As shown it the state-of-the-art work, there is little published evidence on the effectiveness of preventative measures to reduce railway suicides and trespasses and their consequences. Moreover, the effectiveness of some measures has never been assessed. RESTRAIL research aimed to tackle this issue and assess a selection of the most promising measures. The RESTRAIL consortium members selected several measures for in-depth field assessment using the lessons learned during the project, as well as the needs of the corresponding national stakeholders. They developed a series of 11 field pilot tests in different locations (Spain, Finland, Sweden, Turkey, UK, Germany, the Netherlands, and Israel).
Planning and monitoring of the pilot tests
Each pilot test was conducted according to a specific implementation plan in order to monitor the evaluation process and to provide additional empirical evidence for the effectiveness of measures (see  for more information concerning the selection of the measures and their implementation in pilot test planning and execution). The main purpose of these evaluations was to estimate quantitatively the effect of single measures or combination of measures on a specific problem. Some field trials focused on measures to prevent suicides only, others on means to prevent trespass, while others addressed the consequences of incidents. Table 3 displays the final list of tested measures, their target problem and the location of the field test.
Method for implementation of the field trials and evaluation of the tested measures
The methodologies used in the evaluation of these measures depended on the nature of the measures. In most cases, results have compared the data obtained from the baseline evaluation with the data collected after the application of the measure (i.e. before-and-after measurements). However, for mid-platform fencing a different method was used. A logic map was created to clarify the overall objectives of the intervention and the context in which the intervention was implemented. As result of this, important steps were described as a series of inputs, outputs, outcomes, and impacts (as defined by Hills and Junge ). Other specific methodologies were used in the evaluations targeted at mitigation of consequences of suicide and trespassing accidents. The evaluation of the CBT module was qualitative – assessing the CBT’s capacity to achieve its objectives. In the case of FFCCTV, it was not possible to organise a field trial. Instead, a study was carried out focusing on the numbers and costs to the rail industry of rail fatalities; the application, costs and the effectiveness of FFCCTV; and how, by whom and for what purpose the available information is used. Furthermore, where possible the pilot tests included a socio-economic evaluation of the implemented measure(s) in the form of CEA (Cost-Effectiveness Analysis) and CBA (Cost-Benefit Analysis) (see  for more details). For each possible pilot test, cost data included those related to design, implementation, maintenance, whereas effectiveness data included one or several variables assessed before and after the test period. However, CEA and CBA ratios were not comparable between measures for many structural reasons, thus the main results enabled only to make clear the amount of investment required to achieve the given goal in the circumstances where the measure was implemented.
Results and main conclusions of the field studies
The results of the pilot tests provided altogether new recommendations to improve reduce the number of railway suicide, (fatal) trespassing accidents and post-incident consequences (see  for more details about the results of the pilot tests and lessons learned during the trials). Those results which did not bring new recommendations were in line with the evidence from the literature, and provided new empirical support for the effectiveness of particular measures.
Concerning the reduction of the number of suicides, the results of the both gatekeeper training courses were comparable to those reported in the literature [32, 33]. Gatekeepers are people who have frequent contact with possibly suicidal members of the community on account of their professional status (e.g. railway personnel, security staff, fire-fighters, local charity workers). The gatekeeper training teaches them to identify people at risk by recognising suicidal risk factors, to assess the levels of risk, and to manage the situation appropriately by employing adequate approaching tactics. Both RESTRAIL evaluations (one in Germany and one in the Netherlands) have shown that the training provided a significant improvement in (a) skills such as knowledge about railway suicide, (b) attitude toward railway suicides, and (c) the feeling of competence of staff. These increases occurred for both men and women, for all ages, and particularly for those employees with less than 20 years of job experience. These types of courses are highly recommended since they are cheap and adjusted easily to different circumstances and settings, and prior knowledge on the part of participants is not required. The contents of the course depend on the local culture and in countries where suicide is not accepted this course probably will not work. Finally this type of course can be combined with other measures without any problem.
Another measure aimed at railway suicides was societal collaboration. This measure had not been evaluated before; therefore this field test had a major added value. This measure is a joint venture between the RU, IM, police, fire brigade/rescue services and health authorities from a local area. The responder that first becomes aware of any person in the track area requests and receives a temporary traffic shut-down on the concerned railway line. According to the results, 40 of the 64 persons threatening to commit suicide were found and taken to psychiatric care by the Police. In addition, train services were less disturbed by short traffic stops on more occasions than an actual fatal accident. For example, short traffic stops (involving 25 h for 64 threats of suicide) where people have been saved could be compared to 4 cases where this was not the case involving trains stopped for 30 h. This study identified the great importance of the need for very clear communication among the participants and actions to ensure that effective collaboration is achieved.
The evaluation of mid-platform fencing as anti-suicide measure had been one of the more expected, since there have been no previous studies of the effectiveness of this type of fencing at stations. This has been an extensive trial in RESTRAIL and the findings were encouraging. This measure consists of a fence along the central line of an island platform aiming to block pedestrian access from one edge of the platform to the other edge. It is usually used to separate people from the trains passing at high speeds or to isolate fast lines where trains might not stop from the regular lines which should be easily accessible. This field study compared the number of suicides at 51 stations around London between 1994 and 2014 (22 of these stations equipped with mid-platform fencing as part of recent programmes of suicide prevention). There has been only one fatality at fast lines at a station, after mid-platform fencing has been fitted. This result needs, however, to be interpreted with caution. The monitoring period (post-intervention) has been short in relation to many of the stations, therefore there is need for collection and analysis of statistics over a longer period of time to determine if the fencing of fast lines is potentially contributing to a displacement of incidents to other lines or stations. The evaluation has also shown a positive public perception: people liked the fencing and thought that they worked in preventing incidents. There may also be other benefits, such as increasing perceptions of safety while on platforms and the prevention of unsociable behaviour and access to places where people should not be. This type of fencing can be used in combination with other interventions (e.g. training of staff, improved surveillance) and should not present problems in transferring to other countries. It can be costly and is not a solution that can be applied and every station. However, this can be a realistic option to consider where there is an appropriate station configuration and a high proportion of non-stopping trains at a platform at the station.
