Int. J. Environ. Res. Public Health 2014, 11, 8193-8212; doi:10.

3390/ijerph110808193
OPEN ACCESS

International Journal of
Environmental Research and
Public Health
ISSN 1660-4601
www.mdpi.com/journal/ijerph

Review

E-Health Interventions for Suicide Prevention

Helen Christensen 1,†,*, Philip J. Batterham 2,† and Bridianne O’Dea 3,†
1
School of Medicine, Black Dog Institute, University of New South Wales, Sydney, NSW, 2031,
Australia
2
Centre for Mental Health Research, The Australian National University, Canberra, ACT, 0200,
Australia; E-Mail: philip.batterham@anu.edu.au
3
School of Medicine, Black Dog Institute, University of New South Wales, Sydney, NSW, 2031,
Australia; E-Mail: b.odea@blackdog.org.au
†
These authors contributed equally to this work.

* Author to whom correspondence should be addressed; E-Mail: h.christensen@blackdog.org.au;

Tel.: +61-02-9382-9288; Fax: +61-02-9382-8207.

Received: 16 June 2014; in revised form: 30 July 2014 / Accepted: 30 July 2014 /
Published: 12 August 2014

Abstract: Many people at risk of suicide do not seek help before an attempt, and do not
remain connected to health services following an attempt. E-health interventions are now
being considered as a means to identify at-risk individuals, offer self-help through web
interventions or to deliver proactive interventions in response to individuals’ posts on
social media. In this article, we examine research studies which focus on these three
aspects of suicide and the internet: the use of online screening for suicide, the effectiveness
of e-health interventions aimed to manage suicidal thoughts, and newer studies which aim
to proactively intervene when individuals at risk of suicide are identified by their social
media postings. We conclude that online screening may have a role, although there is a
need for additional robust controlled research to establish whether suicide screening
can effectively reduce suicide-related outcomes, and in what settings online screening
might be most effective. The effectiveness of Internet interventions may be increased if
these interventions are designed to specifically target suicidal thoughts, rather than
associated conditions such as depression. The evidence for the use of intervention practices
using social media is possible, although validity, feasibility and implementation remains
highly uncertain.
Int. J. Environ. Res. Public Health 2014, 11 8194

Keywords: internet; suicide prevention; e-health; social media; screening

1. Introduction

E-health interventions for suicide prevention can be classed into three categories. First, the Internet
can be used to help individuals self-screen: to identify whether they might be at risk for suicide or a
mental health problem. Through screening and feedback, it may be possible to increase service
use, by directing at-risk individuals who would not otherwise seek help to access appropriate
evidence-based online programs or to access traditional mental health services [1]. Second, web
applications, both guided and unguided, have been developed to provide psychological interventions to
assist in reducing suicidal behaviour and lowering suicidal ideation. Guided interventions involve a
therapist or a researcher assisting the user through the program either through email or over the
telephone, whereas unguided are self-help, automated programs which can be initiated and used
directly by the public. The third type of intervention is one where a person is considered to be at risk of
suicide because of the nature of their social media use. Here, tweets, status updates, comments
or posts indicative of suicide ideation are used to classify those at risk. Such content can be identified
in real-time by other users or by computerised language processing and progress in this area
has accelerated. The aims of the present paper are to review the evidence around these three styles of
intervention. Three separate literature reviews were conducted. The first examined the evidence for
whether online screening for suicide might be effective in reducing suicidal ideation and behaviours.
The second reviewed web interventions, updating two recent reviews of web based suicide
prevention [2,3] and distinguishing two approaches: web interventions that target suicidal behaviour
and using depression therapies; and interventions that target suicidal behaviour using suicide-specific
therapies. The final review examined the use of social media platforms for identifying those who may
be at risk of suicide.

2. Methods

Comprehensive literature searches for all three reviews were conducted in March 2014 using the
Medline, PsychInfo, and Cochrane Library databases. Conference abstracts, non-peer reviewed papers,
non-English language papers, and PhD theses were excluded from all three reviews. All articles were
screened for eligibility by two independent researchers. The specifics of each search are outlined below.

2.1. Identifying Screening Programs

Search terms indicative of internet technology (computer or computer-based or cyber or

cyberspace or electronic or “electronic mail” or email or e-mail or internet or internet-based or net or
online or virtual or web or web-based or web-based or “world wide web” or www or “social media”
or “social network” or blog or forum), screening (screen* or assess*) and suicide (suicid*) were
employed. The search resulted in 855 articles, of which 98 were excluded due to being in a
non-English language or animal-based and a further 162 were excluded as duplicates. The remaining
595 abstracts were screened for relevance, with 78 papers identified as potentially meeting the criterion
Int. J. Environ. Res. Public Health 2014, 11 8195

of reporting on the screening of individuals for suicidal ideation or behaviours using the internet.
Full-text copies of these papers were then obtained for further scrutiny and reference checks conducted
to identify papers that may have been missed in the search. After examining the full text of these
articles, 47 did not meet the criteria of the screening review. Hierarchical reasons for exclusion were
not being a peer-reviewed paper (11 papers), not using online data collection (20 papers), not including
a measure of the respondent’s suicidality (eight papers), not reporting on suicidality outcomes (seven
papers), or being a review paper (one paper). One additional relevant paper was identified in reference
checking, resulting in 32 papers from 30 studies that reported on online assessment of suicidal ideation
or behaviours.

