Railway fatalities

Characteristics and Incidence of Railway Suicides

2.1  Prevalence of railway suicide


Railways suicides generally consist of persons intentionally putting themselves in the path of moving trains, although there are some other suicide methods, such as being electrocuted by intentionally touching high voltage rails or wires or unintentionally being electrocuted while in the process of putting oneself in the path of a train. The type of behaviour observed in railway suicides varies. For example:

Dinkel, Baumert, Erazo and Ladwig (2011) found that in Germany from 2002 to 2006, out of 1004 cases, there were near equal frequencies of jumping onto the tracks (32.2%), lying on the tracks (32.6%) and wandering on the tracks (34.2%) and there were no age and sex differences between these suicidal behaviours. In their study of suicides in Canada over a ten year period as part of this same contract with Transport Canada, Mishara & Bardon (2010) found that suicidal people most often run out in front of the train from the embankment (33.0%), lie on the track (26.7%) or stand/sit on the track (26.3%).

Classification of railway suicides

The determination of whether or not deaths in railway and metro systems are accidents or suicides is not easy, since in many instances information necessary to make a definitive classification is lacking. For example, in Canada from 1993 to 1996, 39% of railway “trespassing” fatalities were classified as being due to “apparent suicide,” 10% were classified as accidents and the remaining 51% were classified as “undetermined” (Transport Canada, 1996). In a study commissioned by the US Federal Railroad Administration (2008), it was found that “at least 23%” of railway fatalities could be identified as suicides. Mishara and Bardon (2010) found that in Canada between 1997 and 2008, 37.7% of railway fatalities were clearly suicides (classified as such by the coroner), but 7.1% were classified as undetermined. There are important provincial differences, with 45.5% of railway fatalities being suicides in British Columbia and 11.6% in Saskatchewan. There does not seem to be an obvious explanation for these provincial differences.

As television surveillance systems are becoming more widespread in metro systems, on trains and at railway crossings, it may become easier to determine if victims appear to have intentionally put themselves in a situation where they would be struck by a train or if the circumstances appears to be accidental (for example, they accidentally fell off of the platform or, in some rare cases, were pushed in front of the train). Unfortunately, different countries and transportation systems keep different forms of records and use different criteria to determine the number of suicides and “suspected suicides.” Reynders, Scheerder and Van Audenhove (2010) compared data on railway suicides in 15 European countries in the European Railway Agency (ERA) reports with data from the official government statistics as indicated in the European Detailed Mortality Database. They found 34% more suicides in the ERA statistics than the national statistics. There does not seem to be a single explanation to this discrepancy, but differences in classification of suicides between the ERA and the European Detailed mortality Database (with ERA not classifying any death as undetermined) may be part of the issue.

Numbers of railway suicides

In England the first reported railway suicide was in 1852 and it has been estimated that there have been 15,000 rail deaths since then (Farmer, O’Donnell, & Tranah, 1991). Clarke (1994) felt that since the incidence of railway suicides increased proportionately to the expansion of the railway system in England, this offers some evidence of the relationship between the availability of means of suicide and suicide rates It has been estimated (Clarke, 1994), (Symonds, 1985) that railway suicides account for about 5% of all suicides in England and Wales; 5-6% of male suicides and 3-4% of female suicides. Between 1994-95 and 1998-99 (Hudson, 1999), there were 819 cases of over ground railway suicides and “suspected suicides” during this 5 year period and in England and Wales railway suicides were increasing during this period. In this study “suspected suicides” were classified as such when there was no evidence that the death was accidental or a homicide and at least one of the following criteria was met: 1) there was a suicide note, 2) there was a clear statement of suicidal intent to an informant, 3) the person’s behaviour demonstrated suicide intent, 4) the person had made previous suicide attempts, 5) there was prolonged depression or instability as indicated by marked emotional reactions to a recent stressful event or evidence of failure to cope.

In Sweden between 2000 and 2002, there were 145 suicides involving train-person collisions compared to 15 instances of accidental deaths and 32 deaths for which they could not determine the intent of the victim (Radbo, Svedung, & Andersson, 2005).

In Turkey, there were 326 railway suicides reported in the 7 year period from 1997 to 2003, which accounted for 9.4% of all incidents (whether the person died or not) and 21.8% of all fatalities in the railway system (Ozdogan, et al., 2006). Araki and Murata (1986) reported that in Japan there was a substantial increase in suicides of young individuals being run over by trains in the mid-1950s.

