment/

8 Student health services.mp or exp Student Health Services/

9 Firearms.mp or exp Firearms/


10 Gas poisoning.mp or exp Gas Poisoning/
11 Pesticides.mp or exp Pesticides/

12 Drug overdose.mp or exp Drug Overdose/


13 Mass media.mp or exp Mass Media/
14 Harm reduction.mp or exp Harm Reduction/

15 Emergency service hospital.mp or exp Emergency Service, Hospital/


16 Means restriction.mp

17 Restricted access.mp
18 Media.mp

19 Suicide prevention.mp

20 Exp Antidepressive Agents/pd [Pharmacology/]

21 Exp Serotonin Uptake Inhibitors/pd [Pharmacology/]

22 Serotonin uptake inhibitors.mp

23 Exp Antipsychotic Agents/pd [Pharmacology/]


24 Exp Psychotropic Drugs/pd [Pharmacology/]

25 Exp Antimanic Agents/pd [Pharmacology/]

26 Exp Electroconvulsive Therapy/ or ECT.mp

Page 11 of 23
Systematic Review and Meta-Analysis Protocol

27 Exp Monoamine Oxidase Inhibitors/pd [Pharmacology/]

28 Exp Psychotherapy/

29 Exp Cognitive Therapy/


30 TCA.mp
31 SSRI.mp

32 MAOI.mp

33 CBT.mp

35 Exp Transcranial Magnestic Stimulation/ or rTMS.mp


36 Ketamine.mp or exp.Ketamine/
37 VNS.mp or exp Vagus Nerve Stimulation/

38 Exp Transcutaneous Electric Nerve Stimulation/


39 Vagus nerve stimulator.mp

40 TENS.mp

41 Pc.fs

4 OR 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 OR 12 OR 13 OR 14 OR 15
OR 16 OR 17 OR 18 OR 19 OR 20 OR 21 OR 22 OR 23 OR 24 OR 25
42
OR 26 OR 27 OR 28 OR 29 OR 30 OR 31 OR 32 OR 33 OR 34 OR 35
OR 36 OR 37 OR 38 OR 39 OR 40 OR 41

43 3 and 42 (Citations located = 21,469)


(randomized controlled trial.pt OR controlled clinical trial.pt OR
44 randomized.ab OR placebo.ab OR drug therapy.fs. OR randomly.ab OR
trial.ab OR groups.ab)
45 Exp animals/not humans.sh

46 #44 NOT #45

47 46 and 43 (Citations located = 4,641)


*Of note we applied Cochrane’s recommended highly sensitive search strategy for identifying
randomized controlled trials (Line 44, 45, 46) because our original search yielded unfeasible
number of studies to search manually.13

EMBASE Search:

Dates Searched: Inception thru December 31, 2015

Page 12 of 23
Systematic Review and Meta-Analysis Protocol

Search terms were adapted from our Medline search to conform with Emtree and applied as

follows: (Suicide OR Suicidal Ideation OR Suicide Attempt) AND (health education OR health

promotion OR mass screening OR risk assessment OR health service OR firearms OR gas

poisoning OR pesticide OR drug overdose OR mass media OR harm reduction OR emergency

health service OR means restriction OR restricted access OR media OR suicide prevention OR

suicide prevention and control OR antidepressive agents OR serotonin uptake inhibitor OR

neuroleptic agent OR psychotropic drugs OR tranquilizer OR electroconvulsive therapy OR

monoamine oxidase inhibitor OR psychotherapy OR cognitive therapy OR TCA OR SSRI OR

MAOI OR CBT OR transcranial magnetic stimulation OR ketamine OR rTMS OR VNS OR

vagus nerve stimulation OR vagus nerve stimulator OR transcutaneous nerve stimulation OR

TENS) AND (random$ OR factorial$ OR crossover$ OR cross over$ OR cross-over$ OR

placebo$ OR doubl$ adj blind$ OR singl$ adj blind$ OR assign$ OR allocat$ OR volunteer$ OR

crossover-procedure OR double-blind procedure OR randomized controlled trial OR single-blind

procedure)

