Table 2 Study characteristics and results

Beautrais (2001) [15]; Beautrais et al. (2009) [16]

1

Grafton Bridge, Auckland, New Zealand.

• Restricting access to lethal means

Metal screens fixed above concrete parapets for purposes of suicide prevention, removed in 1996. Reinstallation of a barrier in 2003 with an improved curved glass design.

A-B-A (reversal) study assessing number and rates of suicides in three periods:

• Five suicides from the bridge during period in which original barriers were in place (1.0 per year). This rose to 19 in period when original barriers were removed (3.2 per year). No suicides occurred after the installation of the new barrier.

• 1991–1995 (5-year period in which original barriers were in place);

• 1997–2002 (5 year period in which no barriers were in place); and

• No change in overall number of suicides by jumping in Auckland.

• 2003–2006 (5 year period in which new barriers were in place).

Bennewith, Nowers and Gunnell (2007) [17]; Bennewith, Nowers and Gunnell (2011) [18]

2

Clifton Suspension Bridge, Bristol, United Kingdom.

• Restricting access to lethal means

Two metre high wire fencing installed on main span in 1998.

Before-and-after analysis comparing number of suicides in two periods:

• Suicides dropped from 8.2 per year in period prior to installation of fencing to 4.0 per year in period following it.

• Increasing the likelihood of intervention by a third party

Role of bridge staff expanded to include ensuring individuals’ safety and monitoring incidents. CCTV cameras installed.

• 1994–1998 (5-year pre-intervention period); and

• 90% of suicides from the bridge were by males but there was no evidence of an increase in male suicide by jumping from other sites in Bristol following the installation.

• 1999–2003 (5-year post-intervention period).

Before-and-after analysis comparing number of suicides in two periods:

• 1996–1998 (3-year pre-intervention period); and

• Number of incidents remained stable (39 per year in pre-installation period; 43 per year in post-installation period).

• 1999–2005 (7-year post-intervention period).

• Bridge staff more likely to be involved in incidents after the installation of barriers.

Interviews with 10 of 13 bridge staff.

• Majority of interviewed bridge staff felt that the barriers had been successful in preventing suicide.

Etzersdorfer and Sonneck (1992) [19]; Sonneck, Etzersdorfer and Nagel-Kuess (1994) [32]; Etzersdorfer and Sonneck (1998) [20]; Niederkrotenthaler and Sonneck (2007) [26]

3

Vienna underground railway system, Vienna, Austria

• Providing guidance on responsible media reporting of suicide

Guidelines on media reporting of suicides – with particular reference to railway suicides – developed and disseminated.

Before-and-after analysis comparing number of completed and attempted railway suicides in two periods:

• Suicidal acts on the underground railway system rose dramatically in the latter part of the pre-intervention period (when sensationalist reports of suicide were common), peaking at nine completed suicides and 10 attempted suicides in the first half of 1987. Following the introduction of the guidelines, both completed and attempted suicides dropped dramatically (to two and one incidents, respectively). This level was then sustained for the remainder of the observation period.

• 1 Jan 1980 – 30 June 1987 (7.5 year pre-intervention period); and

• 1 July 1987 – 31 Dec 1996 (9.5 year post-intervention period).

Subsequent interrupted time series analysis that examined trends in overall suicides from 1946/47 to 2004/05 and trends in railway suicides from 1982/83 to 2004/05.

• Some evidence of nationwide impact, with a reduction of 81 overall suicides.

Isaac and Bennett (2005) [22]

4

Beachy Head, Sussex, United Kingdom

• Restricting access to lethal means

Road access blocked from Jan-Jun 2001 due to foot and mouth crisis

Before-and-after analysis comparing number of suicides in two periods:

• Suicides had risen to a high in the pre-intervention period (85% higher than in 1965–1979) but reduced to zero once road access was blocked.

• 1987–2000 inclusive (14-year pre-intervention period); and

• Jan-Jun 2001 (6-month post-intervention period).

Law et al. (2009) [23]

5

Hong Kong underground railway system, Hong Kong. This system is operated by the Mass Transit Railway (MTR) Corporation and the Kowloon-Canton Railway (KCR) Corporation.

• Restricting access to lethal means

Platform Screen Doors (PSDs) installed on 71 platforms in 30 MTR underground stations on three prominent transit lines. Work began in 2002 and ended in 2005, but most of the busiest station platforms were sealed in the first year.

Before-and-after analysis comparing number of suicides in two periods.

• Significant decrease in the number of suicides on the Hong Kong underground railway system from 51 (10.2 per year) in the pre-installation period to 22 (4.4 per year) in the post-installation period.

• 1997–2001 (5-year pre-intervention period); and

• 2003–2007 (5-year post-intervention period).

• No evidence for displacement to other rail platforms; the number of suicides at MTR stations dropped from 38 to seven, whereas the number at KCR stations remained fairly stable (13 in the pre-installation period and 15 in the post-installation period).

Incorporated quasi-experimental design element which considered numbers of suicides over time at stations with and without PSDs.

King and Frost (2005) [21]

6

New Forest, Hampshire, United Kingdom

• Encouraging help-seeking

Signs displaying Samaritans’ national telephone number placed in 26 car parks in 1998.

Before-and-after analysis comparing number of suicides in two periods:

• Car park suicides dropped from 10.0 per year in period prior to installation of signs to 3.3 per year in the period following it.

• 1 Oct 1988 – 30 Sept 1998 (10-year pre-intervention period); and

• Average annual total number of suicides in the district also decreased.