Mishara et Bardon (2010) found that suicides represent between 0.8% (in 2004) and 1.7% (in 2007) of Canadian suicides, with a mean of 1.3% across the 2003 to 2007 period. Suicide in the railway system has been a particularly important problem in Germany. During the 6 years from 1997 to 2002, there were a total of 5,731 suicides on the German central railway register of person accidents. In Germany, railway suicides account for 7% of all suicides (Baumert, Erazo, & Ladwig, 2006), There is an average of 18 railway suicides each week in the German railway system. Railway suicides increased in the German railway system in the 10 years between 1991 and 2000, which contrasts with a decrease overall in suicides in Germany during the same time period.

In The Netherlands, of a total of 1500 suicides registered each year, about 180 are railway suicides (Kerkhof, 2003). Railway suicides account for between 10% and 14% of all suicides (C. A. van Houwelingen & Beersma, 2001). Kerkhof (2003) suggested that since the population of The Netherlands is very dense, with no one having to travel more than 20 kilometres to reach a railroad track, this may explain their higher proportion of total suicide that involve railways. Rates in the Netherlands (12.4%), compared to England (5%), Germany (7.0%), Sweden (6.2%), Japan (6.3%), Austria (5.7%) (Deisenhammer, Kemmler, De Col, Fleischhacker, & Hinterhuber, 1997), Denmark (3.1%) (Lindekilde & Wang, 1985) Hungary (2.7%) (Veress & Szabo, 1980) and Canada (1.3 %) (Mishara & Bardon, 2010).

2.2  Lethality of train suidice attempts

The percentage of suicide attempters who die from their attempts varies greatly, most probably related to operational characteristics of the railway and metro systems, such as the speed at which trains travel and enter stations, their ability to stop to avoid a collision and suicide prevention measures, such as so called “suicide pits,” areas of suspended rails so that the train can pass over a person, such as exist in many London underground stations. In Germany, a small minority (9.4%) of persons survive a suicidal event involving a railway train (Erazo, Baumert, & Ladwig, 2005). However, there are relatively high survival rates in many urban transportations systems. A study of the Boston, USA subway from 1966 to 1972 (Guggenheim & Weisman, 1972), where subway trains entered stations at a relatively slow speed, found that only 32% of suicide attempters actually died. Mishara (1999) reported a 28% fatality rate in the Montreal Metro. The death rates in the Munich subway system from 1980 to 1999 was 64% (Ladwig & Baumert, 2004). Between 1981 and 1986 in London, England 43% of attempters died (Cocks, 1987),, which compares with a 42% death rate in Hong Kong and 48% in Toronto, Canada (Gaylord & Lester, 1994). A comparison of 23 urban transportation systems by O’Donnell and Farmer (1992) found that between 20% and 80% of suicide attempters died, depending in part on the design of the station, the tracks and the train equipment. In the Netherlands, 90% of all collisions result in death and 45% occur at or close by a railway crossing.

A study in Germany (Erazo, et al., 2005), where suicide attempt survival rates are quite low, indicated that the highest risk of dying from an attempt occurs in men, on open track areas (as opposed to station areas), on lines where trains travel fast and during night time.

A study of 192 cases of fatal train-person collisions in Sweden during the 3 years from 2000 to 2002 (Radbo, et al., 2005) found that 30% of suicides occurred in station areas and 55% occurred in other places distant from station areas but still in urban areas. In Sweden, most suicides occur during the day time, which they explain in terms of the higher density of train traffic. As in Germany, they found that there was a higher incidence of fatal outcomes from suicide attempts that occurred during night time.

2.3  Socio-demographic characteristics of railway suicides

Lester (1995) reported a lack of correlation between metro suicide rates and overall suicide rates in 17 cities around the world. In Sweden, the ratio of male to female to suicides on the railway is 2.6 to 1 which is comparable to the ratio in the general Swedish population (2.5 to 1)(Radbo, et al., 2005). However the mean age for railway suicides in Sweden is 43, compared to the mean age for all suicides in Sweden of 51 and three quarters of the victims (74%) were between 20 and 59 years old. Kerkhof (2003) reported that railway suicide victims tend to be generally younger that persons who use other methods, mostly men between age 20 and 59 (74%) and the peak frequency occurs 1.5 to 3 hours after sunset for males, and between 7 and 8 hours before sunset for females (Van Luipen, Kiers, Ubink, Kraker, & Thomas, 2002). The demographics picture of persons who die by suicide in railways is similar in Turkey (Ozdogan, et al., 2006). The majority are between age 25 and 60. The most fatalities occur at level crossings.