Search Strategy: We used the same search approach as described in our search of Medline but as

described above modified the search to conform with Emtree. Furthermore, because our original

search yielded an unfeasible number of citations to search by hand, we applied a sensitive search

strategy recommended by Cochrane for identifying randomized controlled trials.14

Total Citations Located (no limits): 33,832

Total Citations Located (sensitive strategy recommended by Cochrane): 3,297

PsycINFO Search

Dates Searched: Inception thru December 31, 2015

Search Terms: (Suicide OR Suicidal Ideation OR Suicide Attempt) AND (health education OR

health promotion OR mass screening OR risk assessment OR student health services OR firearms

OR gas poisoning OR pesticides OR drug overdose OR mass media OR harm reduction OR

emergency service hospital OR means restriction OR restricted access OR media OR suicide

Page 13 of 23
Systematic Review and Meta-Analysis Protocol

prevention OR antidepressive agents OR serotonin uptake inhibitors OR antipsychotic agents OR

psychotropic drugs OR antimanic agents OR electroconvulsive therapy OR monoamine oxidase

inhibitors OR psychotherapy OR cognitive therapy OR SSRI OR MAOI OR CBT OR TCA OR

transcranial magnetic stimulation OR ketamine OR rTMS OR VNS OR vagus nerve stimulation

OR vagus nerve stimulator OR transcutaneous electric nerve stimulation OR TENS) AND

(SU.EXACT (“Treatment Effectiveness Evaluation”) OR SU.EXACT.EXPLODE (“Treatment

Outcomes”) OR SU.EXACT(“Placebo”) OR SU.EXACT(“Followup Studies”) OR placebo* OR

random* OR “comparative stud*” OR clinical NEAR/3 trial* OR research NEAR/3 design OR

evaluat* NEAR/3 stud* OR prospectiv* NEAR/3 stud* OR (singl* OR doubl* OR trebl* OR

tripl*) NEAR/3 (blind* OR mask*))

Search Strategy: We used the same search approach as described in our search of Medline. We

applied a sensitive search strategy recommended by Cochrane for identifying randomized

controlled trials in PsycINFO.15-16

Total Citations Located (no limits)—19,592

Total Citations Located (with sensitive search strategy recommended by Cochrane)---1,525

CINAHL Search:

Dates Searched: Inception thru December 31, 2015

Search Terms: (Suicide OR Suicidal Ideation OR Suicide Attempt) AND (health education OR

health promotion OR mass screening OR risk assessment OR student health services OR firearms

OR gas poisoning OR pesticides OR drug overdose OR mass media OR harm reduction OR

emergency service hospital OR means restriction OR restricted access OR media OR suicide

prevention OR antidepressive agents OR serotonin uptake inhibitors OR antipsychotic agents OR

psychotropic drugs OR antimanic agents OR electroconvulsive therapy OR monoamine oxidase

inhibitors OR psychotherapy OR cognitive therapy OR SSRI OR MAOI OR CBT OR TCA OR

transcranial magnetic stimulation OR ketamine OR rTMS OR VNS OR vagus nerve stimulation

OR transcutaneous electric nerve stimulation OR vagus nerve stimulator OR TENS) and ((MH

Page 14 of 23
Systematic Review and Meta-Analysis Protocol

“Clinical Trials+”) OR PT Clinical Trial OR TX clinic* n1 trial* OR (TX ((singl* n1 blind*) or

(singl* n1 mask*)) or TX ((doubl* n1 blind*) or (doubl* n1 mask*)) OR TX ((tripl* n1 blind*)

or (tripl* n1 mask*)) or TX ((trebl* n1 blind*) or (trebl* n1 mask*)) OR TX randomi* control*

trial* OR (MH “Random Assignment”) OR TX random* allocate* OR TX placebo* OR (MH

“Placebos”) OR (MH “Quantitative Studies”) or TX allocate* random*)