Regarding the prevention of trespassing incidents, two measures focused on the educational aspects. The education programme in schools for 8–11 year old children implemented in Finland had a positive effect on (a) the level of knowledge related to railway trespassing, (b) reported crossing behaviour, and (c) pupils’ assessment of safety related to crossing railway lines. The detailed results of this RESTRAIL field study have recently been published (see ), suggesting that a 45-min lesson on safe behaviour could even have a positive effect on the future frequency of trespassing. Similarly, the Railway Safety Education Programme implemented in Spain achieved (a) an increase in teacher awareness about the need to cover railway safety at school and (b) greater confidence, skills and commitment to do so in the future. Moreover, students were able to apply this knowledge to explain the repercussions for someone acting dangerously on or near the railway tracks and in a station. The findings from both evaluations are in line with the published evidence [35, 36] but also suggested the children’s knowledge of railway safety and their subsequent behaviour was heavily influenced by the actions observed in the adults around them. For this reason, education outside of schools also plays an important role in communications the safely message. In this sense the railway museums have a crucial role in bringing the society closer to the world trains. These programmes can be applicable in different social contexts, although it is obviously necessary to adapt the contents to the reality of where the measure is being applied and contents should take into account the demographic profile of the target population and the features of the local implementation site.
The use of warning signs and posters proved to be effective in discouraging pedestrians from using illegal crossing places when the displayed messages provided information concerning the possibility of being fined in combination with information about rail safety. The results were consistent with the few published studies, namely that the effect is likely to be significant but not necessarily large [35, 37]; however, the recommendations from this study were threefold: (a) the design and content of the signs/posters should be carefully planned to fit the local context and culture as well as the preferences of the national train operators that might disagree with the displayed message; (b) the amount of signs and posters should be carefully considered, to avoid unnecessary signage; and (c) the period of time the signs and posters are exposed in a specific area should be considered, since the effect of is likely to be reduced over time. The optimal measure would be to combine signage with wider targeted campaigns against trespass.
After the implementation of the video enforcement and sound warning system at two sites in southern Finland, the number of trespassing incidents dropped significantly . The system included a video camera linked to a motion sensor and a loudspeaker. Trespassers were observed by automatic camera that took a series of pictures whenever movement was detected on the illegal crossing route. Upon detection the trespasser was given a warning message by the loudspeaker. The effect on the frequency of trespassing was calculated by comparing trespasser counts before and after the implementation. In the two pilot test sites the reduction in the frequency of trespassing was 18 % and 44 %, supporting the only existing study . However, because of the lack of control sites the effect may have become somewhat overestimated. Those who are planning to implement a similar measure are advised to use an expected effect of the reduction of trespassing between 10 % and 30 %, depending on local circumstances, especially the distance to alternative legal crossing facilities. In general, there were no difficulties regarding the system’s maintenance. It seems likely that adding media campaigns and a true threat of punishment to video enforcement and sound warning, its effect on trespassing could be enhanced, at least on the short term. In order to maintain the effect high, media coverage should be maintained and include also information on issued penalties.
Lastly, the field trial which evaluated a combination of different measures in Turkey included: (a) anti-trespass panels in conjunction with fencing at platform ends and intermediate fencing between the tracks to physically block the access; (b) video surveillance camera; and (c) warning and prohibitive signs informing the public of the dangers and illegality of trespassing. This trial had high added value since it was not known whether similar combinations of measures against trespassing in railway area had been implemented before, nor which was their combined effect on the frequency of trespassing. The results indicated an immediate reduction of almost 95 % on trespass behaviour. Therefore, this combination of measures could be a good option in order to reduce the number of trespassers in specific railway areas. This is also supported by a recent field trial conducted in Belgium , which indicated that a somewhat similar combination of measures applied to a relatively similar context reduced the number of trespassers by 78 %.
Concerning the mitigation of consequences, two studies were conducted and the results of the both measures were innovative. The computer based training (CBT) module was effective in making a positive contribution to the understanding by decision makers handling suicides and fatal trespassing incidents of the manner in which such incidents are handled. Collaboration between decision makers, RUs and IMs for effective incident management and the manner in which it can support them in managing these incidents was emphasised. Altogether this training was considered highly relevant with extremely high effectiveness for RUs and the police. Above all, the means described in the lesson was perceived as valuable to reducing shut-down time as a result of suicides and fatal trespassing incidents.
In addition, the analysis concerning Forward Facing CCTV, has shown that FFCCTV with a wireless link provided real time remote access to images by key decision makers, particularly the police, facilitating the earliest possible decision making on the circumstances involved with rail fatalities. Determining whether a suicide, trespassing accident or homicide is involved has a considerable impact on system shut down time, thus close liaison by RUs and IMs with the police is essential to maximise FFCCTV benefits.