2.4  Behavioural patterns of railway Suicides

In Sweden, most train-person collisions occur in densely populated areas (Radbo, et al., 2005). In terms of the victim’s behaviour immediately before the crash, a study in Sweden showed that 75% of the suicide victims were waiting on or loitering close to the tracks before the train arrived. A small minority (21 of the 145) appeared suddenly in front of the train by jumping or running in front of the train. This finding, that people tended to wait for the train, is consistent with results in an English study (Abbott, et al., 2003). They report that train drivers generally notice a person on the track when they are usually less than 100 meters from the person and it is usually impossible to stop the train because of the short distance.

The situation may differ in Subway systems because of their spatial configuration (enclosed space, delimitated platform). However, here again, people tend to wait for the train. Mishara (1999) observed in the Montreal metro, that suicide victims tended to wait in the station near the side where the train entered and often jumped in front of the train as it was entering the station or went down on the tracks just before the train was coming into the station. Mishara and Bardon (2010) found that Canadian intercity railway suicides very rarely occur in train stations. Over the 10 years period covered by the study, only 12 (6.5%) of suicides took place in stations, yards, tunnels or bridges. Most suicides took place on open tracks (58.4%) and 35.1% were located at a crossing.

2.5  Environmental characteristic of railway suicides

In Germany there are suicide peaks in April and September with lower rates in December for men. However for women no significant seasonal variations were observed (N. Erazo, J. Baumert, & K. H. Ladwig, 2004). In Canada there is a peak in July and August (25.1% of suicides occur during those two months) (Mishara & Bardon, 2010).

One of the challenges for assessing the effectiveness of any railway suicide prevention strategy is the low frequency of occurrence and the fact that suicides are generally spread across a vast geographical area. When there are clusters at one location; they generally do not persist over time. Our study of the location of railway deaths in Canada over 10 years indicated that there are no locations where multiple suicides have occurred consistently from year to year (Mishara & Bardon, 2010). As the table below shows, the number of suicides occurring in a limited area (2 miles of track) varies over time and there are no long term significant clusters.

Similarly, a study of possible patterns of suicide in California commuter rail lines, which was instigated following a series of teen suicides in the San Francisco Bay area, found that when a long period of time (from 1992 to 2009) was considered, the teen suicide cluster did not persist or stand out in their analyses (Botha, Elmasu, & Leitzell, 2010)

2.6  Psychiatric diagnosis and clustering around psychiatric institutions

Researchers who have studied the relationship between psychiatric diagnosis and railway and metro suicides in Western countries have invariably found a high incidence of diagnosed mental illness. In the county of Fyn, Denmark, in a study comparing railway suicides to other methods (Lindekilde & Wang, 1985), the author found a higher proportion of train suicides (81%) being psychiatric patients compare to those who used other suicide methods (38%).

Huisman, A., van Houwelingen, C. A., & Kerkhof, A. J (2010), in a study of 505 suicides in The Netherlands, found that particularly patients with a bipolar disorder and a psychotic disorder used the railways as a suicide method more often, when compared with other diagnoses of mental illnesses. Van Houwelingen and Kerkhoff (2008) reported that train suicides receive mental health care more than general suicides and are more characterized by severe psychopathology.