Search strategy: We used the same search approach as described in our search of Medline. We

then applied a search filter limit to identify randomized trials recommended by the Scottish

Intercollegiate Guidelines Network.17

Total Citations Located (no limits)—8,212

Total Citations Located (with sensitive search strategy from SIGN)--- 2,171

Cochrane Library Search:

Dates Searched: Inception thru December 31, 2015

Search Strategy: Search Terms: (Suicide OR Suicidal Ideation OR Suicide Attempt) AND

(health education OR health promotion OR mass screening OR risk assessment OR student health

services OR firearms OR gas poisoning OR pesticides OR drug overdose OR mass media OR

harm reduction OR emergency service hospital OR means restriction OR restricted access OR

media OR suicide prevention OR antidepressive agents OR serotonin uptake inhibitors OR

antipsychotic agents OR psychotropic drugs OR antimanic agents OR electroconvulsive therapy

OR monoamine oxidase inhibitors OR psychotherapy OR cognitive therapy OR SSRI OR MAOI

OR CBT OR TCA OR transcranial magnetic stimulation OR ketamine OR rTMS OR VNS OR

vagus nerve stimulation OR vagus nerve stimulator OR transcutaneous electric nerve stimulation

OR TENS)

Search Strategy: The Cochrane reports that an RCT filter is not required because all the records

are thoroughly and correctly indexed. Therefore, we only applied a randomized controlled trial

limit in our search. 18

Total Citations Located (no limits)—2,160

Page 15 of 23
Systematic Review and Meta-Analysis Protocol

Total Number (with RCT limits applied): 232

Clinical Trials.gov:

Dates: Inception thru December 31, 2015

Search Strategy: “suicide & prevention”

Results

o Three trials-no usable data

o Thirteen trials-ongoing/actively recruiting

Reference Review

Additional references located: Three

Page 16 of 23
Systematic Review and Meta-Analysis Protocol

Preliminary Abstracts

Abstract

Authors: Vijayakumar L, Jeyaseelan L, Kumar S, Mohanraj R, Devika S, Manikandan S. A

central storage facility to reduce pesticide suicides--a feasibility study from India. BMC Public

Health 2013; 13: 850.

Background: Pesticide suicides are considered the single most important means of suicide

worldwide. Centralized pesticide storage facilities have the possible advantage of delaying access

to pesticides thereby reducing suicides. We undertook this study to examine the feasibility and

acceptability of a centralized pesticide storage facility as a preventive intervention strategy in

reducing pesticide suicides.

Methods: A community randomized controlled feasibility study using a mixed methods

approach involving a household survey; focus group discussions (FGDs) and surveillance were

undertaken. The study was carried out in a district in southern India. Eight villages that engaged

in floriculture were identified. Using the lottery method two were randomized to be the

intervention sites and two villages constituted the control site. Two centralized storage

facilities were constructed with local involvement and lockable storage boxes were

constructed. The household survey conducted at baseline and one and a half years later

documented information on sociodemographic data, pesticide usage, storage and suicides.

Results: At baseline 4446 individuals (1097 households) in the intervention and 3307 individuals

(782 households) in the control sites were recruited while at follow up there were 4308

individuals (1063 households) in the intervention and 2673 individuals (632 households) in the

control sites. There were differences in baseline characteristics and imbalances in the prevalence

of suicides between intervention and control sites as this was a small feasibility study.The results

from the FGDs revealed that most participants found the storage facility to be both useful and

acceptable. In addition to protecting against wastage, they felt that it had also helped prevent

pesticide suicides as the pesticides stored here were not as easily and readily accessible. The

Page 17 of 23
Systematic Review and Meta-Analysis Protocol

primary analyses were done on an Intention to Treat basis. Following the intervention, the

differences between sites in changes in combined, completed and attempted suicide rates per

100,000 person-years were 295 (95% CI: 154.7, 434.8; p < 0.001) for pesticide suicide and 339

(95% CI: 165.3, 513.2, p < 0.001) for suicide of all methods.

