74 Protecting bridges and high buildings in suicide prevention

Suicide by jumping is a relatively uncommon method of suicide in most countries. However, in some places where there is accessible high-rise housing, jumping accounts for a significant proportion of suicides. While, internationally, most suicides by jumping occur from residential housing units, preventive efforts have tended to focus on a relatively small number of sites (often bridges) which have acquired notoriety as sites for suicide. A small number of studies suggest that the installation of safety barriers at these sites is an effective approach to reducing suicides by jumping. The extent to which lessons and principles from these examples may be applied to other jumping sites has yet to be fully explored.

Evidence leads to three major generalizations about the relationships between the accessibility of a specific method of suicide and suicidal behaviour:

Because of the complex relationships between access to methods and suicidal behaviours, it is important that policies aimed at methods reduction are subject to thorough monitoring and evaluation.

The potential reduction of suicide rates by restricting access to means of suicide is addressed in a large body of research (Cantor et al. 1996; Beautrais 2000; Daigle 2005). In this chapter the focus will be one method of suicide: jumping from bridges and high buildings. The prevalence of suicide by jumping, characteristics of those who die by jumping, features of bridges and high buildings which may encourage suicidal people to jump, and evidence for various approaches to prevent suicide by jumping is described.

There are marked national and international differences in rates of suicide by jumping. In part, jumping appears to be more common in some countries, city states or cities in which there are a large number of accessible high-rise buildings. For example, suicide by jumping accounts for approximately 60 per cent of suicides in Singapore (Ung 2003), 45 per cent in Hong Kong (The Hong Kong Jockey Club Centre for Suicide Research and Prevention 2005), 30 per cent in New York City (Fischer et al. 1993), but only 7 per cent in Switzerland (Reisch and Michel 2005), 6 per cent in Norway (Statistics Norway 1982) and Australia (Australian Bureau of Statistics 2001), 5 per cent in England and Wales (Gunnell and Nowers 1997), 4 per cent in Sweden (Statistics 2002), and less than 2 per cent in New Zealand (Ministry of Health 2006). There is little research that has explored reasons for jumping being more common in some cities, but less so in others with equal volume and accessibility to high-rise buildings. In addition to availability, the factors that may contribute to preferences for a particular method in a population include culturally determined views and attitudes to accepted methods of suicide, the role of media influences in shaping population attitudes about methods of suicide, the perceived lethality of various methods and the personal acceptability of the method (Clarke and Lester 1989). There is need for more research about the reasons for choice of jumping as a method of suicide.

Because of the high lethality of the method, there are sound reasons to try to prevent suicides by jumping in those countries in which it is a common method of suicide, and even in countries in which suicides by jumping are relatively uncommon. The tendency of clustering of such suicides to specific sites that consequently gain notoriety as suicide sites justifies preventive interventions.

Some studies have suggested that individuals who choose jumping as a method of suicide tend to be younger, male and more often to have psychotic disorders than those who choose other methods (Cantor et al. 1989; Ku et al. 2000; Beautrais 2001; Lindqvist et al. 2004). However, other studies have not found such differences (Prevost et al. 1996; Gunnell et al. 1997). For example, in Australia, 38 per cent of a series of suicides from a bridge in Adelaide, 45 per cent from a series of deaths from Brisbane bridges and 27 per cent from a series of deaths from a Melbourne bridge had psychotic illnesses (Pounder 1985; Cantor et al. 1989, Coman et al. 2000). In contrast, suicides from the Clifton Suspension Bridge in Bristol included only 10 per cent with psychosis, and suicides from Beachy Head in the United Kingdom included 9 per cent with psychosis, while deaths from the Jacques Cartier Bridge in Montreal included only 13 per cent with psychosis (Nowers and Gunnell 1996; Prevost et al. 1996; Surtees 1982). The inconsistency in findings from studies of jumping might be explained by such factors as the small numbers of suicides in some of these descriptive studies (for example, thirty-two suicides in the Adelaide series described above), the proximity of some jumping sites to psychiatric hospitals—the psychiatric hospital was only metres from the bridge in Auckland, New Zealand, for example (Beautrais 2001), the notoriety of the site in question, and the frequency of suicide by jumping in the countries of study. Nevertheless, in some countries where suicide by jumping is common, reports suggest one quarter of those who chose jumping had psychotic symptoms at the time of their attempt (Ku et al. 2000).