In England, there is evidence of clustering of incidents in the London underground stations that were in proximity to a psychiatric hospital (Farmer, et al., 1991; O’Donnell & Farmer, 1994), and no relationship to the extent of passenger usage of a station. Symonds (1985) determined that persons who died by suicide who were receiving inpatient treatments generally chose the nearest available station to the hospital or to their home. Mishara (1999) in a study of the Coroner’s investigation of the 129 persons who committed suicide in the Montreal metro from 1986 to 1996 found that 73% could be identified as having received psychiatric treatment around the time of death, and at the time of death, 27% resided in a mental health treatment facility. The most common primary diagnosis was Depression (50%), followed by Schizophrenia (25%). In the 100 cases where information was available on expression of suicidal intent, 81 had expressed their desire or intention to kill themselves, usually on several occasions (with no indication of a means), and at least 26 people expressed the intention to kill themselves on the day of their death. He found that 9% of persons for whom there were data on prior attempts had previously attempted suicide in the Montreal Metro. Seventy percent of those who committed suicide chose the metro station closest to their residence, and for individuals who resided in an institution, they generally committed suicide at the station closest to the institution. Similarly, Kerkhof (2003) observed that there are “hot spots” near large psychiatric hospitals, with 20 % of all suicides in The Netherlands occurring in the vicinity of psychiatric hospitals. Battistini, A., Gentile, G., Palazzo, E., & Zoja, R. (2009) found in their analysis of autopsies of 155 railway suicides in Milan, Italy, from 1993 to 2008 that 52% suffered from depression and an additional 8.4% suffered from schizophrenia or another form of psychosis.

A study of 5,731 suicides in Germany (N. Erazo, J. Baumert, & K.H. Ladwig, 2004) focussed upon identifying clusters of increased suicide prevalence. They identified 16 high risk places for railway suicides in Germany, each having 6 to 29 suicides per railway kilometre. Of these, 75% were located within 2 km of psychiatric hospitals. Van Houwelingen, Kerkhof and Beersma (2010) found that in The Netherlands, all high risk locations with 28 or more suicides from 1980 to 2007 were within 800 meters walking distance of a psychiatric hospital. An investigation of a small number of suicides in Brisbane, Australia (Emmerson & Cantor, 1993) found that 57% of victims had been treated for Schizophrenia, 57% were psychiatric inpatients at the time of the suicide and 48% of the deaths occurred close to the regional psychiatric hospital. De Leo and Krysinska (2008) found that between 1990 and 2004, in Queensland, Australia 40.4% of suicide victims had a psychiatric diagnosis and 29.8 had positive blood alcohol content.

Mishara and Bardon (2010) reported on the sub-sample of 253 suicide victims in Canada for whom there was information on mental health status. The most prevalent diagnosis was depression (57.3%), followed by substance abuse problems (30.0%) and psychosis/schizophrenia (17.4%). Also, at the time of death 45.8% had a level of alcohol in their blood over the legal limit to drive. Only 22.1% of suicides were currently in psychological or psychiatric care. They identified 51 areas with 2 suicides or more within the 10 years covered by their study. They concluded that there are no significant suicide clusters in Canada, with all multiple suicides over a ten year period having just a few incidents. However, a higher number of these clusters of two or more suicides had occurred close to a hospital (35.3%), which is substantially greater than for accident cluster sites (10.0%).

Although there are not much data available and research methodologies may not be comparable, it is possible that the prevalence of psychiatric disorders is lower in railway suicide victims in some less developed countries. A study of 22 suicide attempters and 13 completed suicides in the Calcutta, India underground railway (Chowdhury, Dutta, & Chowdhury, 2000) found that only 15% could be identified as having a psychiatric illness.


Media impact on railway and metro suicides

There is evidence that media reports on railway and metro suicides can influence the incidence of these events. The classic example is the dramatic increase in Vienna of the number of subway suicide deaths from 1984 to 1987 when local newspapers competed to print sensational and detailed descriptions whenever subway suicides occurred. The increase during the same period of the overall suicide rate in Vienna was explained by the increase in the subway suicide increase. After the Austrian Suicide Prevention Association convinced newspapers to completely stop publicizing subway suicides there was a 75 % decrease in suicides by this method, no increase in suicides by other methods and this lower rate was maintained for five years (Sonneck, Etzerdorfer, & Nagel-Kuess, 1994), and had remained low 10 years later (Etzerdorfer & Sonneck, 1998).

In Germany, a 6-episode weekly serial that was aired in 1981 and again in 1982 provided dramatic data on the possible effects of a fictional suicide on rates of suicide using a specific method. This fictional television series concerned a 19 year-old male student who committed suicide by being hit by a train and the suicide was shown at the beginning of each episode. During the 70 days after the first episode, railway suicides among 15-19 year old males increased by 175% and there were also increases for males up to age 40 and females 15 to 30 (Menahem, 1988), (Schmidtke & Häfner, 1988). Significant increases were also observed when the series was re-broadcast the following year. Kunrath, Baumert and Ladwig (2010) reported a significant average daily increase of 44% in railway suicides following widespread television and newspaper coverage of the accidental railway deaths of 3 members of a railway investigation team in December 2006 in Bregenz, Austria. Because of the potential for media reports leading to increased metro and railway suicides, newspapers in many cities have stopped reporting on railway and metro suicides.

