The main objective is to study long-term effects of legislation and promotion of child-restraint use in motor vehicles. Subgoals are to describe the level of restraint use in cars and changes in mortality and morbidity patterns in regard to differences among age groups and geographical areas (morbidity). The study areas were; 'The First Four', 'The Six Counties', 'The Rest of Sweden', and 'Sweden as a whole', and the population were two age groups--0-6 and 7-14 years. The incidence of restraint use comes from observations of car-seat-belt usage. Outcome evaluation was based on hospital-discharge data 1978-1996 and mortality statistics 1970-1996. The level of restraint use for children in the front seat was 97% in 1988 and was equal to 1995. Mortality data shows a decrease of 2.8% on an average per year, 76% over the study period. A significant change over time in the two intervention areas was shown (annual changes of -2.8 and -1.8%), but not in the rest of Sweden. The local authorities that started early with preventive programs, such as safety seat loan schemes and those having an organised safety-promotion program showed a much better improvement than the rest of Sweden. There is a need for appropriate information for local action on childhood injury prevention to accompany national legislation.