Conclusions: Suicide by pesticides poisoning is a major public health problem and needs

innovative interventions to address it. This study, the first of its kind in the world, examined the

feasibility of a central storage facility as a means of limiting access to pesticides and, has

provided preliminary results on its usefulness. These results need to be interpreted with caution in

view of the imbalances between sites. The facility was found to be acceptable, thereby

underscoring the need for larger studies for a longer duration.

Confirmation Table

Category Inclusion Criteria Criteria Met

Study Design: RCTs and pooled analysis of RCT

RCTs.

Population Any adult age 18 and older Population in India at risk for

suicide including adults

Intervention: Interventions targeted at Centralized storage facility for

preventing death by suicide (or pesticide

suicidal behavior or ideation)

in a population

Comparison Usual care or placebo Usual care/No storage facility

condition (placebo or sham) or

waitlist

Page 18 of 23
Systematic Review and Meta-Analysis Protocol

Outcomes: Death by suicide Evaluated death by suicide as an

outcome

Page 19 of 23
Systematic Review and Meta-Analysis Protocol

Protocol Changes

Date

Section of Protocol Original version New version

Type of amendment* Justification

October 15, 2016 We searched Based on feedback from

Search Strategy Medline, Cochrane reviewers, we added these

-Included additional databases in Library, and additional databases and

including EMBASE and PsycINFO from revised our search strategy to

CINAHL in addition to Medline, inception through increase the likelihood that we

Cochrane Library and PsycINFO December 31, 2015 would capture all relevant

-expanded our search terms to studies. Because we located an

include the key word “suicide unfeasible number of citations

prevention” in all our electronic to manually search after

database searches (in the case of applying the aforementioned

EMBASE we used suicide search strategies, we also

prevention as an emtree term). incorporate recommended

“Suicide prevention” is not a highly sensitive search strategy

MeSH term in Medline so we for identifying RCTs in our

could only use it as a keyword. various electronic searches.

-We used a highly sensitive

search strategy recommended by

Cochrane to identify RCTs in

Medline, EMBASE and

PsycINFO. We will use a search

strategy for identifying RCTs in

Page 20 of 23
Systematic Review and Meta-Analysis Protocol

CINAHL recommended by the

recommended by the Scottish

Intercollegiate Guidelines

Network.

*Amendments may include additions, deletions, changes, and/or clarifications

References

1. Centers for Disease Control and Prevention. National Center for Injury Prevention and

Control. Suicide facts at a glance, 2012

(http://www.cdc.gov/ViolencePrevention/pdf/Suicide_DataSheet-a.pdf).

2. Centers for Disease Control and Prevention. National Center for Injury Prevention and

Control. Web-based Injury Statistics Query and Reporting System (WISQARS), 2010

( http://wisqars.cdc.gov/:8080/costT/).

3. McLaughlin C, McGowan I, O’Neill S, Kernohan G. The burden of living with and

caring for a suicidal family member. J Ment Health 2014; 23: 236-240.

4. Ernst CL, Goldberg JF. Antisuicide properties of psychotropic drugs: a critical review.

Harv Rev Psychiatry 2004; 12: 14-41.

5. Bagley SC, Munjas B, Shekelle P. A systematic review of suicide prevention programs

for military or veterans. Suicide Life Threat Behav 2010; 40: 257-265.

6. Scott A, Guo B. For which strategies of suicide prevention is there evidence of

effectiveness? World Health Organization HEN Synthesis Report, 2012

(http://www.euro.who.int/__data/assets/pdf_file/0003/168843/HEN-Suicide-Prevention-

synthesis-report.pdf).

Page 21 of 23
Systematic Review and Meta-Analysis Protocol

7. Mann JJ, Apter A, Bertolote J, Beautrais A, Currier D, Haas A, et al. Suicide prevention

strategies: a systematic review. JAMA 2005;294: 2064-74.