Mishara and Bardon (2012) did not find any noticeable impact of the media on Canadian railway suicides. Indeed, guidelines on suicide reports has been provided to the Canadian press and it appears that in the case of metro and railway suicides, these guidelines are generally respected by the industry. Therefore, railway suicides are rarely reported in Canada, and almost never in a sensationalistic manner.

Prevention Strategies

A number of possible measures to prevent people from committing suicide on railways have been proposed. Most prevention measures have focussed upon preventing potential victims from gaining access to the tracks, identifying and intervening with potential victims just before they attempt suicide and technical interventions to minimize harm to suicide attempters. Very little has been done to decrease the risk of a railway or metro suicide by trying to reduce the suicide potential in vulnerable populations or by influencing their choice of a railway or metro system for their suicide method.

An example of a typical approach is the “Project for stopping trespassing on railroad” of the Swedish Transportation Administration (2011). The project is based upon keeping potential suicide attempters away from tracks and detecting them in time to for trains to stop. The Swedish plan includes testing:

  • Fences between tracks (on double or more tracks)
  • Fences to prevent access to tracks
  • Lights at the end of platforms
  • Improved gates at the end of platforms
  • Marking and signs
  • Eliminating “hideaway” places where people can hide close to tracks
  • Camera surveillance with automatic alarm systems
  • Motion detectors

Similarly, in October 2009, a report commissioned by the Belgian Government proposed a series of initiatives to prevent suicides on the rails, again focusing upon limiting access, (Infrabel, 2009), which they hoped to implement in 2012-2015. Their approach was to add more fences to limit access, eliminate many grade crossings, improve visibility for train drivers by cutting vegetation along rails and installations of markers at all unprotected crossings with information about help and a button that connects a person in distress with the railway security services. They also are working with several local partners to advertise suicide prevention services.

The following sections review the scientific evidence for the various prevention strategies that have been tested to date:

4.1  Limiting access to tracks

Although it is generally agreed that severely limiting access to the tracks can virtually stop railway and metro suicides, this prevention measure is often considered to be financially untenable, particularly in railways that have extensive exposed track areas and in existing metro systems. Barriers that effectively stop passengers from gaining access to the tracks currently exist in at least some of the stations and railway and subway lines in: Bangkok, Barcelona, Beijing, Copenhagen, Dubai, Guangzhou, Hong Kong, Kuala Lumpur, Lille, London, Paris, Rome, Saint-Petersburg, Seoul, Singapore, Taipei, Torino, and Toulouse.

ProRail in The Netherlands experimented in 2005 with replacing metal fences near a mental institution near Eindhoven with a flyover and also removed trees and shrubs to eliminate hiding near the tracks (Anonymous, 2010). They reported that “the number of suicides in the area declined, but elsewhere there was an increase”, but did not report specific data or statistical tests of significance.

One of the frequent objections to the installation of barriers in stations to prevent suicides is the concern that suicidal persons will choose other substitute methods and, if not all stations have barriers, potential victims will simply go to stations where barriers have not been installed. However, a study by Law and colleagues in Hong Kong (2009) showed that between 1997 and 2007, when 30 of the 38 stations on the MTR Line had sealed doors installed on all platforms, but the remaining 8 MTR stations and all stations on the KCR network had no barriers, there was an overall reduction of 59.9% in suicides with no indication of substitution by potential attempters using the unprotected stations.

There are great costs to install physical barriers along thousands of kilometres of railway lines and to secure access to pedestrians and vehicles at thousands of railway grade crossings.   The costs of installing secure barriers with automatic doors that open when a train has stopped in the station can be extremely high in existing metro systems that were not originally designed to limit access to the rails.  For example, a task force at the Montreal metro system estimated that the cost of installing such a system in the existing stations could surpass the system’s annual operating budget.  Some metro systems have installed such a system in new metro lines (e.g. Paris, France), but easy access to the tracks is still available elsewhere in the metro system.