• 1 Oct 1998 – 30 Sept 2001 (3-year post-intervention period).

• No changes were found in comparable forest districts.

Incorporated quasi-experimental design element which considered numbers of suicides in same periods in comparable forest districts.

Lester (1993)[24]; O’Carroll and Silverman (1994) [27]

7, 8

Ellington Bridge, Washington DC, United States

• Restricting access to lethal means

Eight foot fence installed in 1986.

Before-and-after analysis comparing number of suicides in two periods:

• Suicides dropped from 3.7 per year in period prior to installation of fencing to 0.2 per year in period following it.

• 1979–1985 (7-year pre-intervention period); and

• Suicides from nearby Taft Bridge remained relatively stable (1.7 in pre-installation period; 2.0 in post-installation period).

• 1986–1990 (5-year post-intervention period).

• Overall mean number of suicides in Washington DC was 76.4 in the pre-installation period and 71.6 in the post-installation period.

Lester (2005) [25]

9

Sunshine Skyway Bridge, St Petersburg, Florida, United States

• Encouraging help-seeking

Crisis emergency telephones installed in 1999 and a police presence on the bridge established at the same time.

Before-and-after analysis comparing number of suicides in two periods:

• Suicides dropped from 25 in pre-intervention period (8.3 per year) to 19 in post-intervention period (6.3 per year).

• 1996–1998 (3-year pre-intervention period); and

• 2000–2002 (3-year post-intervention period).

• Increasing the likelihood of intervention by a third party

Pelletier (2007) [28]

10

Memorial Bridge, Augusta, Maine, United States.

• Restricting access to lethal means

11 foot high fence installed on either side bridge in 1983.

Before-and-after analysis comparing number of suicides in two periods:

• 14 suicides in period prior to installation of fence; none in period following installation.

• 1 Apr 1960 – 31 May 1983 (22-year pre-intervention period); and

• Number of suicides by jumping or drowning at sites in Augusta other than the Memorial Bridge remained unchanged (nine in each period).

• 1 Jun 1983 – 31 Jul 2005 (22-year post-intervention period).

• Overall suicide rate in Augusta dropped by 9.0% (from 26.0/100,000 in pre-installation period to 23.8 per 100/000 in post-installation period).

Reisch and Michel (2005) [29]

11

Muenster Terrace, Bern, Switzerland.

• Restricting access to lethal means

Four metre wide wire mesh net, 7 metres below top of terrace installed in 1998 following high level of media attention.

Interrupted time series analysis assessing expected and observed numbers of suicides in two periods:

• No suicides from the terrace in the period following installation of safety net.

• Overall decrease in suicides by jumping from all sites in Bern (95 expected; 44 observed).

• 1995–1998 (4-year pre-intervention period); and

• 1999–2002 (4-year post-intervention period)

Sinyor and Levitt (2010) [30]

12

Bloor Street Viaduct, Toronto, Canada

• Restricting access to lethal means

Five metre high barrier constructed between April 2002 and June 2003 comprising closely-spaced steel rods supported by an angled steel frame.

Before-and-after analysis comparing number of suicides in pre- and post-intervention periods:

• Annual numbers of suicide from the viaduct dropped from 9.3 to 0.0 after the barrier was installed.

• 1993–2001 (9-year pre-intervention period); and

• No reduction in overall rates of suicide by jumping due to increase in suicides by this method at other Toronto sites.

• 2003–2007 (5-year post-intervention period).

Skegg and Herbison (2009) [31]

13

Lawyers Head Cliff, Dunedin, New Zealand

• Restricting access to lethal means

Road access blocked in 2006 due to maintenance.

Before-and-after analysis comparing number of suicides in two periods:

• 14 deaths in the 10-year period before closure (11 suicides, two open verdicts and one accidental death); none during 2-year closure period.

• 1 Aug 1996 – 31 Jul 2006 (10-year pre-intervention period); and

• 1 Aug 2006 – 31 Jul 2008 (2-year post-intervention period).

• 77 police call outs for threatened or attempted suicide in 4-year period before closure (19.3 per year); 19 call-outs during closure period (9.8 per year).

Before-and-after analysis comparing number of police call-outs in two periods:

• 1 Aug 2002 – 31 Jul 2006 (4-year pre-intervention period); and

• 1 Aug 2006 – 31 Jul 2008 (2-year post-intervention period).

Wong et al. (2009) [33]

14

Cheung Chau, Hong Kong. This is an island which attracted visitors who rented holiday flats in which they took their own lives by charcoal burning.

• Encouraging help-seeking

Integrative suicide prevention program established in 2002 which included telephone hotlines, gatekeeper training and suicide patrols.

Before-and-after analysis comparing number of completed and attempted suicides in two periods:

• Visitor completed suicides dropped from 37 (8.7 per year) in pre-intervention period to 6 (2.0 per year) in post-intervention period.

• Increasing the likelihood of intervention by a third party

• 1 Jan 1998 – 31 Mar 2002 (4.25 year pre-intervention period); and

• Visitor attempted suicides remained relatively stable (27 (6.4 per year) in pre-intervention period; 24 (6.9 per year) in post-intervention period).

• 1 Oct 2002 – 31 Mar 2006 (3.50 year post-intervention period).

Incorporated quasi-experimental design element which considered numbers of completed and attempted suicides in same periods on two islands with similar demographic profiles.

• No comparable change in visitor suicides on comparison islands over study period.