The objectives of the study are to assess the impact of a community-based bicycle-helmet program aimed at children aged 5-12 years (about 140,000). A quasi-experimental design, including a control group, was used. Sex- and age-group-based changes in the risk of bicycle-related head injury leading to hospitalisation were measured, using rate ratios. Compared with the pre-program period, significant risk reductions were observed during the post-program period among both boys (RR = 0.56, 95 percent CI = 0.40, 0.77) and girls (RR = 0.52, 95 percent CI = 0.33, 0.82), and among both younger (RR = 0.46, 95 percent CI = 0.31, 0.68) and older (RR = 0.63, 95 percent CI = 0.44, 0.89) children. A significant reduction was also observable during the program phase among the groups most at risk, i.e. boys (RR = 0.94, 95 percent CI = 0.66, 1.35) and younger children (RR = 1.07, 95 percent CI = 0.70, 1.63). The population-based educational program significantly decreased the risk of head injuries among boys and girls despite observable differences in the voluntary adoption rate of bicycle-helmet wearing. The impact was more pronounced among younger children. (Contains 3 figures and 4 tables.)