The effectiveness of platform screen doors for the prevention of subway suicides in South Korea

• The effectiveness of platform screen doors for suicide prevention is understudied.
• The relative effectiveness of half- and full-height platform doors was compared.
• Installing doors at stations is shown to be an effective strategy to prevent suicides.
• Doors that extend from the platform to the ceiling are extremely effective.
• When these doors are not high, they can be climbed over and thus are ineffective.

BackgroundSubway suicide can significantly impact the general public. Platform Screen Doors (PSDs) are considered to be an effective strategy to prevent suicides at subway stations, but the evidence on their effectiveness is limited.MethodsWe assessed the effectiveness of installing half- and full-height platform screen doors in reducing subway suicides using Poisson regression analysis. Ten-year monthly panel data for 121 subway stations between 2003 and 2012 in the Seoul metropolitan area were used for the analysis.ResultsWe found that installing PSDs decreases fatal suicide cases by 89% (95% CI: 57–97%). We also found that the installation of full-height PSDs resulted in the elimination of subway suicides by completely blocking access to the track area; however, half-height PSDs, which do not extend to the ceiling of the platform, were not as effective as full-height ones.LimitationOur findings were based on the data from a single subway operator for a limited period of time. Accordingly, we did not consider the possibility that some passengers choose to die at a station run by other operators. Our study did not examine the potential substitution effects of other suicide methods.ConclusionInstalling physical barriers at subway stations can be an effective strategy to reduce the number of subway suicides; however, half-height PSDs are not as effective as full-height ones, even when they are as high as the height of an adult. Thus, these barriers should be made high enough so that nobody can climb over them.