2.2. Identifying Web Programs

In addition to the databases stated above, the Centre for Research Excellence of Suicide Prevention
(CRESP) Suicide Prevention Database (http://cresp.edu.au/databases/sprct) was also used for this
review. Search terms indicative of suicide, self-harm and mobile or online applications were employed:
“Self harm OR self-harm OR deliberate self-harm OR deliberate self poisoning OR self cutting OR
self-inflicted wounds OR deliberate self cutting OR suicid* OR suicide gesture OR suicidal behavio*
OR suicidal ideation OR suicide attempt OR self-mutilation OR auto-mutilation OR auto mutilation
OR self-injury OR self-injurious behavio* OR self destructive behavio* OR self-poisoning and
overdose OR drug overdose” in Title, Abstract, Keywords and “trials OR randomised controlled trials
OR randomized controlled trials OR meta-analysis” in Title, Abstract, Keywords and “internet OR
mobile OR web OR email OR e-mail OR online” in Title, Abstract, Keywords.
The search yielded a total of 198 abstracts which was reduced to 109 after the removal of duplicates.
Of these, nine papers met inclusion criteria. Inclusion and exclusion criteria were applied in order to
identify any internet or mobile-based interventions that included a measure of suicidal behaviour. The
trials did not need to explicitly target those experiencing suicidal behaviours, but they were required to
measure participants’ level of suicidality prior to program commencement and following program
completion. Studies examining non-suicidal self-injury were not included. Due to the low numbers of
trials, studies without control or comparison groups were included in addition to trials including
control groups. The control group could consist of a wait-list, treatment-as-usual, or another treatment.
There was no restriction on participant age.
Studies were excluded if they did not include an intervention, if suicidality was not measured as a
primary or secondary outcome, and if the intervention was not internet or mobile based. One paper was
excluded as the research design and sample was identical to another paper written by the same authors.
Further, the study led by Merry [4] employed the Kazdin Hopelessness Scale in place of a suicidal
behaviour measure. Considering this scale is widely used as a proxy for suicidal ideation, the study
was included.
Two of the studies in the review are not discussed further. These are: Marasinghe, et al. [5]; and
Wagner, et al. [6]. The first was eliminated because it did not involve a web component. The second
because it was difficult to discern the type of therapy, and the extent to which the intervention was
delivered online (it was not clear whether both groups received a paper and pencil manual).
Int. J. Environ. Res. Public Health 2014, 11 8196

2.3. Identifying Social Media Interventions

Search terms indicative of suicide, self-harm and social media were employed: suicid* OR suicide
gesture OR suicidal behavio* OR suicidal idea* OR suicide attempt OR self-mutilation OR self harm
OR self-harm OR deliberate self-harm OR deliberate self poisoning OR self cutting OR self-inflicted
wound OR deliberate self cutting AND social media OR internet OR web OR online OR blog* OR
online social network* OR website OR twitter OR Myspace OR Facebook OR social networking site*
OR bebo OR tweet* OR status update OR post*AND screen* OR assess OR prevent* OR track*.
Using Kaplan and Haenlein’s [7] classification of social media, the current review focused explicitly
on blogs/microblogs (e.g., Twitter), and social networking sites (e.g., Facebook, Myspace, Bebo).
Content communities (e.g., YouTube), virtual worlds (e.g., Second life), MMORGs (e.g., World
of Warcraft) and collaborative projects (e.g., Pinterest) were excluded. Articles that related to online
discussion forums, online support groups or internet message boards were also excluded. The search
yielded 1934 following number of articles. Of these, only 13 papers met inclusion criteria.
Papers published prior to 2000 (the onset of social media) were excluded alongside those that focused
on non-suicidal self-injury, murder-suicides or mental illnesses such as depression, which may
increase the risk of suicide. A search for similar articles was conducted on all relevant papers; however,
this returned zero results. Citation lists for each included article was also screened for other related
papers which yielded an additional three papers. The article titled “Mining Twitter for Suicide
Prevention” [8] is not discussed further as the publication source could not be identified. A total of 15
papers were included.

3. Results

3.1. Online Screening for Suicide Ideation

Of the 32 papers that met inclusion criteria, only six (19%) had direct relevance to online screening
programs. The details of these papers are provided in Table 1. Of the remaining papers not detailed in
the table, 23 (72%) focused on assessing risk factors or prevalence of suicidal ideation or behaviours
using cross-sectional web surveys, while one paper reported on the cost-effectiveness of an online
suicide prevention trial, one paper reported on the psychometric properties of an online behavioural
health measure and another reported on a cross-sectional survey assessing reasons why people seek
help for suicide online. All of the six papers on screening were based on studies from the United States.
Five of the papers focused on young people, including four on university students and/or staff. Sample
sizes ranged from 374–13155 (M: 2916, Median: 1010) and 45% of all participants were female. Most
studies used the PHQ-9 [9] to assess suicidal ideation, usually in association with a lifetime suicide
attempt item. The remainder used binary prevalence measures (i.e., presence/absence of suicidal
ideation, suicide plan, or suicide attempt).
Three of the papers [10–12] reported on the use of an online screening program developed by the
American Foundation for Suicide Prevention in the university setting. Overall, these studies reported
that online screening was effective for identifying students with history of suicidal ideation or
behaviours. The direct referral of at-risk students to in-person services, such as campus counsellors,
was also found to be effective, although only a minority (ranging approximately 10%–20%) of
Int. J. Environ. Res. Public Health 2014, 11 8197

at-risk students agreed to service referral and fewer received services. Another study [13] examined
web-based screening among adolescents presenting at the emergency department of a children’s
hospital, finding that 65% of adolescents agreed to screening, resulting in a 70% increase in
identification of psychiatric problems and a 47% increase in assessments by a social worker or
psychiatrist. Lawrence, et al. [14] tested the feasibility of using the PHQ-9 within a web-based
screener administered in outpatient clinics to screen for depression and suicidality among people with
HIV. In this program, 14% reported some level of suicidal ideation and 3% were deemed high-risk and
referred to services. Critically though, none of the identified studies included a control (non-screened)
group for comparison, so they were unable to assess whether screening and referral alone was
sufficient to increase help seeking or to improve mental health outcomes.

Table 1. Studies of online screening for suicidal thoughts or behaviours identified in the review.

Paper Topic N % Female Location Population Measure

Evaluation of Behavioural
Adolescents;
Fein, et al. [13] emergency department 857 56 USA Health
Emergency dept
psychiatric screener Screener
Description of
PHQ-9 +
Garlow, et al. [10] a university 729 72 USA University students
past attempts
screening program
Description of
PHQ-9 +
Haas, et al. [11] a university 1162 70 USA University students
past attempts
screening program
Description of
People with HIV;
Lawrence, et al. [14] a suicide 1216 21 USA PHQ-9
primary care
screening program
Description of
University staff & PHQ-9 +
Moutier, et al. [12] suicide/depression 374 -- USA
students past attempts
screening program
Responses to being National
Whitlock, et al. [15] asked about suicide, 13,155 43 USA University students Comorbidity
self-harm Survey items

One additional study examined the acceptability and risk of online screening for suicidal ideation
and behaviours, also among university students [15]. This study concluded that such screening was
generally acceptable, with few individuals (<3%) reporting negative experiences as a result of the
screening. Individuals with previous suicidality had greater discomfort with the survey but such
individuals also found it more thought-provoking than those without previous suicidality [15].