8. Sareen J, Isaak C, Bolton SL, Enns MW, Elias B, Deane F, et al. Gatekeeper training for

suicide prevention in First Nations community members: a randomized controlled trial.

Depress Anxiety 2013; 10: 1021-29.

9. Husain N, Afsar S, Ara J, Fayyaz H, Rahman RU, Tomenson B, et al. Brief psychological

intervention after self-harm: randomized controlled trial from Pakistan. Br J Pyschiatry

2014; 6:462-70.

10. Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB. Designing Clinical

Research. Lippincott Williams & Wilkins, 2013.

11. Jacobs DG, Baldessarini RJ, Conwell Y, Fawcett JA, Horton L, Meltzer H, et al. Practice

Guidelines for the Assessment and Treatment of Patients with Suicidal Behaviors. APA,

2010

(http://psychiatryonline.org/pb/assets/raw/sitewide/practice_guidelines/guidelines/suicide

.pdf).

12. Glenn CR, Nock MK. Improving the short-term prediction of suicidal behavior. Am J

Prev Med 2014; 47: S176-80.

13. Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching for studies, Box 6.4.c:

Cochrane Highly Sensitive Search Strategy for identifying randomized trials in

MEDLINE: sensitivity-maximizing version (2008 revision); Ovid format. In: Higgins

JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions

Version 5.1.0 (updated March 2011). The Cochrane Collaboration, 2011.

(www.cochrane-handbook.org)

14. Lefebvre C, Manheimer E, Glanville J. Chapter 6.4.11.2 Search filters for identifying

randomized trials in EMBASE. In: Higgins JPT, Green S (editors). Cochrane Handbook

Page 22 of 23
Systematic Review and Meta-Analysis Protocol

for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011). The

Cochrane Collaboration, 2011. (www.cochrane-handbook.org)

15. Watson RJ, Richardson PH. Identifying randomized controlled trials of cognitive therapy

for depression: comparing the efficiency of Embase, Medline and PsycINFO

bibliographic databases. Br J Med Psychol 1999; 72(Pt 4): 535 - 42.

16. Identifying RCTs in PsycINFO. Search strategy, amended to ProQuest format.

(http://work.cochrane.org/psycinfo)

17. Scottish Intercollegiate Guidelines Network Healthcare Improvement. Search Filters.

http://www.sign.ac.uk/methodology/filters.htm#random

18. Cochrane Work. RCT filters for different databases. (http://work.cochrane.org/rct-filters-

different-databases)

Page 23 of 23
Appendix DS2

Excluded Studies

Table 1: Randomized trials comparing an intervention versus active control for suicide prevention. (horizontal position)

Study Number of suicides Study arm


Author, year N Study Population Intervention Comparison Significance
Duration by study arm favored

Psychotherapy

Rogerian
Cognitive
Cottraux, 20085 65 BPD 24 months Supportive No events Neither NS
Therapy
Therapy

BPD, PTSD & self-


Harned, 20141 26 15 months DBT + PE DBT DBT – 1 Neither NS
injury

Breast cancer plus Behavioral


Hopko, 20134 80 24 months PST No events Neither NS
MDD Activation

CBT for Cognitive


Klingberg, 20122 198 Schizophrenia 12 months No events Neither NS
psychosis remediation

Arm 2: DBT-I
BPD plus ≥2 SA
Linehan, 20153 99 24 months DBT-S Arm 3: Standard Standard DBT - 1 Neither NS
&/or NSSI
DBT
Antidepressants
Moderate-Severe
Perroud, 20096 811 Unipolar 3 months Escitalopram Nortriptyline Nortriptyline – 1 Neither NS
Depression
Rucci, 20117 291 Nonpsychotic MDD 4 months IPT Escitalopram No events Neither NS

Arm 2: Bupropion
SR plus
Nonpsychotic
Escitalopram Escitalopram;
Zisook, 20118 665 chronic &/or 7 months No events Neither NS
plus placebo Arm3:
recurrent MDD
Venlafaxine XR
plus Mirtazapine

