Problem: Suicide is a dominating, although hidden, safety problem oil Swedish railroads. The aim of this paper is to describe the epidemiology of fatal train-person collisions as a basis for systems-oriented prevention. Method: Data oil collision circumstances were collected from narrative reports at the Swedish National Rail Administration. Results: The events were evenly distributed by months and weekdays, however, Most Suicides occur during the clay while unintentional events usually occur at night. Most train-person collisions happened in densely populated areas, and 75% of the suicide victims were waiting on the track before the collision. Significance test between types of injury event (Suicide, accident, or unknown intent) showed small or no differences. Conclusion: Traditional approaches to accident prevention by systems modification seem largely applicable to combat railroad Suicide as well. Impact on industry: Our findings show promising preventive potentials. (c) 2005 National Society Council and Elsevier Ltd. All rights reserved.