Suicide is often an impulsive act and it has been shown that most people who survive suicide attempts do not go on to die by suicide (Owens et al. 2002). Studies of people who attempted suicide by jumping but were restrained (Seiden 1978) or survived (Rosen 1975) found that most did not ultimately die by suicide. Researchers who studied individuals who made both fatal and non-fatal suicide attempts by jumping in Hong Kong found that half were described as impulsive by their psychiatrists (Ku et al. 2000). These findings suggest that restricting access to sites for jumping may delay or avert some fraction of impulsive suicide attempts and suicides.

Sites for suicide attempts by jumping include bridges and viaducts, high-rise public, residential and institutional buildings (including hotels, psychiatric hospitals and multi-storey car parks), and cliffs and terraces. In countries in which suicide by jumping is a common method of suicide most people attempt suicide from their home, from relatives' homes, or from nearby familiar places (Ku et al. 2000). However, an issue of particular concern is that specific sites or structures may acquire reputations, symbolic significance or iconic status as places for suicide (Beautrais 2001). Often these sites may be in public places, in attractive locations and may be aesthetically pleasing structures in themselves. These features appear to influence the appeal of the site for suicide. Sites which become notorious as places for suicide by jumping have emerged both in countries where suicide by jumping is common and in those in which it is rare. For example, in Hong Kong, where suicide by jumping accounts for almost half of all suicides, the new Tsing Ma Bridge has quickly acquired a reputation for suicide (see http://www.info.gov.hk/gia/general/2003121/15/1015173.htm for more detail). In New Zealand, where jumping accounts for <2 per cent of all suicides, Grafton Bridge had a long-standing national reputation as a suicide site (Beautrais 2001). Sites which acquire iconic reputations for suicide are sometimes referred to as ‘hotspots’ (Reisch and Michel 2005). All significant suicide hotspots, internationally, are, in fact, jumping sites (Aitken et al. 2006). For example, the Golden Gate Bridge in San Francisco is the most popular public suicide site in the world, with an average of almost 20 suicides per year (Gunnell et al. 2005). It is interesting to note that sites may acquire reputations for suicide in spite of relatively small numbers of suicides from these sites. For example, Grafton Bridge in Auckland had a local reputation as a site for suicide despite having only one suicide per year (Beautrais 2001). Similarly small numbers were associated with other iconic sites—the Bern Muenster Terrace with four deaths per year (Reisch and Michel 2005), and the Bristol Suspension Bridge with an average of eight suicides each year (Bennewith et al. 2007). While there is no clear account of the mechanisms by which such sites acquire iconic status as places for suicide, it seems likely that this process would involve the development of local history, traditions and myths.

The public nature of sites that become notorious for suicide by jumping may lead to the witnessing of suicides by the public or that members of the public may be placed at risk. For these reasons, suicides from such sites tend to attract media coverage which, in turn, tends to promote and perpetuate the notoriety and appeal of the site. There is some evidence that people tend to make their choice of suicide methods based upon their perceptions of what they understand to be certain to achieve death: quick, readily available methods with a low risk of disfigurement (Lester 1988). Jumping fulfils some of these conditions. On the other hand, the symbolism and romanticism associated with an iconic or symbolic suicide site appear to play a decisive additional role for those who chose to jump from such sites (Seiden and Spence 1983–84).