3.2. Web Applications for Suicide Prevention

Outlined in Table 2, six studies examined suicide outcomes arising from the use of a web intervention
which targeted depression. Of these, two studies used adolescent samples. A pre-post study (n = 83) of
general practice adolescent patients with suicidal ideation (but not frequent ideation or actual intent)
found reductions in self-harm thoughts and depressive symptoms at 6 weeks and 12 weeks using the
Int. J. Environ. Res. Public Health 2014, 11 8198

“PROJECT CATCH-IT” web program (CBT, IPT and parent workbook) [16]. A non-inferiority RCT
(n = 94) of psychiatric outpatients with depression compared a computerised self-help program
“SPARX” (CBT) with TAU (face to face therapy) and found that SPARX was non-inferior on levels
of hopelessness (a proxy measure of suicide ideation) [4]. Four studies targeted adults. Two Australian
studies examined changes in suicide ideation using a pre-post design. In the first, 299 general practice
patients with suicidal ideation undertook an internet intervention for depression (CBT, homework and
clinician contact) and the researchers found a reduction in suicidal ideation [17]. A pre-post study
(n = 359) of depressed or suicidal general practice patients, evaluated the “Sadness Program” (internet
based CBT, homework, supplementary resources) and found reduction in suicidal ideation [18].
A third study (n = 105) of depressed patients with suicidal thoughts employed a RCT methodology.
The researchers compared “Deprexis” (online CBT for depression) with wait list controls and found
decreased scores on depression, dysfunctional attitudes and improved quality of life, but no difference
on either suicidal thoughts and behaviour [19]. A second RCT (four arms) (n = 155) of depressed callers
to Lifeline compared web-based CBT, web-based CBT plus telephone call, telephone call back
line and TAU found no differences in the rate at which suicidal thoughts dissipated between the
four conditions [20].
Taken together these findings from both the adult and the adolescent studies show that suicide ideation
drops over time in response to internet interventions (the Deprexis results is the only anomaly). The
two RCT trials [19,20] also demonstrate that depression websites have specific effects on depression
symptoms above those of the control conditions. However, the studies were not able to establish that
suicide ideation falls as a function of the depression CBT provided by the interventions. In other words,
suicide ideation seems to drop in both depression and control conditions. These findings do not rule
out the possibility that websites which target specific characteristics of suicidal thoughts and behaviour
such as rumination disruption or mindfulness may be more effective than either depression
interventions or the “passage of time”. There is only one RCT that has been published which focuses
on a specific intervention for suicidal thoughts. This study, compared an online self-help program
(6 modules of CBT with DBT, PST, MBCT as well as weekly assignments and automated
motivational emails) with a waitlist control group and found reductions in suicidal thoughts and levels
of hopelessness in favour of the self-help program [21] and improved cost effectiveness [22].

3.3. Social Media for Suicide Prevention

Table 3 outlines the relevant papers and associated findings. Of the 15 articles identified, six were
classified as case studies examining either one individual’s social media use (n = 5), or, the social
media use of a single organisation (n = 1). In the single organisation case study, Boyce [23] described
the social media use of a suicide prevention agency (“Samaritans”) in the United States (U.S.). This
case study provided brief recommendations for the future but did not include any data in regards to the
reach, impact or effectiveness of such social media use. The remaining five individual case studies
focused on young males aged between 13–29 years living in either China or the U.S., or location
undisclosed. In 2011, Ruder and colleagues [24] presented commentary on a case involving a suicide
note posted on Facebook by a 28 year old male who died by suicide. No data or social network
analysis was conducted. Lehavot, Ben-Zeev and Neville [25] outlined the ethical considerations
Int. J. Environ. Res. Public Health 2014, 11 8199

involved with a therapist using Facebook as a monitoring tool for postings of suicidal imagery in a
male patient. No data or social network analysis was conducted. Ahuja, et al. [26] briefly discussed the
suicidal postings of a male in his late twenties with a history of mental illness and the potential for
offline social network intervention. No data or social network analysis was conducted. Using sentiment
analysis software, Fu, Cheng, Wong and Yip [27] examined the reactions and patterns of information
diffusion to a self-harm post made by a male using the Chinese social networking site “Sina Weibo”.
Li, Chau, and Wong [28] examined the relationship between blog posting intensity and language use to
explore the suicidal processes of a 13 year old male. The specific blog site was not identified.
Four of the overall papers, including one brief correspondence, were non-systematic literature
reviews discussing the use of social media for suicide prevention. Two of these reviews, published by
Luxton and colleagues [29,30] described the social media platforms currently being used, or those with
potential, for suicide prevention. The most recent review by Luxton, June and Fairall [29] found a total
of 580 Twitter groups and 385 blog profiles on blogger.com designated to suicide prevention, one
social networking site designed for social media prevention in the US (lifeline-gallery.org), and
discussed the application of internal functions that could act as alert systems for potential suicide
behavior. Luxton, et al. [29] also discussed the presence of suicide notes on social media, but did not
refer to any data or particular studies. The third review [31] outlined the potential benefits and
complications of social networking sites as a therapeutic intervention for self-injury but did not include
any data or references to a specific platform. The brief correspondence [32] presented policy initiatives
and internal functions of social media that could be adopted for suicide prevention.
Only five papers specifically examined the utilisation of social media for the tracking of suicide risk.
These studies focused on either Twitter (n = 1), Myspace (n = 2), Facebook (n = 1) or other blog site (n
= 1, Naver blog, Korea). The studies focusing on Myspace included self-reported adolescent samples
with ages ranging between 13–24 years. Age range of participants in the other papers could not be
determined. To our knowledge, the first study published in this area was an exploration of public
Myspace blogs of New Zealand youth using automated sentiment analysis [33]. Cash and colleagues [34]
further explored the content of suicidal statements made on the public Myspace profiles of adolescents
aged between 13–24 years old. Using Twitter, Jashinsky, et al. [35] identified a significant association
between suicide risk factors within Twitter and geographic specific suicide rates in the US. Using
automated sentiment analysis software, Won, Myung, Song, Lee, et al. [36] examined whether two
blog sentiments (suicide-related and dysphoria-related) along with traditional social, economic and
meteorological variables significantly predicted suicide rates in Korea over a three year period (2008–
2010). The final study [37] was a thematic content analysis of the portrayal of suicide and self-harm
within a Facebook group in April 2009. Computerised sentiment analysis was not used.
Int. J. Environ. Res. Public Health 2014, 11 8200