Antipsychotics

Schizophrenia &
Paliperidone Daily oral
Alphs, 20159,a 450 history of 15 months No events Neither NS
palmitate antipsychotics*
incarceration

Low dose Ultra-low dose


Seen in ER after
Battaglia, 199910,b 58 6 months Fluphenazine Fluphenazine No events Neither NS
recent SA
decanoate decanoate

Schizophrenia or Physician’s LAI-R – 1;


Buckley, 201511,c 305 30 months LAI-R Neither NS
SCAD choice oral SGA SGA – 1

Arm 2:
Aripiprazole – 1;
Crespo-Facorro, First Episode Ziprasidone
202 12 months Aripiprazole Ziprasidone – 1; Neither NS
201412 Schizophrenia Arm 3:
Quetiapine - 2
Quetiapine

Meltzer, 2003 Schizophrenia and Clozapine – 5;


980 24 months Clozapine Olanzapine Neither NS
(InterSePT) 13 high risk for suicide Olanzapine – 3

ITT: RR 1.19
Schizophrenia (0.61-2.31)
Strom, 201114 18,154 (seen in naturalistic 12 months Ziprasidone Olanzapine Time on assigned Neither NS
practice) treatment
RR: 1.37 (0.66 – 2.85)
Thomas, 201015 9,858 Schizophrenia 14,147 PY Sertindole Risperidone HR 0.72 (0.36 – 1.41) Neither NS

Mood Stabilizers

Bipolar Disorder Lithium– 0;


Oquendo, 201116 98 30 months Lithium Valproate Neither NS
plus prior SA Valproate - 0

Arm 2:
Thies-Flechtner, Lithium– 0;
MDD, Bipolar Amitriptyline;
1996d 378 30 months Lithium Carbamazepine - 4; Lithium NS
Disorder or SCAD Arm3:
(MAP Study) 17 Other medications – 5
Carbamazepine

BPD = Borderline Personality Disorder; CBT = Cognitive Behavioral Therapy; DBT = Dialectical Behavior Therapy; DBT-S: DBT skills training; DBT-I: DBT individual therapy; ER = Emergency Room
HR = hazard ratio; IPT = Interpersonal psychotherapy; InterSePT = International Suicide Prevention Trial; ITT = Intention to treat; LAI-R = Long-acting injectable risperidone;
MDD = Major Depressive Disorder; NS = Not statistically significant; NSSI = non-suicidal self-injury; PE = Prolonged exposure; PST = problem solving therapy; PTSD = Post-traumatic stress disorder;
PY = person years; RR = Risk Ratio; SA = Suicide Attempt; SCAD = Schizoaffective Disorder; SGA = Second generation antipsychotics (physician’s choice); SR = sustained release; XR = extended
a. Oral antipsychotics may have included aripiprazole, haloperidol, olanzapine, paliperidone, perphenazine, quetiapine, and risperidone
b. Low dose included 12.5 milligram monthly injections of fluphenazine decanoate; Ultra-low dose included 1.5 milligram monthly injections of fluphenazine decanoate
c. SGA may have included aripiprazole, olanzapine, quetiapine, risperidone, ziprasidone, paliperidone, asenapine, and iloperidone
d. The authors report that other medications may have included antidepressants, neuroleptics or no medications
References

1. Harned MS, Korslund KE, Linehan MM. A pilot randomized controlled trial of Dialectical Behavior Therapy with and without the Dialectical Behavior

Therapy Prolonged Exposure protocol for suicidal and self-injuring women with borderline personality disorder and PTSD. Behav Res Ther 2014;55:7-

17.

2. Klingberg S, Herrlich J, Wiedemann G, Wolwer W, Meisner C, Engel C, et al. Adverse effects of cognitive behavioral therapy and cognitive remediation

in schizophrenia: results of the treatment of negative symptoms study. J Nerv Ment Dis 2012;200:569-76.