There are a series of approaches to preventing suicide by jumping. Installing barriers or safety nets to restrict access to specific sites is one preventive approach. A limited number of studies have evaluated the impact of installing barriers or safety nets at sites which acquire notoriety for suicide by jumping. These studies suggest that barriers significantly reduce suicides by jumping at the site and in the surrounding area. It is less clear, however, whether such restrictions reduce overall rates of suicide.

Examples of this approach include the installation of barriers on the Clifton Suspension Bridge in Bristol, England which halved the number of deaths from 8 to 4 per year, with no evidence of an increase in suicide at other sites in the area, and no overall change in the rate of suicide in the surrounding region (Bennewith et al. 2007). The installation of a safety net at the Bern Muenster Terrace in Switzerland eliminated suicides from the site, again with no evidence of substitution of another site (Reisch and Michel 2005). The removal of safety barriers at Grafton Bridge in Auckland, New Zealand, led to a fivefold increase in suicides from the site (Beautrais 2001) and their reinstallation eliminated suicides with no apparent substitution of other sites (Beautrais et al. submitted).

Similar reductions in suicides from sites to which access has been restricted, usually by the installation of barriers but sometimes by prohibiting access to the structure, have been reported (but not formally evaluated) for the Empire State Building, the Ellington Bridge in Washington DC, the Eiffel Tower, the Jacques Cartier Bridge in Montreal, the Bloor Street Viaduct in Toronto, the Bosphoros Bridge in Istanbul and the Sydney Harbour Bridge (Aitken et al. 2006). Barriers may be effective in reducing suicides at these sites by averting impulsive attempts, by preventing access to sites which have symbolic significance for suicide (for which other less attractive sites are not substituted), or by forcing attempters to substitute less lethal methods. Arguments for installing barriers to prevent jumping are strengthened by observations that suicide rates by jumping increased in Singapore as the number of high-rise housing buildings increased (Lester 1994), and decreased after the Kobe earthquake in Japan, perhaps, in part, because the earthquake destroyed many high-rise buildings (Shioiri et al. 1999). Barriers at jumping sites may take various forms but evidence suggests they need to be 155cm or higher to prevent suicides and must be built in such a way that they do not offer a foothold for potential jumpers (Berman 1990; Reisch et al. 2006).

A second preventive approach that is used at known sites for suicide by jumping is to erect signs providing contact details of telephone help lines for suicidal individuals. In some cases telephones are provided from which individuals can make direct calls to helplines. This approach is underwritten by the expectation that many people planning to jump will be ambivalent about wanting to die and, if provided with ready access to a helpline, will use it. There are limited formal evaluations of this intervention and some concerns that such signs may risk promoting the idea of suicide to whose who had not previously contemplated it (Glatt 1987).

A further approach involves using surveillance and patrols at popular sites. Examples of this approach include closed circuit surveillance cameras, and site patrols by police, dedicated suicide patrol officers or unpaid volunteers. For example, the Golden Gate Bridge has security cameras and has been patrolled by a team of dedicated suicide prevention officers since 1996 (Reed 1996). However, there have been no formal evaluations of such measures and in many places in which they have been implemented, suicides have apparently not declined. There are concerns that these measures may be ineffective because patrols cannot monitor all parts of a structure (for example, a long bridge) at one time, even with the assistance of security cameras.

As noted above, suicides from iconic public sites often lead to extensive media coverage. While these incidents may be newsworthy the reports may inadvertently promote the site for suicide, endorse the symbolic status of the site and risk imitative suicide attempts (Pirkis and Blood 2001; Yip et al. 2006). For these reasons an additional approach to reducing suicides by jumping involves muted media reporting of such incidents, and of preventive measures implemented at specific sites, since reports of these may also advertise the site as a place for suicide (King and Frost 2005). Evidence for the effectiveness of this approach is provided by the reduction in suicides by jumping in front of trains in the Viennese subway following the adoption of muted media reporting practices about these incidents (Etzersdorfer and Sonneck 1998). While it may be difficult to establish the extent to which prudent media reporting may lead to reductions in suicides, there may be opportunities to make ‘naturalistic’ evaluations at specific sites. Findings from some studies suggest that opportunities for media reporting are reduced when barriers reduce suicides at specific sites, with this reduction in reporting being associated with fewer suicides by jumping from all sites in the local area (Gunnell et al. 2005).

