Objective: To determine whether injuries to sub-optimally restrained child occupants in real-world crashes were likely to be preventable by alternative restraint usage practices and to assess the usefulness of crash reconstruction for exploring injury mechanisms in child occupants.
Methods: Real-world crashes in which child occupants sustained significant injuries were reconstructed on a laboratory crash sled using the Hybrid III family of child dummies. Alternative restraint scenarios and cases in which children were not seriously injured were also simulated to compare dummy kinematics and dynamic responses in optimal restraint configurations.
Results: Restraint misuse was associated with greater motion of the dummy torso and head during crashes, often allowing contact between the child and the vehicle interior, resulting in injury. Poor pre-crash posture for a child inappropriately restrained in an adult belt appeared to worsen the geometry of the sash (shoulder) belt, resulting in a cervical injury due to direct interaction with the belt. Dynamic dummy data did not appear to discriminate between injury and non-injury cases.
Conclusions: Dummy kinematics suggest that injuries in which inappropriate use and misuse were a factor were less likely if the most appropriate restraint was used correctly. Adequately controlling the head and upper body of the child occupant was seen to prevent undesirable interactions with the vehicle interior and restraint system, which were associated with injury in the real world. Neck forces and moments and injury criteria calculated from these did not predict injury reliably.