Table 2. Internet and mobile programs designed to assist those experiencing suicidal ideation or deliberate self-harm.
Target Group Suicide Behavior: Suicide
Country & Intervention
Paper (n), Age, %, Research Design Setting Baseline Outcome Results
Period of Trial Component/s
Male Suicide Levels Measure
Depressed Significant reduction in
RCT; four arms: (1)
participants who 6 weeks of any suicidal ideation at post for
Web-based CBT
have called 3 intervention conditions. internet only (p = 0.05),
intervention;
Lifeline (score of Web based CBT telephone call back only,
(2) Web-based Callers to
more than 22 condition (1) consisted of p = 0.003, and TAU
CBT intervention + Lifeline
on the K10), psycho-education Suicidal ideation GHQ-28 (p = 0.005); at 6-month
telephone call back; (telephone
Australia; July n = 155 (Internet provided by BluePages, (excluded if acutely (4-items pertain follow-up for internet only
Christensen, (3) proactive call counselling
2007 to January only (n = 38); and suicidal); Mean GHQ to suicidal (p = 0.016, and telephone call
et al. [20] back telephone line; service for
2009 Internet + call back MoodGYM-interactive suicidal ideation ideation back only (p = 0.029);
(4) TAU. people
(n = 45); web application, based on score = 1.73 component). at 12-month
Participants experiencing
Telephone call CBT-5 modules. follow-up for internet only
assessed at pre and crisis.)
back only (n = 37); Condition 2 also included (p < 0.001), internet + call
post intervention,
TAU (n = 35)); weekly 10 min call from a back (p < 0.001), and
and 6 and 12 month
mean age = 41.49; Lifeline counsellor telephone call back only
follow-up.
18.1% male. (p = 0.011).
Patients
undergoing Clinical trials, Phase 1: Recently attempted
treatment 220–380 min suicide, displaying BSSI scores—Intervention
Single-blinded
post-suicide face-to-face component; suicidal intent; baseline to 6-months to
RCT—clinical
Colombo, Sri attempt, mean age Phase 2: Brief weekly Outpatient, Mean BSSI score 12-months (26.1–3.65–3.6);
Marasinghe, trial vs. wait list
Lanka; no dates intervention phone calls/SMS to following for control males BSSI; Primary Control baseline to 6-months
et al. [5] control with post
given (n = 34) = 32 participants for 26 weeks; primary care (21.3), females to 12 months
and 6 month
years; control Control: Usual Care (22.2), intervention (21.75–7.55–3.75);
follow-up
(n = 34) = 30 years, followed by Phase 2 males (26.7), no p-value reported.
50% male in both component. females (25.5)
conditions
Int. J. Environ. Res. Public Health 2014, 11 8201

Table 2. Cont.
Target Group Suicide Behavior: Suicide
Country & Intervention
Paper (n), Age, %, Research Design Setting Baseline Outcome Results
Period of Trial Component/s
Male Suicide Levels Measure
12–19 years Indirect—depression
olds with mild to Randomised severity (excluded Per protocol improvements in
moderate controlled those deemed high hopelessness were
New Zealand;
symptoms of non-inferiority 7 CBT-based risk of suicide or self significantly greater for
May 2009 to Outpatient, had Indirect—Kazdin
Merry, depression; mean trial—Online interactive modules harm) ; ITT participants in the online
July 2010 + sought help for Hopelessness
et al. [4] age online intervention vs. to be completed participants mean intervention.
follow-up in depression scale for children
Intervention TAU (face-to-face in 4–7 weeks score on ITT improvements were
December 2010
(n = 94) = 15.55, therapy). Pre-post hopelessness for non-significantly larger
TAU (n = 93) = + follow-up control (6.15) and than TAU.
15.58 intervention (6.17)
Participants with
Suicidal thoughts
elevated depression
RCT; Online Online self-help and behaviour Significant symptom decline
symptoms; mean
Hamburg. self-help program program for (excluded patients SBQ-R, assesses on depression, dysfunctional
age intervention
Moritz, Germany (online vs. Wait list depression with strong suicidal suicidal thoughts attitudes, improvement in
(n = 105) = 38.0 Online setting
et al. [19] recruitment); no control; pre-post (Deprexis); ideas); mean SBQ-R and behaviour; quality of life and self-esteem.
years, 22.9%
dates given treatment survey 10 modules, score 12.28 Secondary No significant improvement
male, control
(after 8 weeks) CBT based (wait-list controls); on SBQ-R scores.
(n = 105) = 39.13,
11.37 (intervention)
20% male
Int. J. Environ. Res. Public Health 2014, 11 8202

Table 2. Cont.
Target Group Suicide Behavior: Suicide
Country & Intervention
Paper (n), Age, %, Research Design Setting Baseline Outcome Results
Period of Trial Component/s
Male Suicide Levels Measure
6 modules
Mild to moderate
(30 min per day Mild to moderate
suicidal thoughts
over 6 weeks) of suicidal thoughts;
(scores between 1 Significant reduction in
Netherlands CBT with DBT, General public BSSI mean score of
and 26 on the suicidal thoughts for
(Online RCT intervention PST, MBCT + recruited via 14.5 (control) and
Van Spijker, BSSI); mean age intervention group compared
recruitment); group vs. waitlist weekly online and 15.2 (intervention), BSSI; Primary;
et al. [21] intervention to control group (p = 0.036).
October 2009 to control assignments and newspaper 16.8% had
(n = 116) = 40.46 Non-significant reductions in
November 2010 optional exercises advertisements attempted suicide
years; control depressive symptoms
with up to once and 24.1 had
(n = 120) = 41.39,
6 automated multiple attempts
33.9% male.
motivational emails
Self-harm risk
(suicidal ideation)
14 modules based (excluded patients
Significant reduction in
on CBT, IPT, who expressed
self-harm thoughts at 6-weeks
community frequent suicidal
United States of Primary Care (p = 0.04) and 12-weeks
resiliency ideation or actual
Van America; adolescent Pre-post (at 6 Outpatient, PHQ-A— (p = 0.02) and depressive
concept model intent);13% thought
Voorhees, February 2007 patients, (n = 83), and 12 weeks), following self-harm risk; symptoms at 6-weeks
(CATCH-IT); about suicide in past
et al. [16] to November mean age = 17.39 no control primary care Secondary (p < 0.001) and 12-weeks
Additional parent 2 weeks, 7% with
2007 years, 43% male (p < 0.001) and depressive
workbook to serious suicidal
disorder for major depression
support adolescents thoughts in last
at 12-weeks (p = 0.047).
progress month, 16% with
any suicidal
thoughts
Int. J. Environ. Res. Public Health 2014, 11 8203