4.2  Blue lighting, green crossings and charging for cleanup in Japan

The Japanese have experimented with the colors blue and green to prevent suicides.  It has been reported that the installation of blue-hued street lights may reduce suicides at railway stations and crossings.  Unfortunately, there are no published experimental validations of the impact of blue lighting with data, statistical tests and comparisons with equivalent control locations. The reports from Japan may appear promising, but without scientific data, the impact of this intervention is impossible to assess.   When bright blue LED ceiling lamps were installed in 29 stations on the Tokyo-Yamanotte Line in 2009 (Yuasa, 2009), the railway was quoted as saying “Blue is said to make people’s minds more serene.  The blue lighting is part of an effort to prevent suicides.”  (Anonymous, 2009). The Keihin Electric Express Railway Co. also changed the colour of lights on the ends of platforms at Gumyoji station in Yokohama to blue.  The company reported that before installing the blue lights “a few people” attempted suicide at the station every year (Stolte, 2008).  In the 10 months after installing the blue lights no suicide attempts had occurred. 

In December 2006, West Japan Railway Company installed blue lights at 38 crossings including the Hanwa Line connecting to Osaka and Wakayama prefectures.  In the two years following, there were no reported accidents or suicides at the 38 sites (Anonymous, 2008). Although these reports seem impressive, the only way to verify if blue lights reduce suicides in to conduct a randomized control study with sufficient number of locations to produce statistically significant findings. 

JR Railway has also experimented with installing mirrors opposite platforms in the hope of deterring people from jumping by the prospect of watching themselves in the mirror (Fernandes, 2003). Similarly, the Metro Railway in Calcutta, India has been working with the Calcutta Samaritans helpline by placing mirrors, having calming music and placing posters inviting people to choose life over death in 17 underground stations (Fernandes, 2003).  JR East (East Japan Railway Company) has also installed mirrors in stations and painted crossings “a dazzling green.” (French, 2000), to “alter the state of mind of would be suicides.” (Fernandes, 2003). They have also sponsored a telephone hotline which, in 2006 received 1300 calls a week from suicidal people.  However, to date there are no published reports on the effectiveness of any of these initiatives.

There are many train and subway companies that operate in the Tokyo area.  All train companies charge the families of suicide victims for damages caused by the suicides.  More suicides have been reported to occur on the JR East railway.  This has been interpreted as possibly linked to the fact that JR East charges families less than the other railway and subway companies, .  The cost to a victim’s family for the clean-up or damages can be from 8 million yen (80,000$) up to 100 million yen (1 million dollars) (McCaffery, 2009),. Due to the JR East’s high rate of suicides on their Chuo line, suicides are sometimes referred to as “chuo-cide.”

4.3  Television surveillance

In a talk by John-Åke Halldén of the Swedish National Railway (Halldén, 2006), camera surveillance was proposed as a potentially effective method of suicide prevention, providing that professional operators who monitor have the right training and competence.  Since television surveillance systems may not only serve to prevent suicides, but also provide for a more secure environment by identifying crimes and accidents, metro systems around the world are installing television monitoring of stations In addition, some railways have television surveillance at grade crossings and in stations and many locomotives have cameras installed to record events from the engineer’s perspective. 

The effectiveness of surveillance in preventing suicides depends upon the availability of vigilant observers monitoring the video screens and the potential to dispatch help to the scene or notify train drivers in time to prevent the suicide.  A metro or railway system with hundreds of cameras to monitor requires many people available throughout the day.  While the installation of camera monitoring and videotaping is technically feasible and usually cost efficient, the personnel costs to monitor the video screens continually can be extremely costly and the ability to identify potential attempters and intervene before they act may be limited, even with the best surveillance system.  An ongoing project by the authors of this report involves reviewing video recordings of suicides in the Montreal Metro subway system in order to try to identify behavioural patterns which may be identified by a computer programme, which would signal personnel to investigate when an individual exhibits potentially “at risk” behaviours.

4.4  Gatekeeper programmes

In metropolitan transportation systems, suicidal individuals have been reported to sometimes express their desire to kill themselves to railway personnel (Mishara, 1999).  In other instances severely disturbed individuals exhibit behaviours which may indicate to staff that they are “disturbed” and signal that it may be worthwhile to investigate of they may potentially be dangerous to themselves and others.  In these instances, appropriate verbal questioning to assess suicide risk and offer help may avert a suicide attempt.  The Washington Metropolitan Area Transit Authority approved the testing of a suicide prevention programme based upon training employees to better identify persons who are potentially at risk of attempting suicide and developing a safety assistance programme for persons identified as being at risk (Customer Service. Operations and Safety Committee. Washington Metropolitan Area Transit Authority, 2010).