A further approach involves collaboration with local or national building industries including architects, town planners and construction companies, to encourage the incorporation of safety features (such as barriers, safety glass in rooftops, enclosed stairwells, restricted access to rooftops and balconies, restricted window apertures) into designs of new buildings. Such efforts should particularly concern residential housing and institutions such as hospitals, prisons and juvenile detention centres. There appear to be no evaluations of such approaches. In addition, it would appear to be prudent for hospitals to move existing psychiatric inpatient facilities, or in the planning of new ones, and position them on lower floors of buildings from which patients are not able to jump to their deaths.

A useful adjunct to the practical measures outlined above includes the development of systems to monitor overall trends in suicide mortality with particular responsibility for identifying jumping sites which emerge as common sites for suicide, and implementing appropriate measures to minimize suicides from these sites.

In countries in which jumping accounts for a significant proportion of all suicides there is clear justification for implementing measures to reduce suicides by this method. Even in places where jumping plays only a small part of the overall suicide rates, it is justifiable to attempt to reduce these suicides. As discussed above, a range of preventive approaches exist. Thus far, however, the best evidence for the effectiveness of any approach comes from a small number of evaluations of the impact of installation (or removal) of safety barriers at sites which have become popular as places from which to die by suicide by jumping. The principles and lessons learned from these examples may be applicable to other jumping sites such as high-rise residential housing units and car park buildings.

Although barriers at jumping sites can help to reduce suicides, efforts to install barriers are frequently met with resistance. Often, the public view is that erecting barriers will decrease the aesthetic appeal of a site, or that suicide is inevitable in suicidal individuals and therefore barriers will be ineffective. In a study of public attitudes to suicide by jumping, Miller (2006) asked 2770 people to estimate how many people attempting to jump from the Golden Gate Bridge would have eventually died by suicide if they had been prevented from jumping off the bridge. More than a third of respondents replied that all would eventually die by suicide, and a further 40 per cent replied that ‘most’ would eventually die by suicide. As noted above, this perception is not supported by the research evidence. Another common reason for not installing barriers is cost. While advocates for suicide prevention argue that a socially responsible society would invest in measures to prevent a relatively small number of suicides by vulnerable members, opponents argue that the costs and lack of benefit for most of the society's members do not warrant the expense. A further reason advanced for not installing barriers is that other sites or methods would inevitably be substituted. There is no strong evidence to support this argument. First, suicide by jumping is a highly lethal method—if substitution occurs it is likely to be towards a less lethal method. Second, recent studies suggest that installing barriers at a popular site may, in fact, reduce suicides at other local sites, presumably because of the reduced media coverage of suicides at the notorious site (Bennewith et al. 2007). Third, people who survived suicide attempts by jumping from the Golden Gate and Oakland Bay Bridges in San Francisco said they would not have used any other method if the bridge had not been available (Rosen 1975).

Paradoxically, it seems that the best evidence for reducing suicides by jumping comes from a small number of studies in countries in which suicide by jumping is relatively uncommon. Studies of interventions to reduce suicides by jumping in countries where jumping accounts for a significant fraction of all suicides are lacking. There is a clear need for more investment in implementing and evaluating a range of preventive measures in these countries and, indeed, wherever suicides by jumping are a significant problem.

There is increasing evidence that suicide by jumping can be prevented at specific sites by the installation of appropriate safety barriers and that, in some cases, this measure may decrease suicides by jumping in the surrounding area. This evidence suggests that installing barriers is the current ‘best practice’ and socially responsible approach to reducing suicides by jumping. However, there is a need to explore the extent to which the principles learned from these examples can be applied to a range of preventive approaches to reduce method-specific mortality and morbidity, and overall rates of mortality and morbidity.