BACKGROUND: Motor vehicle collisions are the leading cause of death among children older than 1 year. Use of appropriate restraint systems is associated with reductions in morbidity and mortality in this age group. No studies have evaluated the association between specific injury patterns and restraint use among children. The purpose of this study was to evaluate differences in risks of injuries in different body regions according to restraint use among children 0 to 11 years of age. METHODS: The 1995 to 1999 National Automotive Sampling System data files were used. The National Automotive Sampling System is a national probability sample of passenger vehicles involved in police-reported tow-away collisions. Information on occupant (seating position, restraint use), collision (change in velocity, vehicle intrusion), and outcome characteristics was evaluated. Risks of injuries in different body regions (Abbreviated Injury Scale, 1990 Revision score > or = 2) were calculated and compared according to restraint use. RESULTS: Between 1995 and 1999, there were approximately 1.5 million children 0 to 11 years of age involved in police-reported tow-away MVCs who met the inclusion criteria for this study. Compared with unrestrained children, properly restrained children had significantly lower overall injury risk (risk ratio [RR], 0.37); significant risk reductions were also observed for injuries to the head (RR, 0.18), thorax (RR, 0.35), and lower extremities (RR, 0.26), and mortality (RR, 0.26). Significant risk reductions were not noted when comparing improperly restrained children with unrestrained children. CONCLUSION: Proper restraint use among children is associated with lower risk of injury. Educational initiatives should focus not only on encouraging restraint use but also on ensuring that parents know the appropriate age-dependent restraint method and how to use it properly.