In the ongoing project by the authors of this report which involves analyzing video recording of the behaviours of suicide victims from the time they entered the station until the time of their suicide in order to try to identify behavioural patterns that may be identified automatically by computer technology, it may be possible to then dispatch personnel to talk to the person with “suspicious” behaviours in order to assess suicide risk and, if necessary, provide help.  Results from this study should be available by the end of 2013. 

The Toronto Transit Commission, in collaboration with the University of Toronto Arthur Sommer Rotenberg Chair in Suicide developed a “Gatekeeper” training programme to train personnel to better identify distressed and at-risk persons using the Toronto Transit Commission trains (Fortin, 2008). The effects of this programme, which began in 2008, has not yet been reported.

In the Montreal metro system (with which the authors are currently collaborating), there are an average of 30 suicide prevention interventions each week in the metro, but almost all interventions are initiated due to reports by metro staff members or passengers who report “suspicious” activities by telephone or radio to the metro control dispatchers.  In addition, some interventions occur when the local suicide prevention centre informs the metro system that a person who is calling from a specific station is intending to attempt suicide in the metro. Several years ago, the Montreal metro initiated a somewhat controversial poster campaign encouraging passengers to use the red emergency phones in metro stations to report when someone seemed disturbed. This publicity campaign did appear to result in an increase in suicide prevention interventions.  However, the campaign was contested by groups concerned with the rights of mental patients who were concerned that targeting disturbed people would stigmatize the mentally ill.  This concern was not evident in the reports of the interventions in which the “disturbed” individuals often said that they appreciated the help from the metro employee who expressed concern for their wellbeing when they arrived and talked to them about their feelings and intentions.  However, this programme was discontinued before its impact could be evaluated.

4.5  Signs, posters and telephones

There is some evidence that signs displaying a helpline number near a public telephone may decrease suicides in suicide hot spots.  In a project in New Forest, Hampshire, England, the Samaritans (a national volunteer-based helpline) posted signs with their telephone number in selected car parks in an area where there were high numbers of car exhaust suicides.  Over three years, there was a significant decrease in both the number of car park suicides and the number of total suicides in the New Forest district (King & Frost, 2005).

Similarly, placing dedicated telephone lines connected to a suicide prevention hotline at a 24-hour Psychiatric Emergency Service on the Mid-Hudson Bridge in New York showed that over a 2 year period, 30 out of 39 persons who went to the bridge to commit suicide used the phone to call for help and, of these, only one subsequently jumped from the bridge.  However, 5 of the 9 who did not use the phone jumped to their death (Glatt, 1987).  In the San Francisco area, Caltrain started a pilot project of new signs with a hotline number on a 10 mile segment of the right of way in September 2010 (Caltrain, 2010) 

In the Montreal Urban transit system, there have been several publicity campaigns with posters advertising that help is available at the local suicide prevention centre, Suicide Action Montreal. It is evident the people have phoned telephone help lines and say that they called after seeing a poster or sign in a station or near the rails.  However, there is no systematic study of the level of suicide risk of the callers.  To date, there are not enough data in the scientific literature to convincingly prove indicating that signage has an impact on reducing suicides, although the findings cited in this section indicate that this may prove to be a potentially effective intervention.  However, the effects of signage with helpline numbers needs to be evaluated with a comparison between signed locations and comparable locations without signage, over a long enough period of time to determine if there are statistically significant effects upon the incidence of suicides.

4.6  Technical Interventions: “Suicide Pits”

Other physical changes to the railway and metro systems may decrease fatalities in suicide attempters.  In London, subway stations with so-called “suicide pits” (suspended rails so the train can pass over a person who has fallen) have fewer deaths among those who attempted suicide (O’Donnell & Farmer, 1994). Coats and Walker (1999) reported that the presence of a pit reduces overall mortality by 57%. This technical modification appears to save lives and could be incorporated into many more metro systems and areas at high risk of railway suicides.   It is also possible that other technical changes, such as having trains enter stations more slowly, re-designing the front of trains to be less dangerous upon impact, better signals when a person or vehicle is on a track at a grade crossing and providing better lighting and more “obvious” television surveillance cameras may reduce fatalities and harm.  Again, scientific evidence from well-designed studies is needed.