3. Linehan MM, Korslund KE, Harned MS, Gallop RJ, Lungu A, Neacsiu AD, et al. Dialectical behavior therapy for high suicide risk in individuals with

borderline personality disorder: a randomized clinical trial and component analysis. JAMA Psychiatry 2015; 72: 475-82.

4. Hopko DR, Funderburk JS, Shorey RC, McIndoo CC, Ryba MM, File AA, et al. Behavioral activation and problem-solving therapy for depressed breast

cancer patients: preliminary support for decreased suicidal ideation." Behav modif 2013; 37: 747-67.

5. Cottraux J, Note ID, Boutitie F, Milliery M, Genouihlac V, Yao SN, et al. Cognitive therapy versus rogerian supportive therapy in borderline personality

disorder. Psychother Psychosom 2008; 78: 307-16.

6. Perroud N, Uher R, Marusic A, Rietschel M, Mors O, Henigsberg N, et al. Suicidal ideation during treatment of depression with escitalopram and

nortriptyline in Genome-Based Therapeutic Drugs for Depression (GENDEP): a clinical trial. BMC Medicine 2009; 7: 60.

7. Rucci P, Frank E, Scocco P, Calugi S, Miniati M, Fagliolini A, et al. Treatment-emergent suicidal ideation during 4 months of acute management of

unipolar major depression with SSRI pharmacotherapy or interpersonal psychotherapy in a randomized clinical trial. Depress Anxiety 2011; 28: 303-9.
8. Zisook S, Lesser IM, Lebowitz B, Rush AJ, Kallenberg G, Wisniewski SR, et al. Effect of antidepressant medication treatment on suicidal ideation and

behavior in a randomized trial: An exploratory report from the combining medications to enhance depression outcomes study. J Clin Psychiatry 2011; 72:

1322-32.

9. Alphs L, Benson C, Cheshire-Kinney K, Lindenmayer J-P, Mao L, Rodriguez SC, et al. Real-world outcomes of paliperidone palmitate compared to daily

oral antipsychotic therapy in schizophrenia: A randomized, open-label, review board-blinded 15-month study. J Clin Psychiatry 2015; 76: 554-61.

10. Battaglia J, Wolff TK, Wagner-Johnson DS, Rush AJ, Carmody TJ, Basco MR. Structured diagnostic assessment and depot fluphenazine treatment of

multiple suicide attempters in the emergency department. Int Clin Psychopharmacol 1999;14:361-72.

11. Buckley PF, Schooler NR, Goff DC, Hsiao J, Kopelowicz A, Lauriello J, et al. Comparison of SGA Oral medications and a long-acting injectable SGA:

the PROACTIVE study. Schizophr Bull 2015; 41: 449-59.

12. Crespo-Facorro B, Ortiz-Garcia De La Foz V, Mata I, Ayesa-Arriola R, Suarez-Pinilla P, Valdizan EM, et al. Treatment of first-episode non-affective

psychosis: a randomized comparison of aripirazole, quetiapine and ziprasidone over 1 year. Psychopharmacology 2014; 231: 357-66.

13. Meltzer HY, Alphs L, Green AI, Altamura AC, Anand R, Bertoldi A, et al. Clozapine treatment for suicidality in schizophrenia: international suicide

prevention trial (InterSePT). Arch Gen Psychiatry 2003;60:82-91.

14. Strom BL, Eng SM, Faich G, Reynolds RF,D’Agostino R, Ruskin J, et al. Comparative mortality associated with ziprasidone and olanzapine in real-

world use among 18,154 patients with schizophrenia: the ziprasidone observational study of cardiac outcomes (ZODIAC). Am J Psychiatry 2011; 168:

193-201.

15. Thomas SHL, Drici MD, Hall GC, Crocq MA, Everitt B, Lader MH, et al. Safety of sertindole versus risperidone in schizophrenia: principal results of the

sertindole cohort prospective study (SCoP). Acta Psychiatr Scand 2010; 122: 345-355.
16. Oquendo M, Galfalvy HC, Currier D, Grunebaum M, Sher L, Sullivan GM, et al.Treatment of suicide attempters with bipolar disorder: a randomized

clinical trial comparing lithium and valproate in the prevention of suicidal behavior. Am J Psychiatry 2011; 168: 1050-1056.