Table 2. Cont.
Target Group Suicide Behavior:
Country & Research Intervention Suicide Outcome
Paper (n), Age, %, Setting Baseline Results
Period of Trial Design Component/s Measure
Male Suicide Levels
People
experiencing Randomised
Internet based CBT No between group differences
depression Controlled
intervention for any pre-post treatment
Zurich, (score of at least Non-inferiority General public Suicidal ideation
including measurements. Significant
Switzerland; 12 on the BDI-II); Trial; pre-post; recruited via (excluded if high
Wagner, structured writing pre-post reduction in suicidal
November 2008 mean age online Internet online and risk of suicide); BSI; Secondary
et al. [6] assignments with ideation (p < 0.05) for
to February (n = 32) = 37.25 intervention vs. newspaper BSI = 3.24 (online);
individualized face-to-face treatment group,
2010. 22% male, face-to-face advertisements = 4.87 (face-to-face).
therapist feedback; but not for iCBT group
face-to-face CBT
8 weeks (p = 0.24).
(n = 30) = 38.73; intervention
50% male.
6 CBT-based Suicidal ideation
Primary Care lessons + (excluded “actively PHQ-9 using Q9
Sydney, Significant reduction in
patients (n = 299), Clinical audit; homework with Outpatient, suicidal” patients); as measure of
Watts, Australia; April suicidal ideation scores
mean age = pre-post, clinician making following 54% mild, 30% frequency of
et al. [17] 2009 to May (p < 0.001) and depression
43 years, no control contact at least primary care moderate, 15% suicidal
2011 scores (p < 0.0001).
44% male twice during the severe, 9% ex. ideation;Primary
course severe
Int. J. Environ. Res. Public Health 2014, 11 8204

Table 2. Cont.
Target Group Suicide Behavior: Suicide
Country & Intervention
Paper (n), Age, %, Research Design Setting Baseline Outcome Results
Period of Trial Component/s
Male Suicide Levels Measure
Primary care
Significant reductions in
patients enrolled in iCBT- The Sadness Suicidal ideation;
suicidal ideation for Ss
Australia; October the Sadness Program: 6 online PHQ9 scores =
experiencing suicidal
2010 to November Program, who were lessons within (17% severe, 8% very
Outpatient, ideation (p = 0.001) and for Ss
2011; either severely Quality assurance 10 weeks; regular severe). 53% (n = 189)
Williams, following PHQ-9 Suicide experiencing severe depression
54% of depressed and/or study; pre-post, homework endorsed suicidal
et al. [18] primary item; Primary and suicidal ideation
participants from expressing suicidal no control assignments, thoughts during
care (p < 0 001). 54% of patients
rural or remote ideation, access to the 2-week time period
who completed all 6 lessons
community (n = 359), supplementary prior to commencing
evidenced clinically significant
mean age = 41.59; resources the program
change in depression.
41% male
CBT—Cognitive Behavioural Therapy; PHQ-9—Patient Health Questionnaire—9 item; RCT—Randomised Controlled Trial; SBQ-R—Suicide Behaviors Questionnaire-Revised;
PHQ-A—Patient Health Questionnaire-Adolescent; IPT—Interpersonal Psychotherapy; iCBT—internet-based Cognitive Behavioural Therapy; SMS—Short Message
Service; BSSI—Beck Scale for Suicidal Ideation; BDI-II—Beck Depression Inventory-Revised; DBT—Dialectical Behaviour Therapy; PST—Problem Solving Therapy;
MBCT—Mindfulness Based Cognitive Therapy; K10—Kessler’s Psychological Distress Scale-10 items; TAU—Treatment As Usual; GHQ-28—General Health
Questionnaire-28-item; ITT—Intention To Treat.
Int. J. Environ. Res. Public Health 2014, 11 8205

Table 3. Articles related to social media for suicide prevention.

Type Paper Design/Methods Sample, Location & Platform Findings
Samaritans U.S. Facebook page. Time: not relevant. Argued that social media behavior can
Boyce [23] Descriptive commentary
Data: nil. help determine the path that suicidal people take online.
Suicide notes posted via social media may allow for
A suicide note posted on Facebook by a 28 year old timely suicide intervention by alerting other network users
Ruder, et al. [24] Descriptive commentary
male who died by suicide. Time: not reported. Data: nil. immediately, although understanding the relationship between
online suicide notes and copycat suicides is important to consider.
Male, late 20’s, history of mental illness, location Several ethical issues, including beneficence and maleficence;
Lehavot,
Descriptive case study unknown, posted suicidal imagery on his Facebook privacy and confidentiality; multiple relationships; clinical
et al. [25]
profile. Time: not reported. Data: nil. judgement; and informed consent, were discussed.
Responses were classified as caring (37%), negative (23%),
A self-harm post made by a male on the social
Case Quantitative content shocked (20%) or unemotional reposts (20%). Significant clustering
Fu, et al. [27] networking site Sina Weibo. Time: March 2011.
studies analysis was identified in the repost network in which the speed of diffusion
Data: 5971 microblog responses were included.
was faster when compared to the random network.
The ratio of positive to negative emotion words was associated with
Male, 13 years old, located in China. Microblog site
greater posting trend. There was greater use of negative emotion
Computerised language unidentified. Time: not reported. Data: 193 blog
Li, et al. [28] over time. Progressive self-referencing appeared to be a predictive
processing entries made in the year preceding the participant’s
sign of suicide, although, the comparison did not show other clearly
suicide were analysed.
consistent patterns.
Male, late 20’s, history of mental illness, location
General discussion of how social media can assist in screening for
unknown, posted suicidal ideation his Facebook
Ahuja, et al. [26] Descriptive Case Study suicidality as well as preventative methods when individuals
profile. Time: not reported. Data: 3 posts taken from
display suicidal thoughts via social media.
Facebook page.
Int. J. Environ. Res. Public Health 2014, 11 8206