4.7  Changing the desirability of rail suicides

None of the above interventions try to prevent persons from being suicidal or more specifically wanting to commit suicide in the metro or a railway.  Studies in Boston of survivors of suicide attempts indicated that attempters thought that they would certainly die and that they would die instantly and painlessly (Guggenheim & Weisman, 1972).  However this belief is not founded in fact.  In most urban metro systems there are more survivors than fatalities and those who die may experience minutes, hours, days or weeks of agony before succumbing.  For instance in the Montreal metro system, about half of attempters do not die (Mishara, 1999) and in the Canadian railway network, 5.3% of those who die from their injuries do so after being transported to hospital (Mishara and Bardon, 2012). Publicizing these facts may prevent some people from choosing this method, but it is the experience of the authors of this paper that this suggestion usually meets with resistance, since it also involves publicizing suicides by this method.  In an ongoing study involving in-depth interviews with Montreal metro suicide attempters who are seen in a hospital emergency room, the preliminary results suggest that the belief that this suicide method results in a certain and painless death was omnipresent and several attempters were surprised when they learned from experience or were told by the researcher that this is not the case. Whether or not publicizing the fact that railway and metro suicides do not result in a certain, painless and immediate death can have a dissuasive effect will have to be verified by research investigations.

4.8  Prevention Education in Psychiatric Facilities near Rails

The consistent finding that most railway and metro suicide victims are in treatment for psychiatric disorders, that they often reside in a psychiatric facility and commit suicide in the station or on the tracks in the vicinity of a psychiatric institution, provides important information to identify both high risk populations and high risk locations in metro and railway lines.  It is clear that if cost-effectiveness is a concern, it is best to focus technical modifications on these stations, grade crossings and tracks areas of greatest risk, which are close to psychiatric institutions.  Similarly, it may be useful to develop specific training programmes for staff of psychiatric facilities near tracks on the identification of suicide risk and the need to educate patients about the false beliefs that rail suicides provide for a certain, painless and immediate death.  Mishara (1999) found that most metro suicide victims threatened to commit suicide beforehand and they were often not taken seriously by psychiatric personnel.  This would imply that it is important that staff in psychiatric institutions be well educated in the assessment of the suicide potential of their patients.  Although it has been suggested that potential victims could be informed of the devastating effects of suicide on train drivers (Kerkhof, 2003), one would have to provide some empirical evidence of the dissuasive effect of such a programme before recommending its implementation.

At the present time there have been no published studies that report on the effectiveness of railway and metro suicide prevention programmes that target specific high risk populations.  We need better information to permit more precise identification of those at greatest risk of committing suicide in the railways and metros.  To date, we know that persons who receive inpatient treatment for psychiatric problems are at greatest risk, but we know little about which psychiatric patients are most at risk and how their suicides may be best predicted.  However, we do have some indications of at least one high risk group:  Survivors of railway and metro suicide attempts are at particular risk of repetition and eventual completed suicides.  Thus, persons intercepted in the process of an attempted may be targeted for specific interventions in order to prevent future attempts.  Railway and metro personnel may profit from developing better collaborations with psychiatric emergency services and institutions.  However, it is most important to conduct research to learn more about the specific motivations of people who choose to attempt and commit railway and metro suicides and to evaluate the effectiveness of prevention programmes.

4.9  Public Education on Safety

There is some evidence that public education, such as the publicity campaigns by Operation Lifesaver in Canada and the United States may decrease the number of train collisions at rail-highway crossing (Savage, 2006). Savage found that, based upon the experience of 46 states from 1996 to 2002, there is a significant correlation between the amount of education activities and the number of collisions.  However, a statistically significant effect on the number of deaths could not be concluded, nor was there evidence that public education campaigns may specifically affect rail suicide rates.

4.10  Educating Media and reducing reporting

It is evident from the research literature on media effects on suicide and studies of railway and metro media reports and fictional representations, that media presentations of suicides by this method in news as well as fictional television dramas may greatly increase the incidence of railway and metro suicides.  Therefore, it is important that media reports and depictions of railway and metro suicides be limited as much as possible.