17. Thies-Fletchner K, Mueller-Oerlinghausen B, Seibert W, Walther A, Greil W. Effect of prophylactic treatment on suicide risk in patients with major

affective disorders: data from a randomized prospective trial. Pharmcopsychiat 1996; 29: 103-107.
PRISMA 2009 Checklist Riblet et al. Suicide Prevention Meta-Analysis

Reported
Section/topic # Checklist item
on page #
TITLE
Title 1 Identify the report as a systematic review, meta-analysis, or both. 1
ABSTRACT
Structured summary 2 Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, 2
participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and
implications of key findings; systematic review registration number.

INTRODUCTION
Rationale 3 Describe the rationale for the review in the context of what is already known. 3
Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, 3
outcomes, and study design (PICOS).
METHODS
Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide 3
registration information including registration number.
Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, 3-4
language, publication status) used as criteria for eligibility, giving rationale.
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify 4
additional studies) in the search and date last searched.
Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be App 1
repeated.
Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, 5
included in the meta-analysis).
Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes 5
for obtaining and confirming data from investigators.
Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and 5
simplifications made.
Risk of bias in individual 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was 5
studies done at the study or outcome level), and how this information is to be used in any data synthesis.
Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). 5-6
Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency 5-6
(e.g., I2) for each meta-analysis.
Page 1 of 2
PRISMA 2009 Checklist Riblet et al. Suicide Prevention Meta-Analysis

Reported
Section/topic # Checklist item
on page #
Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective 7
reporting within studies).
Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating 6-7
which were pre-specified.
RESULTS
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at 7
each stage, ideally with a flow diagram.
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and 7, Table1
provide the citations. eTable2
Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). eTable2
Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each 8-9,
intervention group (b) effect estimates and confidence intervals, ideally with a forest plot. eTable2
Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. 8-9, Fig 2
and Fig3
Risk of bias across studies 22 Present results of any assessment of risk of bias across studies (see Item 15). 8,
eTable2
Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). 8-10
DISCUSSION
Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to 11-13
key groups (e.g., healthcare providers, users, and policy makers).
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of 13
identified research, reporting bias).
Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. 14
FUNDING
Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the 1
systematic review.

From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097.
doi:10.1371/journal.pmed1000097
For more information, visit: www.prisma-statement.org.
Page 2 of 2
Strategies to prevent death by suicide: meta-analysis of
randomised controlled trials
Natalie B. V. Riblet, Brian Shiner, Yinong Young-Xu and Bradley V. Watts
BJP published online April 20, 2017 Access the most recent version at DOI: 10.1192/bjp.bp.116.187799

Supplementary Supplementary material can be found at:


Material http://bjp.rcpsych.org/content/suppl/2017/04/10/bjp.bp.116.187799.DC1

References This article cites 0 articles, 0 of which you can access for free at:
http://bjp.rcpsych.org/content/early/2017/04/07/bjp.bp.116.187799#BIBL
Reprints/ To obtain reprints or permission to reproduce material from this paper, please
permissions write to permissions@rcpsych.ac.uk

P<P Published online 2017-04-20T00:05:10-07:00 in advance of the print journal.

You can respond /letters/submit/bjprcpsych;bjp.bp.116.187799v1


to this article at
Downloaded http://bjp.rcpsych.org/ on April 20, 2017
from Published by The Royal College of Psychiatrists

Advance online articles have been peer reviewed and accepted for publication but have not yet
appeared in the paper journal (edited, typeset versions may be posted when available prior to
final publication). Advance online articles are citable and establish publication priority; they are
indexed by PubMed from initial publication. Citations to Advance online articles must include the
digital object identifier (DOIs) and date of initial publication.

To subscribe to The British Journal of Psychiatry go to:


http://bjp.rcpsych.org/site/subscriptions/