Table 3. Cont.
Type Paper Design/Methods Sample, Location & Platform Findings
Social media provides opportunities for effective outreach and
Non-systematic literature
Luxton, et al. [30] NA suicide prevention but cannot replace careful clinical case
review
management. Further evaluation necessary.
Messina & Non-systematic literature A discussion of the internet uses associated with self-injury.
NA
Iwasaki [31] review No reference to particular social media platforms.
Reviews Social media has the potential to be used for suicide prevention within
Non-systematic
Luxton, et al. [29] NA a public health framework although more research is needed on the
literature review
degree and extent of the influence of social media for such purposes.
Suggested that social networking sites could help prevent suicides
Cheng, et al. [32] Brief Correspondence NA by deleting pro-suicide groups re and automatically delivering
private messages to those at risk.
Overall, 3.7% and 5% of active bloggers were potentially suicidal:
35% were identified as positive hits. 638 users out of the 4273
received a score of 1 or higher indicating that at least one match
Computerised sentiment Participants: 15,000 Myspace users aged 15–24
was found with the dictionary phrases. Using the exact phrases,
Huang, et al. [33] analysis with manual living in New Zealand. Time: not reported. Data:
612 bloggers received a score of 1 or higher. Although the ability
inspection 4273 unique blogs were examined.
Sentiment to definitively identify bloggers with suicidal tendencies is limited,
Research the study demonstrates that computerised data mining can be used
to identify users at potential risk.
Themes identified: normalization, nihilism, glorification, ‘us vs.
Thematic content Participants: Facebook users, presumed to be
Zdanow & them’, acceptance, reason, mockery. Facebook groups were found
analysis of user teenagers. Time: 27 April 2009. Data: Group 1:
Wright [37] to encourage and promote positive perceptions of suicidal
statements 15,201 members, Group 2: 228 members.
behavior.
Int. J. Environ. Res. Public Health 2014, 11 8207

Table 3. Cont.
Type Paper Design/Methods Sample, Location & Platform Findings
Participants: Myspace users located in the
United States, public profile, not self-identified as
musicians, comedians or movie makers, had between
Researchers were able to categorise ‘at-risk of suicide’ bloggers
2–1000 friends. Time: 3–4 March 2008 and
with up to 35% success and demonstrated a 14% automated
Computerised sentiment downloaded again in December 2008. Sample was
identification rate. Many of these posts were related to a breakdown
Cash, et al. [34] analysis with manual reduced in 4 stages. Data: 1762 comments collected:
in personal relationships (42.2%) with some references to mental
inspection 1038 met criteria, reduced to 490 comments. 105
health problems (6.3%); however, for the most part, context of the
comments mentioned suicide but referred to the
statement could not be established.
suicide of another. Final coding revealed 64
comments related to a serious comment made by the
commenter about potential suicidality.
Participants: Twitter users located in the U.S. Time:
15 May 2012–13 August 2012. Data: 1,659,274
A total of 2.3% (n = 37,717) of users were identified as at risk
Computerised sentiment tweets from 1,208,809 users over a 3 month period.
Jashinsky, for suicide. A strong correlation was observed between
analysis with manual Exclusion criteria resulted in 733,011 tweets from
et al. [35] state Twitter-derived data for suicide and actual state age-adjusted
inspection 594,776 users: 37,717 identified as suicidal.
suicide data.
A specific state location could be identified for
37,717 tweets from 28,088 users.
Both sentiments were associated with suicide frequency. The
Computerised sentiment
Participants: Korean microbloggers using Naver suicide sentiment displayed high variability and were found to be
analysis comparing
Blog. Time: 1 January 2008–31 December 2010. reactive to celebrity suicide events, while the dysphoria sentiment
Won, et al. [36] national, economic and
Data: 153,107,350 posts on 5,093,832 blogs collected showed longer, secular trends with lower variability. In the final
meteorological data with
over three years. multivariate model, the two sentiments displaced consumer price
blog posts
index and unemployment rate as significant predictors of suicide.
Int. J. Environ. Res. Public Health 2014, 11 8208

4. Discussion

To date, there exists no controlled study testing whether online suicide screening can effectively
increase levels of help seeking or reduce suicidal ideation or behaviours. Online screening for suicidal
thoughts and behaviours appears to be acceptable among young people [15], supporting previous
findings from a randomised controlled trial that screening for suicide risk using paper surveys does not
increase the risk of suicidal thoughts or behaviours [38,39]. Although there is some evidence that
screening is a feasible way to increase identification of suicidality and increase service referrals,
research of online screening programs has not rigorously evaluated whether screening programs are
effective compared to not screening. Furthermore, most of the programs reported in the literature have
been conducted in selective populations, specifically, among university students in the United States.
There is some evidence from paper-and-pencil assessments that suicide screening programs can be
effective in specific settings with the availability of appropriate referral sources, with most of this
research also conducted among young people (e.g., [40,41]). However, there is a need for additional
robust controlled research to establish whether suicide screening can effectively reduce suicide-related
outcomes, and in what settings screening might be effective [42–44]. None of the identified studies had
a control condition, so they were unable to robustly assess whether online screening directly increased
help seeking or led to improved mental health outcomes. This evidence gap appears to be especially
conspicuous for online screening.
With respect to online suicide prevention programs, there is some evidence to suggest that suicide
interventions via the web may be effective, but only if they specifically target suicidal content, rather
than the associated symptoms of depression through CBT. There is no evidence CBT web-based
programs do harm, so excluding those with suicide ideation from participation does not seem
warranted for online programs in general. Further research targeting specific suicidal content on the
web is warranted.
Although limited, there is evidence to suggest that social media platforms can be used to identify
individuals or geographical areas at risk of suicide. The few studies conducted in this area have
demonstrated that it is possible to use computerised sentiment analysis and data mining to identify
users at risk of suicide; however, these studies have been conducted in publically available networks.
Little is known about private online social networks such as those within Facebook. Furthermore, the
prevalence and response rate of suicide notes posted on social media is yet to be clearly understood.
The repost networks within social media may be a potential preventative method that could be
activated quickly for effective intervention in emergency situations; although, current methodologies
cannot fully disentangle whether suicidal postings always receive a response and if so, the nature of
such responses. It is unknown if sharing thoughts and feelings this way is beneficial to an individual or
whether they are placed at further risk. While evidence suggests that social media may be a viable
tool for real-time monitoring of suicide risk on a large scale, future studies are needed to further
validate this method, in addition to determining the underlying mechanisms for providing support via
these channels.
Int. J. Environ. Res. Public Health 2014, 11 8209

5. Conclusions

This review highlights that there is currently limited evidence for the effectiveness of e-health
interventions for suicide prevention. Whilst feasible, the reliability and preventative capacity of online
screening for suicidality is yet to be clearly determined. Research in this area is incomplete.
Furthermore, online therapeutic interventions do not appear to be effective unless content is targeted to
suicidal thoughts and behavior; however, addressing risk factors, such as depression, through online
CBT does not appear to cause harm. Lastly, social media shows significant potential for identifying
those at risk of suicide in addition to mapping suicide contagion, although further validation research
in this area is also needed.

Acknowledgments

The authors thank Jacqueline Brewer and Angeline Tjhin for screening abstracts for the screening
review, and to Catherine King for the social media review. Thank you to both Catherine King and
Alistair Lum for preparing Table 2. Philip J. Batterham is supported by NHMRC fellowship 1035262.
Helen Christensen is supported by NHMRC Fellowship 1056964.

Author Contributions

Helen Christensen had the original idea for the review with all co-authors contributing to the
structure and design of the review. Philip J. Batterham was responsible for the review of online
screening for suicide, Helen Christensen for the review on web programs and Bridianne O’Dea on
social media for suicide prevention. All authors have read, edited and approved the final manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Ybarra, M.L.; Eaton, W.W. Internet-based mental health interventions. Mental Health Serv. Res.
2005, 7, 75–87.
2. Christensen, H.; Petrie, K. Online mental health programs—Promising tools for suicide
prevention. Med. Today 2014, 15, 66–68.
3. Lai, M.H.; Maniam, T.; Chan, L.F.; Ravindran, A.V. Caught in the web: A review of
web-based suicide prevention. J. Med. Internet Res. 2014, 16, e30.
4. Merry, S.N.; Stasiak, K.; Shepherd, M.; Frampton, C.; Fleming, T.; Lucassen, M.F. The effectiveness
of sparx, a computerised self help intervention for adolescents seeking help for depression:
Randomised controlled non-inferiority trial. Br. Med. J. 2012, 344, doi:10.1136/bmj.e2598.
5. Marasinghe, R.B.; Edirippulige, S.; Kavanagh, D.; Smith, A.; Jiffry, M.T.M. Effect of mobile
phone-based psychotherapy in suicide prevention: A randomized controlled trial in sri lanka.
J. Telemed. Telecare 2012, 18, 151–155.
Int. J. Environ. Res. Public Health 2014, 11 8210

6. Wagner, B.; Horn, A.B.; Maercker, A. Internet-based versus face-to-face cognitive-behavioral

intervention for depression: A randomized controlled non-inferiority trial. J. Affect. Disord. 2014,
152, 113–121.
7. Kaplan, A.M.; Haenlein, M. Users of the world, unite! The challenges and opportunities of social
media. Bus. Horiz. 2010, 53, 5968.
8. Abboute, A.; Boudgeriou, Y.; Entringer, G.; Aze, J.; Bringay, S.; Poncelet, P. Mining Twitter
for Suicide Prevention. In Natural Language Processing and Information Systems; Springer:
New York, NY, USA, 2014; pp. 250–253.
9. Spitzer, R.L.; Kroenke, K.; Williams, J.B. Validation and utility of a self-report version of
prime-md: The PHQ primary care study. Primary care evaluation of mental disorders. Patient
health questionnaire. J. Am. Med. Assoc. 1999, 282, 1737–1744.
10. Garlow, S.J.; Rosenberg, J.; Moore, J.D.; Haas, A.P.; Koestner, B.; Hendin, H.; Nemeroff, C.B.
Depression, desperation, and suicidal ideation in college students: Results from the American
foundation for suicide prevention college screening project at Emory University. Depress. Anxiety
2008, 25, 482–488.
11. Haas, A.; Koestner, B.; Rosenberg, J.; Moore, D.; Garlow, S.J.; Sedway, J.; Nicholas, L.; Hendin, H.;
Mann, J.J.; Nemeroff, C.B. An interactive web-based method of outreach to college students at
risk for suicide. J. Am. Coll. Health 2008, 57, 15–22.
12. Moutier, C.; Norcross, W.; Jong, P.; Norman, M.; Kirby, B.; McGuire, T.; Zisook, S. The suicide
prevention and depression awareness program at the University of California, San Diego School
of Medicine. Acad. Med. 2012, 87, 320–326.
13. Fein, J.A.; Pailler, M.E.; Barg, F.K.; Wintersteen, M.B.; Hayes, K.; Tien, A.Y.; Diamond, G.S.
Feasibility and effects of a web-based adolescent psychiatric assessment administered by
clinical staff in the pediatric emergency department. Arch. Pediatrics Adolesc. Med. 2010, 164,
1112–1117.
14. Lawrence, S.T.; Willig, J.H.; Crane, H.M.; Ye, J.; Aban, I.; Lober, W.; Nevin, C.R.; Batey, D.S.;
Mugavero, M.J.; McCullumsmith, C.; et al. Routine, self-administered, touch-screen, computer-based
suicidal ideation assessment linked to automated response team notification in an hiv primary care
setting. Clin. Infect. Dis. 2010, 50, 1165–1173.
15. Whitlock, J.; Pietrusza, C.; Purington, A. Young adult respondent experiences of disclosing
self-injury, suicide-related behavior, and psychological distress in a web-based survey. Arch.
Suicide Res. 2013, 17, 20–32.
16. Van Voorhees, B.W.; Fogel, J.; Reinecke, M.A.; Gladstone, T.; Stuart, S.; Gollan, J.; Bradford, N.;
Domanico, R.; Fagan, B.; Ross, R. Randomized clinical trial of an internet-based depression
prevention program for adolescents (project catch-it) in primary care: 12-week outcomes. J. Dev.
Behav. Pediatrics 2009, 30, 23–37.
17. Watts, S.; Newby, J.M.; Mewton, L.; Andrews, G. A clinical audit of changes in suicide ideas
with internet treatment for depression. BMJ Open 2012, 2, doi:10.1136/bmjopen-2012-001558.
18. Williams, A.D.; Andrews, G. The effectiveness of internet cognitive behavioural therapy (icbt) for
depression in primary care: A quality assurance study. PLoS One 2013, 8, e577447.
19. Moritz, S.; Schilling, L.; Hauschildt, M.; Schröder, J.; Treszl, A. A randomized controlled trial of
internet-based therapy in depression. Behav. Res. Ther. 2012, 50, 513–521.
Int. J. Environ. Res. Public Health 2014, 11 8211

20. Christensen, H.; Farrer, L.; Batterham, P.J.; Mackinnon, A.; Griffiths, K.M.; Donker, T.
The effect of a web-based depression intervention on suicide ideation: Secondary outcome from a
randomised controlled trial in a helpline. BMJ Open 2013, 3, doi:10.1136/bmjopen-2013-002886.
21. Van Spijker, B.A.; van Straten, A.; Kerkhof, A.J. Effectiveness of online self-help
for suicidal thoughts: Results of a randomised controlled trial. PLoS One 2014, 9,
doi:10.1371/journal.pone.0090118.
22. Van Spijker, B.A.; Majo, M.C.; Smit, F.; van Straten, A.; Kerkhof, A.J. Reducing suicidal
ideation: Cost-effectiveness analysis of a randomized controlled trial of unguided web-based
self-help. J. Med. Internet Res. 2012, 14, doi:10.2196/jmir.1966.
23. Boyce, N. Pilots of the future: Suicide prevention and the internet. Lancet 2010, 376, 1889–1890.
24. Ruder, T.D.; Hatch, G.M.; Ampanozi, G.; Thali, M.J.; Fischer, N. Suicide announcement on
facebook. Crisis 2011, 32, 280–282.
25. Lehavot, K.; Ben-Zeev, D.; Neville, R.E. Ethical considerations and social media: A case of
suicidal postings on facebook. J. Dual Diagn. 2012, 8, 341–346.
26. Ahuja, A.K.; Biesaga, K.; Sudak, D.M.; Draper, J.; Womble, A. Suicide on facebook. J. Psychiatr.
Pract. 2014, 20, 141–146.
27. Fu, K.-W.; Cheng, Q.; Wong, P.W.C.; Yip, P.S.F. Responses to a self-presented suicide attempt in
social media: A social network analysis. Crisis 2013, 34, 406–412.
28. Li, T.M.; Chau, M.; Yip, P.S.; Wong, P.W. Temporal and computerized psycholinguistic analysis
of the blog of a chinese adolescent suicide. Crisis 2014, 1–8.
29. Luxton, D.D.; June, J.D.; Fairall, J.M. Social media and suicide: A public health perspective.
Am. J. Public Health 2012, 102, S195–S200.
30. Luxton, D.D.; June, J.D.; Kinn, J.T. Technology-based suicide prevention: Current applications
and future directions. Telemed. J. E-Health 2011, 17, 50–54.
31. Messina, E.S.; Iwasaki, Y. Internet use and self-injurious behaviors among adolescents and young
adults: An interdisciplinary literature review and implications for health professionals.
Cyberpsychol. Behav. Soc. Netw. 2011, 14, 161–168.
32. Cheng, Q.; Chang, S.S.; Yip, P.S. Opportunities and challenges of online data collection for
suicide prevention. Lancet 2012, 379, e53–e54.
33. Huang, Y.; Goh, T.; Liew, C.L. Hunting suicide notes in Web 2.0—Preliminary findings.
In Proceedings of Nonth IEEE International Symposium on Multimedia, Los Almitos, CA, USA,
10–12 December 2007.
34. Cash, S.J.; Thelwall, M.; Peck, S.N.; Ferrell, J.Z.; Bridge, J.A. Adolescent suicide statements on
myspace. Cyberpsychol. Behav. Soc. Netw. 2013, 16, 166–174.
35. Jashinsky, J.; Burton, S.H.; Hanson, C.L.; West, J.; Giraud-Carrier, C.; Barnes, M.D.; Argyle, T.
Tracking suicide risk factors through twitter in the us. Crisis 2014, 35, 51–59.
36. Won, H.H.; Myung, W.; Song, G.Y.; Lee, W.H.; Kim, J.W.; Carroll, B.J.; Kim, D.K.
Predicting national suicide numbers with social media data. PLoS One 2013, 8,
doi:10.1371/journal.pone.0061809.
37. Zdanow, C.; Wright, B. The representation of self injury and suicide on emo social networking
groups. Afr. Sociol. Rev. 2012, 16, 81–101.
Int. J. Environ. Res. Public Health 2014, 11 8212

38. Crawford, M.J.; Thana, L.; Methuen, C.; Ghosh, P.; Stanley, S.V.; Ross, J.; Gordon, F.; Blair, G.;
Bajaj, P. Impact of screening for risk of suicide: Randomised controlled trial. Br. J. Psychiatry
2011, 198, 379–384.
39. Gould, M.S.; Marrocco, F.A.; Kleinman, M.; Thomas, J.G.; Mostkoff, K.; Cote, J.; Davies, M.
Evaluating iatrogenic risk of youth suicide screening programs: A randomized controlled trial.
J. Am. Med. Assoc. 2005, 293, 1635–1643.
40. Gould, M.S.; Marrocco, F.A.; Hoagwood, K.; Kleinman, M.; Amakawa, L.; Altschuler, E.
Service use by at-risk youths after school-based suicide screening. J. Am. Acad. Child.
Adolesc. Psychiatry 2009, 48, 1193–1201.
41. Wintersteen, M.B. Standardized screening for suicidal adolescents in primary care. Pediatrics
2010, 125, 938–944.
42. Gaynes, B.N.; West, S.L.; Ford, C.A.; Frame, P.; Klein, J.; Lohr, K.N.; U.S. Preventive Services
Task Force. Screening for suicide risk in adults: A summary of the evidence for the U.S.
Preventive services task force. Ann. Intern. Med. 2004, 140, 822–835.
43. O’Connor, E.; Gaynes, B.; Burda, B.U.; Williams, C.; Whitlock, E.P. Screening for Suicide Risk
in Primary Care: A Systematic Evidence Review for the United States Preventive Services Task
Force; Agency for Healthcare Research and Quality: Rockville, MD, USA, 2013.
44. Pena, J.B.; Caine, E.D. Screening as an approach for adolescent suicide prevention. Suicide Life
Threat. Behav. 2006, 36, 614–637.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).